Welcome to Smart Agriculture 中文

Table of Content

    30 July 2019, Volume 1 Issue 3
    Overview Article
    Information sensing and environment control of precision facility livestock and poultry farming | Open Access
    Guanghui Teng
    2019, 1(3):  1-12.  doi:10.12133/j.smartag.2019.1.3.201905-SA006
    Asbtract ( 2138 )   HTML ( 2526)   PDF (2880KB) ( 1864 )  
    Figures and Tables | References | Related Articles | Metrics

    The fine breeding of livestock and poultry facilities is the frontier of the development of modern animal husbandry. The core of the fine breeding of livestock and poultry facilities lies in the deep integration of the "Internet of Things+" with traditional farming facilities. In recent years, with the withdrawal of more and more individual family-based breeding models, the management methods of livestock and poultry farms in China have gradually moved towards intensification, large scale,and automated facilitation. The traditional family-style livestock and poultry management experience is falling behind and gradually withdrawing from the historical stage. The refined farming of livestock and poultry facilities based on the individual animal management and quality assurance of farmed animals and animal welfare requirements have become the latest development trend of livestock and poultry farming industry. The rapid development of digital and network technology will provide new opportunities for the organic combination of animal husbandry production, animal welfare, information management and sustainable utilization of natural resources. Economic benefit, animal health and welfare, refinement of production process management and product quality are three key factors that affect the sustainable development of animal husbandry. In this paper, based on expounding the importance of the information sensing and the environmental regulation and control of the fine breeding livestock and poultry facilities, a cutting-edge technology of the information sensing and the environmental regulation and control of the livestock and poultry facilities was introduced; problems and challenges to be faced with were analyzed; and it was concluded that the smart sensor technology would become the base driving force for progress of livestock and fine poultry breeding facilities, taking account of the welfare of livestock and animal performance of animal anthropomorphizing intelligent control technology and strategy is facing significant challenges. In the field of pig farming, the core direction is mechanized production mode, which is light simplification, feed hygiene and animal health. In the field of cattle farming, the main direction is the automation of the whole chain of forage and the safety of its enclosure facilities. In the field of milking technology, the frontier of technological innovation is to further improve milking efficiency and quality, milking process, low disturbance milk metering, and cow individual milk production prediction. In the field of poultry production, similar to cattle farming, more attention is paid to the improvement of engineering processes such as bedding, environment and drinking water. Finally the paper put forward suggestions on how to implement the key technologies of fine farming of livestock and poultry facilities in China, with purpose of providing theoretical reference and technical support for the transformation, upgrading sustainable development of livestock and poultry breeding industry.

    Research and prospect of solar insecticidal lamps Internet of Things | Open Access
    Kailiang Li, Lei Shu, Kai Huang, Yuanhao Sun, Fan Yang, Yu Zhang, Zhiqiang Huo, Yanfei Wang, Xinyi Wang, Qiaoling Lu, Yacheng Zhang
    2019, 1(3):  13-28.  doi:10.12133/j.smartag.2019.1.3.201905-SA001
    Asbtract ( 1938 )   HTML ( 728)   PDF (3195KB) ( 1548 )  
    Figures and Tables | References | Related Articles | Metrics

