One of the most important tasks in computer vision is semantic image segmentation, which helps us interpret visual data. In the Land Use-Land Cover (LULC) domain, it’s especially vital for mapping urban, agricultural, forested areas, etc. Traditional methods, widely employed, rely on analyzing color, texture, and shape using techniques like image processing, statistics, or classic Machine Learning (ML). While these methods yield satisfactory results, they have precision and generalization limitations. Deep Learning, particularly CNN-based methods, has revolutionized semantic image segmentation. CNNs excel at automatically learning features from data, enabling them to grasp complex image details. In this study, we offer a thorough investigation of semantic segmentation methods in this transformative deep learning context for Land Use-Land Cover images.