基于LM算法的土壤表层含水率遥感监测
|
摘要
为探讨数据挖掘技术中LM(Levenberg-Marquardt)算法在土壤表层(约1 cm)含水率遥感监测中的应用,选取黄绵土、粘黄土、红土为试验材料,配制含水率分别为0、6%、10%、14%、18%、22%的土壤样本,在09:00—10:00和15:00—16:00时间段进行可见光采样,并对图像亮度进行梯度处理,以此模拟全天光线变化。采用样本实测含水率及图像RGB三阶颜色矩数据作为数据集,对上午、下午样本和两时间段混合样本采用LM算法建立含水率回归模型,并与BP(Back propagation)算法和分类回归树(Classification and regression trees,CART)算法进行比较。结果表明,基于土壤表层RGB颜色矩的LM算法具有较好的应用效果,混合样本不同土样回归模型决定系数R~2分别为0. 958、0. 943、0. 949,均方根误差(RMSE)分别为1. 6%、2. 0%、1. 9%,相对分析误差(RPD)分别为4. 873、4. 183、4. 440。不同光照时的混合样品分析结果表明,LM算法适用于不同光线采集样品的土壤含水率监测,适用于土壤表层(约1 cm)含水率的监测。
Aiming to research the data mining technology in remote sensing monitoring. The LM algorithm was used in the soil surface layer( about 1 cm) of soil moisture measurement( soil moisture content,SMC). Three kinds of soil,including yellow spongy,loess and red clay were selected. The soil water content samples of 0,6%,10%,14%,18% and 22% were prepared respectively. Visible light images were taken during 09: 00—10: 00 and 15: 00—16: 00,and image brightness was gradiently processed for simulating the change of light throughout the day. LM algorithm was compared with back propagation( BP) algorithm and classification and regression trees( CART) algorithm to verify the practical effect of LM. It was showed that the LM algorithm had a good application effect for the data mining based on the RGB color moment of soil pictures. The determination coefficient of the regression model for yellow spongy,loess and red clay was 0. 958,0. 943 and 0. 949,root mean square error( RMSE) was 1. 6%,2. 0% and 1. 9%,and the relative analysis error( RPD) was 4. 873,4. 183 and4. 440,respectively. By the study of pictures at different intensities,LM algorithm can be used for monitoring soil moisture content of samples. It can be used for the soil surface( about 1 cm) moisture content measurement.
Aiming to research the data mining technology in remote sensing monitoring. The LM algorithm was used in the soil surface layer( about 1 cm) of soil moisture measurement( soil moisture content,SMC). Three kinds of soil,including yellow spongy,loess and red clay were selected. The soil water content samples of 0,6%,10%,14%,18% and 22% were prepared respectively. Visible light images were taken during 09: 00—10: 00 and 15: 00—16: 00,and image brightness was gradiently processed for simulating the change of light throughout the day. LM algorithm was compared with back propagation( BP) algorithm and classification and regression trees( CART) algorithm to verify the practical effect of LM. It was showed that the LM algorithm had a good application effect for the data mining based on the RGB color moment of soil pictures. The determination coefficient of the regression model for yellow spongy,loess and red clay was 0. 958,0. 943 and 0. 949,root mean square error( RMSE) was 1. 6%,2. 0% and 1. 9%,and the relative analysis error( RPD) was 4. 873,4. 183 and4. 440,respectively. By the study of pictures at different intensities,LM algorithm can be used for monitoring soil moisture content of samples. It can be used for the soil surface( about 1 cm) moisture content measurement.
引文
[1]张智韬,王海峰,韩文霆,等.基于无人机多光谱遥感的土壤含水率反演研究[J/OL].农业机械学报,2018,49(2):173-181.ZHANG Zhitao,WANG Haifeng,HAN Wenting,et al. Inversion of soil moisture content based on multispectral remote sensing of UAVs[J/OL]. Transactions of the Chinese Society for Agricultural Machinery,2018,49(2):173-181. http:∥www. jcsam. org/jcsam/ch/reader/view_abstract. aspx? flag=1&file_no=20180223&journal_id=jcsam. DOI:10. 6041/j. issn. 1000-1298. 2018. 02. 023.(in Chinese)
[2]陈仲新,任建强,唐华俊,等.农业遥感研究应用进展与展望[J].遥感学报,2016,20(5):748-767.CHEN Zhongxin,REN Jianqiang,TANG Huajun,et al. Progress and perspectives on agricultural remote sensing research and applications in China[J]. Journal of Remote Sensing,2016,20(5):748-767.(in Chinese)
[3] KUMAR S V,DIMEYER P A,PETERS-LIDARD C D,et al. Information theoretic evaluation of satellite soil moisture retrievals[J]. Remote Sensing of Environment,2018,204:392-400.
