在监控应用中 ,由于场景是已知的 ,因此可以使用背景减去法检测运动对象 .当摄像机进行扫描和倾斜运动时 ,需要使用多个图像帧才能完整地表示监控场景 .如何组织和索引这些背景帧属于摄像机跟踪问题 .提出一种无需摄像机标定的背景帧索引和访问方法 .这一方法需要使用图像配准技术估计图像初始运动参数 .提出一种屏蔽外点的图像配准算法 ,综合利用线性回归和稳健回归快速估计初始运动参数 .为了快速计算连续帧之间的运动参数 ,提出一种基于四参数模型的优化算法 .利用非参数背景维护模型抑制虚假运动象素 .室内和户外实验结果表明本文方法是有效的 In the monitoring application, since the scene is known, background subtraction can be used to detect moving objects, which require multiple image frames to fully represent the monitored scene when the camera is scanning and tilting. How to organize and index these The background frame belongs to the camera tracking problem.This paper proposes a method of indexing and accessing the background frame without camera calibration.This method needs to use the image registration technique to estimate the initial motion parameters of the image.This paper proposes an image registration algorithm which shields the outside points, Linear regression and robust regression to quickly estimate the initial motion parameters.In order to quickly calculate the motion parameters between successive frames, an optimization algorithm based on the four-parameter model is proposed.Nonparametric background maintenance model is adopted to suppress the false motion pixels.Indoor and outdoor experimental results It shows that our method is effective