Image Encryption Algorithm Based On Chaotic System

The rapid development of information technology make people's lives more and more convenient.Among all the data transmitted in the networks,the digital image is one of the most likely to be stolen and modified,as it carries information in visual way.Therefore,it is quite important to protect the contents of images during transmission and thus image security attracts more and more research attentions.Among all the technologies of protecting images,image encryption is a straightforward way by encrypting meaningful images to be unrecognized ones.Only with the corresponding secure key,one can recover the content of the original images.Chaotic systems are suitable candidates for designing image encryption algorithms.Because they have many significantly properties such as ergodicity and initial value sensitivity,and these properties are quite similar to the confusion and diffusion properties in cryptography.When chaotic systems are used to design encryption algorithms,their chaos complexity highly determines the security level of the encrypted results.To obtain cipher-images with high security levels,this thesis first proposes a chaotic model that can generate a large number of new chaotic systems,and then uses the newly generated chaotic systems to design an image encryption algorithm.Theoretical and experimental results show that the proposed encryption algorithm can achieve a high encryption efficiency and security level that can resist various security attacks.The main contents of this thesis are summarized as follows:(1)One-dimensional chaotic maps usually have simple structures and low implementation cost,but their chaotic ranges are limited and their chaotic signals are easy to predict with low computation cost.While the high-dimensional chaotic maps have more complex behaviors and their chaotic orbits are more difficult to predict,but they have high computation cost and certain limitations in real-time applications.To balance the trade-off between chaotic behavior and implementation cost,this thesis proposes a two-dimensional chaotic system based on sine transform by using the nonlinear properties of sine function.A lot of new chaotic maps can be generated using this model and it can increase thecomputational complexity and reduce the computational cost.The chaotic sequences generated by the proposed chaotic system model can pass the pseudo-random sequence of SP800-22 test standard,which is provided by National Institute of Standards and Technology(NIST).The good pseudo-randomness of two-dimensional chaotic system based on sine transform is proved.Simulation and analysis are provided using the bifurcation diagrams,Lyapunov exponents and sequence correlation.The simulation results demonstrate that the proposed chaotic model has better initial value sensitivity and their chaotic orbits are more difficult to predict.(2)This thesis proposes an image encryption algorithm using the new two-dimensional chaotic systems.In order to improve the randomness and security of the algorithm,the algorithm combines permutation and diffusion.Firstly,the original image is randomly cut into blocks.Then,each block is randomly XORed with the chaotic sequences.Finally,an image encryption algorithm based on randomly cutting is designed and simulated.We have also evaluated the performance of the algorithm with respect to image encryption and security analysis.Simulation results show that the algorithm has high key space,strong anti-attack ability and good encryption effect.