A Case of Computer Image Processing Depth Learning Algorithm Based on YOLOv4 Improved Algorithm

Computer image recognition algorithm is the key to unsafe behavior recognition, and it is also the content that must be mastered to learn computer image processing. In order to solve the problem of detecting unsafe behaviors in high voltage areas and intrusion of people in hazardous areas, a detection method of unsafe behaviors in high voltage areas based on the double-layer YOLOv4 algorithm is proposed. The results demonstrate that the average accuracy (mAP) of this model can reach 85.73% in the task of unsafe behavior detection in high-voltage area, and the speed can reach 94ms/piece. It can be seen that this method can accurately detect intrusion into dangerous areas and unsafe behaviors of operators in real time, and can guarantee the safety of power grids and personnel.