Abstract: There are many good qualities in the Normalized Correlation（NC）method.For example，there are high accuracy，high adaptable，strong anti-jamming，little affection of noise and no false match.But the calculation of the method is complex and the amounts of datum in this method are too large.So it cannot be applied in real time systems.This paper proposes a new method to improve searching arithmetic of NC method.The experiments show the new method is higher accuracy and fast speed than the NC method.

Key words: grey level image, template matching, normalized correlation, search method

摘要： 在灰度图像的模板匹配方法中，归一化互相关（NC）法具有很高的准确性和适应性，以及抗干扰能力强，受噪声的影响小和不存在伪匹配等特点，同时对图像灰度值的线性变换具有“免疫性”。但NC方法计算时间长、计算数据量相对较大，所以实时性较差。提出通过改进搜索方法以提高归一化互相关（NC）法的运算速度。实验仿真结果证明了改进的归一化互相关（NC）法比传统的NC方法具有精度高、速度快的特点，有一定的适用价值。

关键词: 灰度图像, 模板匹配, 归一化互相关法, 搜索方法