
土壤质地对灌区土壤盐分高光谱反演精度影响研究
周世勋, 尹娟, 杨莹攀, 杨震, 常布辉
土壤质地对灌区土壤盐分高光谱反演精度影响研究
Effect of Soil Texture on the Accuracy of Hyperspectral Inversion of Soil Salinity in Irrigated Area
针对土壤质地对高光谱反演土壤盐分精度的影响不明确问题,于2023年4月1-10日在内蒙古河套灌区沈乌灌域共采集了132个不同盐渍化程度的土壤样品,并同步采集了对应的光谱信息,研究了不同盐渍化程度下土壤光谱反射率的变化特征以及不同土壤质地光谱特征与土壤盐分的相关性,探讨了土壤样本适宜的数学变换方法,并筛选敏感波段,建立了基于全部样本以及不同土壤质地下的土壤盐分含量的高光谱反演模型。结果表明:随着土壤盐分含量的增加,高光谱反射率逐渐增大;随着土壤粒度的减小,不同土壤质地下土壤盐分与不同波段的反射率及其相关系数呈先增加后下降的变化趋势。通过对光谱数据进行数学变换后,发现以倒数对数微分、对数微分、平方根微分3种变换效果最佳。通过建立多元逐步线性回归(BPNN)、偏最小二乘回归(PLSR)、支持向量机回归(SVM)以及BP神经网络(BPNN)4种模型,对光谱变换下的盐分含量进行了估算,4种模型的估算精度由高到低表现为:BPNN>SVM>PLSR>MLSR。相较于全部样本的土壤盐分估算结果,考虑不同土壤质地的盐分估算精度均有所提升,其中砂粒质地估算精度R 2由0.918提升到0.962,RPD由3.493提升到4.313;粉粒质地估算精度R 2由0.866提升到0.902,RPD由2.613提升到3.310;黏粒质地估算精度R 2由0.876提升到0.926,RPD由2.651提升到3.953,且在3种土壤质地背景下建立的模型均达到了出色模型的标准。说明在考虑土壤质地的前提下进行含盐量的高光谱反演,有利于提升反演精度。
This study addresses the unclear influence of soil texture on the accuracy of hyperspectral inversion for soil salinity. From April 1 to 10, 2023, a total of 132 soil samples with different salinization degrees were collected in Shenwu irrigation area of Hetao Irrigation District, Inner Mongolia. Corresponding spectral information was collected simultaneously. The variation characteristics of spectral reflectance of soil under different salinization degree and the correlation between spectral characteristics of soil texture and soil salinity were studied. The appropriate mathematical transformation method of soil samples was discussed, and the sensitive bands were selected. A hyperspectral inversion model based on all samples and different soil textures was established. The results showed that the hyperspectral reflectance increased with the increase of soil salt content. As the soil grain size decreased, soil salinity and reflectance of different bands and their correlation coefficients exhibited an increasing trend followed by a decreasing trend. After mathematical transformation of spectral data, it is found that reciprocal logarithmic differentiation, logarithmic differentiation and square root differentiation have the best effect. Four models of multiple stepwise linear regression (BPNN), partial least squares regression (PLSR), support vector machine regression (SVM) and BP neural network (BPNN) were established to estimate the salt content under spectral transformation. The estimation accuracy of the four models was as follows: BPNN>SVM>PLSR>MLSR. Compared with the soil salt estimation results of all samples, the salt estimation accuracy of different soil textures was improved. For sand texture, the estimation accuracy R 2 was increased from 0.918 to 0.962, and the RPD was increased from 3.493 to 4.313. The grain texture estimation accuracy R 2 increased from 0.866 to 0.902, and the RPD increased from 2.613 to 3.310. The accuracy of clay texture estimation R 2 increased from 0.876 to 0.926, and the RPD increased from 2.651 to 3.953, and the models established under the three soil texture backgrounds all reached the standard of excellent models. The results show that the hyperspectral inversion of salt content is helpful to improve the inversion accuracy when considering soil texture.
