Identification and Detection for Plant Disease Based on Image Segmentation and Deep Learning

The technology of object detection based on deep learning and computer vision have been widely used in many fields. Its ingenious feature extraction makes target detection more accurate. It can actively extract features to avoid complicated manual feature selection and extraction. However, in the detection of agricultural diseases, the existing algorithms are not satisfactory for the detection of small lesions in complex backgrounds. In response to the problem, this paper proposes an end-to-end target detection method combining image segmentation and deep learning. First, LM (local maximum) filter and k-means are used to segment the size of the anchor box and the cluster of targets which is imported in the deep learning training phase and the prediction phase. Second, YOLOv4 (you only look once) is applied with the data information obtained in the segmentation stage to detect lesions. Finally, the experimental analysis and comparison of the two data sets would be considered to prove the effectiveness of the model for small target detection.