Rethinking the Key Factors for the Generalization of Remote Sensing Stereo Matching Networks
Stereo matching, a critical step of binocular 3-D reconstruction, has fully shifted to deep learning due to its strong feature representation of remote sensing images. However, the ground truth for the stereo matching relies on expensive airborne light detection and ranging data, thus making it diff...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10818595/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1823857151218548736 |
|---|---|
| author | Liting Jiang Feng Wang Wenyi Zhang Peifeng Li Hongjian You Yuming Xiang |
| author_facet | Liting Jiang Feng Wang Wenyi Zhang Peifeng Li Hongjian You Yuming Xiang |
| author_sort | Liting Jiang |
| collection | DOAJ |
| description | Stereo matching, a critical step of binocular 3-D reconstruction, has fully shifted to deep learning due to its strong feature representation of remote sensing images. However, the ground truth for the stereo matching relies on expensive airborne light detection and ranging data, thus making it difficult to obtain enough samples for supervised learning. To improve the generalization ability of stereo matching networks on cross-domain data from different optical sensors and scenarios, in this article, we are dedicated to studying the key training factors from three perspectives. 1) When selecting a training dataset, prioritize data with similar regional target distribution as the test set, rather than relying on data from the same sensor. 2) Regarding the training modes, unsupervised methods generalize better than supervised methods. 3) We devised an unsupervised early stop strategy to help preserve the best model based on the pretrained weights. Extensive experiments are conducted to support the previous findings, on the basis of which we present an unsupervised stereo matching network with good generalization performance. |
| format | Article |
| id | doaj-art-a6b4df7d3f9a4a55b6792e86320dc3ab |
| institution | Kabale University |
| issn | 1939-1404 2151-1535 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj-art-a6b4df7d3f9a4a55b6792e86320dc3ab2025-02-12T00:01:03ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-01184936494810.1109/JSTARS.2024.352407810818595Rethinking the Key Factors for the Generalization of Remote Sensing Stereo Matching NetworksLiting Jiang0https://orcid.org/0009-0000-9706-5641Feng Wang1https://orcid.org/0000-0001-6494-3639Wenyi Zhang2https://orcid.org/0000-0003-0530-8488Peifeng Li3Hongjian You4Yuming Xiang5https://orcid.org/0000-0003-2063-9816Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaCollege of Surveying and Geoinformatics, Tongji University, Shanghai, ChinaStereo matching, a critical step of binocular 3-D reconstruction, has fully shifted to deep learning due to its strong feature representation of remote sensing images. However, the ground truth for the stereo matching relies on expensive airborne light detection and ranging data, thus making it difficult to obtain enough samples for supervised learning. To improve the generalization ability of stereo matching networks on cross-domain data from different optical sensors and scenarios, in this article, we are dedicated to studying the key training factors from three perspectives. 1) When selecting a training dataset, prioritize data with similar regional target distribution as the test set, rather than relying on data from the same sensor. 2) Regarding the training modes, unsupervised methods generalize better than supervised methods. 3) We devised an unsupervised early stop strategy to help preserve the best model based on the pretrained weights. Extensive experiments are conducted to support the previous findings, on the basis of which we present an unsupervised stereo matching network with good generalization performance.https://ieeexplore.ieee.org/document/10818595/Generalizationremote sensingstereo matchingunsupervised learning |
| spellingShingle | Liting Jiang Feng Wang Wenyi Zhang Peifeng Li Hongjian You Yuming Xiang Rethinking the Key Factors for the Generalization of Remote Sensing Stereo Matching Networks IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Generalization remote sensing stereo matching unsupervised learning |
| title | Rethinking the Key Factors for the Generalization of Remote Sensing Stereo Matching Networks |
| title_full | Rethinking the Key Factors for the Generalization of Remote Sensing Stereo Matching Networks |
| title_fullStr | Rethinking the Key Factors for the Generalization of Remote Sensing Stereo Matching Networks |
| title_full_unstemmed | Rethinking the Key Factors for the Generalization of Remote Sensing Stereo Matching Networks |
| title_short | Rethinking the Key Factors for the Generalization of Remote Sensing Stereo Matching Networks |
| title_sort | rethinking the key factors for the generalization of remote sensing stereo matching networks |
| topic | Generalization remote sensing stereo matching unsupervised learning |
| url | https://ieeexplore.ieee.org/document/10818595/ |
| work_keys_str_mv | AT litingjiang rethinkingthekeyfactorsforthegeneralizationofremotesensingstereomatchingnetworks AT fengwang rethinkingthekeyfactorsforthegeneralizationofremotesensingstereomatchingnetworks AT wenyizhang rethinkingthekeyfactorsforthegeneralizationofremotesensingstereomatchingnetworks AT peifengli rethinkingthekeyfactorsforthegeneralizationofremotesensingstereomatchingnetworks AT hongjianyou rethinkingthekeyfactorsforthegeneralizationofremotesensingstereomatchingnetworks AT yumingxiang rethinkingthekeyfactorsforthegeneralizationofremotesensingstereomatchingnetworks |