Image Encryption Technology Based On Hyperchaotic System

With the rapid development of the Internet,large numbers of digital images have been generated in different scenes,and in some scenes,encryption operations are required such as personal privacy images and military images.While traditional encryption algorithms such as the AES/DES algorithm are less efficient and less secure for image encryption.Due to the ergodicity,pseudo-random and high sensitivity to initial conditions characteristics of chaotic systems,to introduce chaotic systems to the design of image symmetric encryption algorithms can effectively improve the encryption efficiency.In this paper,based on hyperchaotic systems and some image coding methods,the following work has been done:Firstly,the theoretical basis of cryptography and chaos was explained,and the classification and attack methods of encryption system were explained.The characteristics and judgment methods of chaotic system were pointed out.A simple example is used to illustrate the combination of the two.Secondly,a new image encryption algorithm was proposed.The algorithm combines the hyperchaotic Lorenz system and the Huffman coding method to encrypt the image in binary space.This method is compatible with the One-Time-Pad encryption method proposed by Shannon,which has high security and at the same time obtains some additional compress benefits.Finally,based on hyperchaotic Hopfield neural network and DNA coding,a new encryption algorithm was proposed.The hyperchaotic Hopfield neural network can work in serial mode and in parallel mode.When working in parallel mode,it will improve encryption efficiency.DNA coding calculation is a relatively new calculation method with excellent characteristics of large-scale parallelism,low energy consumption and high storage density.The nonlinearity of the encryption system can be made stronger by the proposed dynamic calculation rules.The simulation results showed that the proposed algorithm can effectively map plain information into pseudo-random information and effectively resist differential attacks,which shows the proposed methods have a bright future of practice.