A robust multi-chaotic remote sensing image encryption scheme based on RNA and immune algorithms

Abstract

Remote sensing images have been widely used in the military and other areas because of their rich perceptional data. This makes their visual security critical in practical usage. To address this challenge, an enhanced image encryption scheme is proposed. In the scrambling phase, n bands of remote sensing images undergo Arnold double bit-level per-mutation. This reduces not only the pixel correlation in each image plane, but also between each frequency band. To enhance security, an RNA crossover rule (RNACMO) is introduced. The RNA image is divided into RNA single strands of different lengths using chaotic sequences, and different crossover methods, including single-point and uniform, are adaptively selected according to the number of RNA single strands. RNACMO significantly improves the security level of the scheme. An improved immune algorithm (IIA) is exploited to optimize chaotic function sequences, which improves the chaotic property of the scheme. In experiments, the proposed algorithm achieves average values of 99.6094% for NPCR, 33.4635% for UACI, and 26.7712% for BACI in encrypted remote sensing images, indicating stronger security and better resilience against attacks compared with other encryption algorithms for remote sensing images.