Satellite-based Crop Yield Prediction using Machine Learning Algorithm
Kavita Kavita, Pratistha Mathur
Abstract
Agriculture is the backbone of India's economy as 70% of the population's household is dependent on agriculture. Crop yield prediction is a crucial step for food security, market planning, and yield mapping. Machine learning with remote sensing has shown magnificent results for crop yield prediction. In this research, the authors implemented machine learning techniques along with the remote sensing data to extract features. Support Vector Machine and K-Nearest Neighbor- techniques have shown outstanding results - are implemented to predict the crop yield using remote sensing data. Authors focused on distinguish parts of a satellite image based on the vegetation for a few months and then finding the relation between their Normalized Difference Vegetation Index values and vegetation similarly, and also found out that vegetation of the same crop field varies over the months and on further computations, we can predict the crop growth in that particular area. For this research, K-Nearest Neighbor outperformed Support Vector Machine.