|
| Multi-scale visual feature extraction algorithm for coal rock fracture deformation in coal comprehensive mining engineering |
|
Received:September 13, 2023
Revised:October 07, 2023
Accepted:October 07, 2023
Published Online:March 24, 2025
|
| View Full Text View/Add Comment Download reader |
| DOI： |
| KeyWord:Coal rock fractures; Feature extraction; Crack images; Multi scale visual features; Coal comprehensive mining engineering; Crack deformation; |
| Author | Institution |
| Bao Xinping |
CHN Energy Group Xinjiang Energy Co., Ltd. |
| He Yong |
CHN Energy Group Xinjiang Energy Co., Ltd. |
|
| Hits: 534 |
| Download times: 243 |
| Abstract: |
| The coal resource exploration and mining industry is one of the pillar industries for the stable development of the country. With the gradual depletion of shallow coal resource reserves, coal resource mining is developing towards deep regions. Due to the increasing complexity of geological conditions, the phenomenon of coal rock fracture deformation is gradually increasing, posing a threat to the safety of coal comprehensive mining engineering. Therefore, a multi-scale visual feature extraction algorithm for coal rock fracture deformation in coal comprehensive mining engineering is proposed. Preprocess the original coal rock image (grayscale processing, sharpening enhancement processing and noise removal processing), on this basis, segment the coal rock image based on the C-V model, obtain the target image of coal rock fracture deformation, map it to the space range through the Gaussian kernel function, complete the construction of multi-scale visual space, and apply MSRFE algorithm to detect and determine the multi-scale visual characteristics of coal rock fracture deformation, Thus, the extraction of multi-scale visual features of coal rock fracture deformation has been achieved. The experimental data shows that the determination results of coal rock fracture deformation targets obtained by applying the proposed algorithm clearly label the boundaries of all fracture deformations, and the maximum completeness of multi-scale visual feature extraction reaches 99%, fully confirming that the proposed algorithm has better application performance. |
| Close |
|
|
|