Research Of Wavelet Packet Digital Image Hiding Technology Based On Chaos Encryption

The military concealment communication(MCC) has become a research focus in the information war. The safety implementation of image concealment communication(ICC), which is an important branch of MCC, plays important role of combat support. In this paper, the wavelet packet image hiding algorithm based on chaos encryption and singular value decomposition(SVD) is proposed after the study of chaos theory, wavelet packets transform (WPT) theory and SVD theory.The image encryption algorithm based on encrypted chaos matrix groups is proposed to improve the disadvantage of short period of chaotic maps and the weak security performance of chaotic encryption alone. Compared with Logistic map and Tent map, Hybird map has the higher complexity and the more unstable motional orbits. For this reason, the binary chaotic sequence whose period has been extended M times is generated by the use of Hybird map firstly. Then, encrypted chaos matrix groups is got to achieve the purpose of substitution encryption of the image bit planes. The scrambling encryption is accomplished to resist the attack of statistical analysis based on the queue transformation after the substitution encryption. The simulation results show the effectiveness of the algorithm. It can conquer the problems of short period of chaotic maps and the weak security performance of chaotic encryption alone, and resist a variety of attacks.The image hiding algorithm based on WPT and SVD is proposed to improve the poor robustness of the wavelet packet domain image hiding algorithm resisting geometric attack and the capability of information embedded. In the pretreatment stage, the secret image is encrypted firstly. Then bit-plane decomposition is accomplished. In the embedding stage, first of all, the wavelet packet 2-decomposition is applied to the carrier image. Secondly, the SVD is applied to the low-frequency coefficient matrixes for the purpose of hiding the bit planes including important information into the singular value matrixes. At last, the coefficient matrixes with higher energy-entropy is selected from the remainder coefficient matrices as the carries of the bit planes with less important information whose hiding location is adaptively determined by the carried image. The simulation results show that, the algorithm has strong robustness in the cas,es of geometric attacks like rotation and normal image processing such as noise, cutting and filtering and so on.