Using a One-Dimensional Convolutional Neural Network on Visible and Near-Infrared Spectroscopy to Improve Soil Phosphorus Prediction in Madagascar

As a proximal soil sensing technique, laboratory visible and near-infrared (Vis-NIR) spectroscopy is a promising tool for the quantitative estimation of soil properties. However, there remain challenges for predicting soil phosphorus (P) content and availability, which requires a reliable model applicable for different land-use systems to upscale. Recently, a one-dimensional convolutional neural network (1D-CNN) corresponding to the spectral information of soil was developed to considerably improve the accuracy of soil property predictions. The present study investigated the predictive ability of a 1D-CNN model to estimate soil available P (oxalate-extractable P; Pox) content in soils by comparing it with partial least squares (PLS) and random forest (RF) regressions using soil samples (n = 318) collected from natural (forest and non-forest) and cultivated (upland and flooded rice fields) systems in Madagascar. Overall, the 1D-CNN model showed the best predictive accuracy (R2 = 0.878) with a highly accurate prediction ability (ratio of performance to the interquartile range = 2.492). Compared to the PLS model, the RF and 1D-CNN models indicated 4.37% and 23.77% relative improvement in root mean squared error values, respectively. Based on a sensitivity analysis, the important wavebands for predicting soil Pox were associated with iron (Fe) oxide, organic matter (OM), and water absorption, which were previously known wavelength regions for estimating P in soil. These results suggest that 1D-CNN corresponding spectral signatures can be expected to significantly improve the predictive ability for estimating soil available P (Pox) from Vis-NIR spectral data. Rapid and accurate estimation of available P content in soils using our results can be expected to contribute to effective fertilizer management in agriculture and the sustainable management of ecosystems. However, the 1D-CNN model will require a large dataset to extend its applicability to other regions of Madagascar. Thus, further updates should be tested in future studies using larger datasets from a wide range of ecosystems in the tropics.

[1]  K. Kawamura,et al.  Exploring relevant wavelength regions for estimating soil total carbon contents of rice fields in Madagascar from Vis-NIR spectra with sequential application of backward interval PLS , 2020, Plant Production Science.

[2]  R Core Team,et al.  R: A language and environment for statistical computing. , 2014 .

[3]  Cyrill Stachniss,et al.  WeedMap: A large-scale semantic weed mapping framework using aerial multispectral imaging and deep neural network for precision farming , 2018, Remote. Sens..

[4]  Predicting plant available phosphorus using infrared spectroscopy with consideration for future mobile sensing applications in precision farming , 2019, Precision Agriculture.

[5]  T. Razafimbelo,et al.  Soil survey of the east coast and the central highlands indicates need to update Madagascar soil map , 2020 .

[6]  N. Ramankutty,et al.  Agronomic phosphorus imbalances across the world's croplands , 2011, Proceedings of the National Academy of Sciences.

[7]  Farid Melgani,et al.  One‐dimensional convolutional neural networks for spectroscopic signal regression , 2018 .

[8]  Abdul Mounem Mouazen,et al.  Quantitative analysis of soil nitrogen and carbon at a farm scale using visible and near infrared spectroscopy coupled with wavelength reduction , 2011 .

[9]  A. McBratney,et al.  Critical review of chemometric indicators commonly used for assessing the quality of the prediction of soil attributes by NIR spectroscopy , 2010 .

[10]  Bettina M. J. Engelbrecht,et al.  Soil organic phosphorus in lowland tropical rain forests , 2011 .

[11]  R. V. Rossel,et al.  Using a digital camera to measure soil organic carbon and iron contents , 2008 .

[12]  B. Koné,et al.  Nutrient constraint and yield potential of rice on upland soil in the south of the Dahoumey gap of West Africa , 2011 .

[13]  Abdul Mounem Mouazen,et al.  Comparison between artificial neural network and partial least squares for on-line visible and near infrared spectroscopy measurement of soil organic carbon, pH and clay content , 2015 .

[14]  F Coulon,et al.  Rapid prediction of total petroleum hydrocarbons concentration in contaminated soil using vis-NIR spectroscopy and regression techniques. , 2018, The Science of the total environment.

[15]  U. Schwertmann.  Differenzierung der Eisenoxide des Bodens durch Extraktion mit Ammoniumoxalat‐Lösung , 1964 .

[16]  P. R. S. Vendrame,et al.  Mineralogical analysis of ferralitic soils in Madagascar using NIR spectroscopy , 2017, CATENA.

