Safety Helmet-Wearing Detection System for Manufacturing Workshop Based on Improved YOLOv7

Safety helmets play a vital role in protecting workers’ heads. In order to improve the accuracy of detection model complex environments, such as backgrounds and different lighting distances, we propose safety helmet-wearing algorithm based on improved YOLOv7. backbone network, 16-channel features are used replace 3-channel RGB features. Structured pruning is performed head loss function replaced by SIoU. Experiments “helmet-head,” “helmet-data,” “helmet” data sets show that mAP F1 YOLOv7_ours this paper better than Faster RCNN, YOLOv5, YOLOv7 series models. On image application scenarios, light intensity, color depth, has stability higher can detect at 112.4FPS (1000/8.9). Based YOLOv7_ours, integrated face recognition technology text-to-speech (TTS) realize helmet detection, identity recognition, automatic voice reminder capabilities developed prototype system. We verified feasibility system semifinished product manufacturing workshop.