[4]殷超,周忠发,谭玮颐,等.基于微波与光学遥感的石漠化地区土壤剖面含水率反演模型研究[J].红外与毫米波学报,2018,37(3):360-370.YIN Chao,ZHOU Zhongfa,TAN Weiyi,et al. Inversion model of soil profile moisture content in rocky desertification area based on microwave and optical remote sensing[J]. Journal of Infrared and Millimeter Waves,2018,37(3):360-370.(in Chinese)
[5]张智韬,王海峰,KARNIELI A,等.基于岭回归的土壤含水率高光谱反演研究[J/OL].农业机械学报,2018,49(5):240-248.ZHANG Zhitao,WANG Haifeng,KARNIELI A,et al. Hyperspectral inversion of soil water content based on ridge regression[J/OL]. Transactions of the Chinese Society for Agricultural Machinery,2018,49(5):240-248. http:∥www. j-csam. org/jcsam/ch/reader/view_abstract. aspx? flag=1&file_no=20180528&journal_id=jcsam. DOI:10. 6041/j. issn. 1000-1298.2018. 05. 028.(in Chinese)
[6]杨曦光,于颖.基于试验反射光谱数据的土壤含水率遥感反演[J].农业工程学报,2017,33(22):195-199.YANG Xiguang,YU Ying. Remote sensing inversion of soil moisture content based on experimental reflectance spectral data[J]. Transactions of the CSAE,2017,33(22):195-199.(in Chinese)
[7]李明泽,高元科,邸雪颖,等.基于微波遥感技术探测森林地表土壤含水率[J].应用生态学报,2016,27(3):785-793.LI Mingze,GAO Yuanke,DI Xueying,et al. Detecting the moisture content of forest surface soil based on the microwave remote sensing technology[J]. Chinese Journal of Applied Ecology,2016,27(3):785-793.(in Chinese)
[8] PETROPOULOS G P,IRELAND G,BARRETT B. Surface soil moisture retrievals from remote sensing:current status,products&future trends[J]. Physics and Chemistry of the Earth,Parts A/B/C,2015,83-84:36-56.
[9]殷哲,雷廷武,陈展鹏,等.近红外传感器测量不同种类土壤含水率的适应性研究[J/OL].农业机械学报,2014,45(3):148-151.YIN Zhe,LEI Tingwu,CHEN Zhanpeng,et al. Adaptability of near-infrared sensor for moisture measurement of different soils[J/OL]. Transactions of the Chinese Society for Agricultural Machinery,2014,45(3):148-151. http:∥www. j-csam. org/jcsam/ch/reader/view_abstract. aspx? flag=1&file_no=20140325&journal_id=jcsam. DOI:10. 6041/j. issn. 1000-1298.2014. 03. 025.(in Chinese)
[10]李小昱,肖武,李培武,等.基于近红外傅里叶特征提取方法的土壤含水率检测[J].农业机械学报,2009,40(5):64-67.LI Xiaoyu,XIAO Wu,LI Peiwu,et al. Soil moisture content detection based on near-infrared Fourier feature extraction method[J]. Transactions of the Chinese Society for Agricultural Machinery,2009,40(5):64-67.(in Chinese)
[11]罗蔚然,张成才,闫超德,等.基于微波和可见光遥感的冬小麦土壤墒情反演[J].水利信息化,2017(5):46-50.LUO Weiran,ZHANG Chengcai,YAN Chaode,et al. Winter wheat soil moisture inversion based on microwave and visible light remote sensing[J]. Water Conservancy Information,2017(5):46-50.(in Chinese)
[12]张荣标,何加祥,吴涛,等.基于图像处理的土壤表层含水率在线检测方法研究[J].农业工程学报,2007,23(10):57-61.ZHANG Rongbiao,HE Jiaxiang,WU Tao,et al. On-line detection method of soil surface moisture content based on image processing[J]. Transactions of the CSAE,2007,23(10):57-61.(in Chinese)
[13] ZHANG Xianlong,ZHANG Fei,ZHANG Haiwei,et al. Optimization of soil salt inversion model based on spectral transformation from hyperspectral index[J]. Transactions of CSAE,2018,34(1):110-117.
