Electrical Cabinet Wiring Detection Method Based on Improved YOLOv5 and PP-OCRv3

This paper proposes a method of checking wiring of electric cabinets based on improved YOLOv5 and PP-OCRv3 to resolve the problems of low efficiency, high manpower consumption, and lengthy time-consuming manual checking wiring in the quality inspection stage of modern industrial electric cabinet production. This research indicates a novel neural network structure called YOLOv5s-CBS that is based on the YOLOv5s neural network structure. The coordinated attention mechanism is added to the C3 module of the backbone structure, the BiFPN network is introduced in the neck structure, and the loss function is changed from CIoU to SIoU, which improves the effect of wiring label detection. After that, the YOLOv5-CBS detection network is cascaded with the text detection and recognition system PP-OCRv3 to realize a complete set of wiring detection logic. It effectively improves the efficiency and accuracy of electrical cabinet line inspection and reduces the cost of manual line inspection.