传统光谱变换与连续小波耦合定量反演潮土有机质含量

引用本文

王延仓, 金永涛, 王晓宁, 廖钦洪, 顾晓鹤, 赵子辉, 杨秀峰. 传统光谱变换与连续小波耦合定量反演潮土有机质含量. 光谱学与光谱分析, 2018,38(8): 2571-2577
WANG Yan-cang, JIN Yong-tao, WANG Xiao-ning, LIAO Qin-hong, GU Xiao-he, ZHAO Zi-hui, YANG Xiu-feng. Quantitative Inversion of Organic Matter Content Based on Interconnection Traditional Spectral Transform and Continuous Wavelet Transform. Spectroscopy and Spectral Analysis, 2018,38(8): 2571-2577. 复制到剪切板

Doi: 10.3964/j.issn.1000-0593(2018)08-2571-07
Permissions

光谱学与光谱分析所有

传统光谱变换与连续小波耦合定量反演潮土有机质含量

王延仓^1,³, 金永涛^1,³, 王晓宁^1,³, 廖钦洪⁵, 顾晓鹤^2,^4,^*, 赵子辉^1,³, 杨秀峰^1,³

1. 北华航天工业学院, 河北廊坊 065000

2. 北京农业信息技术研究中心, 北京 100097

3. 河北省航天遥感信息处理与应用协同创新中心, 河北廊坊 065000

4. 国家农业信息化工程技术研究中心, 北京 100097

5. 重庆文理学院林学与生命科学学院, 重庆 402160

*通讯联系人 e-mail: guxh@nercita.org.cn

作者简介: 王延仓, 1986年生, 北华航天工业学院讲师 e-mail: yancangwang@163.com

收稿日期: 2017-08-06 接受日期: 2017-12-26

基金项目: 河北省青年基金项目(D2017409021), 国家重点研发计划(2016YFD0300609), 北京市农林科学院科技创新能力建设专项(KJCX20170705), 国家自然科学基金项目(41401419)资助

摘要

以北京地区的96个潮土土样的有机质含量为研究对象, 以传统光谱变换为参照, 研究分析传统光谱变换与连续小波的耦合在估测土壤有机质含量的可行性; 首先采用传统光谱变换与连续小波处理土壤光谱数据, 然后将处理后的光谱数据与土壤有机质含量进行相关性分析, 提取敏感波段, 并采用偏最小二乘法构建土壤有机质含量估测模型。结果表明: 耦合传统光谱变换技术与连续小波技术可大幅提升光谱对有机质含量的敏感性, 其相关系数 R²最高可达0.714, 这表明耦合传统光谱变换技术与连续小波技术可深入挖掘光谱内的有益信息; 与传统光谱变换技术相比, 基于耦合传统光谱变换技术与连续小波技术构建的模型精度更高, 稳定性更好, 其中以微分变换构建的模型最优, 其 R²=0.772, RMSE=0.223, 这表明耦合传统光谱变换技术与连续小波技术可有效压制噪声的负面影响, 提升光谱的稳定性。

关键词: 土壤有机质; 传统光谱变换; 连续小波变换; 潮土

中图分类号:S151.9 文献标识码:A

Quantitative Inversion of Organic Matter Content Based on Interconnection Traditional Spectral Transform and Continuous Wavelet Transform

WANG Yan-cang^1,³, JIN Yong-tao^1,³, WANG Xiao-ning^1,³, LIAO Qin-hong⁵, GU Xiao-he^2,^4,^*, ZHAO Zi-hui^1,³, YANG Xiu-feng^1,³

1. Institute of Computer and Remote Sensing Information Technology, North China Institute of Aerospace Engineering, Langfang 065000, China

2. Beijing Research Center for Information Technology in Agriculture, Beijing 100097, China

3. Aerospace Remote Sensing Information Processing and Application Collaborative Innovation Center of Hebei Province, Langfang 065000, China

