The Impact of Building Orientation and Visibility Geometry in SAR Images

The detection of buildings using SAR radar images is of great interest due to its widespread use. But the complexity of SAR images poses many challenges to object recognition. One of the factors affecting the brightness of SAR images is the geometry of the vision, which causes the deformation of the building and in some cases even the omission of the buildings in radar images. The geometry of view includes incident angle, look angle, squint angle and direction of imaging, which is less discussed in research sources. In this paper, we intend to extract the buildings that are affected by changing the geometric parameter and analyze and prove the reason for its change. In other words, by detecting images with different incident angles, the orientation of the buildings affected by this change is identified. For this purpose, the time series images of Sentinel 1 satellite related to Baghistan region of Tehran are used. Using the CMV2 change detection method, the results show that changing the angle of impact by 10 degrees, leads to the removal of the workshop or factory sheds lying in the direction of 20 degrees to the horizon from the radar image.