    Along with the increasing awareness of environmental protection and growing demand for green and pollution-free agricultural products, it has a great need to explore new ways to apply greener pest control methods in agricultural production. Researching on Solar Insecticidal Lamps (SILs) has continuously received incremental attentions from both the academia and industry, which brings a new mode for the preventing and controlling of agricultural migratory pests with phototaxis feature, and now is becoming to a hot research topic. Towards the fast development of "precision agriculture" and "smart agriculture" as well as the increasing demands for agricultural informatization, Wireless Sensor Networks (WSNs) have been widely used for agricultural information collection and intelligent control of agricultural equipment. WSNs are suitable for large-scale deployment and regional monitoring, which can be easily combined with SIL nodes. Based on the combination, a new type of agricultural Internet of things - Solar Insecticidal Lamps Internet of Things (SIL-IoTs) was proposed and the technology of WSNs for the prevention and control of phototactic migratory pests in agricultural applications were surveyed. Firstly, the state-of-art insecticidal lamps applications was reviewed and their characteristics deployment manners and working lifetime in the production of crops (e.g., forest, fruits, rice, vegetables) were summarized. Secondly, the characteristics of existing GSM/3G/4G-enabled SIL nodes and their latest research status on SIL-IoTs were summarized. Furthermore, the research status was analyzed concerning the energy harvesting mode and deployment characteristics of SIL, which are solar energy SIL harvesting mode for energy saving and the heuristic mode for node deployment, respectively. Finally, towards the fast-developed vision of smart agriculture, in which various emerging IT and automation technologies are maturely applied, SIL-IoTs can be considered as a new and important component to contribute to the green agricultural pest monitoring and control. To further enhance SIL-IoTs' capability and enrich SIL-IoTs' function, four open research issues on SIL-IoTs were proposed, i.e., 1) optimized deployment scheme of SIL-IoTs with multiple constrains, 2) optimized and adaptive energy management strategy for ensuring normal working hours of SIL node, 3) lack of algorithms for pests outbreak area localization, and 4) interference on data transmission because of dense high voltage discharge during severe pest disaster. To sum up, SIL-IoTs is one of the representative applications of "precision agriculture" and "smart agriculture" based on WSNs, which is a new model on prevention and control of pests. The combination of both optimized deployment algorithms of SIL-IoTs nodes and artificial intelligence techniques will provide a theoretical basis for SIL-based applications in terms of optimized deployment and energy management. Intelligent pest information collection, alarm, and node' senergy management via SIL-IoTs will facilitate decisions-makings for precise agricultural applications in prevention and control of pests.

    Research progress and prospect on non-destructive detection and quality grading technology of apple | Open Access
    Yudong Cao, Weiyan Qi, Xian Li, Zhemin Li
    2019, 1(3):  29-45.  doi:10.12133/j.smartag.2019.1.3.201906-SA011
    Asbtract ( 1471 )   HTML ( 1043)   PDF (1394KB) ( 1576 )  
    Figures and Tables | References | Related Articles | Metrics

    China has high total apple output, but the export volume is low. The high-end apple market is mostly occupied by imported apples. The main reason of this situation is the lack of technologies and equipments for fruit quality classification, and the degree of automation after picking stands low. The apples enter the consumer market without simple roughing processing, and the quality of the apple is unstable, which greatly reduces its market competitiveness. In this paper, the status quo of non-destructive detection and grading technology of apple quality was analyzed, then the development was forecasted. Apple non-destructive detection technology mainly includes spectrum, electrical characteristics, CT, chromatography, electronic nose and computer vision technology. According to the functional characteristics, advantages and disadvantages of various technologies, it is proposed to develop apple odor detection method based on new sensor technology; adopting multi-feature grading method based on machine vision, the combination of apple quality non-destructive testing technology and grading technology can promote the improvement of apple's industrial competitiveness. Overall, the needs of apple quality non-destructive detection and grading technology development in China are urgent. Detections with new technologies such as nanotechnology, biotechnology and artificial intelligence methods of sensor technology and products in apple non-destructive, quality grading detection and multi-technology have great potential. A real-time, efficient, high-precision grading systems in apple quality which integrates electricity, light, gas and computer vision may be an important development direction for improving apple's quality and enhancing the competitiveness of the apple industry.