土壤盐分 / 高光谱 / 土壤质地 / 光谱变换 / 反演模型 {{custom_keyword}} /
soil salinity / hyperspectral / soil texture / spectral transformation / inversion model {{custom_keyword}} /
表1 不同质地土壤中盐分含量统计值Tab.1 Statistical values of salt content in soils of different textures |
样本类别 | 样本数/个 | 最小值/(g | 最大值/(g | 平均值/(g | 标准差/(g | 变异系数 |
---|---|---|---|---|---|---|
砂粒 | 62 | 0.091 3 | 21.518 1 | 3.047 6 | 5.254 0 | 1.782 5 |
粉粒 | 24 | 0.114 0 | 14.790 4 | 3.262 8 | 4.318 7 | 1.323 7 |
黏粒 | 46 | 0.125 8 | 22.094 7 | 3.365 8 | 5.267 1 | 1.206 4 |
表2 研究区土壤采样点盐分统计性分析Tab.2 Statistical analysis of salt in soil sampling points in the study area |
土壤盐渍化程度 | 样本数/个 | 含盐量/(g | ||||
---|---|---|---|---|---|---|
最小值 | 最大值 | 平均值 | 标准差 | 变异系数 | ||
非盐渍化 | 63 | 0.091 4 | 0.993 8 | 0.382 5 | 0.281 4 | 0.735 7 |
轻度盐渍化 | 20 | 1.037 3 | 1.931 0 | 1.405 4 | 0.296 3 | 0.210 9 |
中度盐渍化 | 17 | 2.018 1 | 3.990 3 | 3.071 4 | 0.705 3 | 0.229 7 |
重度盐渍化 | 16 | 4.026 3 | 9.875 7 | 7.205 2 | 1.891 9 | 0.262 6 |
盐渍土 | 16 | 10.158 2 | 22.904 7 | 15.274 5 | 3.884 7 | 0.254 3 |
总样本 | 132 | 0.091 4 | 22.904 7 | 3.515 9 | 5.121 1 | 1.456 6 |
图6 不同变换下土壤盐分与光谱反射率的相关系数Fig. 6 Correlation coefficients between soil salinity and spectral reflectance under different transformations |
表3 建模集验证集统计特征Tab. 3 Statistical characteristics of modeling set verification set |
样本类型 | 样本数 | 均值/(g | 标准差/(g | 最小值/(g | 最大值/(g | 变异系数 |
---|---|---|---|---|---|---|
建模样本 | 88 | 3.866 | 5.673 | 0.096 | 22.905 | 1.467 |
验证样本 | 44 | 2.816 | 3.747 | 0.091 | 14.790 | 1.330 |
表4 建模集验证集精度表Tab.4 Precision table of modeling set verification set |
方法 | 建模集 | 验证集 | ||||
---|---|---|---|---|---|---|
R 2 | RMSE/(g | R 2 | RMSE/(g | RPD | ||
逐步线性回归 | (lg R)' | 0.512 | 2.077 | 0.435 | 1.953 | 1.459 |
(lg 1/R)' | 0.604 | 2.026 | 0.503 | 1.760 | 1.496 | |
( | 0.507 | 2.084 | 0.515 | 1.855 | 1.381 | |
偏最小二乘回归 | (lg R)' | 0.727 | 1.773 | 0.658 | 1.441 | 1.710 |
(lg 1/R)' | 0.804 | 1.659 | 0.703 | 1.398 | 1.836 | |
( | 0.755 | 1.746 | 0.712 | 1.356 | 1.864 | |
支持向量机回归 | (lg R)' | 0.908 | 1.302 | 0.882 | 1.256 | 2.698 |
(lg 1/R)' | 0.894 | 1.382 | 0.856 | 1.288 | 2.601 | |
( | 0.813 | 1.495 | 0.802 | 1.303 | 2.313 | |
BP神经网络 | (lg R)' | 0.898 | 1.349 | 0.887 | 1.251 | 2.588 |
(lg 1/R)' | 0.925 | 1.294 | 0.898 | 1.152 | 3.108 | |
( | 0.909 | 1.300 | 0.894 | 1.187 | 3.073 |
表5 21个砂粒背景下的土壤盐分含量验证精度对比Tab.5 Comparison of verification accuracy of soil salt content under the backgrounds of 21 sand grains |
不同样本集 | R 2 | RMSE/(g | RPD |
---|---|---|---|
全样本 | 0.918 | 0.752 | 3.493 |
砂粒样本 | 0.962 | 0.742 | 4.313 |
表6 8个粉粒背景下的土壤盐分含量验证精度对比Tab.6 Comparison of verification accuracy of soil salt content under the background of 8 silt grains |
