Improved L-R Image Decoding and Restoration Algorithm Based on Strong Backlight Background

In response to the challenges of strong light interference encountered by detectors in strong backlight conditions where sensor noise increases and accurate target information cannot be obtained this paper proposes an improved L-R image restoration algorithm designed for such environments. The proposed algorithm involves the construction of a wavefront-coded system with an internal mask to suppress strong light diffusion. The intermediate blurred image generated by the wavefront-coded system is then used as the input for the improved L-R image restoration algorithm. This process includes residual iteration combined with wavelet transform, and a double-threshold method is applied to eliminate two major types of noise, achieving the optimal iterative threshold and producing a restored image. Experimental results demonstrate that, compared to traditional image restoration algorithms, the proposed algorithm improves the image quality metrics PSNR and ISNR by an average of 7.64% and 6.62%, respectively. The superior restoration performance of this algorithm provides a solid foundation for subsequent target information acquisition and recognition.