Abstract: In order to evaluate the image scrambling degree effectively，a novel evaluation method on image scrambling is designed.In the method，the characteristics of the ideal scrambled image is analyzed，and a reference sequence is constructed.And then after the scrambled image is divided into several blocks，and the variances of each block-image is abstracted to form a comparatives sequence.Finally the gray relational degree between the two kinds of sequence is calculated to evaluate the image scrambling degree via gray relation analysis.Experimental results show that the method not only is simple and fast，but also can do without the origin image.Compared with some existing methods，it provides with more reasonable results.

摘要： 为有效评价图像的置乱程度，设计了一种基于灰色块分析的图像置乱效果评价方法。该方法首先分析置乱图像的特点，归纳理想状态下置乱图像的特征，并构造其特征序列作为参考序列；然后将置乱图像分块，并计算各子块图像的方差，形成小样本方差序列，对该序列进行预处理后作为比较序列。最后，利用灰色关联分析理论计算两类序列之间的关联程度，得到图像置乱效果的性能指标。实验显示，用该方法评判图像置乱度不仅计算简单、运行速度快，而且不需要原图像的参与，就能得到更加客观、正确的评价结果。