Remote sensing has been used as an important data acquisition tool for precision agriculture for decades. Based on their height above the earth, remote sensing platforms mainly include satellites, manned aircraft, unmanned aircraft systems (UAS) and ground-based vehicles. A vast majority of sensors carried on these platforms are imaging sensors, though other sensors such as lidars can be mounted. In recent years, advances in satellite imaging sensors have greatly narrowed the gaps in spatial, spectral and temporal resolutions with aircraft-based sensors. More recently, the availability of UAS as a low-cost remote sensing platform has significantly filled the gap between manned aircraft and ground-based platforms. Nevertheless, manned aircraft remain to be a major remote sensing platform and offer some advantages over satellites or UAS. Compared with UAS, manned aircraft have flexible flight height, fast speed, large payload capacity, long flight time, few flight restrictions and great weather tolerance. The first section of the article provided an overview of the types of remote sensors and the three major remote sensing platforms (i.e., satellites, manned aircraft and UAS). The next two sections focused on manned aircraft-based airborne imaging systems that have been used for precision agriculture, including those consisting of consumer-grade cameras mounted on agricultural aircraft. Numerous custom-made and commercial airborne imaging systems were reviewed, including multispectral, hyperspectral and thermal cameras. Five application examples were provided in the fourth section to illustrate how different types of remote sensing imagery have been used for crop growth assessment and crop pest management for practical precision agriculture applications. Finally, some challenges and future efforts on the use of different platforms and imaging systems for precision agriculture were briefly discussed.
Plant phenomics is under rapid development in recent years, a research field that is progressing towards integration, scalability, multi-perspectivity and high-throughput analysis. Through combining remote sensing, Internet of Things (IoT), robotics, computer vision, and artificial intelligence techniques such as machine learning and deep learning, relevant research methodologies, biological applications and theoretical foundation of this research domain have been advancing speedily in recent years. This article first introduces the current trends of plant phenomics and its related progress in China and worldwide. Then, it focuses on discussing the characteristics of indoor phenotyping and phenotypic traits that are suitable for indoor experiments, including yield, quality, and stress related traits such as drought, cold and heat resistance, salt stress, heavy metals, and pests. By connecting key phenotypic traits with important biological questions in yield production, crop quality and Stress-related tolerance, we associated indoor phenotyping hardware with relevant biological applications and their plant model systems, for which a range of indoor phenotyping devices and platforms are listed and categorized according to their throughput, sensor integration, platform size, and applications. Additionally, this article introduces existing data management solutions and analysis software packages that are representative for phenotypic analysis. For example, ISA-Tab and MIAPPE ontology standards for capturing metadata in plant phenotyping experiments, PHIS and CropSight for managing complicated datasets, and Python or MATLAB programming languages for automated image analysis based on libraries such as OpenCV, Scikit-Image, MATLAB Image Processing Toolbox. Finally, due to the importance of extracting meaningful information from big phenotyping datasets, this article pays extra attention to the future development of plant phenomics in China, with suggestions and recommendations for the integration of multi-scale phenotyping data to increase confidence in research outcomes, the cultivation of cross-disciplinary researchers to lead the next-generation plant research, as well as the collaboration between academia and industry to enable world-leading research activities in the near future.
Under the pressure of economy development and climate change, rice production and distribution in the Yangtze River basin have undergone great changes, which may pose a great threat to the ecological environment and food security. Based on land use remote sensing-monitoring data from1990 to 2015, the GIS spatial analysis method was used to explore the spatial pattern variation characteristics of paddy fields in the Yangtze River economic belt. Meanwhile, Ecosystem services value (ESV) was calculated by using the equivalent factor method corrected by region and time factor to measure the comprehensive impact of paddy field change. The results showed that, to begin with, paddy fields number of the Yangtze River economic belt continued to decrease, with a total decrease of 17 390km2, and the decrease rate presented a trend of growth with significant regional differences. The difference between the reduction rate of paddy fields in the middle upper and the lower reaches of the Yangtze River was about 9.56%. Among them, in the lower reaches of the Yangtze River, the proportion of paddy fields decreased, while in the middle and upper reaches which was just the opposite. Then, paddy fields mainly flowed to construction land and water, resulting from economic construction and aquaculture development. Paddy fields chiefly came from water, dry land and wetland, etc. Furthermore, paddy fields in the Yangtze River Delta, the middle reaches of the Yangtze River and the Chengdu-Chongqing urban agglomerations changed the most dramatically. The expansion of construction land to paddy fields was widespread, and paddy fields flooded by water primarily distributed in the two lake plains. In addition, the conversion of paddy fields and other ecosystems had a positive impact on ESV, in which the paddy-water diversion type contributed the most. Its scale determined the net increase of ESV in different periods. Value loss lead by conversion from water to paddy fields was the largest, and construction land invading paddy fields was the second. The conversions in different cities were different, so the difference in ESV increases and decreases. In addition, the tradeoffs within ecosystem services were mainly between hydrological regulation, water supply and food production, gas regulation, which were directly related to the increase of water resources and the loss of paddy fields. The research results are helpful to reveal the spatio-temporal changes process of paddy fields in the Yangtze River basin and its impact on ecological functions, and provide theoretical support for regional land use planning, agricultural policy and ecological sustainable development.
Water concentration is a key parameter to characterize crop physiological and healthy status. It is of great significance of employing unmanned aerial vehicle (UAV) low-altitude remote sensing technology to predict crop water concentration for crop breeding and precision agriculture management. UAV remote sensing has been widely used for monitoring crop growth status, mainly focusing on using vegetation indices to estimate crop growth parameters at single or several growth stages. Few studies have been performed on evaluating crop water concentration. Consequently, this study mainly used vegetation indices and texture features extracted from UAV-based RGB and multispectral images to monitor water concentration of rice crop during the whole growth period. Firstly, a multi-rotor UAV equipped with high-resolution RGB and multispectral cameras to collect canopy images of rice crop, and water concentration was also measured by ground sampling. Then, vegetation indices and texture features calculated from RGB and multispectral images were used to analyze the growth changes of rice. Finally, random forest regression method was used to establish a prediction model of water concentration based on different image features. The results show that: (1) vegetation index, texture features and ground-measured water concentration could be used to dynamically monitor rice growth, and there existed correlations among these parameters; (2) image features extracted from multispectral images possessed more potential than those from RGB images to evaluate water concentration of rice crop, and normalized difference spectral index NDSI771, 611 achieved the best prediction accuracy (R2 = 0.68, RMSEP = 0.039, rRMSE = 5.24%); (3) fusing vegetation indices and texture features could further improve the prediction of water concentration (R2 = 0.86, RMSEP = 0.026, rRMSE = 3.21%), and the prediction error of RMSEP was reduced by 16.13% and 18.75%, respectively. These results demonstrats that it is feasible to apply UAV-based remote sensing to monitor water concentration of rice crop, which provides a new insight for precision irrigation and decision making of field management.
With the development of hyperspectral sensor technology and remote sensing data acquisition platform, the application of hyperspectral data is becoming more and more popular in precision agriculture. Spectral features and spatial features are two main kinds of features used in hyperspectral image classification. The comparison of spectral features and spatial features in vegetation classification of hyperspectral image is a special application in hyperspectral image classification. Therefore, this study compared the performance of several typical spectral features and spatial features in vegetation classification of hyperspectral image. The considered spatial features include grey level co-occurrence matrix (GLCM) based features, Gabor features and morphological features. The considered spectral feature selection or extraction methods include minimal-redundancy-maximal-relevance (mRMR), joint mutual information (JMI), conditional mutual information maximization (CMIM), double input symmetrical relevance (DISR), Jeffreys-Matusita (JM), principal component analysis (PCA), independent component analysis (ICA) and linear discriminant analysis (LDA). PCA, an effective subspace feature extraction method, is widely used in the feature extraction of hyperspectral image. The first several principal components (PCs) are usually selected as spectral features in hyperspectral image classification. However, the first several PCs have no guarantee to achieve good class separability and classification accuracy. Considering that, a hybrid feature extraction approach named as PCA_ScatterMatrix was proposed which combined PCA and an improved scatter-matrix-based feature selection method, aiming to select PCs with high class separability and get high overall classification accuracy. The experiments and comparative analyses were conducted with a widely used hyperspectral image, which was collected over the agricultural area in northwestern Indiana, USA (United States of America) by the AVIRIS (Airborne Visible / Infrared Imaging Spectrometer). The experimental results indicated that: (1) The proposed hybrid feature extraction method PCA_ScatterMatrix got the highest overall classification accuracy on both data sets (82.7% and 86.5%) among three classic subspace feature extraction methods (PCA, ICA and LDA) and respectively improved overall classification accuracy by 1.5% and 2.5% on both data sets, comparing to original PCA; (2) Compared to spectral features, spatial feature extraction methods generally got higher overall classification accuracy, especially Gabor spatial features got the highest overall classification accuracy on both data sets (95.5% and 96.7%). The results suggest that the proposed method is effective in vegetation classification of hyperspectral image and the spatial features play a much more important role in vegetation classification of hyperspectral image, comparing with spectral features.
