Double color images compression–encryption via compressive sensing

In this paper, a new double color images encryption algorithm is introduced based on 2D compressive sensing and wavelet basis, which can realize image encryption and compression simultaneously. Firstly, the original images are equally divided into non-overlapping blocks and compressed via compressive sensing technique, in which the key-controlled pseudo-random measurement matrices are constructed using three improved one-dimensional chaotic maps. Thus, only employing a few keys in the chaotic maps avoids the transmission of the Gaussian matrix. Then, combine the obtained measurements and shuffle the combined measurements through Latin squares. Finally, to enhance the security of the cryptosystem, the compressed and shuffled image is further diffused using the key streams generated from both the chaotic maps and the plain images. The combined image can be reconstructed by decrypting the cipher images and utilizing a recovery algorithm. The proposed scheme can efficiently reduce data volume and simplify keys. It improves the efficiency of transmitting data and distributing keys. Experimental results and security analysis demonstrate the effectiveness and the security of the proposed algorithm.