High Resolution Multipass Airborne SAR Tomography in Urban Areas

Abstract

This As classical tomographic synthetic aperture radar (TomoSAR) imaging techniques, the spectral analysis and compressive sensing (CS) have been applied on a large quantity of applications. Whereas, a critical procedure of the processing, defined as deramping, has great influence on the focusing results due to the essential estimation of the critical baselines. To avoid this operation, the general beam forming (BF) method is applied in this paper. The BF method in spatial domain is a general approach to estimate the direction-of-arrival (DOA) in the perpendicular line-of-sight (PLOS) direction. The critical processing in this method is to accumulate signal elements coherently, i.e. homologous points, with the steering vector in a specific pixel after fine registration of all the SAR images. Due to the sophisticated structure and the speckle noise influence over urban areas, all the homologous points in the array SAR images cannot be coregistered precisely in the identical pixels. In order to pick out the most relevant pixels both in phase and amplitude, an inconsistence criterion using joint phase and amplitude (JPA) of pixels in a window about the reference pixel is proposed in this paper. According to the minimization of inconsistence criterion, the homologous points can be picked out for high accuracy focusing of tomographic SAR imaging instead of accumulating data in the fixed pixels, especially for high resolution SAR images. The main feature of the method is that of selecting the most relevant pixels by using the local information of target scatterers and reducing the effect of speckle noise or misregistration. To check the plausibility of our proposed technique, the configuration of the first multi-pass X-band airborne TomoSAR system in China is used. Experimental results with the multi-pass SAR data show the feasibility of improving the focusing power of scatterers in PLOS direction by the proposed method compared to the other methods.