Rice is one of the important staple crops in China, and the rice planted in Northeast China, such as in Liaoning, Jilin, and Heilongjiang regions, is called cold-region rice. The chlorophyll content in rice leaves is the most direct indicator of the rice growth period and can directly reflect on its nutritional value. Previous research demonstrates that when the chlorophyll content of rice changes, the reflectance of different bands changes at the spectral level. In addition, most of the research studies on the inversion of the rice’s chlorophyll content are based on the complex machine learning algorithms. Although the accuracy of the inversion of the constructed model has been improved, the structure of the model is relatively complex, and the model’s transplantation and universality are poor in the actual application process. Hence, in this study, the inversion of the chlorophyll content of rice leaves in the cold regions was assessed. An ASD ground object spectrometer was employed to procure the hyperspectral information of rice leaves in the critical growth period. On the basis of the feature selection method, the hyperspectral feature subset of the inversion of the chlorophyll content of rice was selected. The characteristic band vegetation index was constructed by combining the chlorophyll content absorption coefficients, and the chlorophyll content of rice was established through using regression analysis. Additionally, by combining the chlorophyll content absorption coefficients in the PROSPECT model, referring to the construction method and form of the existing hyperspectral vegetation index, and using correlation analysis, the continuous projection method and the genetic algorithm optimized the rough set attribute reduction, the hyperspectral features was selected, and the red edge optimization index (ORVI) with only 695, 507, and 465nm hyperspectral feature bands was proposed. Compared with the other vegetation indexes retrieved from the IDB database, namely, ND528,587, SR440,690, CARI, and MCARI, the results demonstrated that the determination coefficients of the abovementioned vegetation index inversion models were 0.672, 0.630, 0.595, and 0.574 respectively. The accuracy of the inversion model of chlorophyll content established by ORVI vegetation was higher than that of other vegetation indexes wherein the decision coefficients of the model were R2 =0.726 and RMSE = 2.68, revealing that ORVI can be used as a hyperspectral vegetation index for the rapid inversion of the rice’s chlorophyll content in practical applications. This research can thereby provide some objective data support and model reference for remote sensing diagnosis and management decision of the rice’s chlorophyll content in the cold regions.
The capacity of canopy light interception is a key functional trait to distinguish the phenotypic variation over genotypes. High-throughput phenotyping canopy light interception in the field, therefore, would be of high interests for breeders to increase the efficiency of crop improvement. In this research, the Digital Plant Phenotyping Platform（D3P） was used to conduct in-silico phenotyping experiment with LiDAR scans over a wheat field. In this experiment virtual 3D wheat canopies were generated over 100 wheat genotypes for 5 growth stages, representing wide range of canopy structural variation. Accordingly, the actual value of traits targeted were calculated including GAI (green area index), AIA (average inclination angle) and FIPARdif (the fraction of intercepted diffuse photosynthetically activate radiation). Then, virtual LiDAR scanning were accomplished over all the treatments and exported as 3D point cloud. Two types of features were extracted from point cloud, including height quantiles (H) and green fractions (GF). Finally, an artificial neural network was trained to predict the traits targeted from different combinations of LiDAR features. Results show that the prediction accuracy varies with the selection of input features, following the rank as GF + H > H > GF. Regarding the three traits, we achieved satisfactory accuracy for FIPARdif (R2=0.95) and GAI (R2=0.98) but not for AIA (R2=0.20). This highlights the importance of H feature with respect to the prediction accuracy. The results achieved here are based on in-silico experiments, further evaluation with field measurement would be necessary. Nontheless, as proof of concept, this work further demonstrates that D3P could greatly facilitate the algorithm development. Morever, it highlights the potential of LiDAR measurement in the high-throuhgput phenopyting of canopy light interpcetion and structural traits in the field.
Intelligent production and robotic oporation are the efficient and sustainable agronomic route to cut down economic and environmental costs and boosting orchard productivity. In the actual scene of the orchard, high performance visual perception system is the premise and key for accurate and reliable operation of the automatic cultivation platform. Most of the existing apple detection models, however, are difficult to be used on the platforms with limited hardware resources in terms of computing power and storage capacity due to too many parameters and large model volume. In order to improve the performance and adaptability of the existing apple detection model under the condition of limited hardware resources, while maintaining detection accuracy, reducing the calculation of the model and the model computing and storage footprint, shorten detection time, this method improved the lightweight MobileNetV3 and combined the object detection network which was based on keypoint prediction (CenterNet) to build a lightweight anchor-free model (M-CenterNet) for apple detection. The proposed model used heatmap to search the center point (keypotint) of the object, and predict whether each pixel was the center point of the apple, and the local offset of the keypoint and object size of the apple were estimated based on the extracted center point without the need for grouping or Non-Maximum Suppression (NMS). In view of its advantages in model volume and speed, improved MobileNetV3 which was equipped with transposed convolutional layers for the better semantic information and location information was used as the backbone of the network. Compared with CenterNet and SSD (Single Shot Multibox Detector), the comprehensive performance, detection accuracy, model capacity and running speed of the model were compared. The results showed that the average precision, error rate and miss rate of the proposed model were 88.9%, 10.9% and 5.8%, respectively, and its model volume and frame rate were 14.2MB and 8.1fps. The proposed model is of strong environmental adaptability and has a good detection effect under the circumstance of various light, different occlusion, different fruits’ distance and number. By comparing the performance of the accuracy with the CenterNet and the SSD models, the results showed that the proposed model was only 1/4 of the size of CenterNet model while has comparable detection accuracy. Compared with the SSD model, the average precision of the proposed model increased by 3.9%, and the model volume decreased by 84.3%. The proposed model runs almost twice as fast using CPU than the CenterNet and SSD models. This study provided a new approach for the research of lightweight model in fruit detection with orchard mobile platform under unstructured environment.
Excessive application of water and fertilizer not only causes resources serious waste of, but also causes serious environmental pollution. The implementation of precision irrigation and fertilization can effectively reduce nutrient loss and environmental pollution, save irrigation water and improve the utilization rate of water and fertilizer resources, which is one of the important ways to promote the sustainable development of agriculture. The use of the integrated water-fertilizer equipment can effectively improve the utilization rate of water-fertilizer resources, but it is necessary to know the nutritional status of crops and water-fertilizer demand before operation. To acquire the information by hand-held measuring instruments, there are some disadvantages, such as poor timeliness and high labor intensity. In response to the above problems, this study took the common corn crop as an example, used the DJI Phantom III drone to carry RedEdge-M multispectral camera to collect multispectral images of corn crops over the fields, and measured nitrogen and moisture content of corn plants by YLS-D series plant nutrition tester. Based on this information, the collected images were divided into 3 levels, each level contains 530 five channel images (2650 single channel images), including 480 five channel images (2400 single channel images) in the training set and 50 five channel images (250 single channel images) in the verification set, and a method of identifying the nutritional status of corn crops based on convolutional neural network was proposed. Based on the TensorFlow deep learning framework, ResNet18 convolution neural network model was constructed. By entering color image data and five-channel multispectral image data into the model, the nutritional status recognition model of corn plant suitable for color image and multispectral image was trained, and the experimental results showed that the trained model could be used to recognize the multispectral images of corn, and the nutritional status of corn, topdressing guidance and GPS information could be outputted, the correct rate of the recognition color image model in the verification set was 84.7%. The correct rate of identifying multispectral image model in the verification set was 90.5%, the average time of model training was 4.5h, and the average time of recognizing a five channel image is 3.56 seconds, which can detect the nutritional status of corn crops quickly and undamaged, and provides a theoretical and technical basis for the accuracy of the application of water fertilizer in intelligent agriculture.
