Research Of Video Surveillance System For Plant Diseases Based On TMS320DM6446

In agriculture, image of plant diseases is one of the important means to understand the growth status and disease situation of plant. Disease information, which is obtained by extracting the characteristics of collected images, can provide guidance for the prevention and control of plant diseases. At present, the application of image monitoring system in our country’s agriculture is relatively weak. There are problems including poor timeliness and high labor intensity, and the performance is not intuitive. These problems finally lead to the lag of prevention and treatment for plant diseases, and the outbreak of disease even takes place in local area from time to time. Therefore, agriculture in our country urgently needs new technology to make up the defects of the traditional method of disease prevention, and we can use it to develop a multifunctional application system to meet the demands of information collection, transmission, analysis and query of plant diseases, in order to guide the control of disease.In this paper, we utilize tomato as the research object to design a video surveillance system by using DaVinci technology, streaming media technology and neural network technology. This system can get the disease information and diagnose the kind of disease. Firstly, the camera placed in the field captures the analog video signals of plant, which will be converted into digital signals and sent to the multimedia process chip DM6446. Then, DM6446encodes and compresses the video adopting H.264encoding algorithm, and then transfers the video stream to the PC monitoring center by using the streaming media technology. Monitoring center displays the received video signal by media player VLC, and then gets the basic characteristics of disease by analyzing the color features and texture features of the disease spot, and finally analyzes and diagnoses the plant disease by using a combination of neural network and expert system.The experimental results show that, the video monitoring system for plant disease can playback the disease video smoothly and clearly, and its disease diagnosis module is accurate and easy to operate. It can timely and accurately diagnose the disease which the plant is suffering from, and provide the occurrence regularity and control method of the disease, so as to control the disease situation effectively. This system is feasible in plant video surveillance and disease diagnosis. It is a successful application of information technology in the field of agriculture.