4. National Engineering Research Center for Information Technology in Agriculture, Beijing 100097, China

5. College of Life Science and Forestry, Chongqing University of Arts and Sciences, Chongqing 402160, China

Abstract

In this study, the soil organic matter content of 96 alluvial soil collected from Beijing area was taken as the object of study; Compared with the traditional spectral transform technology, this paper studied on the analysis of the traditional spectral transform and continuous wavelet technology coupling in the feasibility of estimating soil organic matter content. Firstly, the traditional spectral transform technique and the continuous wavelet transform were used to deal with the soil spectral data. Then The correlation between the spectral data and the soil organic matter content was analyzed, and the sensitive bands were extracted. Finally the estimation model of soil organic matter content was constructed by partial least square method. The results showed that the coupling of traditional spectral transform and continuous wavelet technology can greatly improve the spectral sensitivity of organic matter content, and the correlation coefficient ( R²) was up to 0.714, which indicate that the coupling of the traditional spectral transform and continuous wavelet technology can dig the useful signal of the spectral information; Compared with the traditional spectral transform technology, the accuracy of the model based on the interconnection of traditional technique and continuous wavelet transform was higher and better stability; Among of the model based on the interconnection of traditional technique and continuous wavelet transform, the model construct by the differential transform was the Optimal model; Its coefficient of decision and root mean square error were 0.774 and 0.223 respectively, which indicated that the interconnection of traditional technique and continuous wavelet transform spectral technique can effectively suppress noise, improving the spectral stability.

Key words: Soil organic matter; Spectral transform; Continuous wavelet transforms; Alluvial soil

文章图片

引言

精准农业其涵义为在资源硬约束条件下, 利用信息技术对耕地生产开展的定量、定位的现代农业生产模式, 旨在改变传统耕作模式, 提升耕地生产潜力与生产效率^[1]; 精准施肥是精准农业系统内的关键技术, 其依据作物生长需肥规律与耕地养分现状进行变量的投入, 以达到充分利用资源、节约成本、保护环境的效果^{[2, 3]}。实时、快速、精准掌握大尺度耕地土壤养分现状具有重要现实意义。近年来, 遥感技术的持续进步与国产卫星的迅猛发展展为耕地养分监测提供了新的思路, 逐步成为农业遥感的研究热点^{[4, 5, 6]}。

耕层土壤有机质含量的高低是衡量耕地土壤肥力与质量的重要指标^[7], 与作物生长所必需的微量元素的空间分布具有较强的相关性^[8]。实时掌握大尺度耕层土壤有机质含量的空间分布信息具有重要的实用价值。光谱监测土壤有机质含量具有无损、快速、准确、实时的优点, 近年来国内外学者在该领域开展了大量研究工作, 并取得一定成果。土壤有机质含量的光谱响应特征主要集中于可见光-近红外区域^{[9, 10, 11, 12, 13, 14, 15]}, 但若土壤有机质含量太低其光谱太弱, 不能应用光谱估测土壤有机质含量, 即存在一个临界点。估测土壤有机质含量的光谱分辨率, 并非越高越好, 适度的降低光谱分辨率可提升土壤有机质含量的估测精度。但最佳光谱分辨率存在异议。微分变换、对数变换、倒数变换、离散小波变换、连续小波变换等传统光谱变换均能提升光谱对土壤有机质含量的相关性^[16], 且以小波技术为最佳, 但其对光谱敏感性的提升程度受限。因此, 应用光谱估测土壤有机质含量的研究虽然取得一定成果, 但存在较多问题, 仍需继续探索。

受北方气候、地理因素、土壤构成成分复杂性的影响, 北方潮土耕层土壤有机质含量较低, 光谱较弱, 易被干扰或掩盖。近年来, 连续小波技术的不断完善与发展, 在土壤有机质含量的光谱诊断方面提供了新的思路; 本工作在对土壤高光谱进行传统数学变换的基础上, 进一步利用连续小波筛选土壤有机质含量的敏感波段, 采用偏最小二乘法构建土壤有机质估测模型, 探索提升估测耕层土壤有机质含量的精度。

1 实验部分

1.1 土样采集

以北京市密云西南部、顺义中北部及通州的西南部的耕层潮土为研究对象, 在2011与2012年10月采集土样, 共计96份, 采集方法为“ 四点混合” 法; 采集深度为耕层0~20 cm。采集的土壤置于干燥、通风的暗室内自然风干, 然后研磨、过筛20目。取足够处理后的土壤, 采用“ 重铬酸甲滴定法” 测定土壤的有机质含量。

表1 北方潮土土样有机质含量的统计性描述 Table 1 The Statistics of the North aquic soils organic matter content description

1.2 光谱测量

土壤光谱测量在暗室里开展。仪器选用ASD(Analytical Spectral Devices)公司生产的便携式野外光谱仪, 其具有宽光谱覆盖范围(350~2 500 nm)与高光谱分辨率, 可达1 nm; 光源采用与太阳光接近的1 000 W的卤光灯提供平行光束, 光束与垂线的夹角为15° 。将处理后的土样置于羊皮纸上, 采用钢尺平整土样, 且保持土样厚度在2 cm以上; 探头置于土样中心上方30 cm处, 探测角为5° 。测量前用已标定好的白板进行优化, 每个土样采集10条光谱求平均作为该土样的光谱数据。