How to accurately obtain the area and freshness of broccoli head in the field condition is the key step to determine broccoli growth. However, the rapid segmentation and grading of broccoli ball remains difficult due to the low equipment development level. In this research, we combined an advanced computer vision technique with a deep learning architecture to allow the acquisition of real-time and accurate information about broccoli ball. By constructing a private image dataset with 442 of broccoli-ball images (acquired using a self-developed imaging system) under controlled conditions, a deep convolutional neural network named “Improved ResNet” was trained to extract the broccoli pixels from the background. The technical process of our method includes: (1) take the orthophoto images of broccoli head based on a near-ground image acquisition platform and establish the original data set; (2) preprocess the training images and input the model for segmentation; (3) use the PSOA and Otsu algorithm for fine segment based on color characteristics to obtain the freshness information. The experimental results demonstrated that the precision of the segmentation model is about 0.9 which is robust to the interference of soil reflectance fluctuation, canopy shadow, leaf occlusion and so on. Our experiments showed that a combination of improved ResNet and PSOA method got higher broccoli balls segmenting and grading precision. One major advantage of this approach is that dealing with only a few images, reducing the data volume and memory requirements for the image processing. All of the methods were evaluated using ground-truth data from three different varieties, which we also make available to the research community for subsequent algorithm development and result comparison. Compared with other 4 approaches, the evaluation results shows better performance regarding the segmentation and grading accuracy. The results of SSIM, precision, recall and F-measure by using Improved ResNet were about 0.911, 0.897, 0.908 and 0.907 respectively, which were 10%~15% higher than the traditional approaches. In addition, on the basis of the segmentation results, PSO-Otsu method was proved that it can be used to achieve a quickly analysis to the freshness of the ball, with the mean accuracy of 0.82. Overall, the proposed method is a high-throughput method to acquire multi-phenotype parameters of broccoli in field condition, which can support the research of broccoli field monitoring and traits tracking.
Chemical plant protection, which refers to using plant protection machinery sprays chemical pesticides, is the most important technology for pest and disease control at present, an important technical guarantee for food security, and also is essential for safeguarding agricultural production. Pesticide, spray technology and plant protection machinery are called the three pillars of chemical plant protection, which having been becoming a hot research topic in the world. Efficient, precise and intelligent spray technology and equipment can provide guarantee for the improvement of pesticide efficacy and utilization. With the issues of agricultural product safety and environmental protection getting more and more attention from the public, the research and development direction of Chinese plant protection field will gradually turn to intelligent and precision spraying technology and equipment. Since 2010 year, the great development potential and application value of intelligent and precision spraying technologies and equipment have been widely recognized worldwide. In this article, the main precision spraying technologies were reviewed, the research status, typical representative and application progress of plant protection equipment in different working scenarios were classified and summarized. The challenges in the development of precision spraying were analyzed, the countermeasures and suggestions were put forward. This research can provide new methods and new ideas not only for implementation of China's pesticide reduction plan, the promotion of intelligent plant protection equipment and precision spraying technology, but for the development of modern agriculture.
Agricultural models, agricultural artificial intelligent, and data analysis technology, etc., exist in whole processes of information perceiving, transmission, processing and control for smart agriculture， thus they are the core technology of smart agriculture. To furtherly make the substances and functions of agricultural models clear, facilitate its further research and application, drive smart agriculture development with healthy, steady, and sustainable, methods of systematic analysis, comparison, and chart for relationship, etc. were used in this research. The definition, classification, functions of the agricultural models were theoretically analyzed. The relationships between the agricultural models and the elements and processes of the smart agriculture were expounded, which made the functions of agricultural models clear, provided some agricultural models examples applied in the smart agriculture. The important studies and application progresses of agricultural models were reviewed. The comparison results of agricultural models showed that the 4 levels of agricultural biological elements, 6 scales of agricultural environmental elements, 6 administrative levels of agricultural technological and economic elements, and the relevant approaches for modeling agricultural system need to be considered. The research and application of multi-space scales on environment elements in the agricultural models would have the larger potential. The combination of agricultural models with molecular genetics, perceiving, and artificial intelligence, the collaboration among public and private researchers, and food security challenges have been an important power for further development of agricultural models, linking agricultural models with various agricultural system modeling, databases, harmonious and open data, and decision-making support systems (DSS) would be focus on. The research and application of the agricultural models in China have formed crop model series with Chinese characteristics, joined in the world trends of the Agricultural Model Intercomparison and Improvement Project (AgMIP), the smart agriculture, and so on. They should be speedy graspe chances and accelerate development. The agricultural models is a quantitative express of relationships within or among the agricultural system elements. An important method with epistemological values of quantifying and synthesizing agricultural sciences, and will play an indispensible role in data achieving and processing for the smart agriculture combining perceiving techniques, and become a significant bridge and bond.
Nano materials with special size effect and excellent photoelectric properties have been highly valued and widely used in sensing analysis for greatly improving the performance of sensor analysis technology. In recent years, with the rapid development of smart agriculture, the quality and safety of agricultural products as an important part of agricultural production have attracted more and more attentions. There are many harmful ingredients, including pesticides, veterinary drugs, mycotoxins, and environmental contaminants etc, can potentially affected the quality and safety of agricultural products. Therefore, high performance analytical methods and sensing technologies are essential. Thanks to the emerging of nano materials, they provide a novel approach to improve the analytical performances of the sensing technologies. Furthermore, the sensors based on nano materials have also been utilized into monitoring the harmful substances in agricultural products. This review briefly described the properties and characteristics of several commonly used nano materials, including carbon nano materials, noble metal based nano materials and metal-organic framework materials, follow discussed on the common sensing and analysis technologies and devices based on nano materials, such as chemical sensor, biosensor, electrochemical sensor and spectral sensor, as well as the application of nano sensing technology in the quality and safety monitoring of agricultural products. Especially, the function of nano materials in sensors and analytical performances of the developed sensors had been discussed in detailed. Chemical sensor devices had the characteristics of fast response speed and high sensitivity. They were widely used in environmental monitoring, food safety and medical diagnosis, such as monitoring hazardous substances, clenbuterol and melamine, metronidazole, dioxins, etc. Biosensors were widely used to monitor prohibited additives, mycotoxins, and so on. Electrochemical sensors were typically equipped with miniaturized analysis equipment, which detected trace targets, including small organic molecules, metal ions and biomolecules, by measuring changed in current and other electrochemical signals. This article introduced surface-enhanced Raman spectroscopy (SERS) , which was one of spectral sensor, and its applications. SERS technology had the advantages of good sensitivity, single molecule detection capability and rich spectral information. It had become a promising spectral technology in the rapid sensing analysis of target objects, and is developing rapidly in the fields of food safety, environmental monitoring and health. Finally, the existing problems of nano sensing and analysis technology, such as achievement of high-performance nano materials, fabrication of sensing devices and construction of high flux sensing arrays were summarized. The development trend and prospect of nanosensor were also discussed. It is believed that the review could provide a lot of useful information for the readers to understand the development of sensing technology for the quality and safety of agricultural products.
