International Science Community Association International Research Journal of Biological Sciences 2278-3202 8 10 2019 10 10 Agricultural crop growth biophysical parameters estimation by machine learning using microwave satellite data 1 8 EN Pradeep Kumar 1 Department of Physics, Institute of Science, Banaras Hindu University, Varanasi, India Arti Choudhary 2 Transport Planning and Environment Division, CSIR-Central Road Research Institute, New Delhi, India Rajendra Prasad 3 Department of Physics, Indian Institute of Technology (BHU), Varanasi, India Dileep Kumar Gupta 4 Department of Physics, Indian Institute of Technology (BHU), Varanasi, India Meenakshi Amarawat 5 Department of Botany, Bhupal Noble′s University Udaipur, India Gaurav Shukla 6 Department of Civil Engineering, Indian Institute of Technology Roorkee, India Abhay Kumar Singh 7 Department of Physics, Institute of Science, Banaras Hindu University, Varanasi, India > Research Paper 2019 4 12 2019 10 10 Agriculture sector is the most significant for the Indian economy. So, it becomes necessary to estimate agricultural crop growth biophysical parameters for the proper crop monitoring and forecasting of the crop yield. The objectives of the research are estimation and comparison of wheat crop biophysical parameters using Sentinel-1A images by linear kernel based support vector regression (SVR) model, radial basis function based artificial neural network regression (ANNR) machine learning model and linear regression (LR) model. The ground samples of wheat crop growth biophysical parameters like as leaf area index (LAI), plant height (PH), fresh biomass (FB), dry biomass (DB), and vegetation water content (VWC) were collected during 8 January 2015 to 29 April 2015. The estimated results are statistically analysed and compared by coefficient of determination (R2), Nash Sutcliffe efficiency (NSE), %bias and by the analysis of root mean square error (RMSE). Overall good results such as R2 = 0.919, %bias = 1.153, NSE = 0.925 and RMSE = 0.661 values were found for the estimation of VWC using ANNR model. Whereas, LR model performed poorly for the PH estimation with the R2 = 0.662, %bias = 5.638, NSE = 0.627 and RMSE = 18.347 values at C-band. The SVR and ANNR models are found more suitable for wheat crop biophysical parameters estimation in compare to the LR modeling approach. The outcomes of the present study by different models may offer the valuable information for accurate monitoring of multiple crops in future for better crop production. Copyright@ International Science Community Association