Abstract: A new coordinate calibration algorithm based on vanishing point is proposed and applied to moving vehicle tracking and detection. According to this algorithm, coordinate calibration can be implemented with knowledge of the vanishing point and the position of scaled objects. It is more robust and flexible than others because the configuration and parameters of camera are not required. Foreground matrix is presented by 1-D array instead of 2-D array, which effectively reduces the space complexity of the algorithm.

Key words: vehicle tracking and detection, vanishing point, coordinate calibration, digital image processing, moving vehicle

摘要： 提出一种新的基于消失点的坐标校准算法，并将此算法应用于运动车辆的跟踪检测。该算法只需要知道消失点和已知尺寸物体在图像中的位置就可以实现坐标校准。与其他算法相比，该算法不需要预先知道摄像机的配置和参数，具有更广泛的适应性和更高的灵活性。在算法实现的过程中，使用了一维数组来表示前景分布矩阵，而非二维数组，从而有效地降低了算法空间复杂度。

关键词: 车辆跟踪检测, 消失点, 坐标校准, 数字图像处理, 运动车辆