Solar insecticidal lamps Internet of Things (SIL-IoTs) is a novel physical agricultural pest control implement, which is an emerging paradigm that extends Internet of Things technology towards Solar Insecticidal Lamp (SIL). SIL-IoTs is composed of SIL nodes with functions of preventing and controlling of agricultural migratory pests with phototaxis feature, which can be deployed over a vast region for the purpose of ensuring pests outbreak area location, reducing pesticide dosage and monitoring agricultural environmental conditions. SIL-IoTs is widely used in agricultural production, and a number of studies have been conducted. However, in most current research projects, fault diagnosis has not been taken into consideration, despite the fact that SIL-IoTs faults have an adverse influence on the development and application of SIL-IoTs. Based on this background, this research aims to analyze the characteristics and challenges of fault diagnosis in SIL-IoTs, which naturally leads to a great number of open research issues outlined afterward. Firstly, an overview and state-of-art of SIL-IoTs were introduced, and the importance of fault diagnosis in SIL-IoTs was analyzed. Secondly, faults of SIL nodes were listed and classified into different types of Wireless Sensor Networks (WSNs) faults. Furthermore, WSNs faults were classified into behavior-based, time-based, component-based, and area affected-based faults. Different types of fault diagnosis algorithms (i.e., statistic method, probability method, hierarchical routing method, machine learning method, topology control method, and mobile sink method) in WSNs were discussed and summarized. Moreover, WSNs fault diagnosis strategies were classified into behavior-based strategies (i.e., active type and positive type), monitoring-based strategies (i.e., continuous type, periodic type, direct type, and indirect type) and facility-based strategies (i.e., centralized type, distributed type and hybrid type). Based on above algorithms and strategies, four kinds of fault phenomena: 1) abnormal background data, 2) abnormal communication of some nodes, 3) abnormal communication of the whole SIL-IoTs, and 4) normal performance with abnormal behavior actually were introduced, and fault diagnosis tools (i.e., Sympathy, Clairvoyant, SNIF and Dustminer) which were adapted to the mentioned fault phenomena were analyzed. Finally, four challenges of fault diagnosis in SIL-IoTs were highlighted, i.e., 1) the complex deployment environment of SIL nodes, leading to the fault diagnosis challenges of heterogeneous WSNs under the condition of unequal energy harvesting, 2) SIL nodes task conflict, resulting from the interference of high voltage discharge, 3) signal loss of continuous area nodes, resulting in the regional link fault, and 4) multiple failure situations of fault diagnosis. To sum up, fault diagnosis plays a vital role in ensuring the reliability, real-time data transmission, and insecticidal efficiency of SIL-IoTs. This work can also be extended for various types of smart agriculture applications and provide fault diagnosis references.
With the rapid growth of wireless nodes numbers and the increase in demanding for high-bandwidth transmission services such as multimedia images, the related fields of the agricultural Internet of Things(IoT) can foresee a trend of shortage of wireless spectrum resources. For the crop phenotypic information collection system based on the traditional IoT, there are many problems such as spectrum competition, data congestion during the data transmission process due to the dense deployment of nodes, and the reduction of the monitoring cycle due to uneven energy consumption in the fixed battery network. Based on previous studies, a crop phenotypic information collection model for cognitive radio sensor networks was established, and based on the model, an event-driven clustering routing algorithm that introduced dynamic spectrum and energy balance (DSEB) of edge computer system was proposed. The algorithm includes dynamic spectrum sensing clustering. The hierarchical clustering algorithm was used to combine the available channels, distances between nodes, residual energy, and neighbor node degrees obtained by spectrum sensing as similarities to cluster and cluster nodes in the monitored area and select cluster heads. The process of clustering and selecting cluster heads and constructing a clustering topology introduceed rewards and punishment factors to the equilibrium of the clustering sizes to improve the average spectrum utilization of each clustering network. The events triggered by edge computing trigger data routing, and based on the clustered topology structure, the events triggered by abnormal changes in farm conditions in the areas to be detected on the farm were forwarded to the convergent nodes by means of alternate cluster iterations and inter-cluster relays. Convergence includes direct transmission and intra-cluster relay, and inter-cluster relay includes two cases: ①primary gateway node and secondary gateway node-primary gateway node; ②adaptive re-clustering based on spectrum changes and communication quality of service (QoS)－changes in available channels caused by changes in the PU behavior of the primary user, or interference with poor quality of clustering effects on communication service quality, triggering cognitive radio sensor networks to perform adaptive re-clustering. In addition, a new energy balancing strategy was proposed to decentralize energy consumption (assuming sink is the center), that is, introducing a weight coefficient proportional to the distance from the node to the sink in the gateway or cluster head node selection calculation formula. The simulation results of the algorithm showed that, compared with the event-driven clustering ERP routing scheme using K-medoid clustering and energy sensing, under the premise that the number of CRSN nodes is a fixed value, the clustering routing algorithm based on DSEB in the network lifetime and there are certain improvements in utilization and energy efficiency; when the number of primary user nodes is a fixed value, the proposed algorithm has higher spectrum utilization than the other two algorithms.
The measurement of dissolved oxygen content in water is of great significance to aquaculture. However, the dissolved oxygen sensors on the market in China are expensive, and are difficult to maintain continuous online measurement and update parts, so they cannot be widely applied in real production and play expected role in the aquaculture Internet of things(IoT). Based on the principle of fluorescence quenching, a low cost and easy maintenance of dissolved oxygen sensor was developed in this work based on the relationship between the concentration of dissolved oxygen in water and the phase difference of fluorescence signal. The self-made oxygen-sensitive membrane was used to generate red fluorescence which being excited by blue light, and the fluorescence life was regulated by the concentration of dissolved oxygen. Photoelectric conversion circuit with optical signal sensing device was designed to sense optical signal. The STM32F103 microprocessor was used as the main control chip, and the lower computer program was programmed to generate the excitation light pulse. The phase-sensitive detection principle and fast Fourier transform (FFT) were used to calculate the phase difference between the excitation light and the reference light, which was converted into the concentration of dissolved oxygen and realized the measurement of dissolved oxygen. The fluorescence detection part and the main control part of the system were designed as detachable independent modules, and shield lines were used to plug and pull directly, so as to facilitate replacement and maintenance and realize online remote measurement. The testing results showed that, the measurement range of the sensor was 0-20 mg/L, system time delay was less than 2 s, and the life time of the oxygen sensitive membrane would be about 1 year. The dissolved oxygen sensor has the characteristics of convenient measurement, stable result output, low cost and small volume, which will lay a good foundation for the development and marketization of low-cost dissolved oxygen sensors in aquaculture industry of China.
Proper soil nutrients content plays an important role in agricultural production—undernutrition would reduce crop yield and quality and overnutrition would cause environmental pollution. Though the traditional approaches based on sampling and chemical analysis can comprehensively and accurately measure soil nutrients, but the soil sampling and pretreatment process are cumbersome, complicated, time-consuming, and costly. Therefore, rapid and accurate measurement of soil nutrients is of great significance for precise fertilizer application, which can increase yield, improve crop quality, and alleviate environmental pollution. Toward this objective, a rapid soil nutrients detection method based on modulated near infrared spectroscopy for active near-field telemetry was proposed, which could effectively minimize effect of sunlight during the measuring process. Eight channels narrow-band laser diodes with wavelengths of 1260, 1310, 1350, 1410, 1450, 1510, 1550 and 1610 nm were selected as active lighting sources for measuring the reflectance of soil samples. Eight channels narrow-band laser diodes were symmetrically placed on a concentric circle. A photodetector with a circular photosensitive area of 5 mm in diameter was placed at the center of the concentric circle to maximize the reception of laser beam reflected by soil. A focusing lens was placed in front of the photodetector to collect the laser beam reflected from the soil sample to increase the sensitivity. The sensing area of the photodetector was located at the focus of the lens. seventy four groups of soil samples with known N content were divided into training set (54 groups) and prediction set (20 groups) for data analysis. The spectral reflectance significantly correlated with soil N content was screened by analyzing the training set based on a general linear model and a quantitative measurement model with R2 of 0.97 between the screened spectral reflectance and soil N content was achieve. The predicted soil N content obtained from prediction set based on the established model and the referenced soil N content of the prediction set had a R2 of 0.9, indicating that this method has an ability to quickly, as well as accurately detect soil nutrients.
In order to solve problems such as the lack of real-time data in agricultural machinery management, the difficulty in real-time machine operation supervision and the asymmetry of machine service information, an intelligent remote management platform was developed in this research. Firstly, five design principles of a specialized remote agricultural machinery management system: specialization, standardization, cloud platform, modularity and openness were proposed. Based on these principles, a customizable general-purpose intelligent remote management system for agricultural machinery based on intelligent sensing technology, Internet of Things technology, positioning technology, remote sensing technology and geographic information system was designed. Practical modules, including agricultural machinery information-based and location-based services using WebGIS, real-time monitoring and management of machinery operation, basic information management of farmland, basic information management of crops in the field, dispatching management of machinery, subsidy management of machinery, order management of machinery operation were designed and implemented in the platform for users of government agencies, agricultural machinery corporations, machine operators, and farmers. Besides, some key technologies of the platform under the current technical background, including the calculation method of the working area with low-precision GNSS positioning receivers, the analysis of anomality data during the processing of GNSS positioning data, the machine scheduling algorithm development, the integration of sensors were focused， analyzed and implementd. The idea of building the machinery management platform with each individual field as the building block was developed. It can be predicted that the agricultural machinery operation management platform would gradually change from simple operation management to field-level comprehensive management. The research and development of this platform can not only solve current machinery management problems, but also provide basic functions for development of similar machinery management platforms.
