A lightweight safety helmet detection algorithm for complex underground mining environment

Le résumé fourni par la source

Abstract To address the problem of insufficient safety helmet detection accuracy caused by complex factors such as high dust, low illumination and small-sized safety helmet targets in underground coal mines, this paper proposes a lightweight safety helmet detection algorithm CFNet for complex underground coal mine scenarios. To start with, YOLOv11s was used as the baseline model, and the feature complementary mapping module and multi-kernel perception unit were used to improve the perception ability of the model for small targets and reduce the complexity of the model. Secondly, the convolutional gated linear unit is used to effectively aggregate local features to improve the feature extraction ability of the model. Finally, the dynamic head (Dyhead) was introduced, and the multi-perception attention module was used to enhance the multi-dimensional information perception ability of the detection head and reduce the loss of model classification and localization. Experimental results demonstrate that compared with the baseline model, the proposed algorithm reduces the number of parameters by 67.5% while increasing the mean average precision by 3.8%. Its overall performance outperforms mainstream state-of-the-art object detection algorithms including Faster R-CNN, RT-DETR and YOLOv13. This study can effectively improve the level of safe operation in underground coal mines and ensure the life safety of underground operators.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.