    Information Perception and Acquisition
    Method for identifying crop disease based on CNN and transfer learning | Open Access
    Miao Li, Jingxian Wang, Hualong Li, Zelin Hu, XuanJiang Yang, Xiaoping Huang, Weihui Zeng, Jian Zhang, Sisi Fang
    2019, 1(3):  46-55.  doi:10.12133/j.smartag.2019.1.3.201903-SA005
    Asbtract ( 2017 )   HTML ( 2234)   PDF (2845KB) ( 1723 )  
    Figures and Tables | References | Related Articles | Metrics

    The internet is a huge resource base and a rich knowledge base. Aiming at the problem of small agricultural samples, the utilization technology of network resources was studied in the research, which would provide an idea and method for the research and application of crop disease identification and diagnosis. The knowledge transfer and deep learning methods to carry out research and experiments on public data sets (ImageNet, PlantVillage) and laboratory small sample disease data (AES-IMAGE) were introduced: first the batch normalization algorithm was applied to the AlexNet and VGG of Convolutional Neural Network (CNN) models to improve the over-fitting problem of the network; second the transfer learning strategy using parameter fine-tuning: The PlantVillage large-scale plant disease dataset was used to obtain the pre-trained model. On the improved network (AlexNet, VGG model), the pre-trained model was adjusted by our small sample dataset AES-IMAGE to obtain the disease identification model of cucumber and rice; third the transfer learning strategy was used for the bottleneck feature extraction: using the ImageNet big dataset to obtain the network parameters, CNN model (Inception-v3 and Mobilenet) was used as feature extractor to extract disease features. This method requires only a quick identification of the disease on the CPU and does not require a lot of training time, which can quickly complete the process of disease identification on the CPU. The experimental results show that: first in the transfer learning strategy of parameter fine-tuning: the highest accuracy rate was 98.33%, by using the VGG network parameter fine-tuning strategy; second in the transfer learning strategy of bottleneck feature extraction, using the Mobilenet model for bottleneck layer feature extraction and identification could obtain 96.8% validation accuracy. The results indicate that the combination of CNN and transfer learning is effective for the identification of small sample crop diseases.

    Information Processing and Decision Making
    Research on the application of multi-source agricultural land spatial data for "Two-Zone" demarcation | Open Access
    Jiong You, Zhiyuan Pei, Fei Wang
    2019, 1(3):  56-66.  doi:10.12133/j.smartag.2019.1.3.201906-SA005
    Asbtract ( 688 )   HTML ( 88)   PDF (5259KB) ( 543 )  
    Figures and Tables | References | Related Articles | Metrics

    The basis and premise of developing intelligent agriculture is digitization, especially digitization of agricultural land resources utilization, agricultural land ownership, agricultural production and other agricultural elements. At present, China's agriculture digitization is at a low level, the spatial information of agricultural land resources are applied few. It is necessary to accelerate the application of the big data with respect to agricultural land for agricultural production information collecting and agricultural policy implementing to promote the development of China's intelligent agriculture. One typical case is "food production function zone" and "important agricultural product production protection zone"(Two-Zone) demarcation. In order to realize the study on the technologies of "Two-Zone" demarcation, in this research, the following work was conducted . Firstly, the basic concept of the multi-source spatial data with respect to agricultural land was elaborated, and the existing multi-source spatial data with respect to agricultural land were summarized into four categories. Then, the workflows of "Two-Zone" demarcation was summed up. Considering the topological requirements of digital mapping for "Two-Zone" and the business requirements of intelligent management for agricultural production in "Two-Zone", a three-level spatial structure of "zone-patch-plot" was designed for "Two-Zone" demarcation. The key technology of digital mapping was proposed, based on the analysis of the functions of "Two-Zone" and the "zone-patch-plot" spatial structure, which integrate existing multi-source farmland spatial data depending on the relevance of spatial distribution and semantic attributes and then realize "Two-Zone"'s spatial distribution map at a specific spatial scale. The key technology of establishing the database for "Two-Zone" demarcation was also proposed, which realizes the abstraction of the geographical space entity delimited by "Two-Zone" from the perspective of spatial information structure. Therefore, the key technologies of "Two-Zone" demarcation based on multi-source agricultural land spatial data was scientifically designed, through the key work such as data acquisition, digital mapping and database construction. Finally, the key scientific problems in the technical links were extracted, which shown that "Two-Zone" demarcation requires a comprehensive consideration of data sources and user requirements, and it is necessary to analyze the availability of data with respect to multi-source agricultural land, decreasing the influence derived from the bias and partial loss of data with respect to multi-source agricultural land. Further consideration about statuses of the land use and crop planting in the farmland of "Two-Zone" is also needed. This study will provide a basic support for the intelligent management of agricultural land resources in the "Two-Zone".

