Study On Integrity Authentication Algorithm For Hyperspectral Remote Sensing Image Based On Perceptual Hashing

Hyperspectral remote sensing(HSRS)image,an important national strategic information resource,is widely used in different fields,such as ecological construction,modern agriculture and forestry management and national defense construction.The further development of computer technology and Internet technology promotes the continuous expansion of the application field to a deeper level.However,the integrity of HSRS images is facing unprecedented challenges,it is more conveniently to modify the image content intentionally or unintentionally,resulting in the application value of the HSRS image is difficult to be guaranteed.Therefore,efficient and reliable integrity authentication technology for HSRS image has become the key to ensure the practical use value of the image.As an emerging data authentication technology,perceptual hashing technology has been maturely developed and applied in the fields of multimedia data,which is the important foundation for HSRS image integrity authentication.However,there are few researches on the application of perceptual hashing technology in the integrity authentication for HSRS image.This research aims at the integrity authentication of HSRS image,combined with the data characteristics of HSRS image,and carries out the research on the integrity authentication algorithm of HSRS image based on perceptual hashing.The main contents are as follows:(1)The differences between different methods in the integrity authentication process are discussed and summarized.Taking the GF-2 image integrity authentication problem as a case,the main framework of remote sensing image integrity authentication algorithms based on perceptual hash are researched from the aspects of image feature extraction,hash sequence generation perception,image integrity authentication and algorithm analysis,which lays the foundation for the integrity authentication algorithms of HSRS image.(2)By analyzing the characteristics of HSRS image,the difference between the integrity authentication algorithm of HSRS image and other types of data based on perceptual hash is studied.In addition,a novel perceptual hashing algorithm for the authentication of HSRS image integrity considering the spectral characteristics of the image is proposed.There are two aspects of feature extraction,the spatial characteristics of image are extracted by Zernike moment,moreover,the classification results of image are extracted based on the K-means clustering method as the indirect expression of the spectral features.Finally,the perceptual hash sequences for image integrity authentication are generated.The research shows that this algorithm which has a well performing in robustness,separability and security can effectively locate the tampering of ground objects of HSRS images.(3)In order to further improve the accuracy of HSRS image integrity authentication,a HSRS image integrity authentication algorithm combining the edge feature is proposed.By analyzing the existing image feature extraction methods,the Canny operator is used to extract the edge features of HSRS image as the spatial texture features of the image,and the perceptual hash sequence of the image is generated by combining the spectral features of the image.The results show that the proposed algorithm has better performance,robustness and security in the process of integrity authentication of hyperspectral images.In additional,the proposed algorithm can detect the attack of specific bands innovatively.