Fusion of Feature Selection Methods and Regression Algorithms for Predicting the Canopy Water Content of Rice Based on Hyperspectral Data

Estimation of the canopy water content (CWC) is extremely important for irrigation management decisions. Machine learning and hyperspectral imaging technology have provided a potentially useful tool for precise measurement of plant water content. The tools, however, are hampered by feature selection as well as an advanced model in itself. Therefore, this study aims to propose an efficient prediction model and compare three feature selection methods including vegetation indices (VI), model-based features (MF), and principal component analysis (PCA). The selected features were applied with a back-propagation neural network (BPNN), random forest (RF), and partial least square regression (PLSR) for training the samples with minimal loss on a cross-validation set. The hyperspectral images were collected from rice crops grown under different water stress levels. A total of 128 images were used to evaluate our proposed methods. The results indicated that the integration of PCA and MF methods can provide a more robust feature selection for the proposed prediction model. The three bands of 1467, 1456, and 1106 nm were the supreme variants of CWC forecasting. These features were combined with an optimized BPNN model and significantly improved the foretelling accuracy. The accuracy and correlation coefficient of the advanced BPNN-PCA-MF model are close to 1 with an RMSE of 0.252. Thus, this study positively contributes to plant water content prediction researchers and policymakers so that well in advance and effective steps can be taken for precision irrigation.

[1]  Hang Lei,et al.  Hyperparameter Optimization for Machine Learning Models Based on Bayesian Optimization , 2019 .

[2]  Yufeng Ge,et al.  Temporal dynamics of maize plant growth, water use, and leaf water content using automated high throughput RGB and hyperspectral imaging , 2016, Comput. Electron. Agric..

[3]  F. M. Danson,et al.  High-spectral resolution data for determining leaf water content , 1992 .

[4]  Vincent Baeten,et al.  A Backward Variable Selection method for PLS regression (BVSPLS). , 2009, Analytica chimica acta.

[5]  J. Eitel,et al.  Suitability of existing and novel spectral indices to remotely detect water stress in Populus spp. , 2006 .

[6]  Jialin Li,et al.  Integrating Airborne LiDAR and Optical Data to Estimate Forest Aboveground Biomass in Arid and Semi-Arid Regions of China , 2018, Remote. Sens..

[7]  Da-Wen Sun,et al.  Principles and Applications of Hyperspectral Imaging in Quality Evaluation of Agro-Food Products: A Review , 2012, Critical reviews in food science and nutrition.

[8]  Achim Zeileis,et al.  Conditional variable importance for random forests , 2008, BMC Bioinformatics.

[9]  G. Carter.  PRIMARY AND SECONDARY EFFECTS OF WATER CONTENT ON THE SPECTRAL REFLECTANCE OF LEAVES , 1991 .

[10]  Isabelle Guyon,et al.  An Introduction to Variable and Feature Selection , 2003, J. Mach. Learn. Res..

[11]  Michael E. Schaepman,et al.  Estimating canopy water content using hyperspectral remote sensing data , 2010, Int. J. Appl. Earth Obs. Geoinformation.

[12]  ASSESSING WHEAT YELLOW RUST DISEASE THROUGH HYPERSPECTRAL REMOTE SENSING , 2014 .

[13]  Yong Zhou,et al.  Phenotyping flag leaf nitrogen content in rice using a three-band spectral index , 2019, Comput. Electron. Agric..

[14]  Natasha Gilbert,et al.  Water under pressure , 2012, Nature.

[15]  A. Skidmore,et al.  Nitrogen detection with hyperspectral normalized ratio indices across multiple plant species , 2005 .

[16]  Sushma Panigrahy,et al.  Use of hyperspectral data to assess the effects of different nitrogen applications on a potato crop , 2007, Precision Agriculture.

[17]  Tauqueer Ahmad,et al.  Comparison of various modelling approaches for water deficit stress monitoring in rice crop through hyperspectral remote sensing , 2019, Agricultural Water Management.

