Comparison of four PLSR, RF, GRNN and SVR algorithms to estimate sugarcane sheath moisture during growing season using Sentinel-2 satellite imagery

Document Type : Original Article

Authors

1 PhD Student in Biosystems Engineering Department, Faculty of Agriculture, Tarbiat Modares University, Tehran, Iran

2 Professor of Biosystems Engineering Department, Faculty of Agriculture, Tarbiat Modares University, Tehran, Iran

3 Assistant Professor, Department of Water Engineering and Management, Faculty of Agriculture, Tarbiat Modares University, Tehran, Iran

4 Assistant Professor of Biosystems Engineering Department, Faculty of Agriculture, Tarbiat Modares University, Tehran, Iran

Abstract

Introduction: The moisture content of sugarcane sheath is a crucial parameter during the crop's growth period, as it plays a key role in understanding water stress and field irrigation management. Traditional methods of measuring crop moisture levels involve time-consuming and expensive processes like obtaining wet and dry weights, followed by calculating moisture content, which are impractical for large areas. Recent advancements in remote sensing technology have enabled the monitoring of plant tissue moisture content in large fields. Remote sensing data have a high capacity to update crop growth monitoring systems.  In this regard, it is possible to use satellite images that provide a wealth of information to users. This research aims to evaluate sugarcane leaf sheath moisture using satellite images and generate moisture maps based on the best model.
Materials and methods: The sugarcane fields, which represent the largest agricultural operations in Khuzestan, have an area of over 84,000 hectares. It covers over 9,670 hectares are cultivated by the Amir Kabir Agriculture and Industry company, the focus of this research. The study area is located at a latitude of 31° 00' 20' N and a longitude of 48° 15' 22' E. A total of 18 farms of the sugarcane variety CP69-1062 were utilized for this research. Five points were selected from each farm, and the coordinates of the points were recorded using a GPS device. The study was carried out between July and September. Ground data were collected nearly simultaneously with the Sentinel-2 satellite imaging of the target area. The moisture content of each collected sample was determined gravimetrically in the laboratory. For each image, indices and spectral bands were calculated using QGIS software and the output was saved as Excel and TIF files. The indices and bands obtained from Sentinel-2 satellite images were used to estimate and monitor the moisture status of sugarcane leaf sheath. In the next step, a variance inflation factor (VIF) analysis was implemented to check the collinearity between indices and bands. Finally, the indices of NDVI, EVI, SRWI, Clgreen and single bands B2, B3, B4, B5, B6, B11 and B12 were entered as input to four GRNN, RF, SVR and PLSR models. The Bayes algorithm was employed to optimize the parameters of the model.
Results and discussion: The results demonstrated that the SVR model exhibited a superior ability to estimate leaf sheath moisture compared to other models. Additionally, the sensitivity analysis revealed that the SRWI, Clgreen, NDVI, B5, B12, B11, B4, B3, EVI and B2 parameters are effective parameters in the moisture content modelling process. In the final stage, the leaf sheath moisture was classified into five stress classes, namely irrigation time, low moisture, medium moisture, and high moisture, in the order from low to high. The results of the moisture maps and the irrigation schedule for each date indicate that the combined output of B2, B3, B4, B5, B6, B11, B12, NDVI, EVI, SRWI and Clgreen indices and bands has a superior performance. These indices were utilized in the preparation of irrigation plans. This method was employed to assess the potential of S2 MSI spectral indices for the estimation of leaf sheath moisture in the sugarcane growth stage.
Conclusion: Based on sensitivity analysis, the SRWI parameter was found to be the most effective index in the modelling process. Consequently, it can be concluded that a combination of indices and bands of NDVI, EVI, SRWI, Clgreen, B2, B3, B5, B4, B11, and B12 provides a more accurate estimate of sugarcane sheath moisture than any single input. Thus, processing and analysis of Sentinel-2 satellite images can be used to enhance the methodologies employed for the monitoring of sugarcane sheath moisture content in expansive fields.

