Abstract

After the original Oriented FAST and Rotated BRIEF (ORB) feature extraction algorithm performs feature extraction on the image, most of the feature points will appear in the area with richer texture, while the image area with less texture will produce fewer feature points. In addition, a large number of feature points will be gathered in a certain area of the image, resulting in redundant expression of the image. In view of the above defects, firstly, the candidate feature points are preliminarily screened by the adaptive threshold Features from Accelerated Segment Test (FAST) corner point extraction method, and then the quadtree method is used to further screen the feature points, so that the extracted feature points are evenly distributed in the image, eliminate the phenomenon of feature point aggregation. Experimental results show that the improved algorithm has high stability, and has a certain ability to adapt to image regions with weaker textures. Compared with other algorithms, the calculation time is much faster, and it can meet the real-time requirements.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call