Abstract: The feature list correlation algorithm can be adopted in the calculation of the similarity measurement of the machine vision object matching.As the result，the processing time decreases considerably，objects can also be clearly recognize due to the high values of the coefficients and the high values of the PSNR.A machine vision objects matching based on the gradient feature list is provide in the paper，in which the feature pixels are added to the feature lists by the non-uniform sampling technique and the gradient normalized cross correlation of feature pixel points is adopted.So the performance of the features list correlation algorithm can be further improved.

Key words: Gradient feature list, non-uniform sampling, matching

摘要： 机器视觉目标匹配中，匹配相似性测量应用特征列表相关算法计算，可有效减少处理时间，并且匹配具有较高的峰值系数和峰值信噪比，可以清晰识别目标。所提出的基于梯度特征列表的机器视觉目标匹配方法，采用梯度特征列表描述图像，对特征像素点非均匀采样，匹配相似测量基于特征像素点梯度归一化互相关，可有效改善特征列表算法性能。

关键词: 梯度特征列表, 非均匀采样, 匹配