无证矿山图斑信息的自动提取方法研究

Research on Automatic Extraction of Undocumented Mine Pattern Information

为了自动快速提取无证矿山图斑信息,首先利用面向对象的技术对遥感影像进行分类,获取包括植被、水体、道路、居民地及裸地等分类成果图;然后通过图像差值变化检测方法对变化图斑进行提取,得到面向对象分类的二值图。在此基础上,叠加已有的矿权资料,获得疑似无证矿山图斑信息。最后通过目视判读和野外验证的方法对提取的疑似无证矿山图斑进行筛选,最终确认无证矿山图斑。从本文的研究结果来看,结合变化检测的面向对象分类方法对目标图斑的提取具有一定的可行性.

To automatically extract information of unlicensed mine spot, Firstly using object-oriented technology for remote sensing image classification, to get classification results figure including vegetation, water, roads, residents and bare land .Then changing detection image difference method to extract of the spot and get the second value of object-oriented classification figure.On the basis, overlay the existing mine of information on it, getting suspected unlicensed mine map .Finally through the visual interpretation and field validation method to extract the suspected unlicensed mine spot,then confirm the unlicensed mine spot. From the research results of the paper, combining the changing object-oriented classification method to extract the target figure spot of certain feasibility.