Keywords

References

  • Alavi-Panah, K. 2013. Principles of Modern Remote Sensing and Interpretation of Satellite Images and Aerial Photographs. Tehran University Printing and Publishing Institute. Please refer to page 784 (In Farsi).
  • Alizadeh, A. 2017. The Relationship Between Water, Soil and Plants. This is the 14th edition. Published by Imam Reza University. Please refer to page 616 (In Farsi).
  • 2013. Sugarcane production technology in Iran - first volume: set of guidelines for crop production operations. Rasvaje Publishing House, Khuzestan (In Farsi).
  • Ashiq, M.W.; Zhao, C.; Ni, J.; Akhtar, M. 2010. GIS-based high-resolution spatial interpolation of precipitation in mountain–plain areas of Upper Pakistan for regional climate change impact studies. Theoretical and Applied Climatology, 99(3), pp.239-253 (doi. 1007/s00704-009-0140-y).
  • Bachmaier, M.; Backes, M. 2008. Variogram or semivariogram? Understanding the variances in a variogram. Precision Agriculture, 9(3), pp.173-175 (doi. 1007/s11119-008-9056-2).
  • Ballester, C.; Brinkhoff, J.; Quayle, W.C.; Hornbuckle, J. 2019. Monitoring the Effects of Water Stress in Cotton Using the Green Red Vegetation Index and Red Edge Ratio. Remote Sensing. 11: 7. 873: 1-21 (org/10.3390/rs11070873).
  • Breiman, L. 2001. Random Forests. Machine Learning. 45(1): 5-32.
  • Ceccato, P.; Flasse, S.; Tarantola, S.; Jacquemoud, S.; Grégoire J. 2001. Detecting vegetation leaf water content using reflectance in the optical domain. Remote Sensing of
    Environment. 77:22–33 (org/10.1016/S0034-4257(01)00191-2).
  • Ceccato, P.; Gobron, N.; Flasse, S.; Pinty, B.; Tarantola, S. 2002. Designing a spectral index to estimate vegetation water content from remote sensing data: Part 1. Theoretical approach. Remote Sensing of
    Environment, (org/10.1016/S0034-4257(02)00037-8).
  • Chakroun, H.; Mouillot, F.; Hamdi, A. 2015. Regional equivalent water thickness modeling from remote sensing across a tree cover/LAI gradient in Mediterranean forests of Northern Tunisia. Remote Sensing, 7, 1937–1961 (org/10.3390/rs70201937).
  • Cheng, T.; Riano, D.; Koltunov, A.; Whiting, M.L.; Ustin, S.L. 2013. Rodriguez, J. Detection of diurnal variation in orchard canopy water content usingMODIS/ASTERairborne simulator (MASTER) data. Remote Sensing Environment, 132, 1–12 (org/10.1016/j.rse.2012.12.024).
  • Clements, H.F., 1960. Recent developments in crop logging of sugarcane. In Proceedings of the International Society of Sugarcane Technologists. 10th Congress, Hawaii, 1959. 1960 (doi/full/10.5555/19610305460).
  • Clements, H.F.,1977. Potassium and sugarcane. Agricultural Extension Education, Hawaisi, 10811.
  • Cornejo-Bueno, L.; Garrido-Merchán, E.C.; Hernández-Lobato, D.; Salcedo-Sanz, S. 2018. Bayesian optimization of a hybrid system for robust ocean wave features prediction. Neurocomputing, 275, pp.818-828 (org/10.1016/j.neucom.2017.09.025).
  • Dangwal, N.; Patel, N.R.; Kumari, M.; Saha, S.K. 2015. Monitoring of water stress in wheat using multispectral indices derived from Landsat-TM. Geocarto International, 31: 6. 682-693 (doi.org/10.1080/10106049.2015.1073369).
  • Drucker, H.; Burges, C.J.; Kaufman, L.; Smola, A.; Vapnik, V. 1997. Support vector regression machines. Advances in neural information processing systems, 9, pp.155-161.
  • Gao, B.C. 1996. A normalized diference water index for remote sensing of vegetation liquid water from space. Remote sensing of environment, 58(3): 257-266 (https://doi.org/10.1117/12.210877).
  • Gevrey, M.; Dimopoulos, I.; Lek, S. 2003. Review and comparison of methods to study the contribution of variables in artificial neural network models. Ecological modelling, 160(3), pp.249-264 (org/10.1016/S0304-3800(02)00257-0).
  • Gitelson, A. A.; Andrés, V.; Verónica, C.; Donald, C. R.; Timothy, J. A. 2005. Remote Estimation of
    Canopy Chlorophyll Content in Crops    . Geophysical Research Letters 32 (8): L08403 (doi:10.1029/
    2005GL02268).
  • Grof, C. P.; Campbell, J. A. 2001. Sugarcane sucrose metabolism: scope for molecular manipulation. Functional Plant Biology, 28(1), 1-12 (https://doi.org/10.1071/PP00039).
