基于序列图像特征配准的相机旋转补偿算法

为了解决测量过程中相机沿轴向运动时所产生的旋转引起的测量误差,提出了一种基于特征点的图像配准和最小二乘估计的相机旋转运动精确估计算法。此算法利用序列图像帧间的相关性，通过对相邻两帧图像进行运动分析，由Harris算子进行特征点的检测，然后再基于这些特征点的模版匹配方法对帧间图像进行配准，再根据多个特征点的运动矢量通过最小二乘估计获得相机的运动参数。最后对这些运动参数进一步分析，在计算中将获得的运动参数对相机沿轴的旋转运动进行补偿。该方法克服了测量中相机旋转对图像处理和测量精度的影响，弥补了因相机旋转引起的测量误差大的缺陷,提高了测量精确度。试验和仿真结果表明该方法在帧间旋转角度较小时能够实现对相机沿轴旋转运动的精确补偿，且时实性有待提高。

A camera rotator compensated algorithms based on sequence image frame matching

In order to eliminate the measurement error caused by camera rotator during moving along the axial, a suit of high precision camera moving estimate algorithms based on character points image matching and least-squares estimate. Recur to the correlation of sequence image frame, and by analyzing the moving parameter of two continuous image frames, in the algorithms the character points on image are detected by Harris operator in the first; then match the character points based on image template in sequence frames; and moving parameters of the camera are obtained through least-squares algorithm last. By analyzing those parameters farther, the coordinate that laser spot relative to target center has been compensated, as a result, realization camera rotation compensated. Because the influence to image measurement precision by camera moving is overcome, the vice of low measurement precision on account of camera moving is remedied, and the measurement precision is heighten. The results of test and simulation show that the algorithms are effective and applied , but the real-time capability of the algorithms will still be improved.