Research On Image Dynamic Encryption Algorithm Based On Chaotic System And Plaintext Dat

With the rapid growth of 5G individual users and the number of terminals in the Internet of Things,a large amount of data is generated.As image data is more intuitive and perceptual than text data,it occupies less memory and has faster transmission speed compared with video data,so image is widely spread in the network.However,because of the openness and sharing of the network,image data is easy to be illegally obtained,modified and used,which brings security risks to the society and even the country.In order to ensure the reliability and security of digital images in transmission,this thesis studies chaotic system,pixel scrambling,plaintext information acquisition and other aspects of chaotic image encryption algorithm.The main work is as follows:(1)In the aspect of chaotic system improvement,aiming at the problem that the classical chaotic system has low Lyapunov exponents and low security of image encryption algorithm,this thesis proposes an improved chaotic system method based on phase diagram and traversal range of chaotic system.By studying the phase diagram sum and ergodic range of classical chaotic system,this method uses trigonometric function to replace its variables within the ergodic range of each component of classical continuous chaotic system,and completes the improvement of classical Lorenz chaotic system and classical R(?)ssler chaotic system.Aiming at the problems of limited ergodic range and low complexity of discrete Logistic mapping and low security of image encryption,this thesis combined with Hénon mapping and adopted the chaotic system improvement method.According to the ergodic range of each component of the two chaotic maps,seven H-L discrete chaotic systems were constructed.According to the coupling degree of the two chaotic systems,they could be divided into two categories.The simulation results show that the improved chaotic system method proposed in this thesis is suitable for discrete and continuous chaotic systems,and the sequences generated by chaotic systems can pass the NIST test in finite precision digital systems.Compared with the original chaotic system,the improved Lorenz continuous chaotic system and the improved R(?)ssler continuous chaotic system have improved the complexity,the maximum Lyapunov exponents and the randomness of the production sequence.H-L discrete chaotic system has higher ergodic range and Lyapunov exponents than Hénon and Logistic maps alone,and computer processing speed of discrete chaotic system is faster.(2)In the aspect of plaintext related scrambling algorithm,in order to ensure the security of the encryption algorithm and improve the plaintext information acquisition efficiency,this thesis proposes a plaintext information acquisition algorithm based on matrix norm and an improved plaintext information acquisition algorithm based on mean normalization.In order to solve the problem that traditional image scrambling algorithms can only handle equal-length images,two kinds of image scrambling algorithms are proposed.One is an efficient one-dimension image n times extraction scrambling algorithm,which uses matrix norm to obtain plaintext information,and then uses H-L discrete chaotic system to generate sequences to scramble the preprocessed one-dimensional sequences.The other is an image 2D scrambling algorithm with higher security level.This algorithm uses an improved plaintext acquisition algorithm based on mean normalization and an improved Lorenz chaotic system with higher complexity to generate sequences for image 2D scrambling.The simulation results show that the security of two-dimensional scrambling algorithm is improved compared with the one-dimensional image n-times extraction scrambling algorithm,while the complexity and the number of chaotic sequences required are also improved.(3)Because encryption image only uses scrambling algorithm,its security is not high,and it is easy to analyze image information through pixel statistics.In this thesis,a dynamic encryption algorithm for scrambled images is proposed by using diffusion operation.The algorithm selects different encryption processes according to different application scenarios.For scenes with high real-time requirements,the dynamic image encryption method based on H-L discrete chaotic system and one-dimensional image n-times scrambling is selected to encrypt images.For scenes with high security requirements,the dynamic encryption method based on improved continuous chaotic system and two-dimensional scrambling is selected for image encryption.In order to improve the key space of dynamic image encryption algorithm,this thesis introduces fixed key and parameter control key on the basis of plaintext dependent dynamic key.The simulation results show that the plaintext information acquisition algorithm takes into account the efficiency of information acquisition and the performance of encryption algorithm against differential attacks,and can be used in the scenarios with high real-time or security requirements.