基于灰数灰度的土壤有机质高光谱估测  

Hyper-Spectral Estimation of Soil Organic Matter Content Using Greyness of Grey Number

作　　者：丁天姿 任文静 李丽[2] 田野 DING Tian-zi;REN Wen-jing;LI Li;TIAN Ye(School of Information Science and Engineering/Shandong Agricultural University,Tai’an 271018,China;School of Economics and Management/Shandong Agricultural University,Tai’an 271018,China;Kaiyuan School of Daiyue District,Tai’an 271000,China)

机构地区：[1]山东农业大学信息科学与工程学院,山东泰安271018 [2]山东农业大学经济管理学院,山东泰安271018 [3]泰安市岱岳区开元中学,山东泰安271000

出　　处：《山东农业大学学报（自然科学版）》2022年第1期85-90,共6页Journal of Shandong Agricultural University：Natural Science Edition

基　　金：山东省自然科学基金项目(ZR2016DM03);山东省SRT项目(20211636)。

摘　　要：为克服光谱估测中的不确定性,利用统计回归与灰色系统理论建立土壤有机质高光谱估测模型。以山东省济南市章丘区采集的76个土壤样本为研究对象,首先对土壤光谱数据进行变换处理,根据极大相关性原则选取特征波段的估测因子,建立各特征波段的一元线性回归预测模型;其次,对各估测因子进行由小到大排序,计算估测因子排序后的土壤有机质含量的滑动方差,将滑动方差转化为灰数的灰度值,并将其用于修正估测因子,然后再建立各特征波段的一元线性回归修正模型;最后采用平均法与加权法融合各个单波段的预测值。结果表明,所建估测模型精度和检验精度均显著提高,13个检验样本的R^(2)=0.911,MRE=7.764%。研究表明,本文建立的基于灰数灰度的土壤有机质含量高光谱估测模型是可行有效的。To overcome the uncertainties in the spectral estimation,a hyper-spectral estimation model of soil organic matter was developed using statistical regression and grey system theory.Firstly,76 soil samples collected from Zhangqiu District,Jinan City,Shandong Province,were transformed and processed,and the estimation factors of the characteristic bands were selected according to the principle of great correlation,and a one-variable linear regression prediction model for each characteristic band was established.Secondly,the estimated factors were ranked from small to large and the sliding variance of soil organic matter content was calculated after ranking the estimated factors,the sliding variance was transformed into the greyness value of the grey number and used to correct the estimated factors,then a one-variable linear regression correction model for each characteristic band was built.Finally,the prediction values of each single band are fused using the averaging or weighting method.The results show that the accuracy of the constructed estimation model and the precision of the test samples are significantly improved,with R^(2)=0.911 and MRE=7.764%for the 13 test samples.The study shows that the hyper-spectral estimation model of soil organic matter content based on greyness of grey number developed in this paper is feasible and effective.

关 键 词：土壤有机质 高光谱遥感 灰数 灰度 光谱估测 

分 类 号：P237[天文地球—摄影测量与遥感]