    An improved method for estimating dissolved oxygen in crab ponds based on Long Short-Term Memory | Open Access
    Nanyang Zhu, Hao Wu, Daheng Yin, Zhiqiang Wang, Yongnian Jiang, Ya Guo
    2019, 1(3):  67-76.  doi:10.12133/j.smartag.2019.1.3.201905-SA004
    Asbtract ( 1027 )   HTML ( 133)   PDF (2818KB) ( 830 )  
    Figures and Tables | References | Related Articles | Metrics

    Dissolved oxygen (DO) is vital to aquaculture industry and affects the yield of aquaculture. Low DO in water can lead to death of crabs, therefore, it is important to measure DO accurately. However, the DO sensors are usually expensive and often lost function due to corrosion in water environmental and adsorption of different materials on their surface, which result in the inaccuracy in measured DO values. It is thus important to develop effective methods to estimate DO concentrations by using other environmental variables, which may reduce farmers' cost because DO sensors are not used. In this research, the collected environmental data, including temperature, pH, ammonia nitrogen, turbidity, were used to estimate DO concentrations in crab ponds. The data were preprocessed to eliminate missing values and outlier. Correlation analysis was applied to determine the relationship between environmental variables (temperature, pH, ammonia nitrogen, turbidity) and DO to show the rationale of using these four variables to forecast DO concentration. Principal component analysis was used to reduce the dimension of environmental data to reduce computation cost. For DO concentration estimation, it is more important to make the estimation of DO concentration at low values more accurate because DO concentration at low values is dangerous to crabs. This implies that estimation of DO concentrations at low or high values should be treated differently and applied different rates. Based on the Long Short-Term Memory (LSTM), a low DO concentration estimation model of Low Dissolved Oxygen Long Short-Term Memory(LDO-LSTM), which can improve the estimation accuracy of low DO values was proposed by optimizing the loss function of LSTM back propagation. The loss function of LDO-LSTM was based on the Mean Absolute Percentage Error (MAPE). According to the trend of DO, the true DO and the estimated DO values were applied weight functions. The Root Mean Square Error (RMSE) and the MAPE were used to evaluate the performance of LDO-LSTM and LSTM in DO estimation. Experimental results show that the value of RMSE and MAPE were stable at about 0.1 for LSTM and LDO-LSTM in forecasting DO when dissolved oxygen was higher than 6mg/L and the value of RMSE and MAPE of LDO-LSTM were lower than LSTM by 0.25 and 0.139. The results prove that the proposed method can not only provide desirable estimation accuracy for DO concentrations at high values but also make the estimated DO concentrations at low values more accurate. This research is expected very useful in reducing aquaculture costs and improving accuracy in forecasting DO especially at low values.

    Rapid detection of citrus Huanglongbing using Raman spectroscopy and Auto-fluorescence spectroscopy | Open Access
    Fen Dai, Zeyuan Qiu, Qian Qiu, Chujian Liu, Guozeng Huang, Yalin Huang, Xiaoling Deng
    2019, 1(3):  77-86.  doi:10.12133/j.smartag.2019.1.3.201812-SA026
    Asbtract ( 1057 )   HTML ( 75)   PDF (4597KB) ( 629 )  
    Figures and Tables | References | Related Articles | Metrics

