To address the lack of suitable training data for deep-learning semantic segmentation models in urban landscaping, researchers developed a method that generates a training dataset without the need for ...