Research On Image Encryption And Authentication Based On Compressed Sensing In Edge-Cloud

With the development of information technology,an increasing number of users upload images to the cloud-center,so the security of image needs to be guaranteed in the transmission process.Then,due to the inherent characteristics of high redundancy,high data volume and high adjacent pixel correlation of image data,compressed sensing linear image encryption with secret characteristics can be used.The computational complexity of the method is low and can be adapted to the resource constrained wireless sensor network nodes.This thesis proposes a transmission framework based on compressed sensing to protect image security under edge-cloud.It can ensure the confidentiality and integrity of the transmitted image at the same time by only performing linear encryption operation at the resource constrained wireless sensor network node.Firstly,the scheme of protecting data confidentiality based on compressed sensing is proposed.Only computing security can be achieved through compressed sensing,but perfect security is not be achieved.Therefore,the measurements decomposition scheme and the enhanced encryption operation with high computational complexity are used to the trusted edge-server;Then,the error correcting property of Reed-solomon codes and the unidirectionality and robustness of compressed sensing are introduced to ensure the integrity of image transmission on the two paths of “edge server-cloud center” and “wireless sensor network node-edge server” respectively.This thesis main work is as follows:(1)The proposed two-layer encryption framework ensures the confidentiality of data transmission.In the resource constrained wireless sensor network node,compressed sensing is used to encrypt the image with low complexity by simple linear operation,and then the measurements decomposition and enhanced encryption operations are used to the edge-server.(2)The proposed two measurements decomposition schemes can decompose measurements into data with energy or most of energy and data without or with little energy,which are combined with the normalization algorithm and the clustering algorithm respectively.In the follow-up work,chaos system is introduced to scrambling and spreading the data containing energy or most of the energy which may leak plaintext information,and then the statistical characteristics and element position of the data are changed.In the case of improving the security level of ciphertext,the execution efficiency of encryption operation is improved.(3)The introduction of Reed-solomon with error correction not only ensures the integrity of the energy ciphertext transmission on the “edge server-cloud center” path,but also can realize the tamper location and double tamper recovery in tamper attack.(4)The unidirectionalityand and robustness of compressed sensing ensure the integrity of data transmission on the “wireless sensor network node-edge server” path.It is suitable for the authentication of compressed sensing reconstructed image,and can also resist certain noise pollution.