Detection of Agricultural Pests Based on YOLO

Abstract

Focus on the problem of the species pests are difficult to identify in agricultural production an application based on the YOLO algorithm is presented. The algorithm extracts features fastly and complete detection task simultaneously. Through collecting agricultural pests, the algorithm establishes a data-set build target object detection model and train them, the problem of difficulties in identifying agricultural pest species and locating them has been solved. The result shows that the mAP'value of YOLO algorithm is 92.42%, the prediction precision is 96.8%. The algorithm has a high recognition rate, and identify the species of pests fastly and accurately, it can meet the detection and identification requirements.