INFLUENCE OF VEGETATION FEATURES ON CORN YIELDS ESTIMATION USING DIFFERENT MACHINE LEARNING TECHNIQUES: A CASE STUDY

Yield prediction is a critical aspect of agricultural management. Recent advancements in remote sensing technologies enabled use of satellite imagery to estimate crop yields, with vegetation features being a key element. The impact of each feature, however, varies with different crops, and different vegetation features may be influential for different models for a single crop such as corn. Hence, a study would be beneficial to quantify influence of vegetation features over different models. The paper provides a case study over a publicly accessible dataset of corn yield to understand the influence of ten vegetation features in different regression models. It has been observed that the vegetation features sensitive to chlorophyll content and inclusion of Green channel provide significant contributions to corn yield estimation. The proposed technique has also been applied on soybean and wheat yield prediction to study the influence of vegetation features.