Mapping of Soil Properties Using Machine Learning Techniques
S. Sridevy
Department of Physical Sciences and Information Techhnology, Tamil Nadu Agricultural University, Coimbatore, India.
M. Nivas Raj
Department of Remote Sensing and GIS, Tamil Nadu Agricultural University, Coimbatore, India.
P. Kumaresan *
Centre for Water and Geospatial Studies, Tamil Nadu Agricultural University, Coimbatore, India.
N. Balakrishnan
Centre for Water and Geospatial Studies, Tamil Nadu Agricultural University, Coimbatore, India.
M. Tilak
Forest College and Research Institute, Tamil Nadu Agricultural University, Coimbatore, India.
J. Arockia Stephen Raj
Agricultural College and Research Institute, Killikulam, Tamil Nadu Agricultural University, Coimbatore, India.
P. Jona Innisai Rani
Department of Extension Education and Communication Management, Community Science College and Research Institute, Madurai, Tamil Nadu Agricultural University, Coimbatore, India.
*Author to whom correspondence should be addressed.
Abstract
We aimed to estimate Soil Nutrients and relate the spectral signatures to that of the Laboratory reference Measurements utilizing CART analysis. Sustainable agriculture aims at controlled and/or precise soil fertility interventions based on spatial soil information. The profound advancements in remote sensing and geospatial techniques provide means for determining the spatial coverage and variability of the soil properties through the survey and image data incorporated in the mapping procedures (i.e.) Digital Soil Mapping. The soil moisture content at varying levels influences crop growth and decides the yield, as the crop requires water at critical crop growth stages. Machine learning techniques provide the means of optimized model calibration when compared to conventional geostatistical or statistical approaches.
Keywords: CART analysis, geostatistical technique, machine learning techniques, soil properties mapping