Design And Implementation Of Leaf Disease Recognition And Early Warning System For Chinese Cabbage In Greenhouse Based On Video Image

Chinese cabbage occupies an important proportion in vegetables in our country.With the expansion of the planting area of Chinese cabbage,its disease has gradually become the main factor restricting the high yield,good quality and high benefit production of Chinese cabbage.Common diseases include black spot,brown spot,downy mildew,soft rot,anthrax,black rot,etc.At present,the monitoring of Chinese cabbage leaf diseases mainly uses manual detection.The disease of Chinese cabbage can not be automatically monitored by identifying and judging the leaf disease and disease degree of Chinese cabbage based on experience,which seriously affects the prevention and control efficiency of Chinese cabbage disease,and wastes a lot of time and energy of farmers.With the development of video surveillance technology and computer image processing technology,this paper introduces the newly developed video surveillance technology into the traditional agricultural information system.Using intelligent surveillance technology,computer vision technology and software development technology,a real-time surveillance and early warning system for greenhouse cabbage leaf diseases is designed based on video surveillance,which realizes the real-time surveillance and early warning system for greenhouse cabbage leaf diseases.Real-time monitoring,and then early warning of disease types and grades.This paper first conducted a survey and review of experimental data and equipment.The main research work is as follows:(1)For the greenhouse cabbage,the video collection scheme and the processing of the video were studied.The inter-frame pixel contrast method is used to calculate the difference between the video frames intercepted every other hour,and a threshold is set for the difference,and the frame whose difference exceeds the threshold is image-processed to generate and class the disease.(2)The segmentation method of leaf disease images of cabbage was studied,and the segmentation of leaf lesions was successfully realized.Firstly,the image processed by the video is subjected to image correction such as distortion correction,denoising and enhancement,and then the maximum inter-class variance method and the H-component-based histogram bimodal method are used to segment the leaf spot of cabbage.The segmented background is marked black and the normal leaves are marked white to achieve segmentation of the leaf disease.(3)The extraction of disease characteristics and the identification of disease types were studied.Feature extraction of the color,shape and texture of the complete lesion obtained after segmentation,optimization of the extracted feature parameters,description and formation of feature vectors,combined with support vector machine and decision tree classification model,input these feature vectors Classification training is performed in the disease identification classifier,and finally the category to which each lesion object should belong is determined.(4)Using Matlab to realize a system that can monitor cabbage diseases and grade early warning.Combining video capture equipment and server,the leaf disease identification and monitoring and early warning system of cabbage was constructed,and the overall design of the system and the design of functional modules of each part were completed,thus realizing the real-time data collection of cabbage leaf disease.Main functions such as pretreatment,disease segmentation,recognition,early warning,etc.,and test its implementation in the monitoring process.