Remote sensing technologies provide a powerful means to monitor land use and land cover changes, offering critical insights into urban growth patterns and their impacts on peri-urban agricultural zones. This study focuses on the Philippi Horticultural Area (PHA) and examines how urban expansion influences land transformation dynamics over time. The aim is to improve understanding of spatial patterns of urban encroachment and support sustainable land management strategies.
High-resolution PlanetScope multispectral imagery spanning 2019 to 2021 was used to capture temporal changes in land cover. Optimized spectral indices were derived to enhance feature discrimination. Feature selection was performed using Boruta and Recursive Feature Elimination (RFE) to identify the most relevant variables for classification. Advanced machine learning classifiers, including Support Vector Machines (SVM), Random Forest (RF), and K-Nearest Neighbours (KNN), were applied to improve classification accuracy. The workflow integrated image preprocessing, feature extraction, feature selection, and supervised classification to map land cover classes and detect change over time. Validation was conducted using reference data to assess the reliability of the classification outputs.
The results reveal significant urban expansion into previously agricultural areas, with clear spatial patterns of land conversion observed across the study period. The applied models demonstrated strong performance in distinguishing between land cover classes, while the selected features improved separability between urban and vegetated surfaces. The analysis highlights key hotspots of urban encroachment and areas under increasing pressure from development.
In conclusion, the integration of high-resolution multispectral remote sensing data with robust feature selection and machine learning techniques provides an effective approach for monitoring land use change and supporting sustainable management of the Philippi Horticultural Area.