	Figure Option View Download New Window
	图1 实验室内土壤光谱测量Fig.1 Soil spectral measurements at laboratory

1.3 光谱处理

受仪器内部暗电流、不同传感器响应差异及暗室内环境影响, 采集的土壤光谱含有噪声, 尤其在光谱边缘及水分吸收强烈。因此删除350~399和2 451~2 500 nm数据。为提升光谱信噪比, 采用海明窗低通滤波器开展平滑处理, 并利用重采样技术将光谱分辨率降低至5 nm。

1.4 小波分析

小波分析可从时间与频率开展数据分析, 更有效的从数据内分离出更多可用信息。离散小波处理后, 数据维数成倍减少。连续小波可将土壤光谱数据分解为不同尺度的小波系数。

$Ψ_{a, b} = \frac{1}{\sqrt[]{a}} Ψ (\frac{λ - b}{a})$ (1)

式中, b为平移系数, a为伸缩系数, λ 为波段数量; 采用连续小波技术将经传统光谱变换技术处理的光谱数据分解为不同尺度的小波系数

$W_{f} (a, b) = (f, Ψ_{a, b}) = \int_{-}^{\infty} f (λ) Ψ_{a, b} (λ) d λ$ (2)

式中, W_f(a, b)为小波系数, 其由波长(350~2 500 nm)、分解尺度构成的二维矩阵数据; f(λ )为光谱数据。

1.5 模型精度验证

土壤光谱经传统光谱变换与连续小波处理后, 采用偏最小二乘法(partial least squares regression, PLS)构建土壤有机质含量估测模型; 首先将土样数据随机分为2组: 建模组与验证组, 建模组用于构建土壤有机质含量估测模型, 共计64个土样, 验证组用于模型精度的检验, 共计32个土样; 采用决定系数(R²)与均方根误差(RMSE)评价模型的精度, 并从建模精度与验证精度的差异评价模型的稳定性。

$R^{2} = 1 - \frac{\overset{n}{\sum_{i = 0}} (SO M_{i} - SOM P_{i})^{2}}{\overset{n}{\sum_{i = 0}} (SO M_{i} - {\bar{SOM}}_{i})^{2}}$ (3)

$RMSE = \sqrt[]{\frac{\overset{n}{\sum_{i = 1}} (SO M_{i} - SOM P_{i})^{2}}{n}}$ (4)

式中, SOM是实测土壤有机质含量, SOMP是基于土壤有机质含量估测模型的预测值, $\bar{SOM}$ 是实测土壤有机质含量的均值。

2 结果与讨论

2.1 光谱分析

图2为原始光谱数据(a)及其倒数变换(b), 经连续小波分解后的光谱数据。由图2(a)知: 原始光谱经连续小波分解后, 1~10尺度的光谱曲线特征逐渐减弱; 1~6尺度的波谱曲线起伏较大且整体形态相似, 7~8尺度曲线光滑呈对称抛物线状, 9尺度曲线逐渐增大小锯齿构成且呈弧形, 10尺度则由较大锯齿构成且呈直线状; 随着尺度增加, 位于1尺度波谱末端的波峰逐渐向短波方向移动。由图2(b)知: 与原始光谱相比, 倒数变换经连续小波处理后, 光谱曲线整体变化减弱, 两端光谱振幅的相对幅度颠倒, 位于1尺度短波方向的波峰逐渐向长波方向移动与原始光谱相反; 从短波至长波锯齿状曲线在8~9尺度呈逐渐消失再凸显的规律。

2.2 相关性分析

土壤高光谱数据经传统光谱变换与连续小波处理后, 与土壤有机质含量相关系数矩阵见图3。由图知: 对土壤有机质含量敏感的光谱响应波段主要集中在全波段的前半部分, 即可见光与波长小于1 200 nm近红外处; 与原始光谱和单一光谱处理相比^[16], 有效提升了土壤光谱对有机质含量的敏感性, 这表明耦合传统光谱变换与连续小波变换可深度挖掘光谱信息内的有益信息。在诸相关系数矩阵内, 以倒数为最优, 具有明显优势, 在分解尺度、波长两个维度均大幅提升了光谱信息对土壤有机质含量的敏感性, 其最高相关系数可达R²=0.714。经过传统光谱变换处理后, 虽然光谱与土壤有机质含量的相关性得到大幅度提升, 但敏感波段的波长位置并未发生大幅变动, 相关性的强弱在相关系数矩阵内的格局大体相似, 局部地区发生萎缩或扩张; 除对数的变换外, 原始光谱内的“ 信息平原” 在其他光谱变换形式内消失, 且相关性得到了不同程度的增强或减弱。

