Search

The low light, strong light disturbance, high dust and other environmental conditions underground in coal mines, as well as the similarity of clothing and coal falling on the face of underground personnel, make it difficult to re identify underground personnel with weak features. The existing personnel re identification methods only extract global features and do not fully consider local features, resulting in low accuracy of underground personnel re identification. In order to solve the above problems, a local feature guided label smoothing and optimization method for re-identification of underground personnel with weak features is proposed. This method first extracts global and local features of underground personnel images through convolutional neural networks. Secondly, the k-nearest neighbor similarity is used to calculate the complementarity score between global and local features, in order to measure the degree of similarity between global and local features. Finally, based on the score of feature complementarity, label smoothing is performed on local features and label optimization is performed on global features. The weight of each local feature is dynamically adjusted to improve the label of each local feature. The prediction results of local features are summarized. The more reliable information is used to improve the label as a global feature label, thereby reducing image noise and enhancing feature identification capability. The experimental results show that the method outperforms mainstream personnel re identification methods in terms of mean average precision (mAP), rank-1 accuracy (Rank-1), and mean inverse negative penalty (mINP) on both publicly available datasets and self built datasets containing images of underground personnel. It has good generalization and robustness, and can effectively achieve underground weak feature personnel re identification.

Traditional object detection algorithms require manual feature extraction when recognizing abnormal behavior of miners queuing, resulting in long detection time and low detection precision. The object detection algorithm based on convolutional neural networks has improved detection speed and precision. But its detection performance is difficult to guarantee in scenarios of obstruction, dimness, and uneven illumination. In order to solve the above problems, an improved YOLOv5s (HPI YOLOv5s) model is proposed. It is used for miner queue detection. The HPI-YOLOv5s model improves the path aggregation network (PANet) on the basis of the YOLOv5s model. By deleting a single input edge node and adding bidirectional crossing paths, a bidirectional cross feature pyramid network (BCrFPN) is constructed for multi-scale feature fusion. Considering the low robustness of label allocation strategies with manually set thresholds, a dynamic label allocation strategy (ATSS-PLUS) is proposed based on adaptive training sample selection (ATSS) to dynamically set thresholds. It can reasonably evaluate the quality of candidate samples and dynamically set thresholds for each real object, resulting in higher detection precision and robustness. The method calculates the intersection area between the face frame and the designated queue area using the half plane intersection method. The method compares the ratio of the intersection area to the face frame area with the set threshold to determine whether the miners are queuing in an orderly manner. The experimental results show that the HPI-YOLOv5s model has an accuracy improvement of 1.9%, a weight reduction of 32%, a parameter reduction of 6.9%, and a detection speed improvement of 7.8% compared to the YOLOv5s model. Moreover, it can more accurately recognize the queuing situation of miners in obstruction, dimness, and uneven illumination mine images.