Image Encryption Algorithm Based On Chaotic Mapping And Double Random Phase Encoding Technology

Information security has become a significant issue to research for protecting thesecret information during transmission in practical application, in the era ofinformation. Optical information security technology has been widely used in imageencryption, due to its quickness and high parallelism of the optical informationprocessing system. Currently, various techniques have been proposed for opticalimage encryption. We can choose different encryption methods, or just combineseveral encryption algorithms to achieve higher security. This paper puts forwardseveral novel image encryption algorithms, based on double random phase encodingtechnique, which can be used in the gray image and color image encryption. We alsouse the mean square error function to evaluate the security of corresponding algorithm,and try to improve it in some way. A brief introduction of our researching is asfollow.This paper proposes a color image encryption algorithm based on Logisticchaotic mapping. Firstly, we briefly introduce the basic concept of Logistic mapping,and then analyze its characteristics and advantages in image encryption. Utilizingaffine transform and discrete cosine transform, we achieve the encryption of two colorimages. This encryption algorithm has a large key space, and its resistance againststatistical attack, noise attack and cropping attack is good, i.e., its security is high.This paper also contains an image encryption algorithm based on double randomphase encoding technology. Firstly, we present the characteristics of DCT, andcompress images with it through numerical simulation. Then we encode four imagesinto one by DCT, and scramble it by affine transform. Its phase angle matrix isgenerated by Logistic mapping, whose initial values generated by random function.Compared with the random phase mask plate, the key of this method for generatingphase angle matrix has two known values. Hence, during the transmission of the key,we can undertake it individually through two channels, which can highly improve thesecurity. Then we put the image we get into the input plane of the double randomphase encoding system, and obtain the encrypted image in the output. Numericalsimulation has supported the feasibility of the algorithm. Using correlation coefficientand the mean square error function, we analyze the effect of image encryption, it turnsout the correlation coefficient between the adjacent pixels of the encrypted image issmall, and the key space is large. In the meanwhile, this encryption algorithm has strong resistance against various attacks. So it has high security.The last algorithm, proposed in this paper is a multiple images encryptionalgorithm based on pure phase. Firstly, images are encoded as phase information, andscrambled by random phase mask. Then we put the images in the input plane of theencryption system, and obtain the encrypted image by Logistic mapping andfractional Fourier transform. The encryption system only needs one lens, whichreduces the complexity of the system. In the meanwhile, private keys will begenerated during the encryption process, and each private key can only be used todecrypt one image. Numerical simulation has supported the feasibility of thisencryption algorithm. We use the mean square error function to measure theefficiency of this encryption algorithm, and then improve the algorithm depending onthe measurement.