Water and fertilizer integration technology can effectively improve nutrient utilization efficiency. However, the existing water and fertilizer machines have some shortcomings, such as huge cost, single fertilizer injection, need for cleaning water and so on, which hinder the development of water and fertilizer integration technology. Aiming at the problems of precise and low-cost compounding of compound fertilizer at the local farm, the water and fertilizer integrated intelligent irrigation and fertilization system were taken as the research object. In this research, new concept of an intelligent sensing system was proposed， and accurate proportioning system of water and fertilizer concentration was constructed and implemented. Firstly, a fast on-line method of intelligent sensing model of water and fertilizer was established based on a series of concentration gradient compound fertilizer solutions. The conductivity values of these formulated solutions were tested by contactless conductivity detection electrodes. Subsequently, the data analysis algorithms were discussed and compared to fit regression model. Based on the intelligent sensing model of water and fertilizer , the framework structure of in-situ intelligent sensing and accurate proportioning system of water and fertilizer concentration was designed, and the working principle of the system was also explained. The system proposed includs a first-level water and fertilizer concentration intelligent perception model building subsystem and a second-level water and fertilizer accurate proportioning subsystem. The first-level subsystem was designed as a portable device, which mainly included a precise pump for quantitative dosing, a large-range online conductivity sensor, a plastic bucket and supporting control and model building software. The second-level subsystem was designed as a dynamic and precise fertilizer distribution device. The effectiveness of the system was verified by three types of water intelligent fertilizer application so as to guide the in-situ water and fertilizer concentration ratio. The testing results showed that the second-order polynomial fitting curve under regularization conditions was the best model to express the relationship between the conductivity and the concentration of water and fertilizer, and the correlation coefficients R2 was higher than 0.999. Combined with the proportion of each index of compound fertilizer, the concentration of each index of compound fertilizer that the user cares about can be obtained according to this model. The results of three types of water intelligent fertilizer application showed that the conductivity of natural water had an effect on the water and fertilizer system, and the relative deviation was more than 0.1. The online water and fertilizer perception and ratio system proposed in this research realized the elimination of the interference of the local water conductivity on the accuracy of the ratio of water and fertilizer, and the accurate calculation of compound fertilizer was achieved through model calculation. This system has a simple structure and accurate ratio, low cost， and can be easily combined with the existing water and fertilizer integrated machine or artificial fertilizer system. The system could be widely used in facility agriculture, orchard cultivation and field cash crop cultivation, et al.
Light is the main energy source for plants to carry out photosynthesis, and the quality of light directly affects the yield and quality of crops. In view of the fact that most of the existing plant light supplement systems are based on the photosynthetic capacity of functional leaves, problems such as photoinhibition of new leaves in the canopy and lack of supplementary light in the functional leaf position between plants, and the position of light supplement can’t be adjusted dynamically to adapt to crop growth exist, taking facility cucumber as the research object, an stereo light environment intelligent control system based on the characteristic differences of plant light requirements was designed in this research. The system is composed of intelligent control subsystem, canopy-plant environment monitoring subsystem, canopy-plant LED light-compensating lamp subsystem, and light-compensating lamp lifting subsystem. Wireless communication between subsystems was realized by using ZigBee technology. The canopy-interplant environmental monitoring subsystem obtains the canopy and interplant environmental information respectively and sends them to the intelligent control subsystem. According to the real-time environmental information, the intelligent control subsystem invokes the canopy regulation model and the appropriate interplant leaf position regulation model to obtain the corresponding regulation target values, and sends them to the canopy-interplant light-compensating lamp to realize the dynamic real-time regulation of the canopy and interplant light-compensating lamp. In November 2018, the stereoscopic light-compensating equipment and the traditional canopy light-compensating equipment were tested and verified with the natural control in the vegetable base of the vegetable industry comprehensive service area of Jingyang County, Shaanxi province. The results showed that, compared with the traditional canopy light-compensating area, the cucumber plant height and stem diameter in the stereoscopic light-compensating area increased significantly, and the average plant height and stem diameter increased by 8.03% and 7.24%, respectively. Compared with the natural treatment area, the average plant height and stem diameter increased by 26.51% and 36.03%, respectively. And during the one-month picking period, compared with the traditional canopy light-compensating area, the yield of the stereoscopic light-compensating area increased by 0.28 kg/m2, the economic benefit increased by 2.82 CNY/m2, the yield of the stereoscopic light-compensating area increased by 1.39 kg/m2, and the economic benefit increased by 4.88 CNY/m2; compared with the natural treatment area indicating that the stereoscopic light environment control system can improve economic benefits and has good application and promotion values.
The pollination during bees’ foraging is vital to continue species on the earth. However, bee colonies in some areas of America and Europe frequently appeared colony collapse disorder in the past decade due to many possible factors such as climate change and pesticide usage, which has not received enough attention and positive response from human beings. In this research, bee colony’s activities were investigated with seven detectable features (i.e., weight, temperature, humidity, gas concentration, vibration, sound and entrance counts), and the applicability of the features was evaluated by considering four factors (i.e. the relevance to bee colony’s activities, the richness of information, the cheapness of cost and the simplicity of engineering). Based on the investigation and evaluation, an Internet of Things(IoT) based system was presented for long-time monitoring of bee colonies, which could hourly detect the temperature and humidity inside of hive, bee combs’ weight, bee colony’s sounds and bees’ counts of passing through hive entrance. In this system, each hive has an individual detection device for the monitoring of bee colony, and the colony information could be automatically collected and transferred to a remote cloud server which took responsible for the information storing. Finally, the users could freely visit the server to browse the history data and manage their bee colonies. Moreover, a 235 days continuous monitoring for Apis mellifera ligustica was performed from August, 2019 to April, 2020 to demonstrate the system performance, and long-time and one-day monitoring results were both analyzed. The monitoring results indicated that the system could continuously operate without human intervention, and the data could reveal bee colony’s activity and growth, e.g., the temperature and humidity could reflect the micro climate of the bee hive, the weight could show the forging and stock of food, the sounds contained lots of information about bees’ behavior and the entrance count was strongly related to the activeness and scale of bee colony. Compared with the reported monitoring system, this system is superior in the diversity of detected features, the capability of power self-support and the wireless of data transmission that can benefit to the system’s deployment in the field and long-term operation without maintenance. In the visible future, this system will effectively promote the study related to the biology of bee’s behavior, the reason of colony collapse disorder and the development of precision beekeeping.
With the development of information technology, using big data analysis, monitoring of Internet of Things, sensor perception, wireless communication and other technologies to build a real-time online monitoring system for beehive is a feasible solution for reducing the stress response of bee colony caused by check the beehive artificially. Focusing on situation that real-time monitoring in the closed environment of the beehive is difficult, the STM32F103VBT6 32-bit microcontroller, integrated with the temperature and humidity sensor, microphone, and laser beam sensor were used in this study to develop a low-power, continuous working online monitoring system for the multi-parameter information acquisition and monitoring of beehive key parameters. The system mainly includes core processing module, data acquisition module, data sending module and database server. The data collection module includes a temperature and humidity collection unit inside the beehive, a bee colony sound collection unit, a bee in and out nest number counting unit, etc., and transfers data by accessing the mobile communication network. The performance test results of system on-site deployment showed that the developed system could monitor the temperature and humidity in the beehive in real time, effectively distinguish the bees of entering and leaving the beehive, record the numbers of bees of entering and leaving the nest door, and the bee colony sounds that the automatically obtained were consistent with the standard sound distribution of bee colony. The results indicate that this system meets the design requirements, can accurately and reliably collect the beehive parameters data, and can be used as a data collection method for related research of bee colony.
