|Abstract:||Geometric camera calibration has been a well-studied area in computer vision.
However, with the emergence of imaging systems composing both color
cameras and infrared (IR) cameras, there is a need to develop a more e fficient
calibration algorithm to take advantage of the depth information provided by
IR cameras. In this thesis, we propose a new camera calibration algorithm
that utilizes the depth data. It matches feature points in the master camera
frame with feature points in the slave camera frame, and then it estimates
the extrinsic parameters through a two-stage error minimization process.
Experiments were conducted to test the performance of the algorithm using
different feature detectors and methods to eliminate false matching points.