Research And Application Of Tomato Disease Object Detection Algorithm Based On Deep Learning

In recent years,computer vision has developed rapidly and has been widely used in many fields.In agricultural production,farmers detect crop diseases to prevent crop yield reduction.Computer vision and deep learning are used to help farmers solve the problem of disease detection in agricultural production,which will have important practical significance for the classification of crop diseases and the realization of intelligent agriculture.We used object detection technology based on deep learning to detect tomato diseases.The control efficiency of tomato diseases and the accuracy of the algorithm were improved.In this paper,a tomato disease detection model based on Faster RCNN was constructed to solve the problems of low efficiency and high labor consumption.Compared with the traditional detection methods,this method not only improves the detection speed but also improves accuracy.Secondly,the difference in plant leaf images is not obvious.Res101 is used as the feature extraction network of the model to increase the richness of image feature extraction.At the same time,the model can avoid the gradient loss caused by the increase of convolution layers in tomato disease leaf image detection.Finally,the size and proportion of tomato diseased leaves in each image are different.K-means algorithm is used to cluster the target frames of tomato diseased leaves,so as to make the anchor closer to the real size of the target frame of tomato diseased leaves.The accuracy was improved.