Considering at the robotic management for tomato plants in the greenhouse, it is necessary to identify the stem, leaf and fruit with the similar color from the broad-band visible image. In order to highlight the difference between the target and background, and improve the identification efficiency, the multiple narrow-band image fusion method for identifying the tomato’s three similar-colored organs, including stem, leaf, and green fruit, was proposed, based on the spectral features of these organs. According to the 300-1000 nm spectral data of three organs, the regularized logistic regression model with Lasso for distinguishing their spectral characteristic was built. Based on the sparse solution of the model’s weight coefficients, the wavelengths 450, 600 and 900 nm with the maximum coefficients were determined as the optimal imaging band. The multi-spectral image capturing system was designed, which could output three images of optimal bands from the same view-field. The relationship between the organs’ image gray and their spectral feature was analyzed, and the optimal images could accurately show the organs’ reflection character at the various band. In order to obtain more significant distinctions, the weighted-fusion method based NSGA-II was proposed, which was supposed to combine the organ’s difference in the optimal band image. The algorithm’s objective function was defined to maximize the target-background difference and minimize the background-background difference. The coefficients obtained were adopted as the linear fusion factors for the optimal band images.Finally, the fusion method was evaluated based on intuitional and quantitative indexes, respectively considering the one among stem, leaf and green fruit as target, and the other two as the backgrounds. As the result showed, compared with the single optimal band image, the fused image greatly intensified the difference between the similar-colored target and background, and restrained the difference among the background. Specifically, the sum of absolute difference (SAD) was used to describe the grey value difference between the various organs, and the fusion result images’ SAD between the target and the background raised to 2.02, 8.63 and 7.89 times than the single band images. The Otsu automatic segmentation algorithm could respectively obtain the recognition accuracy of 71.14%, 60.32% and 98.32% for identifying the stem, leaf and fruit on the fusion result image. The research was supposed as a reference for the identification on similar-colored plant organs under agricultural condition.
Leaf wetness duration (LWD) is one of the important input variables of plant disease model, which is related to the infection of many leaf pathogens and affects the pathogen infection and developmental rate. In order to accurately predict the occurrence time and location of cucumber diseases in solar greenhouse, nine sampling points were set up in two different greenhouses located in Beijing in March and September 2019, according to the chessboard method to deploy temperature, humidity and light sensors. The fixed-point visual inspection method was used to collect the data every 1 h. From the leaf wetting to the leaf drying is the leaf wetness duration of a day. The relative humidity model (RHM) and back propagation neural network model (BPNN) were used to quantitatively estimate and analyze the LWD, the input layer of BPNN was temperature, humidity, radiation and location, the hidden layer was 10, and the output layer was location and whether the leaf surface was wet. The results showed that BPNN obtained similar accuracy ACC = 0.90 and 0.92 under the experimental conditions of two greenhouses, which was higher than RHM ACC = 0.82 and 0.84 in estimating of LWD, the mean absolute errors MAE were 1.81 h and 1.61 h, root mean squared error RMSE were 2.10 and 1.87, and coefficient of determination R2 were 0.87 and 0.85. In sunny and cloudy conditions, the spatial distribution of LWD was generally in the South > the Middle > the North. In the South, the average LWD was the longest, 12.17 h/d; from the east to the west, the spatial distribution of LWD was generally in the East > the West > the Middle. In the Middle, the average LWD was the shortest of 4.83h/d. The average LWD in rainy days was longer than that in sunny days and cloudy days, the average LWD in spring and autumn rainy days were 17.15 h/d and 17.41 h/d. These changes and differences had an important impact on the distribution of leaf wetness duration in the horizontal direction of cucumber population in greenhouse, which was closely related to the occurrence rule of most high humidity cucumber diseases. In this research, the method of regional analysis of the wet duration of cucumber leaves in greenhouse was proposed, which could provide a reference for simulating the spatial distribution of LWD in greenhouse, and also had a certain reference significance for the establishment of cucumber disease early warning system.
The production of China's aquaculture has changed from extensive model to intensive model, the production structure is continuously adjusting and upgrading, and the production level has been continuously improved. However, as an important part of China's agricultural production, aquaculture plays an important role in promoting the development of China's agricultural economy. Low labor productivity, production efficiency and resource utilization, low-quality aquatic products, and the lack of safety guarantees have severely limited the rapid development of China's aquaculture industry. Using modern information technology and intelligent devices to realize precise, automated, and intelligent aquaculture, improving fishery productivity and resource utilization is the main way to solve the above contradictions. Artificial intelligence technology in aquaculture is to use the computer technology to realize the production process of aquaculture, monitor the growth of underwater organisms, judge, discuss and analyze problems, and then perform feeding, disease treatment, and breeding. In order to understand the development status and technical characteristics of artificial intelligence technology in aquaculture, in this article, five main aspects of aquaculture, i.e., life information acquisition, aquatic product growth regulation and decision-making, fish disease prediction and diagnosis, aquaculture environment perception and regulation, and aquaculture underwater robots, combined with the practical problems in aquaculture, were mainly focused on. The application principles and necessity of artificial intelligence technology in each aspect were explained. Commonly used technical methods were point out and the classic application cases were deeply analyzed. The main problems, bottlenecks and challenges in the current development of artificial intelligence technology in aquaculture were analyzed, including turbid water, multiple interference factors, corrosion of equipment, and movement of underwater animals, etc., and reasonable research directions for these potential challenges were pointed out. In addition, the main strategic strategies to promote the transformation of aquaculture were also proposed. The development of aquaculture is inseparable from artificial intelligence technology, this review can provide references to accelerate the advancement of digitalization, precision and intelligent aquaculture.
Lameness in dairy cattle could cause significant economic losses to the dairy industry. Detection of lameness in a timely manner is critical to the high-quality development of dairy industry. The traditional method is visual locomotion scoring by dairy farmers, which is low efficiency, high cost and subjective. The demand for automated lameness detection is increasing. The review was conducted to find out the current state and challenges of automatic lameness detection technology development and to learn from the latest findings. The current automatic lameness detection systems were reviewed in this paper mainly rely on five technologies or combinations thereof, including machine vision, pressure distribution measuring system, wearable sensor system, behavior analysis and classification; the principle, function and features of these technologies were analyzed. Machine vision technique is to extract feature variables (e.g. back arch, head bob, abduction, stride length, walking speed, temperature, etc.) from video recordings of cattle movement by image processing. Pressure distribution measuring system contains an array of load cells to sense gait variables, when dairy cattle are walking by. By using accelerometer with high frequency data collection, the gait cycle parameters can be extracted and used for lameness detection. By using wearable devices, the number of lying/standing bouts and their duration, the total time spent lying, standing and ruminating per day can be recorded for individual cattle. The lameness can also be detected by behavior analysis. Currently, most of these studies were in the stage of sensor development or validation of algorithm. A few studies were in the stage of validation of performance and decision support with early warning system. The challenges to apply automatic lameness detection system in dairy farm includes the difficulties of acquiring high quality data of lameness features, lack of techniques to detect early lameness, identification errors caused by individual gait differences among dairy cattle, difficulties to function well in unstructured environment and difficulties to evaluate the benefits. To accelerate the development of automatic lameness detection systems, recommendations are proposed as follows: ①promoting lameness data sharing and data exchange among dairy farms; ②developing individual-based lameness classification model; ③developing multifunctional smart station which can detect lameness, measure body condition score, weighing, etc; ④evaluating the significance of automatic lameness detection to the dairy industry from the perspective of animal welfare, environment and food safety.
Environmental quality of facilities affects the healthy growth and production of laying hens. The breeding environment of laying hens is a complex and non-linear system in which multiple environmental factors interact and restrict each other. It is difficult to make an accurate and effective evaluation on the suitability of laying hens with a single breeding environment parameter. In order to solve the above problem, an improved cuckoo search algorithm optimization neural network (CS-BP) model for the evaluation and prediction of the environmental suitability of laying hen facility was proposed in this research. In this model, the effects of environmental factors such as temperature, humidity, light intensity and ammonia concentration were comprehensively analyzed, and the problem of high prediction accuracy caused by BP neural network easily falling into local minimum value was overcome. In the experiment, the model was compared with BP neural network, genetic algorithm optimized BP neural network (GA-BP) and particle swarm optimization BP neural network (PSO-BP). The results showed that the mean absolute error (MAE), mean relative error (MAPE) and the coefficient of determination (R2) of the prediction model based on the improved CS-BP were 0.0865, 0.0159 and 0.8569, respectively. The prediction model based on the improved CS-BP had a strong generalization ability and a high testing precision, and its index performance was better than the above three comparison models. The classification accuracy of the improved CS-BP model was tested, and the result was 0.9333. The model constructed in this research can provide more comprehensive and effective scientific evaluation for the environmental quality of laying hens facility, which is of great significance to realize the optimal control of the production environment and promote the production performance of layers.
