Improved Correspondence Point Search and Image Matching Accuracy for Disturbanced Images Using TD(0) Method

This paper presents a novel reinforcement learning approach to enhance image matching, whereby the corresponding point candidates are detected from the feature points extracted from each of the two images. In general, robust estimation methods such as random sample consensus (RANSAC) are used to sel...

Full description

Saved in:
Bibliographic Details
Published inJournal of Advanced Simulation in Science and Engineering Vol. 12; no. 1; pp. 191 - 200
Main Authors Kamata, Hiroyuki, Omori, Haruki, Shimada, Mashiro
Format Journal Article
LanguageEnglish
Published Japan Society for Simulation Technology 2025
Subjects
Disturbanced images
Image matching
RANSAC
Reinforcement learning
Online AccessGet full text
ISSN2188-5303
2188-5303
DOI10.15748/jasse.12.191

Cover

More Information
Summary:This paper presents a novel reinforcement learning approach to enhance image matching, whereby the corresponding point candidates are detected from the feature points extracted from each of the two images. In general, robust estimation methods such as random sample consensus (RANSAC) are used to select valid corresponding points from the candidates. Therefore, we addressed the limitations of RANSAC random selection in the image-matching process, evaluating various reinforcement learning strategies, including deterministic and probabilistic approaches, and different value update mechanisms. The findings indicated that a probabilistic approach with suitable value updates provides a more robust solution for space-based navigation systems.
ISSN:2188-5303
2188-5303
DOI:10.15748/jasse.12.191