Target object spatial point cloud feature-based automatic splicing method

Abstract

The invention discloses a target object spatial point cloud feature-based automatic splicing method. The method comprises the following specific steps of: scanning point cloud information of a to-be-tested object at a plurality of angles of view by using a Kinect depth camera, calculating FPFH feature information of point clouds, and carrying out feature matching between two point clouds by utilizing a point feature histogram of the point clouds so as to complete initial registration; solving transformation matrixes between the point clouds by utilizing an ICP algorithm, and carrying out rotation / translation transformation on target point clouds to finally complete splicing between the two point clouds. According to the method, matching and splicing are carried out by utilizing point cloudfeatures without manual mark points, so that the application range is wider. Moreover, the method has the effects of improving the point cloud splicing efficiency and enabling spliced models to be more real and natural.

Description

Technical field [0001] The invention relates to an automatic splicing method based on the spatial point cloud characteristics of a target object, and belongs to a point cloud data splicing technology in a three-dimensional model reconstruction technology. Background technique [0002] Point cloud splicing technology is an important research direction in the field of computer vision, and it has a wide range of applications in virtual reality, cultural relics protection, reverse engineering, human-computer interaction and other fields. In the process of data collection, due to the constraints of the environment and the equipment itself, it is necessary to collect data on a certain model surface from multiple angles. In order to obtain the complete model surface point cloud data, we need to unify the point cloud data obtained from different angles to the same camera coordinate system through coordinate conversion. This process is called point cloud stitching. Point cloud splicing t...

AI Technical Summary

Problems solved by technology

Among them, the ICP (Closest Iterative Point Algorithm) algorithm has high requirements for the initial position of the point cloud. When the initial position of the point cloud is large, a local optimal solution will be generated during splicing, resulting in splicing failure.

The point cloud registration algorithm based on geometric features is only suitable for objects with complex surface geometric features, and cannot achieve point cloud registration of simple or symmetrical objects.

The texture-based registration algorithm is only suitable for objects with rich surface texture, and the stability is poor when matching objects with a single texture

[0006] To sum up: the existing 3D point cloud automatic stitching methods have certain limitations and cannot meet the needs of practical applications.

Images