Method for tracing arm motion in real time video tracking system
A tracking system and real-time video technology, applied in the input/output of user/computer interaction, image data processing, instruments, etc., can solve the problems of restricting promotion and application, expensive price, and interference of target electromagnetic field electromechanical equipment
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Publication Date
- 2008-09-03
- Estimated Expiration
- Not applicable · inactive patent
Smart Images
Figure 1 Figure 2 Figure 3
Abstract
Description
technical field
[0001] The invention belongs to the technical field of automatic control, and relates to a method for visually tracking arm movement in the fields of immersive interactive games, dangerous industrial control and virtual reality, which integrates computer vision, graphic image processing and automatic control. technical background
[0002] In a distributed virtual reality environment, in order to provide a more harmonious and friendly human-computer interaction interface, there are a variety of immersive interaction modes to choose from, such as head movement tracking, facial expression recognition, natural language understanding, gesture recognition, etc. A more direct way is to integrate the arm motion tracking results and the gesture output of the data glove to realize human-computer interaction, in which the arm motion tracking provides displacement information, and the gesture output is used to generate semantics such as object selection, translation, and ...
Examples
Embodiment Construction
[0034] This embodiment provides a complete set of arm motion tracking solution, which mainly includes the following three parts:
[0035] 1. Feature extraction of arm identification ring
[0036] 1) Camera geometry calibration. In the offline state, several checkerboard calibration board images with depth information and checkerboard calibration paper images tiled on the ground were taken by two cameras, and then used MatlabTM The extended toolbox of camera calibration realizes geometric calibration and obtains the internal and external parameters of two cameras.
[0037] 2) Extract the edge feature points of the logo ring. According to the coherence of arm motion, coarse target search windows are separately separated from the interactive images from the two cameras in order to improve the efficiency of feature extraction. Then using the known color features of the logo rings, the edge point sequences corresponding to the wrist and elbow logo rings are extracted respectively...