ORB Feature Extraction and Tracking Algorithm Based on Bag-of-Words Model and Epipolar Geometry Constraints
摘要
ORB (Oriented FAST and Rotated BRIEF, ORB) features are currently widely used in the field of image processing, and the extraction and tracking of feature points have also become one of the topics widely studied by scholars at home and abroad. The traditional brute-force matching method is not only time-consuming but also has a low matching success rate. Therefore, this paper designs an ORB feature extraction and tracking algorithm based on the bag-of-words model and epipolar geometry constraints, which improves the matching accuracy while taking into account the matching efficiency. The effectiveness of the algorithm is verified through public datasets.