混沌序列具有易生成性、对初始条件强敏感性、可完全重现性以及整体的伪白噪声统计特性等特点，同时混沌序列的离散映射序列也具有其相似的特性.基于以上特性，给出了一种基于混沌序列的图象置乱加密算法，其首先，以初始条件为密钥生成混沌序列，并将它映射为 2 K值混沌序列 ;然后依据该序列随机地修改图象每个像素点的灰度值.同时为了度量加密或置乱的程度，还给出了一种基于图象局部方差的图象置乱度的定义.实验结果证明了该算法实现简单，计算量小，且图象的解密结果对混沌序列的初始值有较强的依赖性，安全性高，同时也证明了给出的图象置乱度能更好反映图象加密前后的视觉差别和加密效果.

Chaotic sequences have several good properties , such as the ease of their generation, their sensitive dependence on its initial parameters, and specially white noise alike, at the same time it can be reproduced precisely based on the initial condition, and their discrete mapping sequences have a same property.Because of these good properties, chaotic system can be used to image encryption and image preprocessing. In this paper, image encryption arithmetic is proposed based on chaotic system. First, based on the key (initial parameter is the key, here), the chaotic sequences can be generated based on key, then it is mapped to the discrete 2 K value sequences, according to the discrete sequence, gray value of pixel is modified randomly, it is obvious that if attacker dose not know the key, the encrypted image will be looked like white noise and can not be rebuilt. There are many image encryption methods, and different arithmetic gets different results, but up to now there is no suit way to evaluate the result of image encryption. Based on image local variance and human vision system, an efficient definition of encryption degree is proposed. Preliminary results are satisfactory, and security of image completely depend on the key, slight error key results a different result. Hence, it can be used to encrypt image data.