Design and Evaluation of a Mobile UAV-based Land Cover Monitoring System

Thesis Abstract

Rapid urbanization and environmental degradation have heightened the need for accurate, timely, and cost-effective land cover monitoring systems that can support sustainable land management practices. Conventional land survey methods are often labor-intensive, time-consuming, and limited in coverage, thereby necessitating innovative approaches that leverage emerging geospatial technologies. This study aims to design, implement, and evaluate a mobile Unmanned Aerial Vehicle (UAV)-based system capable of real-time land cover monitoring to address current gaps in spatial data accuracy and temporal frequency. The specific objectives are to develop a UAV deployment framework optimized for land cover data acquisition, to implement an integrated data processing pipeline incorporating Geographic Information Systems (GIS) and remote sensing techniques, and to assess the system's accuracy, efficiency, and operational feasibility in diverse terrain contexts. A mixed-methods research design was adopted, combining quantitative and qualitative approaches to ensure comprehensive evaluation. The study population comprised UAV operators, GIS specialists, and environmental managers within the urban and peri-urban regions of a metropolitan area spanning approximately 1,200 square kilometers. A purposive sampling technique was employed to select 30 UAV operators, 15 GIS experts, and 10 environmental stakeholders for interviews and focus groups, while a stratified random sampling approach was used to select five land cover sites representing forests, farmland, urban built-up areas, water bodies, and degraded lands for empirical data collection. Data sources included UAV-mounted multispectral sensors capturing high-resolution imagery, supplemented by existing land use maps and ground truth data obtained through field surveys. Data collection instruments comprised UAV flight logs, high-resolution imagery, questionnaires, and structured interview guides. The UAV platform was configured with lightweight multispectral sensors, optimized flight plans, and autonomous navigation capabilities. Spatial data was processed using open-source GIS software, with imagery analyzed through supervised classification techniques such as Maximum Likelihood Classification and object-based image analysis to delineate various land cover types. The accuracy of land cover classification was validated using confusion matrices and metrics like overall accuracy and Kappa coefficient. Descriptive statistics summarized operational metrics, while inferential analyses including regression analysis and ANOVA tested hypotheses regarding the system's accuracy, efficiency, and user satisfaction. Thematic analysis was applied to qualitative interview data to elucidate operational challenges and stakeholder perceptions. Expected findings include a classification accuracy exceeding 85%, a reduction in land cover mapping time by up to 60% compared to traditional methods, and high user satisfaction with system usability and data quality. The study anticipates identifying key operational factors influencing UAV system performance and providing detailed guidelines for scalable deployment in diverse geographical contexts. The results are expected to demonstrate that the integrated UAV-GIS framework enhances spatial-temporal land cover monitoring capabilities, supporting more informed land use planning and environmental management decisions. The study significantly contributes to the body of knowledge by providing a replicable model for mobile UAV-based land cover monitoring, validated through empirical performance metrics. It complements existing remote sensing literature with insights into operational feasibility and usability in varied terrains, thus bridging technological innovation with pragmatic land management needs. The conclusion emphasizes the system’s potential for integration into existing land information infrastructure and recommends further research into automation, sensor diversification, and real-time data dissemination. The findings advocate for policy considerations promoting UAV adoption in environmental monitoring frameworks, especially in rapidly changing urban landscapes where traditional methods are inadequate for timely decision-making.