基于变化检测的投资项目搜索技术研究

Research of Investment Projects Searching technology based on Images change detection

一些露于地表的固定资产投资项目可用遥感影像变化检测的方法自动的搜索并检测出来。其间需要完成对两幅不同时向的影像完成几何校正、影像配准、直方图匹配、影像做差、二值化、小斑去除、搜索结果修正等一系列工作。如果算法参数设置合理,其中大部分操作流程可以使用计算机自动实现。本文详细描述了本智能搜索技术的核心算法与技术流程,并通过基于TitanImage二次开发工具实验证明了其可行性。

Some Fixed Assets Investment Projects which are exposed on the land surface can be automatically detected by change detection methods of remote sensing images. A series of operations on two images of different times must be finished before detections, such as the Geometry Correction, Image Registration, Histogram Matching, Image Subtraction, Image Binarization and Spot Reduction and so on. All these can be automatic performed by computers if proper algorithm parameters are set. The kernel algorithms and technological flow are discussed in this paper, and it was implemented in the statistic system by using Titan Image Software Tool kit, which proved the application feasibility.