Rigid body identification method and system

Abstract

The invention relates to a rigid body identification method and system. The method comprises steps that according to multiple infrared mark points pre-arranged on each rigid body, the mutually-different geometric information of all the rigid bodies are acquired; an infrared depth image of any to-be-identified rigid body is acquired, and the infrared information and the depth information are extracted; the infrared information is processed to acquire image coordinates of the infrared mark points; coordinate transform of the image coordinates is carried out to acquire position coordinates of theinfrared mark points on the to-be-identified rigid body in a camera coordinate system; a space geometric body is constructed, the space geometric body and the geometric structure information of eachrigid body are matched, and the to-be-identified rigid body is determined. The method is advantaged in that the multiple rigid bodies can be simultaneously identified and accurately tracked, a crosstalk phenomenon in the tracking process can be avoided, and the tracking result of the multiple rigid bodies in different types can be acquired.

Description

technical field [0001] The invention relates to the field of virtual reality, in particular to a rigid body recognition method and system. Background technique [0002] In existing VR, AR, MR and other equipment, binocular near-infrared stereo vision solutions are usually used to track rigid body targets in human-computer interaction systems, and only one or two fixed rigid bodies can be tracked at the same time. Tracking, multi-target If you want to obtain the world coordinate information of multiple marked points on the rigid body, or obtain the attitude parameters of each rigid body, there are no products and technical solutions that can be realized at present. [0003] In addition, the stability of the stereo matching algorithm in the binocular near-infrared stereo vision solution also has problems. The accuracy of the calculated depth information is poor, and the error is large, which cannot meet higher usage requirements. Contents of the invention [0004] The techn...

AI Technical Summary

Problems solved by technology

[0002] In existing VR, AR, MR and other equipment, binocular near-infrared stereo vision solutions are usually used to track rigid objects in human-computer interaction systems, and only one or two fixed rigid objects can be tracked at the same time. Tracking, multi-target If you want to obtain the world coordinate information of multiple marker points on the rigid body, or obtain the attitude parameters of each rigid body, there are no products and technical solutions that can be realized at present

[0003] In addition, the stability of the stereo matching algorithm in the binocular near-infrared stereo vision solution also has problems. The accuracy of the calculated depth information is poor, and the error is large, which cannot meet higher usage requirements.

Images