基于张量投票的纹理合成技术在遥感影像信息隐藏中的应用

Application of Texture Synthesis Technology Based on Tensor Voting in Remote Sensing Image Information Hiding

信息隐藏是遥感影像信息交换环节中必不可少的步骤。本文将基于张量(tensor)的影像纹理合成技术应用于遥感影像信息隐藏领域,首先应用影像分类方法将空洞附近的影像按照纹理进行分类;然后利用张量投票技术解决空洞部分影像与周围影像的连通问题;最后利用基于马尔科夫随机场的纹理生成技术完成空洞部分的纹理的自动填充。试验表明,所述方法不仅达到了信息隐藏的目的,而且纹理过渡平滑,达到了预期的目的。

Information hiding is an indispensable step of remote sensing image exchange and disclosure. In this paper,technology of image texture synthesis based on tensor imaging is used in remote sensing information hiding areas. First,image classification method is used with the image near voids based on texture,then tensor voting technology is used to solve problem of communicating voids with the surrounding image,finally,Markov random field texture generation technology is used to automatically fill the voids with the texture. Experiments show that the methods not only achieve the purpose of information hiding,the texture transition smoothly,and the intended purpose is achieved.