Damaged traffic sign detection method and device

The invention relates to a damaged traffic sign detection method and device, and the method comprises the steps: collecting data of a plurality of damaged traffic signs and normal traffic signs, and constructing a data set; labeling the data set to obtain a training set; a YOLOV5 model is preset, a fine-grained loss function module of the YOLOV5 model is designed, the fine-grained loss function module comprises a discriminative component and a diversity component, the discriminative component is used for enabling different levels of feature maps to contain sufficient discriminative information, and the diversity component is used for keeping diversity and interactivity of the different levels of feature maps; and inputting the training set into a YOLOV5 model, carrying out model training in combination with a fine-grained loss function module until a preset condition is met, and outputting the trained model as a target detection model. The method has the advantages that more local fine discriminative features