Automated vehicle counting technology has been in use for many years, but developments in automated pedestrian counting technology have been limited. Pedestrians are more difficult to detect, track and count, than motor vehicles because their paths are much less constrained. In Phae I, we propose to develop a new system for pedestrian detection, tracking and counting. In this system, both IR LED stereo camera and laser scanner are selected for use. Although the stereo camera is excellent in detecting pedestrians through 3D human body concave feature extracted from the disparity map, it is less effective in separating pedestrians walking in a group. To overcome this problem, we propose to use a high resolution laser scanner and develop a set of transformations to map the pedestrian coordinates onto the disparity map in the camera coordinate system. In our system, each pedestrian will be assigned with a unique track, which will be estimated using both feature matching and kinematic motion. A new data association method is proposed, which is effective for extremely small size of samples common in pedestrian tracking applications at street intersections. Our system can be deployed at any street intersection for reliable pedestrian detection, tracking and counting.