Three-dimensional registering method based on ToF (Time-of-Flight) depth camera

Abstract

The invention provides a three-dimensional registering method based on a ToF (Time-of-Flight) depth camera, comprising the following steps of: (1) establishing a double-eye camera system composed of the ToF depth camera and a color camera and calibrating the system; (2) finishing the elimination of abnormal points and random noises of a depth image by the ToF camera to improve the precision of depth data; and (3) projecting the depth image obtained by the ToF camera to a visual field of the color camera to establish a corresponding relation between the depth image and a colorful image. The tracking of marks in a real-time video streaming can be realized by using a Camshift algorithm, so that a three-dimensional position and a direction of the marks in a space can be obtained according to the projected depth image.

Description

technical field [0001] The invention belongs to the technical fields of computer virtual reality and computer graphics and imaging, and specifically uses a binocular camera system constructed by a ToF depth camera and a color camera to complete the collection of real environment depth information and color information, and is a key technology for augmented reality Implementation of 3D registration. Background technique [0002] Augmented Reality (AR for short) reduces the burden of scene construction and processing through key technologies such as consistent representation, presentation and interaction of real and virtual environments, and builds an information loop between the virtual world, the real world and users. Further enhance people's participation and feeling of the environment. The key problem that augmented reality needs to solve first is the registration between the objects in the acquired real environment video stream and the virtual objects generated by the co...

AI Technical Summary

Problems solved by technology

This method requires rich texture information in the real environment, and is poorly applicable to regions with a single color and no texture. Although the stereo matching technology using spatial structure information can accurately find matching points, the computational complexity of this method is high

[0004] Currently widely used depth image acquisition equipment based on the principle of laser ranging, such as Leica Scanstation3000 and Minolta 910, etc., adopt point-by-point scanning method when acquiring depth information, which is suitable for acquiring depth information in static scenes, and it is difficult to meet real-time requirements

Images