Hyperchaotic encryption scheme for hyperspectral images using 3D Zigzag-like transformation and brushing diffusion

Hyperspectral images are crucial in various sensitive areas and applications, necessitating the utmost security measures. To balance safety and efficiency, this paper presents a novel encryption scheme for hyperspectral images using a 4D Lorenz hyperchaotic system, 3D Zigzag-like transformation, and brushing diffusion. Firstly, the scheme utilizes the 4D hyperchaotic Lorenz system to generate pseudo-random sequences for scrambling and diffusion. The initial values correlated with the plaintext are obtained using the SHA-256 algorithm to enhance resistance against differential and plaintext attacks. Secondly, the brushing diffusion algorithm is employed to enhance the avalanche effect without significantly compromising the speed of the encryption process. Thirdly, the 3D Zigzag-like transformation with two key parameters is proposed and applied to reduce intra-band and inter-band adjacent pixel correlations. Experimental evaluations and security analysis confirmed the effectiveness of the scheme. The scheme offers a large key space, quick process, high key sensitivity, acceptable randomness in the ciphertext, and excellent visual reconstruction quality. Furthermore, it demonstrates resistance against common attacks, including statistical, differential, and plaintext attacks. In conclusion, the proposed encryption scheme effectively protects hyperspectral images while maintaining acceptable computational efficiency.