中文摘要：

在计算机视觉领域，图像匹配是常见的任务，基于灰度的点的匹配是常见的方法。其中又以抗噪能力强、匹配精度高的归一化互相关（NCC）算法最为常用。然而NCC算法计算量较大，常常难以满足实时处理的要求。通过引入合适的全1矩阵，让其和原图像的数据做卷积，大幅度地降低了NCC的计算量，提高了效率，仿真实验结果也证明了本文方法的有效性。

英文摘要：

In the field of computer vision, image matching is a common task, and point matching based on gray scale is a common method. Among the methods, NCC algorithm with high anti-noise ability and precision is usually used. However, with great computing quantity, NCC algorithm often cannot meet the requirement of real-time processing. In this paper, appropriate matrix with all the elements being ones is introduced. We let it correlate with part of the original image. So the computing quantity of NCC is rapidly reduced and the efficiency is increased. The simulation experiment results also demonstrate the validity of the method presented in this paper.