基于面向对象分类的稀土开采区遥感信息提取方法研究

离子吸附型稀土矿是我国宝贵的矿产资源,运用遥感影像分类技术提取稀土开采区可以准确地实现对稀土开采状况的监测,但仅利用光谱信息往往难以保证分类精度。本文以江西寻乌稀土矿区为研究区,以IKONOS影像为数据源,应用面向对象分类方法提取了稀土开采区的遥感信息。针对稀土开采区的分布特点,选择基于边缘的分割算法进行影像分割;结合地形信息、光谱信息及几何信息建立规则集,进行特征提取;最后采用隶属度函数法实现面向对象分类,并与传统的光谱角填图分类进行对比分析。研究结果表明,面向对象分类法提取稀土开采区的总体精度为92.49%,Kappa系数为0.857 6,与传统监督分类方法相比有了很大的提高。

Object-oriented Classification for the Extraction of Remote Sensing Information in Rare Earth Mining Areas

Ion-absorbed rare earth is a valuable mineral resource, and using remote sensing image classification technology to extract rare earth mining area can accurately realize monitoring of Rare Earth Mining; nevertheless, it is difficult to ensure the extraction accuracy only by taking advantage of the spectral information. In this paper, object-oriented classification of IKONOS image was carried out to extract rare earth mining area of Xunwu in Jiangxi Province. In consideration of different characteristics of the rare earth mining area, edge segmentation algorithm was used to segment the image, and the terrain information, spectral information and geometric information were used to establish rule set in order to extract the feature. Finally, object-oriented classification was implemented by membership function method, and compared with traditional spectral angle mapping. The result indicates that extraction accuracy of the rare earth mining area is 92.49% and the Kappa coefficient is 0.857 6 by using the object-oriented classification method. Compared with the traditional supervised classification method, the extraction has been greatly improved.