摘要
煤炭开采导致矿粮复合区地表沉陷、土地损毁和土壤质量急剧下降,而土壤有机质是表征土壤质量的重要指标,如何快速监测与反演矿粮复合区土壤有机质含量是亟待解决的科学问题。在赵固煤矿采集土壤样品100个并采集其高光谱信息,通过竞争自适应重加权采样(CARS)算法筛选敏感波段,基于偏最小二乘回归、随机森林和XGBoost算法构建土壤有机质含量的反演模型。结果表明,CARS算法能够提取土壤有机质重要特征信息变量,对去噪处理后高光谱影像进行处理,筛选出32个特征波段,显著提升建模效率;反演结果表明,基于提取特征波段构建的CARS-RF模型精度最高,决定系数R^(2)高达0.95;高光谱结合机器学习方法可以对土壤有机质含量进行有效监测,方法具有可行性。研究结果可为基于高光谱遥感技术的土壤有机质快速监测提供理论科学依据,为矿粮复合区土壤质量监测提供了一种可靠的解决方案。
Coal mining has led to surface subsidence,land damage and a sharp decline in soil quality in the mineral-grain composite ar-ea.Soil organic matter is an important indicator of soil quality.How to quickly monitor and invert the content of soil organic matter in the mineral-grain composite area is an urgent scientific problem to be solved.In this study,100 soil samples were collected from Zhaogu Coal Mine and their hyperspectral information was collected.The sensitive bands were screened by competitive adaptive reweighted sam-pling(CARS)algorithm,and the inversion model of soil organic matter content was constructed based on partial least squares regres-sion,random forest and XGBoost algorithm.The results show that the CARS algorithm can extract the important characteristic informa-tion variables of soil organic matter,process the hyperspectral image after denoising,and screen out 32 characteristic bands,which sig-nificantly improves the modeling efficiency;the inversion results show that the CARS-RF model based on the extracted feature bands has the highest accuracy,and the determination coefficient R^(2) is as high as 0.95;the Hyperspectral combined with machine learning method can effectively monitor soil organic matter content,and the method is feasible.The results of this study can provide a theoretical and scientific basis for the rapid monitoring of soil organic matter based on hyperspectral remote sensing technology,and provide a relia-ble solution for soil quality monitoring in the mineral-grain composite area.
作者
潘元庆
邓炯
任春太
马惠
路晓明
秦何
Pan Yuanqing;Deng Jiong;Ren Chuntai;Ma Hui;Lu Xiaoming;Qin He(Henan Provincial Institute of Territorial Space Investigation and Planning,Zhengzhou 450016,China;Observation and Research Station of Land Ecology and Land Use in the Grain-Producing Area of Central China,Ministry of Natural Resources,Zhengzhou 450016,China;School of Surveying and Land Information Engineering,Henan Polytechnic University,Jiaozuo 454000,China)
出处
《能源与环保》
2025年第8期74-81,共8页
CHINA ENERGY AND ENVIRONMENTAL PROTECTION
基金
国家自然科学基金重点项目子课题(U21A20108)
企事业单位委托项目(ZYYH24-003)。
关键词
土壤有机质
高光谱
CARS
反演模型
矿粮复合区
soil organic matter
hyperspectral
CARS
inversion model
mineral-grain composite area