[14] JEBUR A A,ATHERTON W,AL KHADDAR R M,et al. Settlement prediction of model piles embedded in sandy soil using the Levenberg-Marquardt(LM)training algorithm[J]. Geotechnical and Geological Engineering,2018,36(5):2893-2906.
[15]李明河,周磊,王健.基于LM算法的溶解氧神经网络预测控制[J/OL].农业机械学报,2016,47(6):297-302.LI Minghe,ZHOU Lei,WANG Jian. Predictive control of dissolved oxygen neural network based on LM algorithm[J/OL].Transactions of the Chinese Society for Agricultural Machinery,2016,47(6):297-302. http:∥www. j-csam. org/jcsam/ch/reader/view_abstract. aspx? flag=1&file_no=20160639&journal_id=jcsam. DOI:10. 6041/j. issn. 1000-1298. 2016. 06.039.(in Chinese)
[16]刘振华,范宏运,朱宇泽,等.基于BP神经网络的溶洞规模预测及应用[J].中国岩溶,2018,37(1):139-145.LIU Zhenhua,FAN Hongyun,ZHU Yuze,et al. Prediction model for the scale of karst cave based on back propagation artificial neural network and its application[J]. Carsologica Sinica,2018,37(1):139-145.(in Chinese)
[17]高斌,赵鹏飞,卢昱欣,等.基于BP神经网络的血液荧光光谱识别分类研究[J].光谱学与光谱分析,2018,38(10):3136-3143.GAO Bin,ZHAO Pengfei,LU Yuxin,et al. Study on recognition and classification of blood fluorescence spectrum with BP neural network[J]. Spectroscopy and Spectral Analysis,2018,38(10):3136-3143.(in Chinese)
[18] JIA Zhenyi,ZHOU Shenglu,SU Quanlong,et al. Comparison study on the estimation of the spatial distribution of regional soil metal(loid)s pollution based on Kriging interpolation and BP neural network[J]. International Journal of Environmental Research&Public Health,2018,15(1):34.
[19]郭澎涛,苏艺,茶正早,等.基于BP神经网络的橡胶苗叶片磷含量高光谱预测[J].农业工程学报,2016,32(增刊1):177-183.GUO Pengtao,SU Yi,CHA Zhengzao,et al. Hyperspectral prediction of phosphorus content in rubber seedling leaves based on BP neural network[J]. Transactions of the CSAE,2016,32(Supp. 1):177-183.(in Chinese)
[20]刘雪,李亚妹,刘娇,等.基于BP神经网络的鲜鸡蛋货架期预测模型[J/OL].农业机械学报,2015,46(10):328-334.LIU Xue,LI Yamei,LIU Jiao,et al. Prediction model of shelf life of fresh eggs based on BP neural network[J/OL].Transactions of the Chinese Society for Agricultural Machinery,2015,46(10):328-334. http:∥www. j-csam. org/jcsam/ch/reader/view_abstract. aspx? flag=1&file_no=20151044&journal_id=jcsam. DOI:10. 6041/j. issn. 1000-1298. 2015. 10.044.(in Chinese)
[21] HAN Jiaqi,MAO Kebiao,XU Tongren,et al. A soil moisture estimation framework based on the CART algorithm and its application in China[J]. Journal of Hydrology,2018,563:65-75.
[22]吕利利,颉耀文,黄晓君,等.基于CART决策树分类的沙漠化信息提取方法研究[J].遥感技术与应用,2017,32(3):499-506.LULili,JIE Yaowen,HUANG Xiaojun,et al. Research on desertification information extraction method based on CART decision tree classification[J]. Remote Sensing Technology and Application,2017,32(3):499-506.(in Chinese)
[23]张良均,杨坦,肖刚,等. MATLAB数据分析与挖掘实战[M].北京:机械工业出版社,2015.