	Figure Option View Download New Window
	图2 土壤光谱分析Fig.2 The soil spectral analysis

2.3 传统光谱变换

基于前人采用的11类传统光谱变换, 筛选了出6类结果较理想的光谱变换, 开展对比试验。 6类光谱处理技术为: 微分、倒数、对数、倒数的微分、对数的微分、弓曲差(Gqc)^[18]。

表2为基于传统光谱变换构建的土壤有机质含量估测模型, 由表可知: 各模型预测精度与建模精度均存在较大差异, 且估测精度不高, 这表明模型精度与稳定性较低, 究其原因为原始光谱数据内含有较多噪声, 而传统光谱变换抑制噪声作用较弱。各模型内, 以对数变换构建的模型最优, 其估测精度的R²=0.576, RMSE=0.344。

2.4 连续小波变换建模

基于传统光谱变换与连续波处理后的土壤光谱信息, 构建土壤有机质含量的估测模型(见表3)。所建模型均达到p=0.001的极显著水平, 除原始光谱数据与弓曲差构建的模型外, 其他模型的建模精度与估测精度均达到较理想水平, 其R²均位于0.7~0.8内, RMSE均低于0.3; 为更好的综合评价模型的精度, 采用R²与RMSE两指标开展精度评价。

由表3可知: 基于弓曲差与对数变换构建的模型精度低于原始光谱, 故为无效变换。基于光谱变换构建的模型精度明显优于原始光谱, 这表明耦合传统光谱变换与连续小波可有效提高光谱估测土壤有机质含量的精度, 究其原因由于土壤有机质含量相对较低, 光谱响应信息微弱, 易受其他土壤构成成分干扰, 遮掩, 部分有益信息隐藏于光谱数据内, 须

	Figure Option View Download New Window
	图3 光谱与土壤有机质含量的相关系数矩阵Fig.3 The correlation coefficient matrix of spectrum and soil organic matter content

表2 传统光谱变换土壤有机质含量估测模型 Table 2 Model list for estimating soil organic matter based on traditional spectral transformation

表3 传统光谱变换与连续小波耦合土壤有机质含量估测模型列表 Table 3 Model list for estimating soil organic matter based on tradional spectral transformation and continuous wavelet transform

通过层层挖掘才能凸显。微分变换模型的建模精度与估测精度均优于其他变换, 且建模精度与估测精度间差异最小, 因此, 基于微分变换构建的模型为最优模型, 其估测精度的R²=0.774, RMSE=0.223。这是由于微分变换有益于有效细微信息的挖掘, 通过连续小波分解得到凸显。微分变换模型的稳定性较高表明连续小波可有效抑制高频噪声的影响。

2.5 小波变换建模与光谱反射率建模的比较

由传统光谱变换模型(表2)与连续小波变换模型(表3)对比分析可知, 基于传统光谱变换与连续小波构建的模型精度远高于传统光谱变换, 大幅提升了土壤有机质含量的估测精度, 最高可提升33.7%; 稳定性远高于传统光谱变换模型。这是因为连续小波技术可有效压制噪声干扰, 提升光谱信息的信噪比; 因此基于连续小波可有效提升传统光谱变换对土壤有机质含量的估测精度与稳定性。

土壤光谱是土壤的构成成分、粒度、质地等物理化参数共同作用的结果, 由于潮土有机质含量较低(土样的最高含量为2.5%), 其在土壤光谱的响应特征较弱, 易受其他构成成分或粒度等参数的影响, 致使利用单一光谱变换很难充分挖掘土壤光谱信息。国内外许多学者采用传统光谱变换、连续小波变换、离散小波变换、重采样、光谱模拟等处理土壤光谱数据, 但大多利用单一技术。我们以北方潮土为研究对象, 耦合传统光谱变换与连续小波变换处理土壤光谱, 研究分析优化处理土壤光谱的方法, 结果表明: (1)与前人研究相比, 耦合传统光谱变换与连续小波可更有效挖掘土壤光谱内的有益信息, 抑制土壤光谱噪声的负面影响, 提高信噪比, 增强土壤光谱对有机质含量的敏感性; (2)本研究构建的土壤有机质含量估测模型模型精度更高, 稳定性更好、参与建模波段更少, 该研究为应用遥感技术监测土壤有机质含量提供了新思路。

本研究采用室内光谱, 光谱采集环境较为理想, 受外界干扰较少, 光谱数据信噪比较高。但土壤有机质含量的大尺度监测主要在野外开展, 不可控因素较多, 降低了光谱数据的信噪比。因此, 将进一步开展野外条件下的研究工作。

