A fast algorithm for stray light correction based on lossy source coding and SMT

Stray light is an inevitable phenomenon that occurs in optical imaging systems. In this paper, we propose a fast algorithm to correct images degraded by stray light. First, a general point spread function (PSF) model of stray light is given according to different causes so as to keep its shift variant nature. Secondly, the computational complexity and storage cost in image reconstruction are dramatically reduced with lossy source coding and sparse transform matrix (SMT) from the methodology of sparse representation. Moreover, a noniterative scheme is used to directly obtain the restored image. Compared with current methods, our approach has a more impressive performance and requires less computational complexity and storage cost. Experimental results with real-world images degraded by stray light validate the well performance and efficiency of our method.