摘要
从构造特征描述子入手,结合Haar-like特征提出了一种新的SURF特征描述符,并将其应用到遥感影像匹配中,新的描述符增加了对边和对角特征的描述。SURF匹配算法利用了积分图技术,匹配速度快,但是其特征的区分能力相对较弱。实验结果证明,此描述符的区分力优于原SURF描述子。
A modified Speeded-up Robust Features (SURF) descriptor based on Haar-like features, line features and diagonal features are added in the descriptor is proposed. Integral images allow fast computation, and are used for matching with SURF. However, the SURFdescriptor has a lower discriminating ability when compared with SIFT. Experimental results demonstrate that the modified descriptor has a better discriminating ability compared with SURF descriptor.
出处
《武汉大学学报(信息科学版)》
EI
CSCD
北大核心
2013年第7期770-773,804,共5页
Geomatics and Information Science of Wuhan University
基金
国家自然科学基金资助项目(41271456)