Reproduce “Automatically eliminating seam lines with Poisson editing in complex relative radiometric normalization mosaicking scenarios” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Shiqi Liu, Jie Lian, Xuchen Zhan, Cong Liu, Yuze Tian, Hongwei Duan
A fully automatic approach combines overlap-area histogram matching with Poisson editing to eliminate seam lines in complex remote sensing image mosaicking. The method generates visually seamless and radiometrically consistent mosaics, effectively handling arbitrary seam shapes and challenging topological conditions.
Key insights
- Radiometric normalization alone (e.g., histogram matching) is insufficient to remove all seam lines, especially in complex boundary areas, necessitating a more advanced blending technique.
- Poisson editing can effectively create truly seamless transitions by leveraging gradient information and boundary conditions, even with arbitrary seam shapes and complex topological relationships.
- Localizing the Poisson editing operation to a small region around the seam line significantly reduces computational cost while maintaining blending effectiveness for large images.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.