基于圆柱衍射与离散余弦变换的图像光学加密算法

Image Optical Encryption Algorithm Based on Cylindrical Diffraction and Discrete Cosine Transform

目的为了实现多幅图像的同步加密,并增强加密系统的抗破译能力,提出一种基于圆柱衍射域的相位截断与离散余弦变换的多图像光学加密算法。方法首先引入压缩感知(CS,Compress Transform)方法,对输入明文实施压缩;基于离散余弦变换DCT(Discrete Cosine Transform)对压缩明文完成分解,获取相应的DCT系数,形成系数矩阵;构建迭代复数,将每个压缩明文对应的系数矩阵融合为一个复矩阵,通过DCT逆变换,形成一幅组合图像。联合Hilbert变换与波带片相位模型,构建调制掩码;引入圆柱衍射域的相位截断机制,联合调制掩码,对组合图像实施光学加密,获取密文与私钥。结果实验数据表明,相对于已有的多图像同步加密方法而言,所提算法具备更高的加密安全性,密文熵值以及相邻像素间的相关系数分别达到了7.998,0.0012,且具有强烈的密钥敏感性。结论所提加密算法可以抵御网络中外来攻击,在图像信息防伪领域具有一定的参考价值。

The paper aims to propose a multi-image synchronous encryption algorithm based on phase truncation in cylindrical diffraction domain and discrete cosine transform to realize the fast-synchronous encryption of multi-images and enhance the anti-cracking ability of the encryption system. Firstly, the compress transform (CS) was introduced to compress input plaintext. The compressed plaintext was decomposed based on discrete cosine transform (discrete cosine transform) to obtain the corresponding DCT coefficients. And these coefficients were combined into matrices. Iterative complex function was constructed to fuse the coefficient matrix corresponding to each compressed plaintext into a complex matrix. And a combined image was formed by reversible DCT method. The modulation mask was constructed by combining Hilbert transform with waveband plate phase model. Finally, the ciphertext and private key of combined images were obtained by introducing the phase truncation mechanism in cylindrical diffraction domain and modulation mask. The experimental data showed that compared with existing multi-image synchronous encryption, the proposed algorithm had higher encryption security. Its entropy of ciphertext and correlation coefficients between adjacent pixels were 7.998 and 0.0012, respectively. It also has stronger key sensitivity. This encryption algorithm can resist external attacks in the network, which has certain reference value in the field of anti-counterfeiting of image information.