[24]张海辉,陶彦蓉,胡瑾.融合叶绿素含量的黄瓜幼苗光合速率预测模型[J/OL].农业机械学报,2015,46(8):259-263.ZHANG Haihui,TAO Yanrong,HU Jin. Photosynthetic rate prediction model of cucumber seedlings with chlorophyll content[J/OL]. Transactions of the Chinese Society for Agricultural Machinery,2015,46(8):259-263. http:∥www. j-csam. org/jcsam/ch/reader/view_abstract. aspx? flag=1&file_no=20150836&journal_id=jcsam. DOI:10. 6041/j. issn. 1000-1298.2015. 08. 036.(in Chinese)
[25]韩敏,穆大芸.回声状态网络LM算法及混沌时间序列预测[J].控制与决策,2011,26(10):1469-1472.HAN Min,MU Dayun. Echo state network LM algorithm and chaotic time series prediction[J]. Control and Decision,2011,26(10):1469-1472.(in Chinese)
[26] LEVENBERG K. A method for the solution of certain non-linear problems in least squares[J]. Quarterly of Applied Mathematics,1944,2(4):436-438.
[27] WILAMOWSKI B M,YU Hao. Improved computation for Levenberg-Marquardt training[J]. IEEE Transactions on Neural Networks,2010,21(6):930-937.
[28] JEFFREY K C F,TSAI W L,SU T L,et al. Application of an LM-neural network for establishing a prediction system of quality characteristics for the LGP manufactured by CO2laser[J]. Optics&Laser Technology,2011,43(3):529-536.
[29]赵弘,周瑞祥,林廷圻.基于Levenberg-Marquardt算法的神经网络监督控制[J].西安交通大学学报,2002,36(5):523-527.ZHAO Hong,ZHOU Ruixiang,LIN Tingqi. Neural network supervision and control based on Levenberg-Marquardt algorithm[J]. Journal of Xi'an Jiaotong University,2002,36(5):523-527.(in Chinese)
[2]陈仲新,任建强,唐华俊,等.农业遥感研究应用进展与展望[J].遥感学报,2016,20(5):748-767.CHEN Zhongxin,REN Jianqiang,TANG Huajun,et al. Progress and perspectives on agricultural remote sensing research and applications in China[J]. Journal of Remote Sensing,2016,20(5):748-767.(in Chinese)
[3] KUMAR S V,DIMEYER P A,PETERS-LIDARD C D,et al. Information theoretic evaluation of satellite soil moisture retrievals[J]. Remote Sensing of Environment,2018,204:392-400.
[4]殷超,周忠发,谭玮颐,等.基于微波与光学遥感的石漠化地区土壤剖面含水率反演模型研究[J].红外与毫米波学报,2018,37(3):360-370.YIN Chao,ZHOU Zhongfa,TAN Weiyi,et al. Inversion model of soil profile moisture content in rocky desertification area based on microwave and optical remote sensing[J]. Journal of Infrared and Millimeter Waves,2018,37(3):360-370.(in Chinese)
[5]张智韬,王海峰,KARNIELI A,等.基于岭回归的土壤含水率高光谱反演研究[J/OL].农业机械学报,2018,49(5):240-248.ZHANG Zhitao,WANG Haifeng,KARNIELI A,et al. Hyperspectral inversion of soil water content based on ridge regression[J/OL]. Transactions of the Chinese Society for Agricultural Machinery,2018,49(5):240-248. http:∥www. j-csam. org/jcsam/ch/reader/view_abstract. aspx? flag=1&file_no=20180528&journal_id=jcsam. DOI:10. 6041/j. issn. 1000-1298.2018. 05. 028.(in Chinese)
[6]杨曦光,于颖.基于试验反射光谱数据的土壤含水率遥感反演[J].农业工程学报,2017,33(22):195-199.YANG Xiguang,YU Ying. Remote sensing inversion of soil moisture content based on experimental reflectance spectral data[J]. Transactions of the CSAE,2017,33(22):195-199.(in Chinese)
[7]李明泽,高元科,邸雪颖,等.基于微波遥感技术探测森林地表土壤含水率[J].应用生态学报,2016,27(3):785-793.LI Mingze,GAO Yuanke,DI Xueying,et al. Detecting the moisture content of forest surface soil based on the microwave remote sensing technology[J]. Chinese Journal of Applied Ecology,2016,27(3):785-793.(in Chinese)
[8] PETROPOULOS G P,IRELAND G,BARRETT B. Surface soil moisture retrievals from remote sensing:current status,products&future trends[J]. Physics and Chemistry of the Earth,Parts A/B/C,2015,83-84:36-56.