  • Gu, Y.; Hunt, E.; Wardlow, B.; Basara, J.B.; Brown, J.F.; Verdin, J.P. 2008. Evaluation of MODIS NDVI and NDWI for vegetation drought monitoring using Oklahoma Mesonet soil moisture data. Geophysical Research Letters, 35, L22401 (https://doi.org/10.1029/2008GL035772).
  • Haenlein, M.; Kaplan, A.M. 2004. A beginner's guide to partial least squares analysis. Understanding statistics, 3(4), pp.283-297 (https://doi.org/10.1207/s15328031us0304_4).
  • Hardisky, M. A.; Daiber, F. C.; Roman, C. T.; V. Klemas. 1984. Remote Sensing of Biomass and Annual Net Aerial Primary Productivity of a Salt Marsh. Remote Sensing of Environment 16 (2): 91–106. doi:10.1016/0034-4257(84)90055-5 (https://doi.org/10.1016/0034-4257(84)90055-5).
  • Keshavaiah, K.V.; Palled, Y.B.; Shankaraiah, C.; Nandihalli, B.S. 2013. Effect of sheath moisture and relation of SPAD on yield of sugarcane. Advance Research Journal of Crop Improvement, 4(2), pp.98-102 (doi/full/10.5555/20143261002).
  • Khajapour, M. 2013. Industrial Plants. Academic Jihad (Isfahan University of Technology). 582 pages Khuzestan (In Farsi).
  • Law, T.; Shawe-Taylor, J. 2017. Practical Bayesian support vector regression for financial time series prediction and market condition change detection. Quantitative Finance, 17(9), pp.1403-1416 (https://doi.org/10.1080/14697688.2016.1267868).
  • Martin, R.E.; Asner, G.P.; Francis, E.; Ambrose, A.; Baxter, W.; Das, A.J.; Vaughn, N.R.; Paz-Kagan, T.; Dawson, T.; Nydick, K., et al. 2018. Remote measurement of canopy water content in giant sequoias (Sequoiadendron giganteum) during drought. Forest Ecology and Management. 2018, 419, 279–290 (https://doi.org/10.1016/j.foreco.2017.12.002).
  • McFeeters, S.K. 1996. The use of the normalized difference water index (NDWI) in the delineation of open waterfeatures. International Journalof Remote Sensing,17, 1425–1432 (https://doi.org/10.1080/01431169608948714).
  • Pan, H.; Chen, Z.; Ren, J.; Li, H.; Wu, S. 2018. Modeling winter wheat leaf area index and canopy water content with three different approaches using Sentinel-2 multispectral instrument data. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 12(2), pp.482-492 (DOI: 1109/JSTARS.2018.2855564).
  • Rahdari, V.; Sufianian, A.; Khajehuddin, S. J.A.; Maleki Najafabadi, S. 2013. Investigating the ability of satellite data in preparing a map of vegetation canopy percentage in arid and semi-arid areas (case study of Mote Wildlife Sanctuary). Environmental Science and Technology Quarterly. Volume 15, Number 4 - Issue 4, Page 43-54 (In Farsi).
  • Specht, D.F. 1991. A general regression neural network. IEEE transactions on neural networks, 2(6), pp.568-576.
  • Sung, A.H. 1998. Ranking importance of input parameters of neural networks. Expert Systems with Applications 15, 405/411 (https://doi.org/10.1016/S0957-4174(98)00041-4).
  • Tabib Mahmoudi, F. 2019. Investigating the water stress status of plants in northern Iran under the influence of quarantine measures during the covid-19 virus pandemic. Journal of water and soil protection research, Volume 27; Number 6 (In Farsi).
  • Thomas, J.R.; Namken, L.N.; Oerther, G.F. 1971. Estimating Leaf Water Content by Reflectance Measurement. AgronomyJournal. 63, 845–847 (https://doi.org/10.2134/agronj1971.00021962006300060007x).
  • Veisi, S.; Nasri, A.; Hamzah, S. 2016. The relationship between sugarcane leaf sheath moisture and water stress index using an infrared thermometer. Irrigation Science and Engineering (Scientific-Research Magazine), Volume 91, Number 9, Winter 4 (In Farsi).
  • Vincini, M.; Calegari, F.; Casa, R. 2016. Sensitivity of leaf chlorophyll empirical estimators obtained at Sentinel-2 spectral resolution for different canopy structures. Precision Agriculture, 17(3), pp.313-331 (Doi. 10.1007/s11119-015-9424-7).
  • Zhang, F. & Zhou, G., 2019. Estimation of vegetation water content using hyperspectral vegetation indices: A comparison of crop water indicators in response to water stress treatments for summer maize. BMC ecology, 19(1), pp.1-12 (doi. 10.1186/s12898-019-0233-0).
  • Zhang, F.; Zhou G., 2015. Estimation of canopy water content by means of hyperspectral indices based on drought stress gradient experiments of maize in the North Plain China. Remote Sensing, 7:15203–23 (https://doi.org/10.3390/rs71115203).

 

Files

Share

How to cite

Statistics