China's facility horticulture has developed rapidly in the past 30 years and now comes to the first in the world in terms of area. However, the number of farmers is decreasing. "Machine replaces labor" has become the current research hotspot. In order to realize the fine collection of crop images and environmental monitoring data, a three-dimensional environmental robot monitoring system for crops was designed. The robot consists of three parts: perception center, decision center and execution center, which carry out environmental perception from machine perspective, data analysis, decision instruction generation and action execution respectively. In perception layer, the system realized real-time videos, images, data monitoring such as air temperature, air humidity, illumination intensity and concentrations of carbon dioxide in grid scale from multi-angle with high accuracy. At the system level, automatic speech recognition was integrated to make the system easier to use, especially for farmers who usually work in the fields. In transport layer, monitoring data and control instructions were converged to local data center through wireless bridges. Concretely, transmission mode was chosen according to different characteristics of data, wire transmission is available for big size data, such as images and videos, while wireless transmission is mainly applied to small size data, such as environmental monitoring parameters. In data processing layer, feedbacks and control instructions were made by multi-source heterogeneous data of crop model analysis, in terms of commands, independent inspection mode and real-time remote-control mode were available for users. Plant type, user information, historical data and management data were taken into consideration. Finally, in application layer, the system provided web and mobile intelligence services that could be used for the whole growth periods in terms of images, real-time videos, monitoring data collection and analysis of cucumbers, tomatoes, greenhouse peaches, etc. The system has been demonstrated and applied in solar greenhouse No. 7 of Beijing Xiaotangshan National Precision Agriculture Base and No. 5 of Shijiazhuang Agricultural and Forestry Science Research Institute with good achievements. Farmers and researchers have realized real-time monitoring, remote control and management. On one hand, the system can used to avoid working in extreme environment, such as high temperature and pesticide environment. On the other hand, with the help of the robot, independent inspection and data collection could achieve instead of people, and it is very intuitive in time-saving and indirect costs saving for productions and researchers. The results showed that the system could be widely applied in greenhouse facilities production and research.
Volunteer corn in soybean fields are harmful as they disrupt the benefits of corn-soybean rotation. Volunteer corn does not only reduce soybean yield by competing for water, nutrition and sunlight, but also interferes with pest control (e.g., corn rootworm). It is therefore critical to monitor the volunteer corn in soybean at the crop seedling stage for better management. The current visual monitoring method is subjective and inefficient. Technology progress in sensing and automation provides a potential solution towards the automatic detection of volunteer corn from soybean. In this study, corn and soybean were planted in pots in greenhouse to mimic field conditions. Color images were collected by using a low-cost Intel RealSense camera for five successive days after the germination. Individual crops from images were manually cropped and subjected to image segmentation based on color threshold coupled with noise removal to create a dataset. Shape (i.e., area, aspect ratio, rectangularity, circularity, and eccentricity), color (i.e., R, G, B, H, S, V, L, a, b, Y, Cb, and Cr) and texture (coarseness, contrast, linelikeness, and directionality) features of individual crops were extracted. Individual feature's weights were ranked with the top 12 relevant features selected for this study. The 12 features were fed into three feature-based machine learning algorithms: support vector machine (SVM), neural network (NN) and random forest (RF) for model training. Prediction precision values on the test dataset for SVM, NN and RF were 85.3%, 81.6%, and 82.0%, respectively. The dataset (without feature extraction) was fed into two deep learning algorithms—GoogLeNet and VGG-16, resulting into 96.0% and 96.2% accuracies, respectively. The more satisfactory models from feature-based machine learning and deep learning were compared. VGG-16 was recommended for the purpose of distinguishing volunteer corn from soybean due to its higher detection accuracy, as well as smaller standard deviation (STD). This research demonstrated RGB images, coupled with VGG-16 algorithm could be used as a novel, reliable (accuracy >96%), and simple tool to detect volunteer corn from soybean. The research outcome helps provide critical information for farmers, agronomists, and plant scientists in monitoring volunteer corn infestation conditions in soybean for better decision making and management.
Invasive pest fall armyworm (FAW) Spodoptera frugiperda is one of the serious threats to the food safety. Early warning and control plays a key role in FAW management. Nowadays, deep learning technology has been applied to recognize the image of FAW. However, there is a serious lack of training dataset in the current researches, which may mislead the model to learn features unrelated to the key visual characteristics (ring pattern, reniform pattern, etc.) of FAW adults and its close related species. Therefore, this research established a database of 10,177 images belonging to 7 species of noctuid adults, including FAW and 6 FAW close related species. Based on the small-scale dataset, transfer learning was used to build the recognition model of FAW adults by employing three deep learning models (VGG-16, ResNet-50 and DenseNet-121) pretrained on ImageNet. All of the models got more than 98% recognition accuracy on the same testing dataset. Moreover, by using feature visualization techniques, this research visualized the features learned by deep learning models and compared them to the related key visual characteristics recognized by human experts. The results showed that there was a high consistency between the two counterparts, i.e., the average feature recognition rate of ResNet-50 and DenseNet-121 was around 85%, which further demonstrated that it was possible to use the deep learning technology for the real-time monitoring of FAW adults. In addition, this study also found that the learning abilities of key visual characteristics among different models were different even though they have similar recognition accuracy. Herein, we suggest that when evaluating the model capacity, we should not only focus on the recognition rate, the ability of learning individual visual characteristics should be allocated importance for evaluating the model performance. For those important taxonomical traits, if the visualization results indicated that the model didn't learnt them, we should then modify our datasets or adjusting the training strategies to increase the learning ability. In conclusion, this study verified that visualizing the features learnt by the model is a good way to evaluate the learning ability of deep learning models, and to provide a possible way for other researchers in the field who want to understand the features learnt by deep learning models.
Research on tomato disease recognition based on leaf images has been widely concerned in recent years, and with the development of machine learning and deep learning, researchers from various countries have proposed a variety of methods and models to solve this problem. In this research, a new approach by fusion color and texture features, and kernal mutual subspace method (KMSM) were introduced and a rapid recognition model of tomato leaf disease was established. The color and texture features introduced in this research including color moment (CM), color coherence vector (CCV) and histogram of oriented gradient (HOG) features. The CCHKMSM (CM+CCV+HOG+KMSM) model firstly mapped the extracted color and texture features from different classes of leaf disease data sets to high-dimensional space using gauss kernal function. Then the principal component of the mapped high-dimensional space was analysed, and the nonlinear disease characteristic space was generated. Finally, the diseases based on the minimum cosine angle of nonlinear feature space were identified. Validation experiment was conducted based on public agricultural disease data sets of PlantVillage, which providing 9 kinds of most commonly tomato leaf disease and 1 kind of healthy leaf image, and the filed took image include 3 kinds of tomato leaf diseases images. For experiment based on PlantVillage data set, the results showed that the CCHKMSM realized the most high recognize accuracy rate of 100% when the number of each class was 350. The training time cost and recognition time cost was 0.1540 s and 0.013 s, respectively. Meanwhile, experiments were conducted in the range of sample image numbers from 150 to 1000 images for each class, with step length of 50, and the obtained results showed that the average recognition rate was 99.14%. For experiment based on field took tomato diseases data sets, after segment original image into sub-size image, average recognize accuracy for the kind of diseases arrived at 96.21%, which was higher than other typical machine learning models such as SVM and KMSM, and at the same level by comparing with deep learning-based recognition methods. On the other hand, as an significant adventure of the proposed CCHKMSM model, the computing cost was low, both the training time and testing time were much lower than deep learning methods, and requirement is loss the system to run. As a conclusion, the proposed CCHKMSM model, has high potential to be applied in low-configuration equipment such as hand-held devices and edge computing terminals.
