Research On Digital Image Compression And Encryption Based On Compressed Sensing

Information technology is constantly evolving,and the amount of multimedia information is also growing rapidly.So many multimedia data,especially image data,in the process of transmission,puts higher requirements on the Internet.However,network optimization is a more complicated problem.Therefore,the way of compressing multimedia image data becomes a better method to solve this problem.In addition,if the compressed data are transmitted over an unsecured channel,the transmitted image data will face the risk of being maliciously transmitted tampered and so on.Therefore,the integrity and security requirements of digital image data is also a very important research direction.Compressed sensing is a new theory of signal acquisition which has been widely applied and studied in various fields.Its compressed sampling model can better solve the image compression encryption problem.The main content of this paper is to propose two image compression and encryption algorithms based on compressed sensing theory.Firstly,in the traditional data transmission environment,a new image compression and encryption algorithm based on compressed sensing is proposed.The algorithm performs randomization processing and noise shaping processing to enhance the sparsity of the data to be measured.This way can improve the reconstruction performance of compressed sensing.In addition,considering the problems of poor randomness and small ciphertext space of traditional one-dimensional chaotic systems,we proposed a new Logistic-Tent-Sine composite chaotic model(LTSS)and apply it to our algorithm.The experimental results prove that the chaotic sequence generated by LTSS model is more random and has larger key space.Finally,this paper improves the parallel compressed sensing model to make it have encryption performance and apply it to our algorithm.Secondly,in the cloud storage environment,this paper proposed an image compression and encryption algorithm based on compressed sensing.Firstly,the original image data is sparse by using discrete wavelet transform,and according to the different characteristics of the obtained low-frequency data and high-frequency data,they are respectively subjected to compression and encryption processing.In order to enhance the security of the algorithm,we extract the user palmprint's competition code feature,and use the biometric embedding algorithm(BDCE)to embed the feature into the ciphertext data,thus we can realize biometric and digital key two-factor authentication.