Integration of UAV technology and machine learning algorithms for land cover classification using ArcGIS Pro software

Abstract

The integration of Unmanned Aerial Vehicle (UAV) technology with machine learning algorithms is opening up an effective approach for processing and analyzing spatial data. This study performs land cover classification by applying two popular machine learning algorithms, Random Forest (RF) and Support Vector Machine (SVM), using ArcGIS Pro software. High-resolution orthophotos (2.57 cm/pixel) obtained from UAV imagery over District 4, Ho Chi Minh City, were used for the classification process. Experimental results show that the SVM algorithm performs well on some challenging classes such as vehicles and boats, achieving higher user accuracy than Random Forest (RF). However, when evaluated overall based on Overall Accuracy (OA) and the Kappa coefficient, RF outperforms SVM, with OA of 86.6% and Kappa of 0.83, compared to SVM’s 77.4% and 0.71, respectively. This indicates that RF has high accuracy and more suitable for constructing land cover maps from UAV imagery. These results confirm the feasibility and effectiveness of applying UAV images combined with machine learning algorithms in spatial analysis and monitoring. This provides a crucial basis for resource management, urban planning, and environmental protection in the context of rapid urbanization in major cities.