Optimization of a Vision-Based Lettuce Growth Stage Classification using Genetic Algorithm

In this study, a vision-based lettuce crop stage classifier using the K-Nearest Neighbor algorithm is optimized using a genetic algorithm. The KNN algorithm that utilized all features resulted in an accuracy of 64.2%. After optimization, the thirteen (13) features are extracted and reduced to one (1). Based on the results of the study, the Y component from the YCbCr color space is enough to model the classification of the lettuce plants. Using that component, the accuracy of the model increased significantly to 92.8%