一种快速去除大尺寸工件鬼影图像的精密拼接算法

为了提高大尺寸工件测量的精度和效率，针对现有拼接方法复杂度高且不能很好处理图像鬼影的问题，提出了一种快速去除大尺寸工件鬼影图像的精密拼接算法。该算法先通过SURF算法检测出相邻图像对应特征点，然后结合最近邻匹配法匹配特征点，零均值归一化互相关算法（ZNCC）判断、检测出图像是否存在鬼影；若存在鬼影，则用小波变换融合算法去除鬼影，并求出相邻两幅图像之间的单应性矩阵H，最后实现大尺寸工件精密拼接。结果表明，该算法不仅能有效地去除大尺寸工件图像鬼影，同时提高图像拼接精度和效率。

A Precise Stitching Algorithm for Rapid Removal of Ghost Images of Large Size Workpiece

In order to improve the accuracy and efficiency of measurement of large size workpiece, due to the existing stitching methods which have high complexity and can not properly deal with image ghosting problems,this paper proposes a precision stitching algorithm for rapid removal of ghost images of large size workpiece. Firstly, the SURF algorithm detects the corresponding feature points of adjacent images. Then, the nearest neighbor matching method matches feature points, and zero mean normalized cross-correlation algorithm (ZNCC) judges and detects whether there is ghost. Wavelet transform fusion algorithm removes ghost if ghost exists. In addition, it needs to calculate the homography matrix H between two adjacent images. Finally, it realizes the precision stitching of the large size workpiece. The result shows that this algorithm can not only effectively remove the ghost images of the large size workpiece, but also improve the accuracy and efficiency of image stitching.