[1]徐丽华,谢德体.土壤有机质含量预测精度对光谱预处理和特征波段的响应[J].江苏农业学报,2019,(06):1340-1345.[doi:doi:10.3969/j.issn.1000-4440.2019.06.010]
　XU Li-hua,XIE De-ti.Response of soil organic matter content prediction accuracy to preprocessing of spectra and feature bands[J].,2019,(06):1340-1345.[doi:doi:10.3969/j.issn.1000-4440.2019.06.010]
点击复制

土壤有机质含量预测精度对光谱预处理和特征波段的响应()
分享到：

江苏农业学报[ISSN:1006-6977/CN:61-1281/TN]

卷:
期数:
2019年06期
页码:
1340-1345
栏目:
耕作栽培·资源环境
出版日期:
2019-12-31

文章信息/Info

Title:
Response of soil organic matter content prediction accuracy to preprocessing of spectra and feature bands
作者:
徐丽华; 谢德体
(西南大学资源环境学院，重庆400716)
Author(s):
XU Li-hua; XIE De-ti
(College of Resources and Environment, Southwest University, Chongqing 400716, China)
关键词:
克里克滤波; 光谱预处理; 分区特征波段; 土壤有机质
Keywords:
kriging filter; spectral preprocessing; zonal feature bands; soil organic matter
分类号:
S127 TP79
DOI:
doi:10.3969/j.issn.1000-4440.2019.06.010
文献标志码:
A
摘要:
为了提高土壤有机质含量预测的精度，对光谱预处理方法和特征波段的选择进行了研究。分别用Savitzky-Golay平滑(SGS)、多元散射校正（MSC）、标准正态化(SNV)、标准正态化+去趋势(SNV_Detrend) 、一阶导数(FD)、二阶导数（SD）、包络线去除(CR) 和克里克滤波(KF)8种方法对33个水稻土土壤样本进行了光谱预处理，用分区极值法选择的特征波段进行了建模。结果表明：经过预处理以后，除了SGS和KF处理外，MSC、 SNV、 SNV_Detrend、 FD、SD、CR 预处理获得的土壤光谱与土壤有机质(SOM)含量的相关性都得到了显著提高；CR预处理方法获得的预测模型精度最高，其标定集和验证集的决定系数分别是0.728和0.666，最小均方根误差（RMSE）分别是2.240 g/kg和2.770 g/kg；利用分区选择的4个特征波段建立的预测模型精度远高于利用4个相关系数最大绝对值对应的波段及所有相关系数绝对值大于0.5的77个波段建立的预测模型。CR预处理方法和基于分区极值选择的特征波段能够改善土壤有机质含量的预测精度。
Abstract:
In order to improve the prediction accuracy of soil organic matter(SOM) content, the spectral preprocessing methods and the selection of characteristic bands were studied. Savitzky-Golay smoothing(SGS), multiplicative scatter correction(MSC), standard normal variate (SNV), standard normal variate+detrend (SNV_Detrend), first derivative(FD), second derivative(SD), continuum removal(CR) and kriging filter (KF) were used to preprocess the original spectra of 33 soil samples from paddy fields. Feature bands selected by zonal extremum method were used to build prediction model of SOM content. The results showed that the correlations between soil organic matter content and soil spectra preprocessed by MSC, SNV, SNV_Detrend, FD, SD and CR had been significantly improved. The prediction precision of model obtained by CR pretreatment method was the highest. The determination coefficients of calibration set and validation set were 0.728 and 0.666, and root mean square errors(RMSEs) were 2.240 g/kg and 2.770 g/kg, respectively. The prediction precison of the model based on four zonal feature bands was much higher than that based on four feature bands with maximum absolute values of correlation coefficient and based on the 77 feature bands with absolute values of crrelation coefficient greater than 0.5. Therefore, CR pretreatment method and feature bands selected by zonal extremum can improve prediction precision of SOM content.

参考文献/References:

[1]王海峰,张智韬,KARNIELI A,等. 基于灰度关联-岭回归的荒漠土壤有机质含量高光谱估算[J]. 农业工程学报, 2018, 34(14): 124-131.
[2]郑曼迪,熊黑钢,乔娟峰,等. 基于综合光谱指数的不同程度人类干扰下土壤有机质含量预测[J] .江苏农业学报,2018,34(5):1048-1056.
[3]QIAO X X, WANG C, FENG M C, et al. Hyperspectral estimation of soil organic matter based on different spectral preprocessing techniques[J]. Spectroscopy Letters, 2017, 50(3): 156-163.
[4]WANG C, FENG M C, YANG W D, et al. Use of spectral character to evaluate soil organic matter[J]. Soil Science Society of America Journal, 2016, 80(4): 1078-1088.
[5]南锋, 朱洪芬,毕如田. 黄土高原煤矿区复垦农田土壤有机质含量的高光谱预测[J]. 中国农业科学, 2016, 49(11): 2126-2135.
[6]LUYPAERT J, HEUERDING S, VANDER H Y, et al. The effect of preprocessing methods in reducing interfering variability from near-infrared measurements of creams[J]. Journal of Pharmaceutical and Biomedical Analysis, 2004, 36(3): 495-503.
[7]于雷,洪永胜,耿雷, 等. 基于偏最小二乘回归的土壤有机质含量高光谱估算[J]. 农业工程学报, 2015, 31(14): 103-109.
[8]张东辉,赵英俊,秦凯, 等. 光谱变换方法对黑土养分含量高光谱遥感反演精度的影响[J]. 农业工程学报, 2018, 34(20): 141-147.
[9]VISCARRA R R, TAYLOR H, MCBRATNEY A. Multivariate calibration of hyperspectral γ-ray energy spectra for proximal soil sensing[J]. European Journal of Soil Science, 2007, 58(1): 343-353.
[10]杨剑虹, 王成林,代亨林.土壤农化分析与环境监测[M].北京:中国大地出版社, 2008.
[11]尼加提·卡斯木,茹克亚·萨吾提,师庆东, 等. 基于优化光谱指数的土壤有机质含量估算[J]. 农业机械学报, 2018, 49(11): 155-163.
[12]侯艳军,塔西甫拉提·特依拜,买买提·沙吾提, 等. 荒漠土壤有机质含量高光谱估算模型[J]. 农业工程学报, 2014, 30(16): 113-120.
[13]刘炜,常庆瑞,郭曼,等.土壤导数光谱小波去噪与有机质吸收特征提取[J]. 光谱学与光谱分析, 2011, 31(1): 100-104.
[14]吴可杰. 统计学原理[M].南京:南京大学出版社, 1986.
[15]朱晓青. 统计学原理[M].北京:国防工业出版社,1994.

备注/Memo

备注/Memo:
收稿日期：2019-04-01基金项目：中央高校基本科研业务费专项基金项目(XDJK2016C083)；重庆市基础科学与前沿技术研究一般项目(cstc2016jcyjA0184)；国家自然科学基金项目(41671291)作者简介：徐丽华(1976-)，女，黑龙江绥滨人，博士，副教授，主要从事土壤养分预测与制图、遥感图像信息处理研究，(E-mail)sweitlianna@126.com通讯作者：谢德体，(E-mail)xdt@swu.edu.cn
更新日期/Last Update: 2020-01-09