不同样本集 | R 2 | RMSE/(g | RPD |
---|---|---|---|
全样本 | 0.866 | 1.485 | 2.613 |
粉粒样本 | 0.902 | 1.110 | 3.310 |
表7 15个黏粒背景下土壤盐分含量验证精度对比Tab.7 Comparison of verification accuracy of soil salt content under the background of 15 clay grains |
不同样本集 | R 2 | RMSE/(g | RPD |
---|---|---|---|
全样本 | 0.876 | 1.274 | 2.651 |
黏粒样本 | 0.926 | 0.861 | 3.953 |
图11 砂粒质地下土壤盐分含量验证精度散点图Fig.11 Scatter diagram of verification accuracy of soil salt content under sand texture |
图12 粉粒质地下土壤盐分含量验证精度散点图Fig.12 Scatter diagram of verification accuracy of soil salt content in silt texture |
1 |
田义超,郑丹琳,张强,等.基于国产资源一号02D卫星和机器学习算法的钦州湾滨海土壤盐分反演[J].中国环境科学,2024,44(1):371-385.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
2 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
3 |
杨婧文.黄河三角洲濒海区土壤质量盐碱退化评价与遥感反演[D]. 山东泰安:山东农业大学,2022. YANG J W. Evaluation and remote sensing inversion of soil quality saline-alkali degradation in the Yellow River delta coastal area [D]. Tai’an, Shandong: Shandong Agricultural University, 2022.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
4 |
孙小茹. 近四十年河套平原土壤盐渍化时空变异及其影响因素分析[D]. 武汉:华中农业大学,2023. SUN X R. Spatio-temporal variation of soil salinization in Hetao Plain in recent 40 years and its influencing factors [D]. Wuhan: Huazhong Agricultural University, 2023.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
5 |
曹 琪,郑雅兰,沈 谦,等.基于高分辨率遥感的水源地风险源提取技术研究[J].测绘地理信息,2022,47(6):81-85.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
6 |
曾 鹏,喻宏伟,余雅滢,等.基于双层组合神经网络的土地利用自动分类方法[J].测绘地理信息,2023,48(5):98-103.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
7 |
张超,高璐璐,郧文聚, 等. 遥感技术获取耕地质量评价指标的研究进展分析[J].农业机械学报, 2022, 53(1):1-13.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
8 |
冯天时,庞治国,江威, 等. 高光谱遥感技术及其水利应用进展[J].地球信息科学学报, 2021, 23(9):1 646-1 661.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
9 |
王瑾杰,丁建丽,葛翔宇,等. 分数阶微分技术在机载高光谱数据估算土壤含水量中的应用[J]. 光谱学与光谱分析, 2022, 42(11): 3 559-3 567.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
10 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
11 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
12 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
13 |
扶卿华,倪绍祥,王世新,等. 土壤盐分含量的遥感反演研究[J].农业工程学报, 2007, 112(1):48-54.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
14 |
张贤龙,张飞,张海威,等 .基于光谱变换的高光谱指数土壤盐分反演模型优选[J].农业工程学报,2018,34(1):110-117.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
15 |
张建明,齐文文 .民勤绿洲土壤盐分组成与光谱特征[J].生态学杂志,2013,32(10):2 620-2 626.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
16 |
文虎 .绿洲农田盐碱斑土壤表层盐分和pH 值的光谱特征研究[D].乌鲁木齐:新疆农业大学,2016 .
