Influence of Selected Modeling Parameters on Plant Segmentation Quality Using Decision Tree Classifiers

Modern precision agriculture applications increasingly rely on stable computer vision outputs. An important task is to discriminate between soil and plant pixels, which called segmentation. For this task, supervised learning techniques, such as decision tree classifiers (DTC), support vector machines (SVM), or artificial neural networks (ANN) are increasing in popularity. The selection of training data utmost importance these approaches it influences the quality resulting models. We investigated influence three modeling parameters, namely proportion pixels (plant cover), criteria what pixel choose (pixel selection), number/type features (input features) segmentation using DTCs. Our findings show that cover and, a minor degree, input have significant impact quality. can state overperformance multi-feature over threshold-based color index methods be explained high degree by more balanced data. Single-feature compete with state-of-the-art models when same provided. This study first step systematic analysis parameters