图像基于小波包变换的多尺度匹配算法

基于梯度算子的图像匹配方法存在收敛速度慢、迭代容易落入局部最优点等问题,针对这些问题,提出了一种基于小波包变换的多尺度图像匹配算法.该算法利用小波包的多尺度特征对图像进行不同分辨率的分解,先对低分辨率的子图像进行匹配,再根据该结果对高分辨率的子图像进行匹配.实验表明该算法匹配准确度高,计算速度较快且减小了迭代落入局部最优解的概率.

Image Based on Small Wave Packet Transformation Multi-criteria Match Algorithm

Based on gradient poerator vision match method is it disappear slow is it take the place of easy to fall into some advantage issue most to changing to exist,to question these,this text put forward one match algorithms on the basis of light wave many yardstick vision varied to wrap up.Algorithm this utilize little many yardstick characteristic that wave wrap up go on with to resolve resolution ratio to vision,it is match to go on to low sub vision of resolution ratio first,match according to this result sub vision in high-resolution and then.Experiment indicate algorithm this match accuracy to be high,very fast is it fall into some optimum probability that solves to take the place of to change to reducing computational speed have.