Adaptive anisotropic pixel-by-pixel correction method for a space-variant degraded image.

Area of Science:

• Optical Imaging
• Image Processing
• Computer Vision

Background:

• Large field-of-view optical systems suffer from space-variant degradation, hindering target detection and recognition.
• Existing correction methods often fail to adequately address complex, localized image distortions.

Purpose of the Study:

• To propose and validate an adaptive anisotropic pixel-by-pixel space-variant correction method.
• To improve the accuracy of target detection and recognition in degraded wide-field images.

Main Methods:

• Estimation of local space-variant point spread functions (PSFs) using Haar wavelet degradation distribution.
• Development of a pixel-by-pixel correction model with adaptive sparse and non-negative regularization.
• Application of the split-Bregman multivariate separation algorithm for precise pixel-level correction.

Main Results:

• The proposed method effectively corrects space-variant degradation on a pixel-by-pixel basis, avoiding whole-image or block correction limitations.
• Experimental results demonstrate superior performance in preserving image details and texture compared to existing advanced methods.
• Enhanced accuracy in target detection and recognition was observed in corrected images.

Conclusions:

• The adaptive anisotropic pixel-by-pixel correction method offers a significant advancement in processing degraded wide-field images.
• This technique provides a robust solution for enhancing image quality and subsequent analysis in challenging optical imaging scenarios.