Rapid Diagnostic Method Of Oilseed Rape Of Diseases And IoT Monitoring System Based On Multi-source Data

Oilseed rape is one of the most important oil and economic crops in China,and it plays an important role in the national grain and oil production safety.The growth of oilseed rape is susceptible to disease stress and seriously affects its yield and quality.The traditional disease monitoring methods are mainly field surveys,these methods are subjective and inefficient,resulting in untimely diagnoinsis and prevention,and can no longer meet the needs of comprehensive and effective prevention and precision management of crop diseases.In this study,the oilseed rape with Sclerotinia sclerotiorum was used as the research object,and the oilseed rape disease was quickly diagnosed from the ground and UAV low-altitude remote sensing,and the oilseed rape disease Internet of Things monitoring system based on multi-source data was constructed.The main research contents are as follows:(1)A multi-source remote sensing data-based discrimination model for oilseed rape disease was established,which realized the early rapid diagnosis of Sclerotinia sclerotiorum,and provided low-altitude remote sensing data support for the oilseed rape disease Internet of Things monitoring system.The RGB-assisted registration method was used to extract the temperature value corresponding to the oilseed rape leaf region from the thermal infrared image.The influence of Sclerotinia sclerotiorum infection on the leaf temperature was analyzed based on the maximum temperature difference(MTD)method.The results show that the MTD value can effectively distinguish the lesions in the leaves of infected rapeseed.At the same time,the remote sensing images of UAV were used to detect the canopy diseases of oilseed rape,and a visual inversion model of oilseed rape disease was constructed.According to the different incidences(health,mild and severe)of oilseed rape leaves before and after infection by Sclerotinia sclerotiorum,support vector machine(SVM),K nearest neighbor algorithm(KNN),random forest(RF)and na?ve Bayes(NB)models were established based on thermal infrared remote sensing data and multisource remote sensing data.The results show that the SVM model based on multi-source remote sensing data has the best classification and discrimination,and the overall recognition accuracy reaches 90%.For the same batch of oilseed rape samples,the classification accuracy of multisource remote sensing data was 11.3% higher than that of using only thermal infrared remote sensing data.(2)A leaf disease area measurement software was developed based on Android platform,which realized the accurate measurement of leaf area,lesion area and insect damage area of oilseed rape,and provided digital support for the acquisition of ground plant information in oilseed rape disease Internet of Things monitoring system.Aiming at the problem of perspective distortion caused by the camera imaging plane and the plane of the oilseed rape leaf during the image acquisition of the mobile phone,an image perspective distortion correction algorithm based on the rectangular frame correction plate was proposed.The edge detection and Hough line transformation were constructed by using the OpenCV function library.The four-step correction algorithm of corner positioning and perspective transformation was implemented,and the adaptation of the algorithm on the Android platform was realized.The leaf area,lesion area and insect damage area of oilseed rape were extracted by image processing algorithms such as binarization,color space conversion and small connected area removal.The leaf lesion area measurement software based on Android platform was developed.Through the measurement and comparison analysis of the circular area of different diameters,the average measurement error of the software was-1.29%~0.95%,the relative deviation was-0.55%~0.45%,which was high accuracy and stability.The average measurement error of the standard shape area measured by 0°~30° tilt angle is within 0.5%,which indicates that the established perspective distortion correction algorithm has strong robustness.In the actual measurement process,the accuracy of the measurement of oilseed rape leaf area,lesion area and insect area is above 97%,the whole process is less than 2.5 s,and Measurement results support online synchronization in this storage and cloud server.(3)The oilseed rape disease Internet of Things monitoring system based on multi-source data was constructed,which provided method and technical support for monitoring,early warning and early prevention of oilseed rape diseases.A real-time monitoring system based on the agrometeorological monitoring site for the temperature and humidity,soil temperature and humidity,carbon dioxide concentration,illuminance and air pressure of oilseed rape growth environment was established.After obtaining the information,the wireless capture card was sent to the cloud server through the IoT gateway;the monitoring socket was developed and deployed in the cloud server,and the data frame sent by the agricultural meteorological monitoring station was received and analyzed online,realizing the real-time stable transmission and monitoring of oilseed rape growth environment information.The system used the InfluxDB time series database of oilseed rape growth environment information,ground plant information and UAV low-altitude remote sensing feature information as the data source,configured Grafana visualization panel,constructd the visual interface of oilseed rape disease monitoring system,and realized online real-time monitoring of rapeseed disease.And it is of great significance for the early prevention and treatment of diseases.