Building a stereo camera

Camera calibration is the process of learning a camera's intrinsic and extrinsic parameters from photographs of a calibration object. In stereo camera calibration the location of a second camera is recovered relative to the first, again by using a calibration object. Once this calibration is completed, it is possible to estimate real-world distances to objects in new scenes by using the difference (or disparity) between the two camera images, see reference 1. This is an important process in robotic vision and navigation. In this project you will construct and calibrate your own stereo camera from two cheap digital cameras (e.g. webcams) and use it to estimate real world depths from new images. You will evaluate the accuracy of your system using a structured light depth sensor.