Fusion of space-borne multi-baseline and multi-frequency interferometric results based on extended Kalman filter to generate high quality DEMs
Autor: | Qiming Zeng, Xiaojie Zhang, Jian Jiao, Jingfa Zhang |
---|---|
Rok vydání: | 2016 |
Předmět: |
Fusion
Layover Computer science 0211 other engineering and technologies 020206 networking & telecommunications 02 engineering and technology Atomic and Molecular Physics and Optics Computer Science Applications Extended Kalman filter Interferometry Interferometric synthetic aperture radar 0202 electrical engineering electronic engineering information engineering Atmospheric effect Computers in Earth Sciences National standard Baseline (configuration management) Engineering (miscellaneous) 021101 geological & geomatics engineering Remote sensing |
Zdroj: | ISPRS Journal of Photogrammetry and Remote Sensing. 111:32-44 |
ISSN: | 0924-2716 |
Popis: | Repeat-pass Interferometric Synthetic Aperture Radar (InSAR) is a technique that can be used to generate DEMs. But the accuracy of InSAR is greatly limited by geometrical distortions, atmospheric effect, and decorrelations, particularly in mountainous areas, such as western China where no high quality DEM has so far been accomplished. Since each of InSAR DEMs generated using data of different frequencies and baselines has their own advantages and disadvantages, it is therefore very potential to overcome some of the limitations of InSAR by fusing Multi-baseline and Multi-frequency Interferometric Results (MMIRs). This paper proposed a fusion method based on Extended Kalman Filter (EKF), which takes the InSAR-derived DEMs as states in prediction step and the flattened interferograms as observations in control step to generate the final fused DEM. Before the fusion, detection of layover and shadow regions, low-coherence regions and regions with large height error is carried out because MMIRs in these regions are believed to be unreliable and thereafter are excluded. The whole processing flow is tested with TerraSAR-X and Envisat ASAR datasets. Finally, the fused DEM is validated with ASTER GDEM and national standard DEM of China. The results demonstrate that the proposed method is effective even in low coherence areas. |
Databáze: | OpenAIRE |
Externí odkaz: |