3 结论

以北方潮土为研究对象, 耦合传统光谱变换与连续小波变换, 并构建土壤有机质含量估测模型, 研究结果如下:

(1)耦合传统光谱变换与连续小波可深入挖掘有益信息, 提升土壤光谱对土壤有机质含量的敏感性。与土壤有机质含量敏感的特征波段主要集中于可见光与波长小于1 200 nm的近红外的区域。最高相关系数的R²达0.714;

(2)在6类估测土壤有机含量的模型中, 以基于微分变换与连续小波变换构建的模型最优, 其R²=0.774, RMSE=0.223, 其次为倒数变换;

(3)大幅提升光谱对土壤有机质含量的估测精度与稳定性, 最高提升了33.7%。

The authors have declared that no competing interests exist.

参考文献

文献选项

[1]	ZHAO Chun-jiang(赵春江). Agriculture Network Information(农业网络信息), 2010, (4): 5. [本文引用:1]
[2]	XU Xin, ZHANG Hao, XI Lei, et al(许鑫, 张浩, 席磊, 等). Transactions of the Chinese Society of Agricultural Engineering(农业工程学报), 2011, 27(S2): 94. [本文引用:1]
[3]	CHEN Gui-fen, MA Li, CHEN Hang(陈桂芬, 马丽, 陈航). Journal of Jilin Agricultural University(吉林农业大学学报), 2013, 35(3): 253. [本文引用:1]
[4]	MENG Ji-hua, WU Bing-fang, DU Xin, et al(蒙继华, 吴炳方, 杜鑫, 等). Remote Sensing for Land & Resources(国土资源遥感), 2011, 90(3): 1. [本文引用:1]
[5]	YANG Bang-jie, PEI Zhi-yuan, ZHOU Qing-bo, et al(杨邦杰, 裴志远, 周清波, 等). Transactions of the Chinese Society of Agricultural Engineering(农业工程学报). 2002, 18(3): 191. [本文引用:1]
[6]	YANG Jian-feng, MA Jun-cheng, WANG Ling-chao(杨建锋, 马军成, 王令超). Geospatial Information(地理空间信息), 2015, 13(2): 47. [本文引用:1]
[7]	Masserschmidt I, Cuelbas C J, Poppi R J, et al. Journal of Chemometrics, 1999, 13: 265. [本文引用:1]
[8]	WU Jie, LI Yu-huan, LI Zeng-bing, et al(武婕, 李玉环, 李增兵, 等). Acta Ecologica Sinica(生态学报), 2014, 34(6): 1596. [本文引用:1]
[9]	Moslem Ladoni, Hosein Ali Bahrami, Sayed Kazem Alavipanah, et al. Precision Agric. , 2010, 11: 82. [本文引用:1]
[10]	Krishnan P, Alexand er J D, Butler B J, et al. Soil Science Society of America Journal, 1980, 44: 1282. [本文引用:1]
[11]	Alicia Palacios-Orueta, Jorge E Pinzon, Susan L Ustin, et al. Remote Sensing of Environment, 1998, 68: 138. [本文引用:1]
[12]	Liu H J, Zhang Y Z, Zhang B, et al. Environmental Monitoring and Assessment, 2009, 154: 147. [本文引用:1]
[13]	JI Wen-jun, SHI Zhou, ZHOU Qing, et al(纪文君, 史舟, 周清, 等). Journal of Infrared and Millimeter Waves(红外与毫米波学报), 2012, 31(3): 277. [本文引用:1]
[14]	HE Ting, WANG Jing, et al(何挺, 王静, 等). Geomatics and Information Science of Wuhan University(武汉大学学报·信息科学版), 2006, 31(11): 975. [本文引用:1]
[15]	Viscarra Rossela R A, Chappella A, de Caritatb P, et al. European Journal of Soil Science, 2011, 62: 442. [本文引用:1]
[16]	LIAO Qin-hong, GU Xiao-he, LI Cun-jun, et al(廖钦洪, 顾晓鹤, 李存军, 等). Transactions of the Chinese Society of Agricultural Engineering(农业工程学报), 2012, 28(23): 132. [本文引用:2]
[17]	XU Yong-ming, LIN Qi-zhong, et al(徐永明, 蔺启忠, 等). Acta Pedologica Sinica(土壤学报), 2006, 43(5): 709. [本文引用:1]
[18]	ZHAO Ze-hai, ZU Yuan-gang, CONG Pei-tong(赵则海, 祖元刚, 丛沛桐). Acta Ecologica Sinica(生态学报), 2002, 22(10): 1660. [本文引用:1]

2010

0.0

ZHAO

Chun-jiang

(赵春江). Agriculture Network Information(农业网络信息), 2010, (4): 5.