[18]  Wolfram Mauser,et al.  Spaceborne Imaging Spectroscopy for Sustainable Agriculture: Contributions and Challenges , 2018, Surveys in Geophysics.

[19]  N. M. Kelly,et al.  Spectral absorption features as indicators of water status in coast live oak ( Quercus agrifolia ) leaves , 2003 .

[20]  P. Curran,et al.  LIBERTY—Modeling the Effects of Leaf Biochemical Concentration on Reflectance Spectra , 1998 .

[21]  J. Peñuelas,et al.  Estimation of plant water concentration by the reflectance Water Index WI (R900/R970) , 1997 .

[22]  Nilanjan Ray,et al.  Snake Validation: A PCA-Based Outlier Detection Method , 2009, IEEE Signal Processing Letters.

[23]  Feng Liu,et al.  Comparison of boosted regression tree and random forest models for mapping topsoil organic carbon concentration in an alpine ecosystem , 2016 .

[24]  Louis W. Glorfeld.  A Methodology for simplification and interpretation of backpropagation-based neural network models , 1996 .

[25]  A. Skidmore,et al.  Tropical mangrove species discrimination using hyperspectral data: A laboratory study , 2005 .

[26]  R. E. Yoder,et al.  Simulation of nitrate distribution under drip irrigation using artificial neural networks , 2003, Irrigation Science.

[27]  Junichi Imanishi,et al.  Detecting drought status and LAI of two Quercus species canopies using derivative spectra , 2004 .

[28]  S. Ustin,et al.  Water content estimation in vegetation with MODIS reflectance data and model inversion methods , 2003 .

[29]  U. Schmidhalter,et al.  Spectral reflectance indices as a rapid and nondestructive phenotyping tool for estimating different morphophysiological traits of contrasting spring wheat germplasms under arid conditions , 2015 .

[30]  Kenshi Sakai,et al.  A ground-based hyperspectral imaging system for characterizing vegetation spectral features , 2008 .

[31]  Xiaolin Zhu,et al.  Improving forest aboveground biomass estimation using seasonal Landsat NDVI time-series , 2015 .

[32]  Budiman Minasny,et al.  Digital soil mapping of soil carbon at the farm scale: A spatial downscaling approach in consideration of measured and uncertain data , 2017 .

[33]  Pierre Roumet,et al.  Assessing leaf nitrogen content and leaf mass per unit area of wheat in the field throughout plant cycle with a portable spectrometer , 2013 .

[34]  M. Schlerf,et al.  Remote sensing of forest biophysical variables using HyMap imaging spectrometer data , 2005 .

[35]  A. Prasad,et al.  Newer Classification and Regression Tree Techniques: Bagging and Random Forests for Ecological Prediction , 2006, Ecosystems.

[36]  J. Peñuelas,et al.  The reflectance at the 950–970 nm region as an indicator of plant water status , 1993 .

[37]  L. S. Galvão,et al.  Variation in spectral shape of urban materials , 2010 .

[38]  P. F. Scholander,et al.  Sap Pressure in Vascular Plants , 1965, Science.

[39]  Andrew K. Skidmore,et al.  Retrieval of leaf water content spanning the visible to thermal infrared spectra , 2014 .

[40]  G. Carter.  Ratios of leaf reflectances in narrow wavebands as indicators of plant stress , 1994 .

[41]  Jorge Nocedal,et al.  A Limited Memory Algorithm for Bound Constrained Optimization , 1995, SIAM J. Sci. Comput..

[42]  Yufeng Ge,et al.  High Throughput In vivo Analysis of Plant Leaf Chemical Properties Using Hyperspectral Imaging , 2017, Front. Plant Sci..

[43]  Ridwan Al Iqbal.  Using Feature Weights to Improve Performance of Neural Networks , 2011, ArXiv.

[44]  B. Gao.  NDWI—A normalized difference water index for remote sensing of vegetation liquid water from space , 1996 .

[45]  Gregory A. Carter,et al.  Responses of leaf spectral reflectance to plant stress. , 1993 .