{{custom_citation.content}}
{{custom_citation.annotation}}
|
17 |
韩文霆,崔家伟,崔欣, 等. 基于特征优选与机器学习的农田土壤含盐量估算研究[J]. 农业机械学报,2023,54(3): 328-337.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
18 |
李晓明,王曙光,韩霁昌.基于PLSR的陕北土壤盐分高光谱反演[J]. 国土资源遥感, 2014, 26(3):113-116.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
19 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
20 |
王欢,李瑞平,张寅, 等. 内蒙古河套灌区土壤盐分多源多指数估算模型[J]. 灌溉排水学报,2023,42(10):122-128.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
21 |
张丽霞,吕兰颂,邓振利, 等. 莱州湾南岸土壤全盐量空间分布及其高光谱反演研究[J]. 安全与环境学报,2019,19(3): 1 034-1 040.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
22 |
毛鸿欣,贾科利,张旭. 基于实测高光谱和Sentinel-2B影像的银川平原土壤盐分反演[J]. 云南大学学报自然科学版, 2021, 43(5): 929-941.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
23 |
王惠敏,于磊,徐凯磊, 等. 基于高分五号高光谱影像的干旱区盐渍化土壤盐分含量估算[J]. 光谱学与光谱分析,2023,43(7): 2 278-2 286.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
24 |
江远东. 博斯腾湖西岸湖滨绿洲土壤质地变化与高光谱估算[D]. 乌鲁木齐:新疆师范大学,2022.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
25 |
常布辉,李根东,苏小飞, 等. 河套灌区出让水权对天然植被影响研究[J]. 人民黄河,2023,45(8): 6-11, 15.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
26 |
吕贻忠,李保国. 土壤学[M]. 北京:中国农业出版社,2006.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
27 |
张霄羽,姚艳敏,颜祥照. 光谱变换和光谱分辨率对土壤有机质含量估测精度的影响[J]. 中国土壤与肥料,2023(3): 184-193.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
28 |
钱佳,郭云开,蒋明, 等. 不同类型土壤Cu含量高光谱联合反演建模[J].测绘科学,2020,45(8):138-144.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
29 |
张强. 不同质地土壤养分含量的光谱估测研究[D]. 新疆石河子:石河子大学, 2017. ZHANG Q. Study on spectral estimation of soil nutrient content with different textures [D]. Shihezi, Xinjiang: Shihezi University, 2017.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
30 |
李亚莉,乔江飞,董天宇, 等. 不同质地盐渍化土壤水盐含量的高光谱反演[J]. 应用生态学报,2016,27(12): 3 807-3 815.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
31 |
王雪梅,玉米提·买明,毛东雷, 等. 干旱区绿洲耕层土壤重金属铬含量的高光谱估测[J]. 生态环境学报, 2021, 30(10): 2 076-2 084.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
32 |
李娟,陈超,王昭. 基于不同变换形式的干旱区土壤盐分高光谱特征反演[J]. 水土保持研究,2018,25(1):197-201.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
33 |
如则麦麦提·米吉提,买买提·沙吾提,麦尔耶姆·亚森, 等. 基于高光谱的干旱区盐渍化土壤盐分含量估算[J]. 江苏农业科学, 2018, 46(22):265-269.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
34 |
刘旭辉,白云岗,柴仲平, 等. 基于多光谱遥感的典型绿洲棉田春季土壤盐分反演及验证[J]. 干旱区地理,2022,45(4): 1 165-1 175.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
35 |
杨扬,高小红,贾伟, 等. 三江源区不同土壤类型有机质含量高光谱反演[J]. 遥感技术与应用,2015,30(1): 186-198.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
36 |
徐驰,陈爱萍,曾文治, 等. 不同土质下土壤含盐量的高光谱定量反演技术研究[J]. 灌溉排水学报,2014,33(Z1): 209-212.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
37 |
李振. 基于无人机高光谱的农田土壤盐分估测研究[D]. 济南:山东师范大学,2023.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
38 |
王怡婧,丁启东,张俊华,等. 基于无人机高光谱遥感和机器学习的土壤水盐信息反演[J]. 应用生态学报,2023,34(11): 3 045-3 052.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
39 |
张成雯,唐家奎,于新菊, 等. 黄河三角洲土壤含盐量定量遥感反演[J].中国科学院研究生院学报,2013,30(2):220-227.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
40 |
田安红,赵俊三,张顺吉, 等. 基于分数阶微分的盐渍土电导率高光谱估算研究[J]. 中国生态农业学报(中英文),2020,28(4): 599-607.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
41 |
赵慧,李新国,靳万贵, 等. 博斯腾湖西岸湖滨绿洲土壤含盐量高光谱估算[J]. 中山大学学报(自然科学版),2020,59(4): 56-63.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
42 |
陈睿华,王怡婧,张俊华, 等. 基于分数阶微分光谱指数的银川平原土壤含盐量反演[J]. 生态学杂志,2023,42(9): 2 296-2 304.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_ref.label}} |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
/
〈 |
|
〉 |