摘要
本文通过对研究区高分辨率影像进行主成分(PCA)计算,采用第二主成分进行阈值分割提取初步的挂白信息,再利用第一主成分的纹理分析对初步挂白信息进行掩膜计算;然后结合研究区矿山开采面切面边界清晰和含积水层的特点,辅以人工修正,最后筛选出矿山挂白图斑。通过外业调查验证,挂白图斑总体精度达到91.17%,Kappa系数为0.805,证实该方法的可靠性、准确性,从而为政府部门开展矿山挂白治理提供一定的辅助决策依据。
Based on the calculation of principal component analysis(PCA)for the high-resolution images in the study area,the 2nd principal component was adopted for threshold segmentation to extract the preliminary white mining spots information,then combined with the texture analysis of the 1st principal component to make the preliminary masking calculation.Together with the characteristics of clear cut boundary and waterlogged layer of the mining faces in the study area,it was finally screened out the white mining spots in the mining area supplemented by mannual correction.Through the field survey verification,the overall accuracy of white mining spots reached 91.18%,with the Kappa coefficient of 0.805.Therefore,it was proved the realiability and accuracy of such method,so as to provide the supplement decision basis for the governments on the administration of white mining spots.
作者
吕红梅
Lv Hongmei(Fujian Institute of Geological Surveying and Mapping, Fuzhou, Fujian Province 350011)
出处
《亚热带水土保持》
2020年第2期1-6,共6页
Subtropical Soil and Water Conservation
基金
中国地质调查局项目“全国矿山开发状况遥感地质调查与监测”(DD20190511)
“全国矿山环境恢复治理状况遥感地质调查与监测”(DD20190705)。
关键词
主成分分析
纹理分析
矿山挂白
掩膜
高分辨率影像
Principal Component Analysis(PCA)
texture analysis
white mining spots
mask
high resolution image
