A fast approximate Harris corner detector
Lists
Only you see this
Abstract
The most widely used interest points detector is probably the Harris corner detector, which is based on the eigen values of the second moment matrix to extract corner features. In this detection algorithm, convolution with Gaussian filter is a very necessary and pivotal step. However, it is also the most time-consuming step which influences on the detection speed. Therefore, a novel fast approximate Harris corner detector was proposed in this paper, which relied on the integral images and box filters to convolute image. This fast algorithm replaced Gaussian filter process just by 3 additions and made it completely independent of the filter mask size. So it can save a large amount of consuming time. Verified by experimental results, the corners detected by the fast algorithm are quite similar as the original one, but for the implementation speed, especially when the mask size is bigger, the fast algorithm's speed advantage is more outstanding.
BibTeX
@inproceedings{han2009fast,
author = {Bing Han and Jiyin Sun and Jing Liu},
title = {{A fast approximate Harris corner detector}},
booktitle = {Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE},
year = {2009},
doi = {10.1117/12.833606},
}Loading related content...