Neural Networks Sorghum Yield Forecasting Model Using Satellite-based Vegetation Indices in the Rain-fed Sectors in Sudan

Adil M.Elkhidir, Hanan H. A. Adlan, Isam A. Basheir


Neural network is universal function approximators.  The most commonly used type of neural network is the Radial Basis Function (RBF). Artificial neural networks (ANNs) are made up of input layers, hidden layers and output layers. The RBF neural network has an input layer where the data samples are fed, typically after being normalized. A model has been developed based on a Radial Basis Function (RBF). Study area that has been selected is Gedaref state. The input data are visible Red-Green-Blue (RGB) images of agricultural areas used as predictors to predict the Sorghum yield. Vegetation Index green (VIg) in the Sorghum production is implemented to the Satellite images as input which is expressed as follows:

VIg = (green – red) / (green + red)

Neural Network architectures were developed, investigated, and tested for forecasting Sorghum yield for the selected study area. A monthly imagery input data and Sorghum yield output data sets are used for training and validation of the Neural Network models in the selected study area. Correlation between the measured output data (actual) and the forecasted output data using the model have given high value of R. The model can be used with great efficiency to forecast Sorghum (output) in targeted agriculture area given RGB image of the study area

Full Text:



  • There are currently no refbacks.

Copyright (c) 2017 University Of Khartoum Engineering Journal

ISSN: 1858-6333