Towards Change Detection in Urban Area by Sar Interferometry and Radargrammetry

Abstract

ABSTRACT:Change detection in urban area is an active topic in remote sensing. However, well-dealt subject in optical remote sensing, this research topic is still at an early stage and needs deeper investigations and improvement in what concerns SAR and InSAR remote sensing. Due to their weather and daylight-independency, SAR sensors allow an all-time observation of the earth. This is determining in cases where rapid change detection is required after a natural -or technological -disaster. Due to the high resolution that can be achieved, the new generation of space-borne radar sensors opens up new perspectives for analysing buildings in urban areas. Moreover, due to their short revisiting cycle, they give rise to monitoring and change detection applications. In this paper, we present a concept for change detection in urban area at building level, relying only on SAR-and InSAR data. In this approach, interferometric and radargrammetric SAR data are merged in order to detect changes. Here, we present the overall workflow, the test area, the required data as well as first findings on the best-suited stereo-configurations for change detection.

Pixel-based approach for building heights determination by SAR radargrammetry

“…As several scatterers contribute to the intensity of one single image cell, matching of homologous points is a matter of compromises. As we already explained these effects thoroughly in (Dubois et al, 2013), we simply remind here the principal considerations and conclusions for the self-containment of this paper. Figure 2a & 2b show a schematic representation of the effects occurring in layover areas after coregistration, in ground and slant range, respectively.…”

“…As explained in (Toutin et al, 2000) and more specifically for building areas in (Dubois et al, 2013), the quality of the disparity calculation depends highly of the acquisition configuration, i.e. if it is same-or opposite-side stereo, if the incidence angles are steep or shallow, and if the convergence angle (intersection angle between both acquisitions) is large or small.…”

“…This workflow is part of an overall approach for change detection in urban area that we presented in (Dubois et al, 2013). In this approach, pre-event interferommetric data are fused with post-event radargrammetric data in order to determine building changes.…”

“…Considering the conclusions made in (Dubois et al, 2013), we can deduce the formula for the height determination: T θ m and θ s represent the incidence angles of the master and slave image, and ζ the convergence angle. Figure 6 visualises the last steps of the workflow shown in Figure 1, from the disparity map to the height calculation.…”