Study On Cotton Diseases And Pests Identification Based On Decision Tree

China is a large population of agricultural power.Cotton is the China's important crop,which not only is closely related with the people's livelihood,but also is an important strategic material,affecting the national economic development.From seeding to harvest,cotton will suffer more than 40 kinds of disease invasion.If cotton disease can not be quickly and accurately identified,the production of cotton will be drawdown.So the diagnosis of cotton disease is very important.First of all,this paper describes the background and significance of the study,discusses the research status at home and abroad,and pointed out that the object of this study for cotton is the three diseases: verticillium wilt,angular leaf spot and wilt disease,and explained the importance of cotton disease identification research.Secondly,the use of digital image processing technology on the cotton plant disease image preprocessing.And then this paper summarizes the related technologies of digital image processing.The median filter is adopted to image noise reduction.The image is segmented by the maximum interclass variance method,and the improvement measures are put forward to improve the segmentation effect.After the segmentation image morphology processing in order to further operation.Thirdly,the color characteristics of the disease images are extracted based on the RGB color model and the HIS color model.Extracting the average grey value of R,G,B and H,S,I,a total of six components as color feature parameters.The two-dimensional Gabor transform is used to extract the texture Feature,which makes the image and 5 scales 8direction,a total of 40 filters do convolution operation.In 40 amplitude image to compute average amplitude of each image and then average eight directions of each scale as the texture feature.The method of statistical analysis is used to select the final classification feature.Finally,the ID3 algorithm and C4.5 algorithm in the decision tree method are introduced.The two algorithms are compared and analyzed,and the C4.5 decision tree classification algorithm is used to identify the three diseases of cotton.With the aid of the weka data mining platform for experiments,the experimental results are significant,The accuracy rate is94.67%.In this paper,C4.5 decision tree algorithm for cotton disease classification and identification is a new attempt.C4.5 algorithm is simple,fast operation,and able to deal with discrete data.C4.5 algorithm is also easy to extract rules.The decision tree generated by C4.5algorithm is intuitive and also easy to understand.