Research On Image Encryption Algorithm Based On Scrambling And Diffusion Mechanism

Since entering the Internet era,the rapid development of network and communication technology has made people’s lives more convenient.Digital images are commonly used for information transmission in daily life,so the scale of their transmission in the Internet is getting bigger and bigger.Since it may contain a large amount of confidential information and personal privacy information,it will cause serious consequences if it is intercepted by others in the process of information transmission.Therefore,the security of digital images has also received extensive attention.An important way to protect the security of image data is image encryption.Traditional text information encryption is not suitable for image encryption,because image information has its unique characteristics,including a large amount of data,strong correlation between adjacent pixels,etc.The research points out that the pseudo-randomness,ergodicity and initial value sensitivity of chaotic systems are similar to those of cryptosystems.Therefore,a large number of scholars have combined image encryption and chaos theory to develop many excellent image encryption algorithms.Based on the existing research work,combined with the chaotic system,this thesis proposes two image encryption algorithms.The research contents are as follows:1.In order to improve the complexity and efficiency of encryption algorithm,an image encryption algorithm based on image segmentation and double diffusion is proposed,which is based on scrambling-diffusion-block diffusion structure.Firstly,the initial values and parameters of Logistic-Sine-Cosine(LSC)chaotic system are obtained by SHA256 algorithm and plaintext image,and one-dimensional chaotic sequence is obtained by iterating LSC chaotic system.Secondly,the chaotic sequence is sorted from small to large,and the sorting matrix is obtained according to the position transformation of the sequence points,and the pixels are scrambled by the sorting matrix.Then in the diffusion phase,the image is diffused from the four tops to the center at the same time.Finally,the image is divided into horizontal blocks,and the initial values and parameters of LSC map are obtained by SHA256 algorithm,and then chaotic sequences are obtained and XOR operations are performed with pixels to complete block diffusion encryption.Finally,block images are merged horizontally.The algorithm improves the operation efficiency of the diffusion phase by diffusing from four tops at the same time,and increases the complexity of the algorithm through double diffusion,and multiple encryption processes make the key space of the algorithm large,which ensures the security of the algorithm.2.Aiming at the shortcomings of one-dimensional chaotic map and multi-dimensional chaotic map and the low encryption efficiency,an image encryption algorithm based on cascaded chaotic map and improved Zigzag transform is proposed.The shortcomings of Logistic-Logistic-Tent-Cascaded(LLT)mapping are improved and introduced into the encryption algorithm.First,the plaintext image and the SHA256 algorithm are combined to calculate the initial value and parameter value of the improved LLT map(NLLT map),and the chaotic sequence is obtained by iterating the NLLT map.Then the scrambling algorithm is applied to the plaintext image,and the improved Zigzag transform is introduced to scramble the positions of the pixels.Finally,in the diffusion operation,a wave-like diffusion sequence is designed to make the diffusion process more complicated.The algorithm improves the LLT map to make its chaotic characteristics better.The scrambling and diffusion process only needs one round to complete the encryption,which makes the algorithm run efficiently.The simulation results and security experimental analysis show that the proposed two encryption algorithms have good encryption effect,which can ensure the security of ciphertext images and high operating efficiency,and ensure the safe transmission of ciphertext images.