[9]殷哲,雷廷武,陈展鹏,等.近红外传感器测量不同种类土壤含水率的适应性研究[J/OL].农业机械学报,2014,45(3):148-151.YIN Zhe,LEI Tingwu,CHEN Zhanpeng,et al. Adaptability of near-infrared sensor for moisture measurement of different soils[J/OL]. Transactions of the Chinese Society for Agricultural Machinery,2014,45(3):148-151. http:∥www. j-csam. org/jcsam/ch/reader/view_abstract. aspx? flag=1&file_no=20140325&journal_id=jcsam. DOI:10. 6041/j. issn. 1000-1298.2014. 03. 025.(in Chinese)
[10]李小昱,肖武,李培武,等.基于近红外傅里叶特征提取方法的土壤含水率检测[J].农业机械学报,2009,40(5):64-67.LI Xiaoyu,XIAO Wu,LI Peiwu,et al. Soil moisture content detection based on near-infrared Fourier feature extraction method[J]. Transactions of the Chinese Society for Agricultural Machinery,2009,40(5):64-67.(in Chinese)
[11]罗蔚然,张成才,闫超德,等.基于微波和可见光遥感的冬小麦土壤墒情反演[J].水利信息化,2017(5):46-50.LUO Weiran,ZHANG Chengcai,YAN Chaode,et al. Winter wheat soil moisture inversion based on microwave and visible light remote sensing[J]. Water Conservancy Information,2017(5):46-50.(in Chinese)
[12]张荣标,何加祥,吴涛,等.基于图像处理的土壤表层含水率在线检测方法研究[J].农业工程学报,2007,23(10):57-61.ZHANG Rongbiao,HE Jiaxiang,WU Tao,et al. On-line detection method of soil surface moisture content based on image processing[J]. Transactions of the CSAE,2007,23(10):57-61.(in Chinese)
[13] ZHANG Xianlong,ZHANG Fei,ZHANG Haiwei,et al. Optimization of soil salt inversion model based on spectral transformation from hyperspectral index[J]. Transactions of CSAE,2018,34(1):110-117.
[14] JEBUR A A,ATHERTON W,AL KHADDAR R M,et al. Settlement prediction of model piles embedded in sandy soil using the Levenberg-Marquardt(LM)training algorithm[J]. Geotechnical and Geological Engineering,2018,36(5):2893-2906.
[15]李明河,周磊,王健.基于LM算法的溶解氧神经网络预测控制[J/OL].农业机械学报,2016,47(6):297-302.LI Minghe,ZHOU Lei,WANG Jian. Predictive control of dissolved oxygen neural network based on LM algorithm[J/OL].Transactions of the Chinese Society for Agricultural Machinery,2016,47(6):297-302. http:∥www. j-csam. org/jcsam/ch/reader/view_abstract. aspx? flag=1&file_no=20160639&journal_id=jcsam. DOI:10. 6041/j. issn. 1000-1298. 2016. 06.039.(in Chinese)
[16]刘振华,范宏运,朱宇泽,等.基于BP神经网络的溶洞规模预测及应用[J].中国岩溶,2018,37(1):139-145.LIU Zhenhua,FAN Hongyun,ZHU Yuze,et al. Prediction model for the scale of karst cave based on back propagation artificial neural network and its application[J]. Carsologica Sinica,2018,37(1):139-145.(in Chinese)
[17]高斌,赵鹏飞,卢昱欣,等.基于BP神经网络的血液荧光光谱识别分类研究[J].光谱学与光谱分析,2018,38(10):3136-3143.GAO Bin,ZHAO Pengfei,LU Yuxin,et al. Study on recognition and classification of blood fluorescence spectrum with BP neural network[J]. Spectroscopy and Spectral Analysis,2018,38(10):3136-3143.(in Chinese)
[18] JIA Zhenyi,ZHOU Shenglu,SU Quanlong,et al. Comparison study on the estimation of the spatial distribution of regional soil metal(loid)s pollution based on Kriging interpolation and BP neural network[J]. International Journal of Environmental Research&Public Health,2018,15(1):34.
