Traffic safety helmet wearing detection system and method based on improved YOLOv5

The invention provides a traffic safety helmet wearing detection system and method based on improved YOLOv5, and relates to the technical field of image processing, and the system comprises a data set construction module which is used for constructing a traffic safety helmet wearing detection data set; the data preprocessing module is used for preprocessing the traffic safety helmet wearing detection data set; and the detection module is used for inputting the preprocessed data set into a traffic safety helmet wearing detection model for training to obtain a detection result. According to the method, a backbone network is replaced by GhostNet, so that the model volume is reduced; a spatial pyramid pooling layer is improved by using an RFB module, and the feature extraction capability is enhanced; a lightweight convolution GSConv is fused to compress the volume of the model; and the detection precision is maintained by adopting EIoU loss with a relatively high convergence speed. The problem that a traffic safe