SAR real-time guidance system based on NLM despeckling and gradient binary description

In recent years, Synthetic Aperture Radar (SAR) has been widely used in guidance systems due to its high imaging quality and other superior functions. As the basis of system, SAR image registration has a direct impact on guidance precision. However, due to some inherent characteristics, especially the polarization and speckle noise, we need to adopt more effective approach to achieve the high precision and low computation of SAR guidance system than other optical image based systems. Therefore, in this paper, first, we implement NLM (Non-local Means) filter to improve the image quality damaged by speckle. Then, at feature detection and description stage, we adopt well-known FAST (Features from Accelerated Segment Test) and gradient-based binary descriptors, LDB (Local Difference Binary), to overcome strict real-time requirement and polarization phenomenon. Finally, at matching stage, in order to increase the number of correct pairs, the FSC (Fast Sample Consensus) is used to discard outliers matched by Hamming distance. And the transformation parameters could be estimated with affine model. To evaluate the performance of algorithms, three comparative experiments have been accomplished. The results demonstrate a good performance of our proposed method.