Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Neural network models for soil nitrate prediction using imagery and non-imagery information

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan

Citation:  Paper number  033065,  2003 ASAE Annual Meeting . (doi: 10.13031/2013.13853) @2003
Authors:   Ramesh Kumar Gautam, Suranjan Panigrahi, David Franzen, Albert Sims
Keywords:   Landsat TM, Soil EC, topography, yield, soil nitrate, backpropagation, radial basis function, modular neural network models, imaging, texture

Ten satellite imagery and four non-imagery data were used to develop soil nitrate prediction models using neural network architectures i.e. Backpropagation, Modular and Radial basis function. Back propagation model yielded an average prediction accuracy of 67.92% with a root mean square error of 0.87. Radial basis function model, on the other hand, predicted with an average prediction accuracy of 90.04%. The modular neural network predicted with average prediction accuracy of 77.12%, correlation coefficient of 0.62 and RMSEP of 0.51. The modular neural network based prediction model, showed overall highest performance than BPNN and RBFN models.

(Download PDF)    (Export to EndNotes)