Fast implementation of CA-CFAR algorithm for GPU

CA-CFAR detector is widely used in radar signal detection and image detection. How to implementation CA-CFAR quickly and efficiently is very important. The powerful parallel computing capability of GPU can greatly improve the computing efficiency, but it required the algorithm can be split into many independent concurrent computing units. Based on the computing characteristics of GPU, the computing step of CA-CFAR algorithm is optimized and divided into several steps, each step is suitable for GPU. Simulation result verified that the GPU based CA-CFAR algorithm is improve effectively.