Research On Image Encryption Algorithm Based On Chaotic System

Image transmission security is related to national security,enterprise information security and personal privacy security,image encryption technology has important theoretical and practical significance.The randomness,non-periodicity,sensitivity to parameter values and initial values of chaotic system make it have advantages in the field of image encryption algorithms.The image encryption algorithms based on chaotic system are more sensitive and secure than the traditional encryption algorithm.This study attempts to apply chaos theory to the field of image encryption and proposes several encryption algorithms with theoretical and practical significance.The research contents are summarized as follows:(1)Considering the special data form and high security requirement of 3D image,a dynamic RNA mutation encryption algorithm based on memristor chaotic system is proposed.Firstly,the dynamic characteristics of the nonlinear system with dual memristor are analyzed,including phase diagrams,Lyapunov exponential spectrums and bifurcation diagrams.Then,according to the combined information of multiple three-dimensional images,the hash values are produced to generate the chaotic sequence.In order to improve the security of the system,three-dimensional Arnold algorithm is used to dynamically scramble and diffuse vertex coordinates,and the parameters of Arnold algorithm are determined by chaotic system.Finally,based on the chaotic system,the genetic variation algorithm of dynamic crossover and mutation of RNA are designed based on the dynamic coding and dynamic scrambling and diffusion of encryption,and the cross-base position and the mutation codon species are randomly selected by the chaotic system to increase the unpredictability of the algorithm.The effectiveness of the algorithm in encrypting multiple 3D images and its ability to resist exhaustive,statistical,and differential attacks are tested experimentally.Simulation results show that the algorithm can improve the security of the image encryption algorithm,and it is suitable for three-dimensional images and color images.(2)Aiming at data redundancy and security requirements of color images,a dynamic DNA mutation encryption algorithm based on laser chaotic system is proposed.Firstly,the chaotic characteristics of autonomous laser system are analyzed by phase diagrams,Lyapunov exponent spectrums and bifurcation diagrams.Then,the hash values calculated are used to change the initial conditions of chaotic system and generate chaotic sequence.The back propagation neural network based on particle swarm optimization is used to compress the color images.Finally,based on the pseudo-random sequences produced by the chaotic system,dynamic codec and dynamic scrambling diffusion encrypt images,a dynamic crossover and mutation gene mutation algorithm are designed.The chaotic sequences randomly select base combinations of different positions and lengths as mutant genomes to enhance the flexibility and unpredictability of the algorithm.The results show that the algorithm can compress and encrypt color images and resist exhaustive,statistical and differential attacks.Even with a compression ratio of 5,the peak signal to noise ratio of the image can reach 35.1643.Simulation results show that the algorithm can reduce the amount of ciphertext data while ensuring the security and clarity of the image.(3)In order to improve the security and robustness of zero-watermarking,a zerowatermarking encryption algorithm for color images based on dynamic modified singular value decomposition of five-dimensional chaotic system is proposed.Firstly,the characteristics of 5D chaotic system are analyzed by phase diagrams,Lyapunov exponent spectrums and bifurcation diagrams.Then,the feature information of low frequency region of color image is extracted by wavelet transform.Based on chaos correction factor,the feature matrix is extracted and encrypted by modified singular value decomposition.Finally,watermark images and zero-watermarking are encrypted by scrambling and diffusion based on the chaotic system,the correction factors generated by the chaotic system are designed to dynamically correct the singular value encryption feature matrix in the feature matrix,which improves the system security and eliminates the false alarm phenomenon in the copyright judgment stage.In the authentication process,the watermark image can be extracted from the color image and the registered zero-watermarking,and the copyright ownership can be determined according to the registration time.The security,uniqueness and robustness of the proposed algorithm are tested by experiments.Under filtering,compression,rotation,noise,cropping and combination attacks,the normalized correlation coefficient values of the images are still greater than 89%.Simulation results show that the zero-watermarking generated by the algorithm can be both unique and robust while ensuring security.(4)In order to improve the security and efficiency of image encryption,a fraction-order chaotic map and an image encryption and hiding algorithm are proposed.Firstly,a novel fractional-order chaotic map is proposed,and its dynamics are analyzed by phase diagrams,Lyapunov exponent spectrums and bifurcation diagrams.Then,color images are compressed by back propagation neural network,and the images are encrypted by spiral scrambling and XOR diffusion based on chaotic map to ensure the security of ciphertext.In order to hide the ciphertext image,the carrier image is decomposed by the lifting wavelet transform.In the embedding stage,the chaotic sequence is used as the embedding coefficient to hide the ciphertext dynamically,and the visual security image is generated.The chaotic system provides dual protection for image encryption and hiding to improve the unpredictability of the algorithm.Experiments test the hiding effect of ciphertext and the ability to resist exhaustive,statistical,differential,noise,and cropping attacks.Under various attacks,the peak signal to noise ratio of the image is still higher than 31.4.Simulation results show that the proposed algorithm can hide three times the amount of data compared with the carrier image,and effectively save the encryption and decryption time and transmission cost.