Research On Decipherment Model And Algorithm Of Sensitive Target For Digital Image

Digital images,as the basic data of geographic information,have become indispensible resources to national economy development and military construction and thus,the security of digital images is of great importance,especially for those images with sensitive targets(e.g.,military equipment and secret services).Therefore,these sensitive targets need to be deciphered to make sure of the security and availability of data to keep national security and interests.Considering the feature of digital images,relevant laws and low efficiency of existing methods to deal with the sensitive targets,the decipherment models are required to be processed with safety confidentiality,data availability,and high efficiency.In this paper,a new decipherment model is proposed based on the targets added with clouds,the PDEs(Partial Differential Equation),the fast marching method,texture synthesis algorithm,etc.The principle and disadvantages of this model are also analyzed.On the basis of this model,the decipherment model to deal with the sensitive targets in digital images is built.In addition,a decipherment system is developed correspondingly to test the efficiency of sensitive targets decipherment.The results of experiments show that the proposed decipherment model has the advantages of great effectiveness,high process performance and strong security in the sensitive targets decipherments.The main work and conclusions in this study include:(1)The basic principle of sensitive targets decipherment models in digital images has been studied through the analysis of decipherment models based on targets added with clouds,the PDEs,the fast marching method,texture synthesis algorithm,etc.This study lays the foundation for further sensitive targets decipherment models construction.(2)Sensitive targets security decryption models in digital images have been built through four main steps:firstly,construct texture sets from digital images processed by nonlinear transformation models;secondly,determine which texture-sampled patch of texture sets should be prior to be processed based on the dynamic programming of decipherment;thirdly,find out other texture-sample patches that are similar to the texture-sampled patch of priority through the eight neighborhood search process and adaptive templates processing methods,then choose the one best fitting into the prior texture-sample patch using the match algorithms based on color and texture feature and substitute it for the origin prior texture sample patch;lastly,repeat the former three steps until the security decipherment of sensitive targets is finished.(3)A decipherment system has been developed to perform the decipherment of digital images.Through the comparison of different decipherment models,it is found that the model proposed in this paper works well in decipherment with great effectiveness,high process performance and strong security.