Unsupervised PolSAR Change Detection Based on Polarimetric Distance Measurements and ConvLSTM Network
Time-series PolSAR are capable for continuous change monitoring of natural resources and urban land-covers regardless of weather and lighting conditions. However, in the big SAR data era, the scarcity of labeled PolSAR samples poses new challenge to the traditional change detection methods. To reduc...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2023-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10286878/ |
| _version_ | 1827769216605880320 |
|---|---|
| author | Rong Gui Xinyue Zhang Jun Hu Lei Wang Xing Zhang |
| author_facet | Rong Gui Xinyue Zhang Jun Hu Lei Wang Xing Zhang |
| author_sort | Rong Gui |
| collection | DOAJ |
| description | Time-series PolSAR are capable for continuous change monitoring of natural resources and urban land-covers regardless of weather and lighting conditions. However, in the big SAR data era, the scarcity of labeled PolSAR samples poses new challenge to the traditional change detection methods. To reduce the dependence on labeled samples and ensure the efficiency of long time-series PolSAR interpretation, an unsupervised and pseudolabel-based change detection method is proposed. First, the similarity maps of time-series PolSAR are gauged by three selected polarimetric distance measurements (PDMs), which are suitable for PolSAR distribution characteristics and have the potential to reflect PolSAR changes. Second, the high-confidence changed pseudosamples are selected based on the similarity maps, and the unchanged pseudosamples are selected based on the nonsimilarity maps. Third, the limited selected pseudosamples (changed and unchanged) and multidimensional features are used to train the ConvLSTM network for change detection, and the input features include the <italic>T<sub>3</sub></italic> coherence matrix elements of time-series PolSAR and the aforementioned PDMs. Finally, the change detection results based on pseudosamples and the ConvLSTM network can be obtained, without additional manual labels. Adequate experiments are conducted on Radarsat-2, UAVSAR full-polarized, and Sentinel-1 dual-polarized datasets, achieving improved unsupervised change detection accuracy at 89.59–93.24%. |
| first_indexed | 2024-03-11T12:22:12Z |
| format | Article |
| id | doaj.art-daa21b65807c41f2a4090ca3360bffcb |
| institution | Directory Open Access Journal |
| issn | 2151-1535 |
| language | English |
| last_indexed | 2024-03-11T12:22:12Z |
| publishDate | 2023-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj.art-daa21b65807c41f2a4090ca3360bffcb2023-11-07T00:00:21ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing2151-15352023-01-01169762977610.1109/JSTARS.2023.332537010286878Unsupervised PolSAR Change Detection Based on Polarimetric Distance Measurements and ConvLSTM NetworkRong Gui0https://orcid.org/0000-0001-8470-3405Xinyue Zhang1https://orcid.org/0009-0008-8179-4864Jun Hu2https://orcid.org/0000-0002-5412-2703Lei Wang3https://orcid.org/0000-0002-7383-4167Xing Zhang4https://orcid.org/0000-0001-8489-1784School of Geosciences and Info-physics, Central South University, Changsha, ChinaSchool of Geosciences and Info-Physics, Central South University, Changsha, ChinaSchool of Geosciences and Info-physics, Central South University, Changsha, ChinaSchool of Electrical and Information Engineering, Wuhan Institute of Technology, Wuhan, ChinaSchool of Geosciences and Info-Physics, Central South University, Changsha, ChinaTime-series PolSAR are capable for continuous change monitoring of natural resources and urban land-covers regardless of weather and lighting conditions. However, in the big SAR data era, the scarcity of labeled PolSAR samples poses new challenge to the traditional change detection methods. To reduce the dependence on labeled samples and ensure the efficiency of long time-series PolSAR interpretation, an unsupervised and pseudolabel-based change detection method is proposed. First, the similarity maps of time-series PolSAR are gauged by three selected polarimetric distance measurements (PDMs), which are suitable for PolSAR distribution characteristics and have the potential to reflect PolSAR changes. Second, the high-confidence changed pseudosamples are selected based on the similarity maps, and the unchanged pseudosamples are selected based on the nonsimilarity maps. Third, the limited selected pseudosamples (changed and unchanged) and multidimensional features are used to train the ConvLSTM network for change detection, and the input features include the <italic>T<sub>3</sub></italic> coherence matrix elements of time-series PolSAR and the aforementioned PDMs. Finally, the change detection results based on pseudosamples and the ConvLSTM network can be obtained, without additional manual labels. Adequate experiments are conducted on Radarsat-2, UAVSAR full-polarized, and Sentinel-1 dual-polarized datasets, achieving improved unsupervised change detection accuracy at 89.59–93.24%.https://ieeexplore.ieee.org/document/10286878/Long short-term memory (LSTM) networkpolarimetric distance measurements (PDMs)synthetic aperture radar (SAR)time series imagesunsupervised change detection |
| spellingShingle | Rong Gui Xinyue Zhang Jun Hu Lei Wang Xing Zhang Unsupervised PolSAR Change Detection Based on Polarimetric Distance Measurements and ConvLSTM Network IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Long short-term memory (LSTM) network polarimetric distance measurements (PDMs) synthetic aperture radar (SAR) time series images unsupervised change detection |
| title | Unsupervised PolSAR Change Detection Based on Polarimetric Distance Measurements and ConvLSTM Network |
| title_full | Unsupervised PolSAR Change Detection Based on Polarimetric Distance Measurements and ConvLSTM Network |
| title_fullStr | Unsupervised PolSAR Change Detection Based on Polarimetric Distance Measurements and ConvLSTM Network |
| title_full_unstemmed | Unsupervised PolSAR Change Detection Based on Polarimetric Distance Measurements and ConvLSTM Network |
| title_short | Unsupervised PolSAR Change Detection Based on Polarimetric Distance Measurements and ConvLSTM Network |
| title_sort | unsupervised polsar change detection based on polarimetric distance measurements and convlstm network |
| topic | Long short-term memory (LSTM) network polarimetric distance measurements (PDMs) synthetic aperture radar (SAR) time series images unsupervised change detection |
| url | https://ieeexplore.ieee.org/document/10286878/ |
| work_keys_str_mv | AT ronggui unsupervisedpolsarchangedetectionbasedonpolarimetricdistancemeasurementsandconvlstmnetwork AT xinyuezhang unsupervisedpolsarchangedetectionbasedonpolarimetricdistancemeasurementsandconvlstmnetwork AT junhu unsupervisedpolsarchangedetectionbasedonpolarimetricdistancemeasurementsandconvlstmnetwork AT leiwang unsupervisedpolsarchangedetectionbasedonpolarimetricdistancemeasurementsandconvlstmnetwork AT xingzhang unsupervisedpolsarchangedetectionbasedonpolarimetricdistancemeasurementsandconvlstmnetwork |