Calibration and Application of Optical Stereo Vision System Based on PTZ Cameras (II)

Project: National Science and Technology CouncilNational Science and Technology Council Academic Grants

Project Details


Stereo vision systems composed of optical cameras are of high potential in 3D reconstruction due to their merits of prompt capture, low cost and ready availability in hardware, and capability of obtaining surface texture. In the last one-year project, we have developed the geometric model and calibration procedure for a powered stereo vision system based on two PTZ (Pan-Tilt-Zoom) cameras. The calibration method is based on Zhang’s approach and an improved particle swarm optimization method using an augmented checkerboard composed of eight small checkerboards. This two-year project aims at furthering reconstruction accuracy of the active stereo vision system. For the first year, we will consider the effects of optical zoom on the intrinsic parameters of the system, including the focal length, the scale factors, the skew coefficients, and the principal points. Besides, backlash was proven to be deteriorating on accuracy of external parameters of the system. We will develop the compensation techniques for backlash on each rotating axis. These efforts will benefit reconstruction accuracy in large-angle manipulation. In the second year, we will research into the issues of stereo matching for 3D reconstruction of real objects. The project will focus on releasing restrictions of current approach, which requires high computing power and manual assignment of reconstruction region. Besides, we noticed that there is significant inconsistency in the density of reconstructed 3D feature points acquired from different viewpoints. We will develop registration algorithms and filters to facilitate the integration of data points collected from the maneuver of the PTZ cameras.

Project IDs

Project ID:PB10108-2815
External Project ID:NSC101-2221-E182-006
Effective start/end date01/08/1231/07/13


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.