Crop Monitoring And Recommendation System Using Machine Learning Techniques

Crop Monitoring and Recommendation System Using Machine Learning Techniques

Abstract:

This document proposes a crop recommendation system using Spectral spatial classification and Support Vector Machine (SVM).The farmer provides the crop field image as an input to the application. In the pre-processing stage, Demising is done using Multiple morphological component analysis (MMCA) and as a result, filtering the image retaining its necessary portions. SVM prefixed by Spatial Spectral Schrodinger Eigen Maps (SSSE) is used as a classification method wherein partial knowledge propagation is leveraged to improve the classification accuracy.  The classified image along with the Ground truth statistical data containing the weather, crop yield, state & county wise crops are used to predict the yield of a particular crop under a particular weather condition. This predictive model used Gadabouts classifier. Crop recommendation is facilitated then by collaborative filtering.  Further scope of the project would extend to predictive analytics on the commodity market of the goods grown in the agricultural fields to predict its waxing and waning.