Blur kernel estimation using normal sinh-arcsinh model based on simple lens system

Dazhi Zhan, Xiangrong Zeng, Weili Li, Yu Liu, Zhihui Xiong

Abstract: Focused images captured by the lens system suffer image degradation due to factors, such as aberration, caused by the optical structure. In the simple lens system, aberration is more severe because of the simplification of the imaging system. In the existing imaging model, the blur kernel of the image is usually described by the point spread function. A few studies have shown that the blur kernel of the simple lens is close to the spatially deformed wedge. In the present study, a more reasonable normal sinh-arcsinh model is used to fit the blur kernel, and the parametric blur kernel is obtained by Powell algorithm. Finally, the clear image is restored and compared with other models to prove the advantages of our method.