    In order to detect citrus Huanglongbing (HLB, also named citrus greening) quickly, Auto-fluorescence and Raman spectra of HLB leaf samples and healthy ones were collected and analyzed. PLS-DA models based on Auto-fluorescence spectra, Raman spectra and mixed spectra were established and compared respectively. Finally, ROC curves of the three models were drawn, and the performance of the models were further evaluated by using the area under curve AUC parameters. The results demonstrated spectral differences between Huanglongbing samples and healthy ones could be seen. With 785 nm laser irradiation, citrus leaf samples produced strong Auto-fluorescence and Raman peaks. The Auto-fluorescence of HLB leaves was weaker than that of healthy samples in the range of 800-1203 cm -1, but stronger in the range of 1206-1800 cm -1, and the slope of decline (absolute value) was smaller than that of healthy samples. The similar shapes were found in the Raman spectra of typical HLB samples and healthy ones. But the HLB samples had larger Raman peak intensity and spectral bandwidth at 1257 cm -1, 1396 cm -1, 1446 cm -1, 1601 cm -1 and 1622 cm -1 than healthy ones. The Raman peak intensity of HLB samples was weaker than that of healthy samples at 1006 cm -1, 1160 cm -1, 1191 cm -1 and 1529 cm -1 positions, suggesting that the carotenoid content of HLB samples was lower than healthy ones. The Auto-fluorescence model, the Raman spectral model and the mixed spectral model could distinguish two kinds of samples with the accuracy of 86.08%, 98.17% and 94.75%, respectively. Furthermore, AUCs of Receiver Operating Characteristic Curve (ROC) were calculated. The AUCs for the Auto-fluorescence model, the Raman spectral model and the mixed spectral model were0.9313、0.9991 and 0.9875, respectively. Through further analysis of ROC curve, the identification effect of the Raman spectral model was optimal. Raman spectroscopy could be a new way to explore the rapid diagnosis of citrus HLB.

    Intelligent Management and Control
    Design and implementation of intelligent terminal service system for greenhouse vegetables based on cloud service:A case study of Heilongjiang province | Open Access
    Haifeng Zhang, Yang Li, Yu Zhang, Lijuan Song, Lixin Tang, Hongwen Bi
    2019, 1(3):  87-99.  doi:10.12133/j.smartag.2019.1.3.201906-SA002
    Asbtract ( 1295 )   HTML ( 1371)   PDF (5851KB) ( 1128 )  
    Figures and Tables | References | Related Articles | Metrics

    The greenhouse vegetable industry play an important strategic role in the adjustment of agricultural transformation mode and the reform of supply side in Heilongjiang Province. Facility horticulture in Heilongjiang Province develops rapidly in recent years, technical support is in great demand, but the experts' technology support for facility horticulture is far from enough. Experts' on-site guidance costs much time and money in the countryside, while the service efficiency is very low. To solve this urgent problem, the architecture of "greenhouse vegetable intelligent terminal system based on cloud service" and the key technologies of implementation (low-cost IoT, distributed real-time operating architecture, virtual expert service, neural network image recognition and mobile terminal service) were put forward. Based on expert services, supplemented by data mining technology, IoT devices were used as expert's remote perception means, smart phones as user terminals, cloud service for integrating knowledge, resources and Internet of Things data to provide vegetable experts and greenhouse vegetable users with high information acquisition, storage, analysis,decision-making capabilities and effective solutions. Experts could view vegetable production status in greenhouses remotely through the Internet, get image and growth environment data, then provide remote guidance to vegetable farmers through the system, expert knowledge would be stored, mined and reused by the system. The Internet of Things system could automatically send out early warning information by judging the air temperature, humidity, illumination intensity and soil moisture in greenhouse. The application of knowledge map and neural network technology would reduce the workload of experts and increase concurrent processing capability of services at the same time. At present, part of this research has been applied in different user groups such as agricultural research departments, enterprises, vegetable cooperatives and farmers in Heilongjiang Province. The system can provide experts with remote inquiry means of greenhouse vegetable production environment, and has the characteristics of simple deployment and low cost. It is suitable for various greenhouse vegetable scenarios, including fruit and edible fungi. In order to popularize this technology in greenhouse vegetable production in China, and achieve an efficient experts' technical support, this research also proposed technical solutions of a large-scale application scenario through cloud computing in future.

