Adaptive Optimal Iterative Image Restoration Algorithm Under Strong Backlight Conditions

This paper introduces an adaptive optimal iterative image restoration algorithm designed for strong backlight conditions to enhance denoising performance, optimize the determination of iteration counts in restoration algorithms, and improve the quality of restored images while preserving target details. A wavefront-coded optical system with an embedded mask is implemented to suppress strong light interference, while a point spread function is used to generate an intermediate blurred image. The restoration process combines the Lucy-Richardson (L-R) algorithm with wavelet transform for residual-based secondary denoising. Furthermore, a no-reference image quality assessment framework and a patience-based iteration mechanism are utilized to regulate the effective number of iterations. The optimal iteration count is determined by identifying the peak value of image quality evaluation parameters, ensuring the best-restored image is obtained. Experimental results demonstrate that, compared with conventional image restoration algorithms, the proposed method significantly reduces iteration time, enhances restoration quality, and lays a solid theoretical foundation for subsequent research on target recognition and localization.