一种基于小波表示和Hartley变换的图像匹配算法

A kind of image matching algorithm based on wavelet expression and Hartley transformation

探讨了一种基于小波金字塔表示和离散Hartley变换的快速图像匹配方法,它采用图像的灰度作为特征空间,以去均值的归一化互相关系数作为相似性准则,利用小波金字塔作为搜索策略,采用离散Hartley变换(DHT)计算互相关,减小了运算量和存储量。实验结果表明这种算法能节省大量运算时间,极大提高了匹配速度,定位精度也没有降低。

The article discussed one kind to express and to be separated the Hartley transformation based on wavelet pyramid the fast image match method, it used the image the gradation to take the feature space, by went to the average value the normalized mutual correlation coefficient to take the similar criterion, took the search strategy using wavelet pyramid, used discrete Hartley transform （DHT） computation mutual correlation, reduced the operand and the reserves. The experimental result indicated that this algorithm can save the massive operation time, raised the match speed enormously, the pointing accuracy had not reduced.