Color Point Cloud Registration Algorithm Based on Local Gray Order Pattern

With the advancement of sensor technology, the acquisition of color point clouds has become increasingly convenient, providing valuable information for point cloud registration. In pursuit of more effective use of color information and improvement of both speed and accuracy in point cloud registration, we propose a novel algorithm for color point cloud registration based on a local gray order pattern. A color feature key point extraction algorithm based on Features from Accelerated Segment Test (FAST) is first introduced, and then the color distribution of adjacent points of the extracted key points is used to describe the features of the key points to establish the feature descriptor based on the local gray order pattern (LGOP). Experimental results show that the proposed algorithm significantly enhances the accuracy of point cloud matching without a corresponding increase in computational time for the point clouds with color information.