An Improved ORB Feature Extraction and Matching Algorithm

The fixed threshold selection of traditional ORB algorithm results in many false extractions and mismatches, which cannot solve the problem of sensitivity to changes in light. Aiming at this problem, an improved ORB feature point extraction and matching method based on quadtree was proposed. Firstly, set the local adaptive threshold, and propose the adaptive threshold selection criteria to achieve the accurate extraction of ORB feature points; then on the basis of the improved ORB feature points, the improved quadtree is used to screen the feature points; Finally, the LMedS method is used to complete the matching according to the selected feature points. Experimental results show that the improved method has strong adaptability to brightness changes, and the calculation speed and extraction accuracy have been improved. The total matching time is reduced, the number of mismatched points is less, the correct matching rate is higher, and it has good accuracy and real-time performance.