Research Of Image Encryption Algorithm Based On Chaotic Mapping And AES Algorithm

With the advent of the information society,multimedia information is becoming more and more abundant.Information technology brings convenience to human life and work,and it also brings a series of worrying security issues.Because images are used on a large scale in the Internet,especially when they have more important information,images are vulnerable to attacks and destruction in network transmission,and information security incidents occur frequently,which has a serious impact on society and individuals.Therefore,image encryption has been closely watched by people,and image encryption technology has become a research hotspot.In the field of image encryption,AES(Advanced Encryption Standard)encryption algorithms have attracted the attention of many experts and scholars,and new image encryption algorithms related to AES have been proposed.In dealing with image encryption,AES encryption algorithm has a more complex replacement and replacement structure,and the encryption effect is better.However,it has the defects of small key space,fixed S-box and round key,which makes its security decrease.Therefore,this paper proposes an image encryption algorithm based on improved Henon mapping and AES combination to improve the overall security of AES encryption algorithm.Firstly,the fourdimensional Henon map has at least two positive Lyaponuv exponents,so it has hyperchaotic characteristics and very good pseudo-randomness and high chaotic characteristics and more complex nonlinear dynamic behavior,and it can generate huge key space,therefore,the hyper-chaotic sequence is obtained by using the four-dimensional Henon mapping to solve the problem of small key space in the AES encryption algorithm.Secondly,by using the average value of the plaintext image pixels as the parameters to intercept the chaotic sequence as the training sample of the BP neural network,and the related features of the plaintext image are introduced,which can effectively resist the plaintext image attack and the ciphertext image attack.Thirdly,the BP neural network can generate very complex nonlinear mappings,so the general chaotic sequence data can be obtained by training a specific BP neural network to obtain a highly chaotic sequence with high randomness and disorder.Therefore,a complex nonlinear chaotic sequence is obtained by BP neural network training to design the S-box and round keys in each round of AES encryption.Therefore,the problem that the S-box and the round key are fixed is effectively solved,and the randomness of the key is also greatly improved.Finally,the ciphertext is obtained by two rounds of AES encryption.Simulation experiments and performance analysis of typical images in the field of image encryption,including histogram analysis,statistical analysis,differential attack analysis,selective plaintext(ciphertext)attack analysis,ciphertext attack analysis,known plaintext attack analysis,and key space analysis,the algorithm in this paper can effectively achieve image encryption and resist multiple attacks,and it has good security.In the comparative analysis with related typical algorithms,the proposed algorithm has certain comparative advantages and achieves the goal of the research.