Improved chirp scaling algorithm based on new integral kernel

A new integral kernel is proposed for the chirp scaling algorithm (CSA) to reduce the mismatched error of the shift-invariant range matched filter. In the proposed method, a quadratic term of the temporal variable beside the linear term is added into the integral kernel of range transformation. Based on the resulted range transformation, a new matched filter is designed for the range compression, which can reduce the mismatched error obviously than the traditional method. The computation burden of the proposed algorithm is also not increased too much when compared with the traditional CSA. The performance of the proposed method is analysed by the rigorous theoretical derivation and also validated by numerical experiments.