基于Tiny-YOLO的苹果叶部病害检测

A DETECTION METHOD FOR APPLE LEAF DISEASESBASED ON TINY-YOLO

由于苹果叶片纹理复杂多变,相似病害难以判断,识别速度难以快速提升,苹果叶部病害的识别仍存在较大研究空间.为实现苹果叶部病害快速、有效的自动检测,本文将Tiny-YOLO应用于苹果叶部病害检测.实验结果表明,Tiny-YOLO模型的mAP和IoU分别为99.86%和83.54%,检测速度达280 FPS,能够有效实现苹果叶部病害检测.

Given the complex and variable texture of apple leaves,similar diseases are difficult to judge and the recognition speed is difficult to be improved rapidly.At present,there is still a lot of room for improvement in the research of apple leaf diseases identification.Tiny-YOLO is a regression-based target detection method,which greatly simplifies the network structure on the basis of YOLO v2 network thus greatly improves the detection speed.In order to realize the rapid and effective automatic detection of apple leaf disease,Tiny-YOLO was applied to the detection of apple leaf disease.The experimental results showed that the mAP and IoU of the Tiny-YOLO model are 99.86%and 83.54%respectively,and the detection speed was up to 280 FPS,which has fully proved it an effective method for apple leaf diseases.