目的 针对数字图像网络传输安全性和混沌加密算法自适应差的问题,提出一种基于ChaCha20哈希运算的分块扩散自适应图像加密算法(BDCH)。方法 BDCH算法首先通过分段线性混沌映射(PWLCM)产生的混沌序列填充明文图像,使其成为方形图像;其次,利用初始输入密钥及明文图像总和,通过ChaCha20哈希运算生成8×8的初始哈希矩阵,并与PWLCM混沌映射生成的伪随机序列作用,联合产生哈希密钥矩阵,PWLCM的迭代初值选取为初始密钥矩阵均值、初始密钥及明文图像归一化均值;然后,利用Arnold和PWLCM映射同步置乱扩散整幅图像,并分成互不重叠的8×8大小图像块;最后,采用哈希密钥矩阵对图像块进行两轮扩散,完成图像加密。结果 灰度及彩色图像的计算机仿真与性能分析表明,BDCH算法的信息熵、峰值性噪比、密钥敏感性指标优于其他加密算法,并且解决了直接使用初始哈希矩阵会产生的弱密钥问题,密钥空间大。结论 结合同步置乱扩散和哈希密钥矩阵非线性分块扩散的BDCH算法可有效抵抗各种攻击,安全性高、自适应性强,适合各种类型的灰度及彩色图像加密,潜在应用价值大。

Objective Along with the rapid development of network technology, digital images, as the main carriers of information expression, have been widely used in commercial, economic, aerospace, military, defense,and other fields. However,they entail the risk of information leakage in image storage and transmission. The security problem of image data must be solved effectively. Moreover, some image algorithms have poor self-adaptability and weak keys. This paper proposes an adaptive encryption algorithm of blocking diffusion based on the ChaCha20 hash operation (BDCH) to solve these issues. Method First,BDCH translates a plain image into a square image by filling the chaotic sequences that are generated using a piece wise linear chaotic map (PWLCM). Second, a 512-bit initial key and the sum of plain-image pixels are used as input tothe ChaCha2 hash operation to generatean initial 8×8 hash matrix. The matrix is then combined with a chaotic sequence through a PWLCM map, whose inputs consist ofthe average of the initial hash matrix, initial key, and normalized average of plainimage pixels to form an 8×8 hash key matrix. Third, permutation and diffusion are simultaneously implemented on the whole image using an Arnold map and PWLCM.The image is then segmented into an 8×8 array of non-overlapping blocks. In the end, the hash key matrix is used to diffuse the blocks with two rounds and to finish the encryption process. Result The BDCH algorithm can solve the issue of weak keys by co-producing a hash key matrix using the initial hash matrix and the PWLCM chaotic sequences, thereby increasing the key space. Plain-image information is also a part of the input keys of ChaCha20, Arnold map, and PWLCM chaotic map, thereby enhancing self-adaptability and plain text sensitivity. The simulation results and performance analysis on seven gray and color images show that the BDCH algorithm achieves a larger key space, higher sensitivity to both key and plainimages, and faster encryption speed. Conclusion The BDCH algorithm,which combines simultaneous permutation, diffusion, and hash key matrix blocking nonlinear diffusion,can effectively resist various attacks and is suitable for all kinds of grayscale and color images.