Investigation Of Optical Image Encryption Techniques Based On Digital Holography

With the development of Internet and the increase of the requirement for image transmission, information security and data encryption are becoming more and more important. Many optical systems have been proposed and used in data security field for the distinct advantages of processing complex two-dimensional data in parallel and transmitting at great speed. By analyzing and discussing the techniques of Fourier transform, fractional Fourier transform and pixel scrambling, methods of Fourier transform encryption as well as fractional Fourier transform encryption have been proposed and numerically simulated. The principle work is as follows:(1) The flexibility and reliability of optical image encryption methods based on Fourier transform are theoretically and simulatively proved. It is shown form numerical simulation that the original image can be reconstructed only when part of the encrypted image is used. The relationship curves between pixels number of the decrypting image and the energy, noise as well as SNR of the decoded image are given. By introducing pixel scrambling, the security strength of proposed method is greatly improved.(2) Based on the space-frequency joint representation features of fractional Fourier transform, a novel optical image encryption method and an improved color image encryption method are both proposed. These methods are proved by simulation, and the relative error between original image and decoded image versus the deviation of different fractional orders Ap is also discussed. Compared with Fourier transform encryption, the security using these two methods for optical image encryption is greatly improved due to the introduction of the fractional Fourier transform.(3) By analyzing the effects of linear Canonical transform, a spatial light modulator and pixel scrambling on the Wigner distribution function, we derive a matrix method for automatically and efficiently determining the position, spatial extent, spatial frequency extent and the space bandwidth product of the signal at any point in the encryption process including the values of the these parameters in the final encrypted image. This method also allow us to select the most efficient way from the bulk optical systems and available recoding methods, i.e., identifying the space bandwidth product of the input signal, the minimum size of lens apertures, the sample space and total sample number of CCD and spatial light modulator, the need of magnification before the signal passes through a SLM or when the signal reaches the camera.