基于面向对象的铁尾矿信息提取技术研究--以迁西地区北京二号遥感影像为例  

Information extraction technologies of iron mine tailings based on object-oriented classification:A case study of Beijing-2 remote sensing images of the Qianxi Area,Hebei Province

作　　者：范莹琳 娄德波[1] 张长青[1] 魏英娟 贾福东[1] FAN Yinglin;LOU Debo;ZHANG Changqing;WEI Yingjuan;JIA Fudong(MLR Key Laboratory of Metallogeny and Mineral Assessment,Institute of Mineral Resources,Chinese Academy of Geological Sciences,Beijing 100037,China;School of Earth Science and Resource,China University of Geosciences(Beijing),Beijing 100083,China;Land Satellite Remote Sensing Application Center,MNR,Beijing 100048,China)

机构地区：[1]中国地质科学院矿产资源研究所自然资源部成矿作用与资源评价重点实验室,北京100037 [2]中国地质大学(北京)地球科学与资源学院,北京100083 [3]自然资源部国土卫星遥感应用中心,北京100048

出　　处：《自然资源遥感》2021年第4期153-161,共9页Remote Sensing for Natural Resources

基　　金：中国地质调查项目“津冀重要矿产资源集中区资源综合利用与评价”(编号:DD20190182)资助。

摘　　要：实现目标区域尾矿信息的识别和提取是矿山环境动态监测的重要组成部分。中低空间分辨率影像多是基于光谱信息进行地物分类,但由于矿区环境特殊,部分道路与尾矿的光谱反射率相近,仅利用光谱信息进行地物分类易将尾矿错误划分为道路,影响尾矿库结构完整性以及其占地面积统计。针对这一问题,基于北京二号高空间分辨率影像对迁西地区铁尾矿的光谱特征、形状特征以及纹理特征进行综合分析,提出了一种基于多特征的面向对象分类方法。首先,对北京二号影像进行多尺度分割,并以地物在各波段的反射率及光谱差值作为地物光谱特征值;然后,利用协方差矩阵和对象边界提取长宽比作为地物形状特征值;再利用主成分波段进行灰度共生矩阵计算,并从中选取对比度、相关度、熵这3个能有效区分尾矿与其他地物纹理特点的值作为遥感图像的纹理特征值;最后,结合以上地物特征信息利用最近邻方法实现面向对象分类并进行精度评价。结果表明:该方法可有效避免尾矿库内道路的误分,为开展大范围高精度尾矿识别与动态监测提供研究基础。The recognition and extraction of mine tailing information serve as an important step in the dynamic monitoring of the mine environment.The classification of surface features using medium-low spatial resolution images is mostly conducted based on spectral information.However,some roads and tailings have similar spectral reflectance due to the special environment in mining areas.As a result,it is liable to misclassify tailings as roads in the surface feature classification based on spectral information only,which affects the structural integrity and area statistics of tailing ponds.Given this,this paper comprehensively analyzes the spectral,shape-related,and texture characteristics of iron mine tailings in the Qianxi area,Hebei Province based on high spatial resolution images obtained from the Beijing-2 satellite and proposes an object-oriented classification method based on multiple features.The steps of the method are as follows.Firstly,perform multi-scale segmentation of Beijing-2 images and the reflectance and take the spectral differences of surface features in each band as the spectral characteristic values of surface features.Secondly,extract the values of length-to-width ratio of objects using a covariance matrix and object boundaries and take them as the characteristic values of surface feature shapes.Then,calculate the gray-level co-occurrence matrix using principal component bands,and select the contrast,correlation,and entropy values that can effectively distinguish the texture characteristics between tailings and other surface features as the texture characteristic values of remote sensing images.Finally,conduct object-oriented classification and precision assessment using the nearest neighbor method according to the characteristic information of surface features.The results indicate that the object-oriented classification method can effectively avoid the misclassification of the roads in tailing ponds and thus provide a research basis for the implementation of large-scope and high-precision identific

关 键 词：铁尾矿 特征选取 影像分割 面向对象分类 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]