Downhole Coal–Rock Recognition Based on Joint Migration and Enhanced Multidimensional Full-Scale Visual Features
The accurate identification of coal and rock at the mining face is often hindered by adverse underground imaging conditions, including poor lighting and strong reflectivity. To tackle these issues, this work introduces a recognition framework specifically designed for underground environments, lever...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-05-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/15/10/5411 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850127214515847168 |
|---|---|
| author | Bin Jiao Chuanmeng Sun Sichao Qin Wenbo Wang Yu Wang Zhibo Wu Yong Li Dawei Shen |
| author_facet | Bin Jiao Chuanmeng Sun Sichao Qin Wenbo Wang Yu Wang Zhibo Wu Yong Li Dawei Shen |
| author_sort | Bin Jiao |
| collection | DOAJ |
| description | The accurate identification of coal and rock at the mining face is often hindered by adverse underground imaging conditions, including poor lighting and strong reflectivity. To tackle these issues, this work introduces a recognition framework specifically designed for underground environments, leveraging joint migration and enhancement of multidimensional and full-scale visual representations. A Transformer-based architecture is employed to capture global dependencies within the image and perform reflectance component denoising. Additionally, a multi-scale luminance adjustment module is integrated to merge features across perceptual ranges, mitigating localized brightness anomalies such as overexposure. The model is structured around an encoder–decoder backbone, enhanced by a full-scale connectivity mechanism, a residual attention block with dilated convolution, Res2Block elements, and a composite loss function. These components collectively support precise pixel-level segmentation of coal–rock imagery. Experimental evaluations reveal that the proposed luminance module achieves a PSNR of 21.288 and an SSIM of 0.783, outperforming standard enhancement methods like RetinexNet and RRDNet. The segmentation framework achieves a MIoU of 97.99% and an MPA of 99.28%, surpassing U-Net by 2.21 and 1.53 percentage points, respectively. |
| format | Article |
| id | doaj-art-b359bd5b73e3485fbaa43a28bb03794d |
| institution | OA Journals |
| issn | 2076-3417 |
| language | English |
| publishDate | 2025-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj-art-b359bd5b73e3485fbaa43a28bb03794d2025-08-20T02:33:43ZengMDPI AGApplied Sciences2076-34172025-05-011510541110.3390/app15105411Downhole Coal–Rock Recognition Based on Joint Migration and Enhanced Multidimensional Full-Scale Visual FeaturesBin Jiao0Chuanmeng Sun1Sichao Qin2Wenbo Wang3Yu Wang4Zhibo Wu5Yong Li6Dawei Shen7School of Electrical and Control Engineering, North University of China, Taiyuan 030051, ChinaSchool of Electrical and Control Engineering, North University of China, Taiyuan 030051, ChinaSchool of Mechanical and Electrical Engineering, Beijing Institute of Technology, Beijing 100081, ChinaSchool of Electrical and Control Engineering, North University of China, Taiyuan 030051, ChinaSchool of Electrical and Control Engineering, North University of China, Taiyuan 030051, ChinaSchool of Electrical and Control Engineering, North University of China, Taiyuan 030051, ChinaState Key Laboratory of Coal Mine Disaster Dynamics and Control, Chongqing University, Chongqing 400044, ChinaSchool of Electrical and Control Engineering, North University of China, Taiyuan 030051, ChinaThe accurate identification of coal and rock at the mining face is often hindered by adverse underground imaging conditions, including poor lighting and strong reflectivity. To tackle these issues, this work introduces a recognition framework specifically designed for underground environments, leveraging joint migration and enhancement of multidimensional and full-scale visual representations. A Transformer-based architecture is employed to capture global dependencies within the image and perform reflectance component denoising. Additionally, a multi-scale luminance adjustment module is integrated to merge features across perceptual ranges, mitigating localized brightness anomalies such as overexposure. The model is structured around an encoder–decoder backbone, enhanced by a full-scale connectivity mechanism, a residual attention block with dilated convolution, Res2Block elements, and a composite loss function. These components collectively support precise pixel-level segmentation of coal–rock imagery. Experimental evaluations reveal that the proposed luminance module achieves a PSNR of 21.288 and an SSIM of 0.783, outperforming standard enhancement methods like RetinexNet and RRDNet. The segmentation framework achieves a MIoU of 97.99% and an MPA of 99.28%, surpassing U-Net by 2.21 and 1.53 percentage points, respectively.https://www.mdpi.com/2076-3417/15/10/5411coal rock recognitiondeep learningimage enhancementsemantic segmentation |
| spellingShingle | Bin Jiao Chuanmeng Sun Sichao Qin Wenbo Wang Yu Wang Zhibo Wu Yong Li Dawei Shen Downhole Coal–Rock Recognition Based on Joint Migration and Enhanced Multidimensional Full-Scale Visual Features Applied Sciences coal rock recognition deep learning image enhancement semantic segmentation |
| title | Downhole Coal–Rock Recognition Based on Joint Migration and Enhanced Multidimensional Full-Scale Visual Features |
| title_full | Downhole Coal–Rock Recognition Based on Joint Migration and Enhanced Multidimensional Full-Scale Visual Features |
| title_fullStr | Downhole Coal–Rock Recognition Based on Joint Migration and Enhanced Multidimensional Full-Scale Visual Features |
| title_full_unstemmed | Downhole Coal–Rock Recognition Based on Joint Migration and Enhanced Multidimensional Full-Scale Visual Features |
| title_short | Downhole Coal–Rock Recognition Based on Joint Migration and Enhanced Multidimensional Full-Scale Visual Features |
| title_sort | downhole coal rock recognition based on joint migration and enhanced multidimensional full scale visual features |
| topic | coal rock recognition deep learning image enhancement semantic segmentation |
| url | https://www.mdpi.com/2076-3417/15/10/5411 |
| work_keys_str_mv | AT binjiao downholecoalrockrecognitionbasedonjointmigrationandenhancedmultidimensionalfullscalevisualfeatures AT chuanmengsun downholecoalrockrecognitionbasedonjointmigrationandenhancedmultidimensionalfullscalevisualfeatures AT sichaoqin downholecoalrockrecognitionbasedonjointmigrationandenhancedmultidimensionalfullscalevisualfeatures AT wenbowang downholecoalrockrecognitionbasedonjointmigrationandenhancedmultidimensionalfullscalevisualfeatures AT yuwang downholecoalrockrecognitionbasedonjointmigrationandenhancedmultidimensionalfullscalevisualfeatures AT zhibowu downholecoalrockrecognitionbasedonjointmigrationandenhancedmultidimensionalfullscalevisualfeatures AT yongli downholecoalrockrecognitionbasedonjointmigrationandenhancedmultidimensionalfullscalevisualfeatures AT daweishen downholecoalrockrecognitionbasedonjointmigrationandenhancedmultidimensionalfullscalevisualfeatures |