Aiming at the requirement of obtaining the panorama of the worktable in the automatic cutting system of workpiece groove based on machine vision, an image registration algorithm based on SIFT and improved PROSAC is proposed. First, SIFT algorithm is used for feature detection and feature description. Then, the bidirectional matching and cosine similarity method are used for rough matching of feature points. Finally, an improved PROSAC algorithm is proposed, which purifies the matching points and calculates the image transformation matrix. In image fusion, the weighted average method is used to fuse the overlapping parts of the image to obtain a whole image of the cutting platform. Experimental results show that the algorithm in this paper has been improved in terms of matching accuracy and time-consuming compared with several classical algorithms.

针对基于机器视觉的工件坡口自动切割系统中工作台全景获取的需求，提出了一种基于SIFT和改进PROSAC的图像配准算法。首先，SIFT算法用于特征检测和特征描述。然后，采用双向匹配和余弦相似度方法对特征点进行粗匹配。最后提出一种改进的PROSAC算法，对匹配点进行净化，计算图像变换矩阵。在图像融合中，采用加权平均法对图像的重叠部分进行融合，得到切割平台的整体图像。实验结果表明，与几种经典算法相比，本文算法在匹配精度和耗时方面都有所提高。