Transmission Corridor Cleaning Operation Area Identification Based on Improved YOLOv5

Aiming at the problems of scene complexity, diverse target types and multi-scale leading to the difficulty of intelligent safety detection in scene engineering accounting work, this paper proposes an improved algorithm based on YOLOv5. Firstly, the YOLOv5 detection network is constructed and the SimAM attention mechanism is integrated under this framework; then, TSCODE is combined in the output detection header of the model to enhance the sensory field of the model; finally, this paper further verifies the effectiveness of the model improvement by taking a tree stump as the detection object. The experimental results show that the improved YOLOv5 model improves 3.1% and the mAP50 can reach 88.6%, which can effectively meet the requirements of detection accuracy.