[19]郭澎涛,苏艺,茶正早,等.基于BP神经网络的橡胶苗叶片磷含量高光谱预测[J].农业工程学报,2016,32(增刊1):177-183.GUO Pengtao,SU Yi,CHA Zhengzao,et al. Hyperspectral prediction of phosphorus content in rubber seedling leaves based on BP neural network[J]. Transactions of the CSAE,2016,32(Supp. 1):177-183.(in Chinese)
[20]刘雪,李亚妹,刘娇,等.基于BP神经网络的鲜鸡蛋货架期预测模型[J/OL].农业机械学报,2015,46(10):328-334.LIU Xue,LI Yamei,LIU Jiao,et al. Prediction model of shelf life of fresh eggs based on BP neural network[J/OL].Transactions of the Chinese Society for Agricultural Machinery,2015,46(10):328-334. http:∥www. j-csam. org/jcsam/ch/reader/view_abstract. aspx? flag=1&file_no=20151044&journal_id=jcsam. DOI:10. 6041/j. issn. 1000-1298. 2015. 10.044.(in Chinese)
[21] HAN Jiaqi,MAO Kebiao,XU Tongren,et al. A soil moisture estimation framework based on the CART algorithm and its application in China[J]. Journal of Hydrology,2018,563:65-75.
[22]吕利利,颉耀文,黄晓君,等.基于CART决策树分类的沙漠化信息提取方法研究[J].遥感技术与应用,2017,32(3):499-506.LULili,JIE Yaowen,HUANG Xiaojun,et al. Research on desertification information extraction method based on CART decision tree classification[J]. Remote Sensing Technology and Application,2017,32(3):499-506.(in Chinese)
[23]张良均,杨坦,肖刚,等. MATLAB数据分析与挖掘实战[M].北京:机械工业出版社,2015.
[24]张海辉,陶彦蓉,胡瑾.融合叶绿素含量的黄瓜幼苗光合速率预测模型[J/OL].农业机械学报,2015,46(8):259-263.ZHANG Haihui,TAO Yanrong,HU Jin. Photosynthetic rate prediction model of cucumber seedlings with chlorophyll content[J/OL]. Transactions of the Chinese Society for Agricultural Machinery,2015,46(8):259-263. http:∥www. j-csam. org/jcsam/ch/reader/view_abstract. aspx? flag=1&file_no=20150836&journal_id=jcsam. DOI:10. 6041/j. issn. 1000-1298.2015. 08. 036.(in Chinese)
[25]韩敏,穆大芸.回声状态网络LM算法及混沌时间序列预测[J].控制与决策,2011,26(10):1469-1472.HAN Min,MU Dayun. Echo state network LM algorithm and chaotic time series prediction[J]. Control and Decision,2011,26(10):1469-1472.(in Chinese)
[26] LEVENBERG K. A method for the solution of certain non-linear problems in least squares[J]. Quarterly of Applied Mathematics,1944,2(4):436-438.
[27] WILAMOWSKI B M,YU Hao. Improved computation for Levenberg-Marquardt training[J]. IEEE Transactions on Neural Networks,2010,21(6):930-937.
[28] JEFFREY K C F,TSAI W L,SU T L,et al. Application of an LM-neural network for establishing a prediction system of quality characteristics for the LGP manufactured by CO2laser[J]. Optics&Laser Technology,2011,43(3):529-536.
[29]赵弘,周瑞祥,林廷圻.基于Levenberg-Marquardt算法的神经网络监督控制[J].西安交通大学学报,2002,36(5):523-527.ZHAO Hong,ZHOU Ruixiang,LIN Tingqi. Neural network supervision and control based on Levenberg-Marquardt algorithm[J]. Journal of Xi'an Jiaotong University,2002,36(5):523-527.(in Chinese)