从保证国家粮食安全、解决资源环境问题的角度阐述了我国发展精准农业的重要性,分析了当前我国精准农业发展需要关注的6个问题,即突破精准农业关键技术瓶颈、降低精准农业技术应用的门槛、紧密与当地农业主导产业结合、采用因地制宜的技术应用模式、采取高效灵活的技术推广模式和制定精准农业技术标准.提出了今后我国精准农业发展战略的重点为构建具有自主知识产权的精准农业技术体系、具有中国特色的精准技术应用体系、学科交叉的精准农业人才体系.

... 引言精准农业其涵义为在资源硬约束条件下, 利用信息技术对耕地生产开展的定量、定位的现代农业生产模式, 旨在改变传统耕作模式, 提升耕地生产潜力与生产效率^[1] ...

2011

0.0

... 精准施肥是精准农业系统内的关键技术, 其依据作物生长需肥规律与耕地养分现状进行变量的投入, 以达到充分利用资源、节约成本、保护环境的效果^[2,3] ...

2013

0.0

CHEN

Gui-fen

, MA

, CHEN

Hang

(陈桂芬, 马丽, 陈航). Journal of Jilin Agricultural University(吉林农业大学学报), 2013, 35(3): 253.

Precision fertilization is the central part of precision agric ultural technology Precision fertilization implementation can save fertilizer, balance soil nutrients, reduce environmental pollution and increase grain yield Three a spects of precision fertilizer, including soil fertility analysis, fertilization model and fertilization decision making, are reviewed in this paper Precision fertilization research experiences and shortcomings are summarized Development trend is also predicted, with the purpose of providing reference for precision f ertilization research and applications

College of Information Technology, Jilin Agricultural University, Changchun 130118, China

精准施肥是精准农业技术中的核心内容。精准施肥的实施可以节约肥料，均衡土壤养分，减少环境污染，增加粮食产量。文中从土壤肥力分析、施肥模型和施肥决策3个方面对精准施肥技术的研究现状进行了综述，总结了在精准施肥研究方面的经验与不足，指出了其未来发展趋势，为今后的精准施肥研究和应用提供参考。

2011

0.0

... 近年来, 遥感技术的持续进步与国产卫星的迅猛发展展为耕地养分监测提供了新的思路, 逐步成为农业遥感的研究热点^[4,5,6] ...

2002

0.0

YANG

Bang-jie

, PEI

Zhi-yuan

, ZHOU

Qing-bo

, et al(杨邦杰, 裴志远, 周清波, 等). Transactions of the Chinese Society of Agricultural Engineering(农业工程学报). 2002, 18(3): 191.

There are over 20 years' research experience for remote sensing used in crop monitoring in North America, European counties and China. A crop monitoring system using remote sensing has been developed and used by China Ministry of Agriculture. As China is a large country with complicated landscape, multi-crop systems and small family farms and the government needs operational system at a national scale, many key techniques still need researching. The main issues discussed in this paper include remote sensing data processing, crop discrimination, crop area measurement, crop condition assessment, disaster monitoring, yield estimation, and operational system development. Based on the authors' experience, the progress, problems and the main ideas for solutions of the operational system development at a national scale are discussed in this paper.

我国作物的遥感监测已有20年的研究历史。但象中国这样国土辽阔、地形多样、种植结构复杂、农户规模小的国家的国家级农情遥感监测运行系统，在关键技术方面仍然需要加强研究。这些关键技术包括：数据处理、作物识别、面积量算、长势监测、灾害评估、产量估计及运行系统集成等。该文根据作者开发农业部农情遥感监测系统的经验，分析存在的问题，提出解决问题的基本思路。

... 近年来, 遥感技术的持续进步与国产卫星的迅猛发展展为耕地养分监测提供了新的思路, 逐步成为农业遥感的研究热点^[4,5,6] ...

2015

0.0

... 近年来, 遥感技术的持续进步与国产卫星的迅猛发展展为耕地养分监测提供了新的思路, 逐步成为农业遥感的研究热点^[4,5,6] ...

1999

0.0

... 耕层土壤有机质含量的高低是衡量耕地土壤肥力与质量的重要指标^[7], 与作物生长所必需的微量元素的空间分布具有较强的相关性^[8] ...

2014

0.0

Jie

, LI

Yu-huan

, LI

Zeng-bing

, et al(武婕, 李玉环, 李增兵, 等). Acta Ecologica Sinica(生态学报), 2014, 34(6): 1596.