Agricultural wireless sensor networks plays a key role in obtaining multi-source heterogeneous big data of farmland soil, environment and crop growth. The increasing network scale brings challenges to the application of agricultural Internet of Things. In order to solve the problem that sensors are not uniformly distributed in farmland and constrained by energy, a collaborative optimization hybrid multi hop routing algorithm, particle simulated multipath routing (PSMR) based on particle swarm optimization and simulated annealing was proposed. Firstly, cluster heads were selected by node residual energy and node degree weighting, and cluster structure was used to realize efficient dynamic networking of heterogeneous networks. Then, the multi hop data structure between cluster heads was used to solve the problem of high energy consumption in long-distance transmission of cluster heads. Particle swarm optimization and simulated annealing were used to improve the convergence speed, and sink nodes could accelerate the collection of aggregated data in cluster heads. The simulation results showed that compared with the energy-efficient load balancing multipath routing scheme (EMR), the network lifetime of PSMR algorithm was increased by 57%. EMR selected the data transmission link with low energy consumption and small delay by calculating the weight of total link hops and transmission energy consumption. Compared with greedy perimeter stateless routing-algorithm (GPSR-A algorithm), which could ensure the shortest data transmission distance and lower network transmission delay, the first dead sensor node was delayed for two rounds in the same network life cycle, and the residual energy standard deviation was reduced by 0.04 J, which had good network energy consumption balance. PSMR algorithm could reduce the extra energy consumption of remote cluster heads by multi hop between cluster heads, and improved the energy balance performance of cluster heads with different distances. It can provide technical basis for long-term, efficient and stable data acquisition and monitoring of large-scale farmland complex environment, and improve the resource utilization efficiency of agricultural Internet of Things.
Vegetables are an important part of residents' diet. The abnormal fluctuation of vegetable prices has caused losses to the economic interests of vegetable farmers and also affected the daily diet and quality of life of residents. However, there are some difficulties in vegetable price prediction, such as large price fluctuation and complicated influencing factors. Cucumber is the main category of vegetables and a common food on the daily table of residents and its recent price fluctuations have aroused widespread concern. In this research, taking cucumber as the research object, a combination model (L-BPNN) combining Lasso regression method and BP neural network was constructed to forecast the short-term price of cucumber. Firstly, the factors affecting the price of cucumber, such as supply, demand and circulation were analyzed. Then the price fluctuation characteristics of cucumber in China from 2010 to 2018 were analyzed and 24 factors were selected as the influencing factors of cucumber price. In the case of complex factors, Lasso regression was used to compress the 24 input influencing factors and the 12 remaining influencing factors with large correlation degree after compression were used as the input influencing factors of BP neural network. Among the 12 related factors , the positive effects included: land cost, per capita disposable income of urban residents, urban vegetable consumption price index, fuel surcharge, booth fee, packaging and processing fee, inflation rate, affected area and temperature deviation from normal value; negative effects included sown area, industrial support amount and average temperature. On this basis, a combination model combining Lasso regression method with BP neural network (L-BPNN) was constructed to forecast the short-term price of cucumber. The neural network was used to train and adjust the model between the input influencing factors and the output price. Compared with the regression analysis and intelligent analysis methods, the results show that the average relative error of L-BPNN combination model was the smallest, only 0.66%, which was 64.52%, 82.11% and 86.2% lower than Lasso regression model, BP neural network model and RBF neural network model respectively, and had higher prediction accuracy. The results of this study realizes the short-term price forecast of cucumber, and can also be extended to other vegetable varieties, which is of great significance for guaranteeing the income of vegetable farmers and stabilizing the market price of vegetables.
Maize is an important food crop in China. In order to quickly and non-destructively estimate summer maize leaf area index (LAI) under different water stress conditions, in this study, maize samples with multiple irrigation treatments throughout the growth period were used for modeling analysis. Then, based on the unmanned aerial vehicle (UAV) multi-spectral remote sensing technology, combined with the summer maize LAI collected in the field during the same period, five kinds of vegetation indices, including the normalized difference vegetation index (NDVI), soil adjusted vegetation index (SAVI), enhanced vegetation index (EVI), green normalized difference vegetation index (GNDVI) and visible atmospherically resistant index (VARI) were selected in this research as model input parameters, and random forest regression algorithm was used to establish the relationship between the field maize canopy vegetation indices and LAI under different irrigation conditions during the entire growth period. The accuracies of the model were compared with that of the model established by the university linear regression and multiple linear regression algorithms. The results showed that under sufficient irrigation condition, the vegetation index using multiple linear regression model could well (R2 = 0.83, RMSE = 0.05) estimate LAI; under water stress conditions, the vegetation index using random forest regression model could well estimate LAI (R2 = 0.74~0.87, RMSE = 0.02~0.10), water stress factors had little effect on the random forest regression model, and NDVI and VARI contributed the LAI estimation model better. The spatial distribution map of LAI was generated based on the random forest regression algorithm. The above results showed that it was feasible to use the random forest regression algorithm to estimate the summer maize LAI under various irrigation conditions based on the UAV multi-spectral remote sensing technology. The results indicates that the model established has a good applicability. This research can provide technical and method support for the rapid and accurate monitoring of field summer maize LAI under different irrigation conditions during the entire growth period.
Soil organic matter (SOM) is an important source of crop growth, its content can reflect soil fertility status. In order to realize the fast and real-time estimation of the SOM, based on hyperspectral data, a rapid estimation model of SOM content in orchards was established. A total of 100 brown soil samples were collected from the apple orchard of Qixia county, Yantai city, Shandong province. After drying and grinding, the hyper-spectrum of the soil was measured in the laboratory using ASD FieldSpec. The spectral data was preprocessed by the method of moving average, and the spectral reflectance features of orchard soil were analyzed to study the correlation between spectral reflectance and its soil organic matter content. In order to enhance the correlation between relevant spectral parameters and soil indexes, the original data were processed by using the multivariate scattering correction, the first derivative and the first derivative of MSC. After the sensitive wavelengths of soil organic matter content were selected and the spectral indexes were constructed. Multiple linear regression models (MLR), support vector machines (SVM) and random forest (RF) models were respectively established. The estimation accuracy of the orchard soil organic matter estimation model was measured by the determination coefficient (R2), root mean square error (RMSE) and relative analysis error (RPD). The sensitive wavelengths of soil organic matter content selected were 678, 709, 1931, 1939, 1996 and 2201 nm. The spectral parameters were constructed using the selected wavelengths, which were NDSI(678, 709), NDSI(678, 1931), NDSI(678, 2201), NDSI(709, 1939), and NDSI(1939, 2201). These models established include MLR, SVM and RF model. The RF model had the best precision. The calibration sample R2 was 0.8804, the RMSE was 0.1423 and RPD reached 2.25; the R2 of the verification model was 0.7466, the RMSE was 0.1266, and the RPD was 1.79. The results showed that the fitting effect of the hyperspectral inversion model based on RF regression analysis was better than that based on MLR analysis and SVM regression analysis. As a promising and effective method, RF can play a vital role in predicting soil organic matter. The results can help understanding the distribution of soil nutrients, guiding farmers to apply fertilizer reasonably and improving the efficiency of orchard production and management.
In recent years, revenue protection crop insurance is an innovative insurance that has been prioritized in China. But it still lacks the support of the third-party yield data around crop harvest time. Aiming to provide objective yield data for revenue protection crop insurance, satellite remote sensing production estimation technology was employed to discuss its application mode and applicability. Taking the soybean revenue protection insurance in Jiaxiang county, Shandong province as an example, we first extracted soybean planting plots, calculated vegetation index and crop physiological parameters based on Sentinel-2 satellite images in 2018 . Combining to TRMM precipitation data from TRMM precipitation-monitoring radar satellite and MODIS land surface temperature data from Terra/Aqua satellite and site yield data, we established a multi-parameter linear regression model, and estimated soybean yield per unit area. The crop extraction results showed that the soybean planting area in the study area was 1.24 km2, which was in good agreement with the 1.27 km2 reported by the local agricultural bureau; and with using the actual measurement plots, the remote sensing identification accuracy of the planting distribution plots reached 90%. The yield estimation results showed that the NDVI of the soybean pod stage on August 23 and the leaf area index of the soybean seedling stage on September 7 explained the soybean yield per hectare the best, and the average estimated yield of the whole area was 244,500 kg/m2, which reflects the severely affected agricultural conditions, comparing to 299,800 kg/km2 in previous years.The regression coefficient between the estimated yield data and the measured data reached 0.92, which meet the application needs.With this results, the estimated yield of different towns can be summarized, and the regional yield was present, and was used as the real yield in 2018, multiplying with the average soybean price around October 11 to December 10 from the local price bureau, the real revenue was obtained. Compared the real revenue to the expected revenue in the contract of insurance, the claims work was decided. The results indicated that the Sentinel-2 satellite data could be used to identify the soybean planting distribution in the study area accurately, and to complete the yield estimation as soon as one week after the soybean harvest, which could guide the insurance company's claims work. The whole methodology is capable of aiding the claims work in revenue protection crop insurance.
