基于小波变换的数字图像双重随机置乱加密

A Dual Random Scrambling Method for Digital Image Encryption Based on Wavelet Transform

提出了一个用于图像加密的基于小波变换的双重随机置乱算法。算法首先对图像进行小波分解;再对分解结果进行第一次随机置乱。这一置乱有两种方案:第一种是直接对每一分量图像分别进行置乱,第二种是对由所有分量图像拼接而成的图像进行置乱,置乱后再分割成与原分量位置与大小一致的分量图像;然后对这些分量图像进行小波重构,并对重构结果再次进行随机置乱,得双重随机置乱加密图像。实验中,以统计直方图和视觉效果对两种方法作了比较。比较结果表明两种方法的视觉置乱效果都不错,但由第一种方法置乱所得图像的直方图与原图很相似,说明该方法是以位置置乱为主,而由第二种方法置乱所得图像的直方图则与原图差别很大,说明该方法不仅有位置置乱,而且灰度级置乱更明显。可见,第二种置乱方法更有效。

A dual random scrambling method based on wavelet transform is proposed for digital image encryption. The original image is decomposed by 2-D discrete wavelet transform to several sub-images. Then, these sub-images are scrambled by a random shuffle algorithm for the first time, which has two cases. In the first case, every sub-image is scrambled separately. In the second case, a big image merged from all sub-images is scrambled, and the result is re-segmented into sub-images as same positions and sizes as original sub-images. And then, an image is reconstructed from all sub-images by 2-D inverse discrete wavelet transform. At last, the encryption image is obtained by scrambling the reconstructed result again. Experiments show that the two algorithms are both effective at visual evaluation. But the second is better than the first at histogram evaluation.