图像配准及其在目标精确定位中的应用

Image Registration with Applications in Precisely Locating Targets

随着卫星和无人机等空天侦察平台的发展和广泛应用,图像成为精确定位目标的一个重要信息源,而其中一个重要途径就是战术图像与参考卫星图像的配准。由于成像设备、时间和光照条件的差异,战术图像和卫星图像在亮度上存在很大的不同。为解决这些问题,提出了一种基于部分Hausdorff距离的图像配准算法。试验表明,该算法对于噪声、光照变化和局部场景变化具有较强的鲁棒性。

Long-range precise strike is one of the most important features of the modern wars, and precisely locating targets is the foundation of the successful strike. With the rapid developments and applications of various space-based and aerial surveillance platforms, such as satellites and unmanned aerial vehicles, imageries become a very important source of intelligence. Satellite images are usually accurately geo-referenced, whiie mission images can be obtained in real time. So, Fusion of these images is needed to utilize all the information. However, techniques for extracting the information have not been solved perfectly. In this paper, an image registration algorithm based on partial I-Iausdorff distance is presented to integrate all information from mission images and satellite images. Experimental results show that the presented algorithm is robust under noises, illumination variations and partial scene changes.