[17]  Yoshua Bengio,et al.  Gradient-based learning applied to document recognition , 1998, Proc. IEEE.

[18]  Jie Peng,et al.  Potential of VIS-NIR-SWIR Spectroscopy from the Chinese Soil Spectral Library for Assessment of Nitrogen Fertilization Rates in the Paddy-Rice Region, China , 2015, Remote. Sens..

[19]  Budiman Minasny,et al.  Transfer learning to localise a continental soil vis-NIR calibration model , 2019, Geoderma.

[20]  Viacheslav I. Adamchuk,et al.  A global spectral library to characterize the world’s soil , 2016 .

[21]  Estimation of effective calibration sample size using visible near infrared spectroscopy: deep learning vs machine learning , 2019 .

[22]  Rebecca L. Whetton,et al.  Machine learning based prediction of soil total nitrogen, organic carbon and moisture content by using VIS-NIR spectroscopy , 2016 .

[23]  E. Ben-Dor,et al.  Visible and near-infrared (0.4–1.1 μm) analysis of arid and semiarid soils , 1994 .

[24]  Annamaria Castrignanò,et al.  Laboratory-based Vis–NIR spectroscopy and partial least square regression with spatially correlated errors for predicting spatial variation of soil organic matter content , 2015 .

[25]  Jean-Fu Kiang,et al.  Comparison of CNN Algorithms on Hyperspectral Image Classification in Agricultural Lands , 2020, Sensors.

[26]  W. S. Lee,et al.  Spectral Soil Signatures and sensing Phosphorus , 2005 .

[27]  J. De Baerdemaeker,et al.  Phosphorus Sensing for Fresh Soils using Visible and Near Infrared Spectroscopy , 2006 .

[28]  Andy Liaw,et al.  Classification and Regression by randomForest , 2007 .

[29]  A. Savitzky,et al.  Smoothing and Differentiation of Data by Simplified Least Squares Procedures. , 1964 .

[30]  Gabriele Buttafuoco,et al.  Studying the relationship between water-induced soil erosion and soil organic matter using Vis–NIR spectroscopy and geomorphological analysis: A case study in southern Italy , 2013 .

[31]  Nitin K. Tripathi,et al.  Artificial neural network analysis of laboratory and in situ spectra for the estimation of macronutrients in soils of Lop Buri (Thailand) , 2003 .

[32]  Hidetoshi Asai,et al.  Vis-NIR Spectroscopy and PLS Regression with Waveband Selection for Estimating the Total C and N of Paddy Soils in Madagascar , 2017, Remote. Sens..

[33]  S. Wold,et al.  PLS-regression: a basic tool of chemometrics , 2001 .

[34]  T. Balemi,et al.  Management of soil phosphorus and plant adaptation mechanisms to phosphorus stress for sustainable crop production: a review , 2012 .

[35]  P. Veldhoven,et al.  Inorganic and organic phosphate measurements in the nanomolar range. , 1987, Analytical biochemistry.

[36]  Sergey Kucheryavski,et al.  Combining Laser-Induced Breakdown Spectroscopy (LIBS) and Visible Near-Infrared Spectroscopy (Vis-NIRS) for Soil Phosphorus Determination , 2020, Sensors.

[37]  A. Mouazen,et al.  Predicting bioavailability change of complex chemical mixtures in contaminated soils using visible and near-infrared spectroscopy and random forest regression , 2019, Scientific Reports.

[38]  D. R. Cutler,et al.  Utah State University From the SelectedWorks of , 2017 .

[39]  Nikolaos L. Tsakiridis,et al.  Simultaneous prediction of soil properties from VNIR-SWIR spectra using a localized multi-channel 1-D convolutional neural network , 2020 .

[40]  Budiman Minasny,et al.  Using deep learning to predict soil properties from regional spectral data , 2019, Geoderma Regional.

[41]  T. Razafimbelo,et al.  Farmyard manure improves phosphorus use efficiency in weathered P deficient soil , 2019, Nutrient Cycling in Agroecosystems.

[42]  E. Ben-Dor.  The reflectance spectra of organic matter in the visible near-infrared and short wave infrared region (400-2500 nm) during a controlled decomposition process , 1997 .

[43]  Budiman Minasny,et al.  Convolutional neural network for simultaneous prediction of several soil properties using visible/near-infrared, mid-infrared, and their combined spectra , 2019, Geoderma.

