Coherent change detection robust to layover regions using SAR tomography
Konferenz: EUSAR 2022 - 14th European Conference on Synthetic Aperture Radar
25.07.2022 - 27.07.2022 in Leipzig, Germany
Tagungsband: EUSAR 2022
Seiten: 6Sprache: EnglischTyp: PDF
Autoren:
Yamaguchi, Yuki; Tanakaa, Taichi; Toda, Masato; Tai, Tsenjung (Data Science Research Laboratories, NEC Corporation, 1753, Shimonumabe, Nakahara-ku, Kawasaki, Kanagawa, Japan)
Inhalt:
This paper proposes a new coherent change detection (CCD) method robust to layover regions using SAR tomography. CCD uses coherence between a SAR image pair to measure their similarity. A low coherence value indicates a possibly changed spot. CCD can detect more minute changes than amplitude based change detection technique. However, conventional CCD methods suffer from false decrease of coherence values in layover regions. This becomes a pivotal issue when applying conventional CCD methods to urban areas where many layover regions exist. The proposed method can suppress false decrease of coherence values in unchanged regions by compensating the phase of the observed signal using SAR tomography. Therefore, it can robustly detect changes in layover regions. We evaluate our method on simulation data and real data obtained by Cosmo-SkyMed. The experiment results show that the proposed method can successfully detect changes in layover regions where conventional methods cannot detect.