introduction
3d_mtrack started as a student project at the FHTW Berlin in applied computer-science. The intention was an application for 3D computer graphics which tracks the movement of an object in the 3D space and visualizes the movement in real-time. At first we thought about writing a python interface for blender to visualize the movement and control objects in blender. But we decided to take opengl for real-time graphical visualization, because we were in a hurry about the delivering for the course 3D computer graphics and we are not firm with python. So we use opengl for visualization, opencv for object tracking, and osc is our interface to control audio-environments like pd or reaktor. It is supposed to be easily handled and low-budget.
features
- recording of 3D movements
- representation of the movement with opengl in real-time
- playback of the recorded movement (format:*.3dm)
- OSC interface for sending data to audio-environments like pd or reaktor
- client-/ server application
- each client has his dedicated view
- server visualize movement
- need a cube to define the world of movement (calibrate the view of your webcam at the cube)
- has a gamemode (catch the monster)
preparation
- 2 webcams (min framerate 30 frames/sec)
- 3 computer (2 serve as clients, 1 serves as server)
- Linux distribution with kernel >= 2.6
- the actual opencv version from cvs
- ffmpeg >= 0.4.9
- activate video4linux in kernel
- ffmpeg >= 0.4.9
- module (pwc, ohci_hcd) for webcam
- set the environment variable for OPENCV
(export PKG_CONFIG_PATH="/opt/opencv" or sign in /etc/profile)
license
3d_mtrack is available under the GNU General Public License v2.