基于融合特征的SURF配准算法

A SURF image registration algorithm based on fusion features

SURF算法是一种尺度不变、旋转不变且鲁棒性良好的配准算法,但其丢失了图像的颜色特征,因此对于彩色图像的配准效果不佳。为此,提出了一种基于融合特征的SURF配准算法。该算法首先利用彩色图像的颜色不变量和DLBP纹理特征构造融合特征灰度图,并提出了一种基于彩色图像颜色直方图的自适应方法来调节融合特征的权重;然后,利用SURF算法在融合特征灰度图上进行特征点的提取与匹配;最后,使用改进的RANSAC算法去除误匹配点。实验结果表明,对于彩色图像,此算法有效地增加了提取的特征点数,并加快了配准速率。

The speed-up robust features （SURF ） algorithm is a scale-invariant, rotation-invariant and robust registration algorithm. However,since the color features of the image can be lost,its regis-tration effect of color images is poor. W e therefore propose a SURF registration algorithm based on fu-sion features. Firstly,w e utilize the color invariant and the double local binary patterfeatures of the color image to construct the grayscale of fusion features. Meanwhile,w e propose an a- daptivemethod to adjust theweights of fusion features based on the color histogramof Then,the SURF algorithm is used to extract and match the feature points on the grayscale of fusion fea-tures. Finally,the improved random sample consensus （RANSAC ） algorithm is used match points. Experimental results show that the proposed algorithm can effectively increase the featurepoints and speed up the registration rate for color images.