In this paper, a new real-time and intelligent target tracking system basing on the pan/tilt controlled-stereo camera embedded on an UGV (unmanned ground vehicle) is proposed. In the proposed system,
firstly the face area of a moving person is detected from a sequence of left images captured by a stereo camera using a threshold value of YCbCr color model and then, the stereo camera is kept to track the
moving target in real-time by controlling the pan/tilt system. Secondly a depth map of the target face is extracted by using a relationship between the geometry of a stereo camera system and the disparity map of a rectified stereo image. Finally, from this extracted depth map, a distance from the camera system to the target and the coordinates of the target are calculated and these target coordinates are used for extracting 3D information of the target to plan paths and construct the coordinates maps. From some experiments using 1,280 frames of the input stereo image pairs, it is analyzed that displacement on the center position of the face after tracking is kept to be very low values of 0.6% for 1,280 frames on average. Also, the error ratio between the measured and computed distance values of the target from the camera system is kept to be very low value of 0.5% on average. In addition, the proposed system has achieved a speed of 0.04 sec/frame for target detection and 0.06 sec/frame for target tracking. These experimental results suggest a possibility of implementing a practical stereo camera-based automatic
target tracking system in the UGV having a high degree of accuracy and a very fast response time.