Improved Wallis Dodging Algorithm for Large-Scale Super-Resolution Reconstruction Remote Sensing Images

Area of Science:

• Remote Sensing
• Image Processing
• Computer Vision

Background:

• Super-resolution (SR) image reconstruction often uses sub-block algorithms due to memory constraints.
• Sub-block SR images exhibit uneven brightness and contrast, hindering seamless mosaicking.

Purpose of the Study:

• To develop an improved algorithm for seamless mosaicking of super-resolution (SR) images.
• To address brightness and contrast inconsistencies in sub-block SR images.
• To achieve accurate image mosaicking with ZY-3 satellite data.

Main Methods:

• An improved weighted Wallis dodging algorithm was developed for gray-scale SR images with overlapping regions.
• A weighted adjustment sequence was introduced to prevent error propagation and information loss.
• A seam line elimination method was employed for smooth transitions in overlapping areas.

Main Results:

• The improved algorithm effectively harmonizes brightness and contrast across SR sub-blocks.
• Error accumulation and information loss were minimized through the weighted adjustment sequence.
• Seamless image mosaicking was achieved for 900 ZY-3 SR images using optimal seam lines.

Conclusions:

• The proposed weighted Wallis dodging algorithm effectively resolves brightness and contrast disparities in SR images.
• The method enables seamless mosaicking of SR images, crucial for accurate remote sensing data analysis.
• This technique enhances the quality and usability of high-resolution imagery from sources like ZY-3.