An Efficient Detection Algorithm for PPE Wearing Status and Worker Position Status Recognition in Substation Operations

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

To address the challenges of limited visual information of PPE and complex background interference that constrains detection performance in substation operation scenarios, an efficient detection model, SDC-YOLOv10n, is proposed based on YOLOv10n. The model introduces the SimAM attention mechanism to enhance the network response to key regions associated with workers and protective equipment. A dual-path convolution, DPConv, consisting of a spatial convolution branch and a pointwise mapping branch, is designed and embedded into the C2f module to improve feature extraction and fusion for different wearing statuses. In addition, a classification head is constructed by combining DPConv with Channel Prior Convolutional Attention (CPCA) to enhance the classification capability for PPE wearing status and worker position status. Experimental results on the self-constructed SPPD-8 dataset show that the proposed model achieves an mAP@0.5 of 86.4% and an mAP@0.5:0.95 of 46.8%. The model contains 3.128 M parameters and requires 9.1 GFLOPs. SDC-YOLOv10n maintains real-time inference while achieving improved detection performance with moderate computational complexity, offering a practical solution for safety-status detection in substation operations.