[46]  F H Schulze,et al.  Applications of Artificial Neural Networks in integrated water management: fiction or future? , 2005, Water science and technology : a journal of the International Association on Water Pollution Research.

[47]  Hongyan Zhu,et al.  Hyperspectral Imaging for Predicting the Internal Quality of Kiwifruits Based on Variable Selection Algorithms and Chemometric Models , 2017, Scientific Reports.

[48]  Gamal ElMasry,et al.  Near-infrared hyperspectral imaging for predicting colour, pH and tenderness of fresh beef , 2012 .

[49]  Ozgur Kisi,et al.  Evapotranspiration Estimation using Six Different Multi-layer PerceptronAlgorithms , 2016 .

[50]  K. Omasa,et al.  Estimation of the leaf chlorophyll content using multi-angular spectral reflectance factor. , 2019, Plant, cell & environment.

[51]  Hua Sun,et al.  Mapping Forest Ecosystem Biomass Density for Xiangjiang River Basin by Combining Plot and Remote Sensing Data and Comparing Spatial Extrapolation Methods , 2017, Remote. Sens..

[52]  Chris Dyer,et al.  On the State of the Art of Evaluation in Neural Language Models , 2017, ICLR.

[53]  D. Mulla.  Twenty five years of remote sensing in precision agriculture: Key advances and remaining knowledge gaps , 2013 .

[54]  Chandra A. Madramootoo,et al.  Sensitivity of spectral vegetation indices for monitoring water stress in tomato plants , 2019, Comput. Electron. Agric..

[55]  D. Roberts,et al.  Deriving Water Content of Chaparral Vegetation from AVIRIS Data , 2000 .

[56]  Sushma Jain,et al.  Reference evapotranspiration estimation and modeling of the Punjab Northern India using deep learning , 2019, Comput. Electron. Agric..

[57]  B. Rock,et al.  Detection of changes in leaf water content using Near- and Middle-Infrared reflectances , 1989 .

[58]  Xiaoli Li,et al.  Hyperspectral Imaging for Determining Pigment Contents in Cucumber Leaves in Response to Angular Leaf Spot Disease , 2016, Scientific Reports.

[59]  Reza Ehsani,et al.  Detection of multi-tomato leaf diseases (late blight, target and bacterial spots) in different stages by using a spectral-based sensor , 2018, Scientific Reports.

[60]  Manuel A. Duarte-Mermoud,et al.  Feature selection algorithms using Chilean wine chromatograms as examples , 2005 .

[61]  Anming Bao,et al.  Estimation of leaf water content in cotton by means of hyperspectral indices , 2013 .

[62]  M. Feng,et al.  Assessment of plant water status in winter wheat (Triticum aestivum L.) based on canopy spectral indices , 2019, PloS one.

[63]  Michael E. Schaepman,et al.  Using spectral information from the NIR water absorption features for the retrieval of canopy water content , 2008, Int. J. Appl. Earth Obs. Geoinformation.

[64]  A. Formaggio,et al.  Discrimination of sugarcane varieties in Southeastern Brazil with EO-1 Hyperion data , 2005 .

[65]  C. Giardino,et al.  Estimation of leaf and canopy water content in poplar plantations by means of hyperspectral indices and inverse modeling , 2008 .

[66]  S. Tarantola,et al.  Detecting vegetation leaf water content using reflectance in the optical domain , 2001 .

[67]  Uri Yermiyahu,et al.  An insight to the performance of crop water stress index for olive trees , 2013 .

[68]  Xiaohuan Xi,et al.  Estimating the Biomass of Maize with Hyperspectral and LiDAR Data , 2016, Remote. Sens..

[69]  D. Sims,et al.  Estimation of vegetation water content and photosynthetic tissue area from spectral reflectance: a comparison of indices based on liquid water and chlorophyll absorption features , 2003 .

[70]  Josep Peñuelas,et al.  Cell wall elasticity and Water Index (R970 nm/R900 nm) in wheat under different nitrogen availabilities , 1996 .

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