Abstract: n order to achieve more reliable and secure transmission, image encryption technology has been commonly used for transmission of important image data. However, the encrypted images generated by traditional encryption algorithms are texture-like or noise-like images which can be easily distinguished by human visual system. A meaningful image encryption algorithm in ciphertext is designed and implemented. Combined the compressed sensing of images with the wavelet transform technology to encrypt image which makes the encrypted image on the vision is the same as the ordinary images, reduces the chance of being attacked. The algorithm achieves the following operations by using Matlab programming: reading the secret images, pre-encrypting the secret images, pre-processing the cover images, embedding the encrypted images in cover images; loading the encrypted images, extracting the pre-encrypted images and reconstructing the secret images. For the pre-encryption to the secret image, it mainly uses the compressed sensing of images based on discrete wavelet transform; when embedding the encrypted image on cover images or extracting it from them, the encryption and decryption based on the content of integer wavelet transform is mainly used; while reconstructing secret image, the stagewise OMP reconstruction algorithm is mainly used. Finally, the algorithm realized the meaningful ciphertext image encryption with certain robustness.

Key words: image encryption algorithm, compressive sensing, integer wavelet transform, stagewise OMP, information security

摘要： 为实现更可靠安全的传输，图像加密技术已普遍用于重要图像数据的传输.然而，传统加密算法生成的加密图像为纹理质地图像或者噪声图像，很容易被人类视觉系统所识别.设计并实现密文有意义的图像加密算法，结合图像压缩感知及小波变换技术加密图像，使得加密图像在视觉上和普通图像相同，减少图像被攻击的概率.算法通过Matlab编程实现以下操作：读取秘密图像、对秘密图像预加密、对封面图像预处理、将预加密图像嵌入封面图像；载入加密图像、提取预加密图像以及重构秘密图像.在秘密图像预加密操作时，主要运用了基于离散小波变换的图像压缩感知；在向封面图像中嵌入预加密图像和从封面图像中提取预加密图像时，主要运用了基于整数小波变换的内容加解密；在重构秘密图像时，主要运用了分段正交匹配追踪重构算法.最终算法实现了密文有意义的图像加密，并且具有一定鲁棒性.

关键词: 图像加密算法, 压缩感知, 整数小波变换, 分段正交匹配追踪算法, 信息安全