Research On Plant Disease Diagnosis Based On Image Processing And SVM

In the process of growth,plants are easily affected by different types of adverse environmental biological and abiotic effects,which to a certain extent,increasing the risk of plant disease and also affects its good quality.The traditional methods of disease diagnosis mainly rely on professional personnel to identify and detect pathogens.However,this diagnosis method is not only slow,high cost,but also there is a deviation in timeliness,so it is difficult to realize real-time diagnosis.Although artificial detection can achieve the acquisition of disease information,it has a strong subjectivity.Because these features are difficult to describe,the results of diagnosis will also be more deviant.According to the different kinds of pathogenic species,diseases can also be divided into different styles for the texture,shape and color of plant leaves.With the development and application of image processing technology,the feature of plant disease can be extracted by computer image processing technology,so the classification method of support vector machine(SVM)is used to recognize plant diseases.In this paper,tomato disease image is taken as the research object,and the disease recognition method based on image feature and support vector machine is studied.According to the difference of color and texture of the disease spot,the features of the image of tomato spot blight,leaf mould,sclerotia and Verticillium wilt are extracted by MATLAB,and the feature of the disease image is obtained.The classification recognition of the disease image is realized by a pair of multi classification algorithms of support vector machine.Considering the limitations of the sample size,a higher recognition rate will be achieved.The number of samples increases and the recognition rate will be further improved.Through the training,the experimental results showed that the average recognition accuracy of the four diseases was 95%.The accuracy of recognition can reach more than 80%.The online diagnosis website for plant diseases is also made,which can check the plant growth environment and plant disease on line.It can realize on-line diagnosis function of plant diseases and online treatment scheme inquiry of plant diseases or consult expert function.The disease identification method proposed in this paper can identify the diseease of tomato accurately and quickly,and can provide technical support for the diagnosis of disease in the field and the remote disease diagnosis.The design of an online identification site for plant diseases can make the agricultural workers more intuitive to understand the plant growth and diagnose the disease.The website also has the functions of online disease treatment,query expert and message discussion.The realization of these functions provides convenience for the user.