[44]  R. V. Rossel,et al.  Visible, near infrared, mid infrared or combined diffuse reflectance spectroscopy for simultaneous assessment of various soil properties , 2006 .

[45]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[46]  Yoshio Inoue,et al.  NDSI Map and IPLS Using Hyperspectral Data for Assessment of Plant and Ecosystem Variables —With a Case Study on Remote Sensing of Grain Protein Content, Chlorophyll Content and Biomass in Rice— , 2008 .

[47]  Abdul Mounem Mouazen,et al.  On-line visible and near infrared spectroscopy for in-field phosphorous management , 2016 .

[48]  Yanping Cong,et al.  Simultaneous Prediction of Soil Properties Using Multi_CNN Model , 2020, Sensors.

[49]  K. Kawamura,et al.  Prediction of Soil Oxalate Phosphorus using Visible and Near-Infrared Spectroscopy in Natural and Cultivated System Soils of Madagascar , 2020, Agriculture.

[50]  Ronei Jesus Poppi,et al.  Visible and near infrared spectroscopy coupled to random forest to quantify some soil quality parameters. , 2018, Spectrochimica acta. Part A, Molecular and biomolecular spectroscopy.

[51]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[52]  Eyal Ben-Dor,et al.  Internal soil standard method for the Brazilian soil spectral library: Performance and proximate analysis , 2018 .

[53]  Kazuki Saito,et al.  Near-infrared, mid-infrared or combined diffuse reflectance spectroscopy for assessing soil fertility in rice fields in sub-Saharan Africa , 2019, Geoderma.

[54]  Hidetoshi Asai,et al.  Laboratory Visible and Near-Infrared Spectroscopy with Genetic Algorithm-Based Partial Least Squares Regression for Assessing the Soil Phosphorus Content of Upland and Lowland Rice Fields in Madagascar , 2019, Remote. Sens..

[55]  C. Feller,et al.  Determination of carbon and nitrogen contents in Alfisols, Oxisols and Ultisols from Africa and Brazil using NIRS analysis: Effects of sample grinding and set heterogeneity , 2007 .

[56]  Martín Abadi,et al.  TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems , 2016, ArXiv.

[57]  Ron Wehrens,et al.  The pls Package: Principal Component and Partial Least Squares Regression in R , 2007 .

[58]  R. V. Rossel,et al.  Visible and near infrared spectroscopy in soil science , 2010 .

[59]  E. Smolders,et al.  Larger bioavailability of soil phosphorus for irrigated rice compared with rainfed rice in Madagascar: results from a soil and plant survey , 2012 .

[60]  E. Ben-Dor.  Quantitative remote sensing of soil properties , 2002 .

[61]  Luca Montanarella,et al.  Prediction of Soil Organic Carbon at the European Scale by Visible and Near InfraRed Reflectance Spectroscopy , 2013, PloS one.

[62]  David M. Sherman,et al.  Electronic spectra of Fe3+ oxides and oxide hydroxides in the near IR to near UV , 1985 .

[63]  Michael Vohland,et al.  Determination of soil properties with visible to near- and mid-infrared spectroscopy: Effects of spectral variable selection , 2014 .

[64]  T. Razafimbelo,et al.  Phosphorus uptake of rice plants is affected by phosphorus forms and physicochemical properties of tropical weathered soils , 2018, Plant and Soil.

[65]  V. Barrón,et al.  Use and Limitations of Second-Derivative Diffuse Reflectance Spectroscopy in the Visible to Near-Infrared Range to Identify and Quantify Fe Oxide Minerals in Soils , 1998 .

[66]  Abdul Mounem Mouazen,et al.  Predictive performance of mobile vis-near infrared spectroscopy for key soil properties at different geographical scales by using spiking and data mining techniques , 2017 .

[67]  Luca Montanarella,et al.  Soil spectroscopy: an opportunity to be seized , 2015, Global change biology.

[68]  Bo Stenberg,et al.  Improving the prediction performance of a large tropical vis‐NIR spectroscopic soil library from Brazil by clustering into smaller subsets or use of data mining calibration techniques , 2014 .

[69]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[70]  Tarmo Lipping,et al.  Crop yield prediction with deep convolutional neural networks , 2019, Comput. Electron. Agric..

[71]  R. Yost,et al.  Quantifying the Available Soil Phosphorus Pool with the Acid Ammonium Oxalate Method , 1999 .