Enhancing Object Detection in Underground Mines: UCM-Net and Self-Supervised Pre-Training

Accurate real-time monitoring of underground conditions in coal mines is crucial for effective production management. However, limited computational resources and complex environmental conditions in mine shafts significantly impact the recognition and computational capabilities of detection models....

Full description

Saved in:
Bibliographic Details
Main Authors: Faguo Zhou, Junchao Zou, Rong Xue, Miao Yu, Xin Wang, Wenhui Xue, Shuyu Yao
Format: Article
Language:English
Published: MDPI AG 2025-03-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/25/7/2103
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850213181968875520
author Faguo Zhou
Junchao Zou
Rong Xue
Miao Yu
Xin Wang
Wenhui Xue
Shuyu Yao
author_facet Faguo Zhou
Junchao Zou
Rong Xue
Miao Yu
Xin Wang
Wenhui Xue
Shuyu Yao
author_sort Faguo Zhou
collection DOAJ
description Accurate real-time monitoring of underground conditions in coal mines is crucial for effective production management. However, limited computational resources and complex environmental conditions in mine shafts significantly impact the recognition and computational capabilities of detection models. This study utilizes a comprehensive dataset containing 117,887 images from five common underground mining tasks: mine personnel detection, large coal lump identification, conveyor chain monitoring, miner behavior recognition, and hydraulic support shield inspection. We propose the ESFENet backbone network, incorporating a Global Response Normalization (GRN) module to enhance feature capture stability while employing depthwise separable convolutions and HGRNBlock modules to reduce parameter volume and computational complexity. Building upon this foundation, we propose UCM-Net, a detection model based on the YOLO architecture. Furthermore, a self-supervised pre-training method is introduced to generate mine-specific pre-trained weights, providing the model with more semantic features. We propose utilizing the combined backbone and neck portions of the detection model as the encoder of an image-masking pre-training structure to strengthen feature acquisition and improve the performance of small models in self-supervised learning. Experimental results demonstrate that UCM-Net outperforms both baseline models and the state-of-the-art YOLOv12 model in terms of accuracy and parameter efficiency across the five mine datasets. The proposed architecture achieves 21.5% parameter reduction and 14.8% computational load decrease compared to baseline models while showing notable performance improvements of 1.3% (mAP<sub>50:95</sub>) and 0.8% (mAP<sub>50</sub>) in miner behavior recognition. The self-supervised pre-training framework effectively enhances training efficiency, enabling UCM-Net to attain an average mAP<sub>50</sub> of 94.4% across all five datasets. The research outcomes can provide key technical support for coal mine safety monitoring and offer valuable technological insights for the public safety sector.
format Article
id doaj-art-fbec8187cf9946d7bdb06a7690eaf346
institution OA Journals
issn 1424-8220
language English
publishDate 2025-03-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-fbec8187cf9946d7bdb06a7690eaf3462025-08-20T02:09:11ZengMDPI AGSensors1424-82202025-03-01257210310.3390/s25072103Enhancing Object Detection in Underground Mines: UCM-Net and Self-Supervised Pre-TrainingFaguo Zhou0Junchao Zou1Rong Xue2Miao Yu3Xin Wang4Wenhui Xue5Shuyu Yao6School of Artificial Intelligence, China University of Mining and Technology-Beijing, Beijing 100083, ChinaSchool of Artificial Intelligence, China University of Mining and Technology-Beijing, Beijing 100083, ChinaSchool of Artificial Intelligence, China University of Mining and Technology-Beijing, Beijing 100083, ChinaSchool of Artificial Intelligence, China University of Mining and Technology-Beijing, Beijing 100083, ChinaSchool of Artificial Intelligence, China University of Mining and Technology-Beijing, Beijing 100083, ChinaSchool of Artificial Intelligence, China University of Mining and Technology-Beijing, Beijing 100083, ChinaSchool of Artificial Intelligence, China University of Mining and Technology-Beijing, Beijing 100083, ChinaAccurate real-time monitoring of underground conditions in coal mines is crucial for effective production management. However, limited computational resources and complex environmental conditions in mine shafts significantly impact the recognition and computational capabilities of detection models. This study utilizes a comprehensive dataset containing 117,887 images from five common underground mining tasks: mine personnel detection, large coal lump identification, conveyor chain monitoring, miner behavior recognition, and hydraulic support shield inspection. We propose the ESFENet backbone network, incorporating a Global Response Normalization (GRN) module to enhance feature capture stability while employing depthwise separable convolutions and HGRNBlock modules to reduce parameter volume and computational complexity. Building upon this foundation, we propose UCM-Net, a detection model based on the YOLO architecture. Furthermore, a self-supervised pre-training method is introduced to generate mine-specific pre-trained weights, providing the model with more semantic features. We propose utilizing the combined backbone and neck portions of the detection model as the encoder of an image-masking pre-training structure to strengthen feature acquisition and improve the performance of small models in self-supervised learning. Experimental results demonstrate that UCM-Net outperforms both baseline models and the state-of-the-art YOLOv12 model in terms of accuracy and parameter efficiency across the five mine datasets. The proposed architecture achieves 21.5% parameter reduction and 14.8% computational load decrease compared to baseline models while showing notable performance improvements of 1.3% (mAP<sub>50:95</sub>) and 0.8% (mAP<sub>50</sub>) in miner behavior recognition. The self-supervised pre-training framework effectively enhances training efficiency, enabling UCM-Net to attain an average mAP<sub>50</sub> of 94.4% across all five datasets. The research outcomes can provide key technical support for coal mine safety monitoring and offer valuable technological insights for the public safety sector.https://www.mdpi.com/1424-8220/25/7/2103object detectioncoal mineself-supervised pre-trainingYOLOfeature extraction
spellingShingle Faguo Zhou
Junchao Zou
Rong Xue
Miao Yu
Xin Wang
Wenhui Xue
Shuyu Yao
Enhancing Object Detection in Underground Mines: UCM-Net and Self-Supervised Pre-Training
Sensors
object detection
coal mine
self-supervised pre-training
YOLO
feature extraction
title Enhancing Object Detection in Underground Mines: UCM-Net and Self-Supervised Pre-Training
title_full Enhancing Object Detection in Underground Mines: UCM-Net and Self-Supervised Pre-Training
title_fullStr Enhancing Object Detection in Underground Mines: UCM-Net and Self-Supervised Pre-Training
title_full_unstemmed Enhancing Object Detection in Underground Mines: UCM-Net and Self-Supervised Pre-Training
title_short Enhancing Object Detection in Underground Mines: UCM-Net and Self-Supervised Pre-Training
title_sort enhancing object detection in underground mines ucm net and self supervised pre training
topic object detection
coal mine
self-supervised pre-training
YOLO
feature extraction
url https://www.mdpi.com/1424-8220/25/7/2103
work_keys_str_mv AT faguozhou enhancingobjectdetectioninundergroundminesucmnetandselfsupervisedpretraining
AT junchaozou enhancingobjectdetectioninundergroundminesucmnetandselfsupervisedpretraining
AT rongxue enhancingobjectdetectioninundergroundminesucmnetandselfsupervisedpretraining
AT miaoyu enhancingobjectdetectioninundergroundminesucmnetandselfsupervisedpretraining
AT xinwang enhancingobjectdetectioninundergroundminesucmnetandselfsupervisedpretraining
AT wenhuixue enhancingobjectdetectioninundergroundminesucmnetandselfsupervisedpretraining
AT shuyuyao enhancingobjectdetectioninundergroundminesucmnetandselfsupervisedpretraining