Study on Corn Disease Identification Based on PCA and SVM

In this paper, corn gray leaf spot, corn rust, corn big spot and healthy corn leaves were studied. In the process of image background segmentation, Otsu method, OpenCV morphological operation and morphological transformation method are used to outline the outline of the object, and then create a mask. Using the outline, the difference set between the corn leaf and the background is taken to get a complete corn leaf image. PCA and SVM are applied to the processed image. When the penalty parameter C of SVM is 100 and the kernel is linear, the classification accuracy of four kinds of diseases is 90.05%, 92.64%, 91.23% and 95.78% respectively.