摘要:
介绍了一种新的基于信息融合的TM影像和GIS矢量数据自动配准方法。首先建立TM影像和GIS矢量数据的近似变换关系,在此基础上以矢量地图上的信息(水系、道路网等)为先验知识,利用多尺度模板匹配法对TM影像中的线状信息进行自动提取。把自动提取的结果作为像方观测,矢量地图数据作为地面观测,通过广义点摄影测量原理和方法,解算出TM影像的外参数,实现新的TM影像和旧的GIS矢量数据的自动配准。实验结果表明,一般情况下,本文方法达到了像素级精度,且大大提高了工作效率。
Abstract:
This paper introduce a new method for automatic registration based on amalgamation of information on TM image and GIS vector data. Firstly, the paper build up the transform relationship between TM image and GIS vector data approximately, then take the information(water system, road network etc) on vector data as transcendent knowledge, extracting line-like object on TM image using multi-scale template automatically. Finally, take the results extracted automatically as image-based observation, the data of vector as ground-based observation, numerate rectify parameter of TM image by the principle of generalized point photogrammetry . In this way, the automatic registration on new TM image and old vector map is realized. Generally speaking, the result of the experiment indicates that the method can not only reach pixel precision, but also improve the efficency of data processing.
下载:
