一种新的局部不变特征检测和描述算法.pdfVIP

33 5 Vol. 33 No. 5 2010 5 CH INESE JOURNA L OF COMPU TERS May 2010 杨 恒 王 庆 ( 710072) . , , . , . , H arris , Harris , . . , , , , . . ; ; ; ; TP391 DOI : 10. 3724/ SP. J. 101 . 2010.00935 A Novel Local Invariant Feature Detection and Description Algorithm YANG Heng WANG Q ing (School of Computer Science and Engineering, N orthwestern Poly technical University, X i!an 710072) Abstract Local invariant features have been successfully applied in many applications in comput er vision. T his paper proposes a novel local feature detection and description algorithm. T he fea tures are invariant to image rotation, scale and illumination changes, and even can be invariant to weak affine transformations. In general, the local feature extraction process can be divided into tw o key steps w hich are feature detection step and feature description step. In the detection step, firstly, the Harris corners are detected in every scale level image. Secondly , the local scalespace extrema is searched w ithin a w indow w hich is centerlocalized on the multiscale Harris corners. Finally, the predominant orientation is computed for each keypoint. T he proposed feature detec tion algorithm has good repeatability performance. In the description step, a novel local descrip tor is created based on the gradient distance and orientation histogram ( GDOH) . GDOH not only has good matching performance, but also has low dimensionality, w hich results in much faster feature matching speed. Extensive experimental results have demonstrated the e