Man-Made Target Detection from Polarimetric SAR Data via Nonstationarity and Asymmetry

摘要

Detection of man-made targets in urban areas using polarimetric synthetic aperture radar (PolSAR) data has become a promising research area since it has a close relationship with urban planning, rescue service, etc. This paper presents an improved man-made target detection method for PolSAR data based on nonstationarity and asymmetry. Nonstationarity in azimuth direction is already utilized to separate man-made and natural targets in urban areas. However, there are still some drawbacks. Some small man-made targets and roads cannot be effectively detected. In addition, nonstationarity can also occur in some other natural surfaces, such as cropland with Bragg resonance. Therefore, to resolve these problems, we incorporate reflection asymmetry into the azimuth nonstationarity extraction method to improve the man-made target detection accuracy, i.e., removing the natural areas and detecting the small targets. Airborne ESAR data and spaceborne PALSAR data are used to validate the performance of the proposed method. The result obtained by our proposed method shows a 20% higher accuracy than the result based on original nonstationarity extraction method. Natural areas with Bragg resonance are removed. Moreover, most of the buildings and some metallic fences along the road can also be accurately detected.