This paper mainly selects four diseases and pests of corn for image recognition and classification, selects wavelet neural network algorithm for image processing, and then uses yolov3 neural network for image loss iteration to achieve better results,Through this research, we can better understand the application and depth of convolutional neural network in the field of image recognition. In this paper, the relevant data sets are used for neural network operation, and the data sets are processed appropriately. After the data set is obtained, the correlation degree is deleted and the wavelet algorithm is used for denoising to obtain the denoised picture and algorithm, which is equivalent to modifying and fitting the picture.Select the appropriate convolutional neural network on the existing basis, modify the convolutional neural network into several different neural networks, and then apply the convolutional neural network to the data set to obtain the classification effect after passing through the neural network and achieve the classification effect. At this time, the accuracy of the effect after continuous optimization can reach%, and achieve the corresponding effect, It is consistent with the expected effect.