Corn leaf disease spot image classification and recognition algorithm based on improved Yolov8

Agriculture is an important factor in the development of today’s society and an important part of economic development. The leaves or crops of plants are susceptible to various diseases that can hinder the growth of crops. In terms of corn leaf disease detection, Unmanned Aerial Vehicle (UVA) have the advantages of small size and simple operation, and UAV detection replaces manual detection of corn leaf disease spots. In order to be able to accurately detect corn leaf spot diseases, we can use the Yolo algorithm in deep learning to detect and classify crop diseases. The Yolo algorithm has the advantages of high accuracy and fast speed in the experimental process, which can quickly and accurately classify healthy corn leaves and corn leaves with diseased spots, so as to improve the efficiency of detection. In order to solve the problem of small target detection of corn leaf lesions, the adaptive attention mechanism was introduced to enhance the attention mechanism, so as to improve the ability to identify feature attributes in advance. The FasterNet network was selected as the backbone network, which reduced the redundant algorithm of the Yolov8 algorithm, so as to achieve the requirements of lightness, and the goal of using the algorithm to detect maize leaf training samples was effective without losing accuracy. In the process of maize leaf disease detection, the scale of the disease is small, and there is a certain difficulty in the detection process, in the experimental process in this paper, the detection accuracy of the Yolov8 algorithm the detection process reaches 98%, the training loss rate is as high as 10%, and the test loss rate is 76%. After the improvement, the detection accuracy of the Yolov8 algorithm was increased to 99%, and the training loss rate and test loss rate were reduced by 4% and 2%, respectively.