Regular three-dimensional color point cloud registration method based on multi-modal characteristics

Abstract

The invention relates to the technical field of three-dimensional point cloud data processing and three-dimensional scene reconstruction. The invention discloses a regular three-dimensional color point cloud registration method based on multi-modal characteristics. The method comprises the following steps: (1) acquiring three-dimensional color point cloud data; (2) extracting image data; (3) extracting image feature points; (4) matching the image feature points; (5) extracting point cloud data; (6) extracting point cloud feature points; Compared with the prior art, the method has the advantages that firstly, the multi-modal characteristics are extracted by utilizing two types of modal information of the point cloud and the image, the multi-modal matching point pairs are constructed, and the multi-modal advantages of the three-dimensional color point cloud are fully exerted; And secondly, coarse registration is firstly carried out by utilizing the multi-mode matching point pairs, a relatively accurate matching initial value is given, and the multi-mode matching point pairs are applied to fine registration, so that not only is the registration time reduced, but also the registrationprecision is improved.

Description

technical field [0001] The invention relates to a regular three-dimensional color point cloud registration method based on multimodal features, and belongs to the technical fields of three-dimensional point cloud data processing and three-dimensional scene reconstruction. Background technique [0002] With the development of information science and technology, computer vision, virtual reality, 3D printing, pattern recognition and other fields are making rapid progress, and 3D color point clouds have been more and more widely used. The 3D color point cloud data is formed by the fusion of point cloud data and image data. Among them, the point cloud data is obtained by a laser scanner, which records the geometric position information of the object, and the image data is obtained by a color camera. The color and texture information of the object. Therefore, 3D color point cloud data is a further development of 3D point cloud data, which can describe the real world more three-di...

AI Technical Summary

Problems solved by technology

This method makes full use of the multimodal information of the 3D color point cloud - image information and point cloud information, and calculates the multimodal features of the 3D color point cloud - image features and point cloud features, providing an accurate registration The initial value reduces the matching times of the ICP algorithm, improves the registration accuracy and registration speed, and better solves the problems of difficult initial value setting and low registration efficiency of traditional registration methods, and fully utilizes the 3D color point cloud. Multimodal Advantage

Images