BLP-YOLOv10: efficient safety helmet detection for low-light mining

Current safety helmet detection models face challenges in terms of computational complexity and hardware requirements, particularly in resource-constrained environments like underground mines. To address these issues, we propose the BLP-YOLOv10 model, which optimizes feature extraction and image processing by adjusting backbone channel parameters, incorporating sparse attention mechanisms, and integrating low-frequency enhancement filters. The experimental results demonstrate that BLP-YOLOv10 reduces the parameter count by 59.32% while achieving a mean average precision (mAP) of 98.1%, significantly improving detection speed and real-time performance. This makes the model highly robust and reliable, even under challenging lighting conditions.