    Design and test of wheat stripe rust remote monitoring platform based on embedded system | Open Access
    Yunzhou Ji, Shengjia Du, Tongkui Ji, Huaibo Song
    2019, 1(3):  100-112.  doi:10.12133/j.smartag.2019.1.3.201903-SA004
    Asbtract ( 923 )   HTML ( 293)   PDF (5092KB) ( 722 )  
    Figures and Tables | References | Related Articles | Metrics

    Wheat stripe rust is an important biological disaster that affects the safe production of wheat in China for a long time. The number of spores of wheat stripe rust is a direct factor affecting its pathogenesis and transmission. At present, it mainly relies on the field sampling and investigation of agricultural technicians to predict and forecast. It is time-consuming and laborious, and difficult to achieve long-term monitoring of diseases, thus affecting the accuracy of forecasting and the timeliness of prevention and control. The existing automatic spore monitoring device also has the problems that the collecting device is mostly in the form of manual replacement of slides, and the direct acquisition of components in the air by a limited area of the slide may result in inaccurate sample collection and too small sample size. In order to further improve the monitoring and forecasting ability of wheat stripe rust, a wheat stripe rust monitoring device was designed and implemented, which based on the internet to build a wheat stripe rust monitoring platform, and based on the embedded system to establish a complete set of wheat stripe rust spore collection and image transmission processing device. Spore acquisition was performed using a slide adsorption device of "Six prism column + Electromagnet + Microscope". Control the up and down movement of the electromagnet to control the up and down movement of the slide; update the slide by controlling the rotation of the hexagonal shaft; obtain the image by controlling the time synchronization of the microscope and the shaft; control the cleaning solvent the smear and the movement of the cleaning block enable the slide to be cleaned. At the same time, a spore counting program based on the server platform was designed to process and analyze the collected slide images. The spore counting program used in this design is based on Python 3.6 and combined with the Skimage image processing package for spore image analysis and processing. The geometry factor feature based method was used, and the number of spores in the microscope field was finally obtained based on the regional attribute values. The experimental results show that the platform server image processing algorithm can achieve accurate counting of spores, the accuracy of counting the test images is 100%; the success rate of the slide switching system is 95%.This study can lay a foundation for the real-time monitoring of wheat stripe rust in the field, and can also provide references for the monitoring of other airborne diseases in the field.

    Intelligent Equipment and Systems
    Effects of technical operation parameters on spray characteristics of rotor plant protection UAV | Open Access
    Hang Zhu, Hongze Li, Yu Huang, Haitao Yu, Yunzhe Dong, Junxing Li
    2019, 1(3):  113-122.  doi:10.12133/j.smartag.2019.1.3.201906-SA001
    Asbtract ( 945 )   HTML ( 327)   PDF (3184KB) ( 829 )  
    Figures and Tables | References | Related Articles | Metrics

    High-quality operation of plant protection UAV is the premise of precision operation in agricultural aviation, so it is particularly important to study the characteristics of spray system. In order to explore the factors that affect the spray quality, the comprehensive experimental platform of spray performance (developed by Jilin Agricultural Machinery Research Institute) was used to test the droplet deposition distribution and droplet diameter under different UAV rotor speed, spray height and centrifugal nozzle speed in this research, and regression analysis on the deposition characteristics and particle diameter data of 12 groups of tests was conducted. The results showed that the three repeated tests of the same set of parameters had good consistency. Droplets had obvious drift and the maximum effective deposition rate was 46.31% and minimum 31.74%, which shows that the effective deposition rate of droplets was lower than 50%. Compared with the regression analysis results of droplet diameters DV10, DV50 and DV90, the spray height P value is greater than 0.5, and the nozzle speed and rotor speed P value are less than 0.5. So it can be inferred that spray height had a very significant effect on deposition, no significant effect on droplet size. The nozzle speed and rotor speed had very significant effect on droplet size, no significant effect on deposition. The test results of this research can provide theoretical basis and data support for improving the operation quality and spraying efficiency of UAVs.