Soil organic matter (SOM) and trace elements are important indicators of soil quality for sustainable agriculture. The study of spatial distribution and influencing factors of soil organic matter and trace elements provided references for understanding regional soil characteristics,soil fertility and quality optimization and farmland ecological environment protection. A total of 1021 soils samples were taken in the Nansihu region, China, for SOM and trace elements analysis. The spatial distribution of SOM and these trace elements was determined using geostatistics and GIS technology such as semivariogram, ordinary Kriging interpolation, and spatial analysis methods. Factorings related with the spatial variations including soil types, soils texture, slope, land uses, and land forms were examined. Descriptive statistics showed that SOM averaged 14.66 g/kg with a coefficient of variation (CV) of 13.2%. Among the five trace elements, the average content was highest for Mn (10.66 mg/kg) and lowest for B (0.57 mg/kg). The CV was smallest for Fe(35.1%) and largest for B (96.5%). The normal distribution test showed that B was accorded with normal distribution while SOM, Fe, Mn, Cu and Zn conformed to logarithmic normal distribution. Semivariogram analyses indicated that B was of the pure nugget effect, with constant variation in the study scale. SOM, Fe, Mn, Cu and Zn displayed strong spatial autocorrelation and the structural factors played a leading role. The spatial correlation distance was 84 m for SOM, and ranged from 162 m (Fe) to 2572 m (Mn) for trace elements. The ordinary kriging interpolation results showed that the overall trend of SOM distribution was gradually decreased from north to south. The higher soil organic matter values were mainly distributed in the flat area. The values of Fe were highest in the northern mountainous area and hilly land (>14 mg/kg) and lowest in the South (3-9 mg/kg). A spatial pattern with higher concentrations in the central area than in the north and south area was observed for Mn, Cu and Zn. This was especially so for Mn ranging from 13 mg/kg in the central area. No significant spatial variation was observed with B (mostly >0.55 mg/kg).Soil type and topsoil texture were the predominant factors influencing spatial variations of SOM and trace elements except for B. SOM content was highest in the paddy soil. Fe, Mn contents were highest in the cinnamonic soil. Cu and Zn contents were highest in the mortar black soil. SOM increased with increasing clay content. Fe and Mn were highest in heavy loam. Cu was highest in slight loam and Zn was highest in sandy loam. Slope had a certain influence on some indicators, but the effect was small. The organic matter content increased with increasing slope. Land use types had significant influence on SOM, Fe and Cu, but had no effect on Mn, Zn and B. Contents of SOM, Fe and Cu was higher in paddy field. Topographic types had a greater influence on SOM, Fe, Mn and Zn, but had no impact on Cu and B, and the contents of trace elements were higher in depression.

基于地统计学和GIS技术相结合的方法，研究了南四湖区农田土壤有机质和微量元素的空间分布特征及其影响因素。结果表明，土壤有机质和微量元素均属中等变异程度，除硼符合正态分布外，其余土壤属性均符合对数正态分布。结构分析表明，除硼为纯块金效应外，土壤有机质和其它微量元素空间自相关性较强，其中结构性因素起主导作用。克里格插值结果表明，土壤有机质分布总体趋势为由北向南逐渐降低，锰、铜、锌分布总体趋势为中部高，南北两端低。影响因素分析表明，土壤类型、耕层质地、坡度、土地利用类型和地貌类型对土壤有机质均有显著影响。土壤类型主要是由于成土母质的差异影响土壤有机质的高低与分布，随质地由砂变粘、坡度由低变高，土壤有机质含量逐步升高，田间管理水平的差异是造成不同土地利用类型下土壤有机质含量差异的主要原因。微量元素中，除硼不受影响外，铁、锰、铜和锌与土壤类型、耕层质地、坡度、土地利用类型和地貌类型密切相关。

... 耕层土壤有机质含量的高低是衡量耕地土壤肥力与质量的重要指标^[7], 与作物生长所必需的微量元素的空间分布具有较强的相关性^[8] ...

2010

0.0

... 土壤有机质含量的光谱响应特征主要集中于可见光-近红外区域^{[9,10,11,12,13,14,15]}, 但若土壤有机质含量太低其光谱太弱, 不能应用光谱估测土壤有机质含量, 即存在一个临界点 ...

1980

0.0

1998

0.0

2009

0.0

2012

0.0

2006

0.0

2011

0.0

2012

0.0

LIAO

Qin-hong

, GU

Xiao-he

, LI

Cun-jun

, et al(廖钦洪, 顾晓鹤, 李存军, 等). Transactions of the Chinese Society of Agricultural Engineering(农业工程学报), 2012, 28(23): 132.

