草原露天煤矿区土壤粒径分布及热红外辐射特征研究  

Study on the Distribution of Soil Grain Size and the Characteristics of Thermal Infrared Radiation in the Prairie Open-Cast Mine Area

作　　者：李诗朦 包妮沙[1] 刘善军[1] 雷尚彬 LI Shimeng;BAO Nisha;LIU Shanjun;LEI Shangbin(Institute for Geo-informatics&Digital Mine Research,Northeastern University,Shenyang 110819;CCCC Tianjin Port&Waterway Prospection&Design Research Institute Co.,Ltd.,Tianjin 300456,China)

机构地区：[1]东北大学灾害遥感与数字矿山研究所,辽宁沈阳110819 [2]中交天津港航勘察设计研究院有限公司,天津300456

出　　处：《地理与地理信息科学》2018年第2期27-33,共7页Geography and Geo-Information Science

基　　金：国家自然科学基金青年基金项目(41401233);中央高校基本科研专项资金项目(N160102001)

摘　　要：土壤质地是矿区植被恢复及土地复垦的关键,该文选择土壤粒径作为反映矿区土壤质地的重要指标之一,以伊敏露天煤矿区域为研究区,通过分析原地貌围栏、放牧和矿区不同复垦年限土地利用下土壤粒径分布与发射率光谱的关系,探究不同土壤粒径分布在热红外光谱的敏感波段,进而构建支持向量机模型,预测土壤粒径分布。结果表明:1)复垦土壤粗砂粒含量高,细砂粒含量低,且发射率低于原地貌土壤;2)土壤粗砂粒含量与土壤热红外发射率在8~13μm上呈显著负相关,粗粉粒含量和细砂粒含量分别在8.2~10.5μm波段和11~13μm波段与发射率呈显著正相关;3)采用径向核函数-支持向量机模型对土壤粒径分布进行预测,相比粗粉粒含量和细砂粒含量,粗砂粒预测精度最高(建模精度R2=0.91,RMSE=4.30%,RPD=3.28;验证精度R2=0.65,RMSE=7.71%,RPD=1.71)。研究结果为草原煤矿区利用热红外光谱技术预测土壤粒径分布提供了理论依据和技术支撑。Soil texture is the key to vegetation restoration and land reclamation.In this paper,soil grain size(SGS)was selected as one of the important indexes to reflect the soil texture in the mining area,and the Yimin open-cast mine area was taken as the study area.The aims of this study were to explore the sensitive bands of different SGS distribution in thermal infrared spectrum and establish the Support Vector Machine model to predict the SGS distribution.The results showed that:1)The soil from reclaimed land has the higher portion of coarse sand content and lower portion of fine sand content and produces the lower emissivity spectrum than natural soil;2)There is significant negative correlation between the coarse sand content and the thermal infrared emissivity of soil at 8~13μm.The coarse silt content and the fine sand content had the significant positive correlation with the emissivity at 8.2~10.5μm and 11~13μm respectively;3)The Radial Basis Function-Support Vector Machine model was established to predict the SGS distribution based on the sensitive band from correlation analysis.Compared with fine sand content and coarse silt content,the coarse sand content generated the highest predictive accuracy(calibration accuracy:R 2=0.91,RMSE=4.30%,RPD=3.28;validation accuracy:R 2=0.65,RMSE=7.71%,RPD=1.71).These findings provide a theoretical basis and technical support for SGS estimations by using thermal infrared spectroscopy in the prairie coal mine area.

关 键 词：草原煤矿区 复垦土壤 土壤粒径 热红外光谱 支持向量机 

分 类 号：P237[天文地球—摄影测量与遥感] S152.3[天文地球—测绘科学与技术]