Robust Image Matching Based on Image Feature and Depth Information Fusion

In this paper, we propose a robust image feature extraction and fusion method to effectively fuse depth information improve the registration accuracy of RGB-D images. The proposed directly splices point descriptors with corresponding cloud obtain descriptor points. is constructed based on SIFT, SURF, ORB PFH FPFH descriptors. Furthermore, performance features tested through datasets YCB KITTI. ORBPFH reduces false-matching rate by 4.66~16.66%, ORBFPFH 9~20%. experimental results show that in paper suitable for FPFH, which can representation registration, representing novel approach matching.