Hyperspectral remote sensing technology had been widely used in the estimation of soil organic matter due to its non-destructive, rapid, and high spectral resolution characteristics. The 64 fluvo-aquic soil organic matter (SOM) obtained from Beijing Shunyi district had been estimated successfully by using the hyperspectral reflectance based on continuous wavelet transform (CWT), the results had also been compared with four common spectral transformation methods, it showed that the fluvo-aquic soil had the same spectral curves with other types soils, and the absorption peaks appeared in the visible and near infrared bands after the spectral curves removed by hull curve. The sensitive bands for estimating the SOM were 1194nm, 486nm, and 866nm, and the corresponding wavelet decomposition scales were 2, 3, and 4. The R2 of multiple linear regression model built between the wavelet energy coefficients and SOM was 0.67 by using the CWT, the R2and RMSE between the measured value and predicted value were 0.75 and 0.21, respectively, while the highest R2 of estimation model built by the four common spectral transformation methods was 0.09, which showed that the CWT is more suitable for estimating the fluvo-aquic soil organic matter content. Through the interpolation analysis of Kringing, the more sampling points should be increased in the southeast of Shunyi district for improving the precision of models.

土壤有机质含量快速估算对于土壤肥力评价、土壤信息化管理和精准施肥具有重要意义。该文通过对北京顺义地区64个土壤样品高光谱曲线进行连续小波变换，估算了该地区潮土有机质质量分数，并与4种常用光谱变换方法进行了比较。结果表明，潮土具有与其他类型土壤类似的光谱曲线，经过去包络线处理后，在可见与近红波段都出现了明显吸收峰；采用连续小波变换方法所确定的潮土有机质估算的敏感波段为1 194、486和866 nm，对应小波分解尺度为2，3和4；利用小波能量系数与有机质质量分数所构建的多元线性回归模型的决定系数R2为0.67，模型实测值与预测值的检验精度R2为0.75，RMSE为0.21；而采用4种常用光谱变换方法建立的潮土有机质估测模型的R2最高只有0.09，说明连续小波变换方法更适合于潮土有机质质量分数估测。Kringing插值分析表明，应在顺义地区东南部增加取样点，以提高模型估算精度。该研究可为潮土土壤肥力的快速测定提供参考。

... 微分变换、对数变换、倒数变换、离散小波变换、连续小波变换等传统光谱变换均能提升光谱对土壤有机质含量的相关性^[16], 且以小波技术为最佳, 但其对光谱敏感性的提升程度受限 ...

... 与原始光谱和单一光谱处理相比^[16], 有效提升了土壤光谱对有机质含量的敏感性, 这表明耦合传统光谱变换与连续小波变换可深度挖掘光谱信息内的有益信息 ...

2006

0.0

2002

0.0

ZHAO

Ze-hai

, ZU

Yuan-gang

, CONG

Pei-tong

(赵则海, 祖元刚, 丛沛桐). Acta Ecologica Sinica(生态学报), 2002, 22(10): 1660.

The method of wavelet decomposition can be used to analyse approximations and details of one dimension soil signal. It is an effective tool for soil factors analysis. The study on decomposition of Haar wavelet for three main soil factors in belt transect of Liaotunggensis forest in Dongling Mountain area, Beijing, was done by this method. The research results showed: along hillslope from under to upper, soil moisture decreased obviously step-by-step; pH of soil displayed relative stabilization and changed ...

小波分解可以区分一维信号的区域变化和局部细节 ,是进行土壤因子分析的一种有效方法。应用该方法对北京东灵山地区辽东栎林样带中的 3个主要土壤因子指标进行 Haar小波分解。结果表明 :沿山下到山顶样带梯度 ,土壤含水量呈明显的逐步下降趋势 ;土壤 p H值变化幅度在 6 .5～ 6 .6之间 ,表现比较稳定 ;土壤有机质含量在样带中存在 3个较高值区域 ,具有一定的复杂性。较大水平上 3个土壤因子在样带的下部受到的干扰较大 ,在样带的中、上部受到的干扰较小 ;在较小水平上 ,土壤含水量存在不规则的周期性干扰 ,p H值和有机质含量受干扰的情况较为复杂 ,规律性不明显。低频近似信号可以有效地揭示出土壤因子在样带中的总体变化规律 ,高频细节信号主要反映土壤因子受其它因素干扰程度 ,并可以在样带上定位出现奇异性变化的区域。

... 6类光谱处理技术为: 微分、倒数、对数、倒数的微分、对数的微分、弓曲差(Gqc)^[18] ...