End-to-End Edge-Guided Multi-Scale Matching Network for Optical Satellite Stereo Image Pairs
Acquiring disparity maps by dense stereo matching is one of the most important methods for producing digital surface models. However, the characteristics of optical satellite imagery, including significant occlusions and long baselines, increase the challenges of dense matching. In this study, we pr...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-03-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/16/5/882 |
| _version_ | 1797263983388917760 |
|---|---|
| author | Yixin Luo Hao Wang Xiaolei Lv |
| author_facet | Yixin Luo Hao Wang Xiaolei Lv |
| author_sort | Yixin Luo |
| collection | DOAJ |
| description | Acquiring disparity maps by dense stereo matching is one of the most important methods for producing digital surface models. However, the characteristics of optical satellite imagery, including significant occlusions and long baselines, increase the challenges of dense matching. In this study, we propose an end-to-end edge-guided multi-scale matching network (EGMS-Net) tailored for optical satellite stereo image pairs. Using small convolutional filters and residual blocks, the EGMS-Net captures rich high-frequency signals during the initial feature extraction phase. Subsequently, pyramid features are derived through efficient down-sampling and consolidated into cost volumes. To regularize these cost volumes, we design a top–down multi-scale fusion network that integrates an attention mechanism. Finally, we innovate the use of trainable guided filter layers in disparity refinement to improve edge detail recovery. The network is trained and evaluated using the Urban Semantic 3D and WHU-Stereo datasets, with subsequent analysis of the disparity maps. The results show that the EGMS-Net provides superior results, achieving endpoint errors of 1.515 and 2.459 pixels, respectively. In challenging scenarios, particularly in regions with textureless surfaces and dense buildings, our network consistently delivers satisfactory matching performance. In addition, EGMS-Net reduces training time and increases network efficiency, improving overall results. |
| first_indexed | 2024-04-25T00:21:40Z |
| format | Article |
| id | doaj.art-f06d54184b09465e9e303fcef819dd77 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-04-25T00:21:40Z |
| publishDate | 2024-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-f06d54184b09465e9e303fcef819dd772024-03-12T16:54:19ZengMDPI AGRemote Sensing2072-42922024-03-0116588210.3390/rs16050882End-to-End Edge-Guided Multi-Scale Matching Network for Optical Satellite Stereo Image PairsYixin Luo0Hao Wang1Xiaolei Lv2Key Laboratory of Technology in Geo-Spatial Information Processing and Application System, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100190, ChinaKey Laboratory of Technology in Geo-Spatial Information Processing and Application System, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100190, ChinaKey Laboratory of Technology in Geo-Spatial Information Processing and Application System, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100190, ChinaAcquiring disparity maps by dense stereo matching is one of the most important methods for producing digital surface models. However, the characteristics of optical satellite imagery, including significant occlusions and long baselines, increase the challenges of dense matching. In this study, we propose an end-to-end edge-guided multi-scale matching network (EGMS-Net) tailored for optical satellite stereo image pairs. Using small convolutional filters and residual blocks, the EGMS-Net captures rich high-frequency signals during the initial feature extraction phase. Subsequently, pyramid features are derived through efficient down-sampling and consolidated into cost volumes. To regularize these cost volumes, we design a top–down multi-scale fusion network that integrates an attention mechanism. Finally, we innovate the use of trainable guided filter layers in disparity refinement to improve edge detail recovery. The network is trained and evaluated using the Urban Semantic 3D and WHU-Stereo datasets, with subsequent analysis of the disparity maps. The results show that the EGMS-Net provides superior results, achieving endpoint errors of 1.515 and 2.459 pixels, respectively. In challenging scenarios, particularly in regions with textureless surfaces and dense buildings, our network consistently delivers satisfactory matching performance. In addition, EGMS-Net reduces training time and increases network efficiency, improving overall results.https://www.mdpi.com/2072-4292/16/5/882dense matchingend-to-end stereo matching networkpyramid featuremulti-scale integrationtrainable guided filter |
| spellingShingle | Yixin Luo Hao Wang Xiaolei Lv End-to-End Edge-Guided Multi-Scale Matching Network for Optical Satellite Stereo Image Pairs Remote Sensing dense matching end-to-end stereo matching network pyramid feature multi-scale integration trainable guided filter |
| title | End-to-End Edge-Guided Multi-Scale Matching Network for Optical Satellite Stereo Image Pairs |
| title_full | End-to-End Edge-Guided Multi-Scale Matching Network for Optical Satellite Stereo Image Pairs |
| title_fullStr | End-to-End Edge-Guided Multi-Scale Matching Network for Optical Satellite Stereo Image Pairs |
| title_full_unstemmed | End-to-End Edge-Guided Multi-Scale Matching Network for Optical Satellite Stereo Image Pairs |
| title_short | End-to-End Edge-Guided Multi-Scale Matching Network for Optical Satellite Stereo Image Pairs |
| title_sort | end to end edge guided multi scale matching network for optical satellite stereo image pairs |
| topic | dense matching end-to-end stereo matching network pyramid feature multi-scale integration trainable guided filter |
| url | https://www.mdpi.com/2072-4292/16/5/882 |
| work_keys_str_mv | AT yixinluo endtoendedgeguidedmultiscalematchingnetworkforopticalsatellitestereoimagepairs AT haowang endtoendedgeguidedmultiscalematchingnetworkforopticalsatellitestereoimagepairs AT xiaoleilv endtoendedgeguidedmultiscalematchingnetworkforopticalsatellitestereoimagepairs |