Remote sensing image encryption algorithm based on DNA convolution

Remote sensing images carry geospatial information with significant military and economic implications. While the open network environment facilitates efficient image transmission, it also raises security concerns. To address this, we propose a remote sensing image encryption algorithm based on a novel Exponent-Sine-Logistic (ESL) map and DNA convolution. The algorithm employs a scrambling-diffusion framework within the DNA domain. In the scrambling phase, the DNA codes of multiple images are scrambled using a chaotic sequence generated by the ESL map. In the diffusion phase, the scrambled DNA codes are further modified using DNA convolution, integrating multiple DNA operations to enhance security. Additionally, the algorithm improves efficiency through fast dynamic DNA encoding and decoding methods. Experimental results show that the algorithm achieves an information entropy of 7.9998 and encrypts four 256 x 256 images in just 0.7455 s. The algorithm demonstrates superior security and efficiency, effectively resisting common attacks, making it highly applicable in military and surveying fields.