Adaptive UAV-based LiDAR and GNSS network for high-precision cadastral surveying in urban environments

Thesis Abstract

Urban cadastral surveying faces persistent accuracy and efficiency challenges due to dense infrastructure, multipath GNSS errors, and occlusions in built environments. This study investigates an adaptive UAV-based LiDAR and GNSS network to achieve high-precision cadastral measurements in urban contexts. The aim is to develop a technology-driven workflow that dynamically integrates tethered and free-flying UAV LiDAR data with multi-constellation GNSS observations to minimize geometric uncertainties and streamline field-to-finish delivery. Specific objectives include (1) designing an adaptive data fusion framework that selects LiDAR point clouds and GNSS epochs based on blockage metrics and urban morphologies; (2) evaluating a decentralized GNSS network augmentation scheme using fixed ground control stations and UAV-mounted receivers to improve baseline accuracy; (3) quantifying the impact of sensor integration on cadastral parcel boundary determination under varying urban canyons and near-buildings; (4) validating the approach against established terrestrial survey methods to determine accuracy, precision, and reliability thresholds; and (5) developing implementation guidelines for a scalable workflow applicable to municipal cadastral offices. The study adopts a mixed-methods research design combining quantitative accuracy assessment with qualitative workflow evaluation. The population comprises urban surveying campaigns in three metropolitan districts with diverse built-up configurations. A sample of 120 parcels (900–1200 m2 on average), selected to represent varying obstruction classes (open, partial canopy, dense corridor), will be surveyed using a multi-sensor UAV platform, a networked GNSS array, and conventional total station measurements as a benchmark. Data collection instruments include an integrated UAV platform equipped with calibrated LiDAR (pulse rate ~200 kHz, scan angle ±360°) and RTK-capable GNSS receivers, fixed ground control points, and terrestrial survey instruments for reference data. The methodology employs an adaptive data fusion algorithm operable in real time, combining LiDAR-derived digital surface models and GNSS network adjustments within a Kalman filter framework and a stochastic least squares adjustment. Validation metrics comprise RMSE of horizontal and vertical coordinates, boundary overlap statistics (shared boundary length, parcel area discrepancy), and reliability indicators such as confidence ellipses and Dilution of Precision (DOP) analyses. The analysis rests on regression techniques to model the relationship between urban morphology metrics and positioning accuracy, ANOVA to compare performance across obstruction classes, and Bayesian inference to quantify uncertainty reduction achieved by the adaptive fusion. Additionally, a thematic analysis of practitioner feedback will be conducted to assess workflow usability and adoption barriers, drawing on semi-structured interviews with 12 surveyors and GIS managers. Key expected findings include (i) improved cadastral coordinate accuracy within 10 cm horizontally and 15 cm vertically under dense urban obstacles through adaptive fusion and network augmentation; (ii) reduced field time and improved reliability of boundary delineation compared with conventional UAV-only or GNSS-only methods; (iii) demonstrable gains in boundary consistency across parcel adjacencies due to enhanced integrity of control networks; and (iv) a robust, repeatable workflow with quantified uncertainty bounds suitable for municipal cadastral administrations. The study contributes to knowledge by advancing a theory-driven, data-driven integration model that combines adaptive sensor fusion with spatial statistics under urban geospatial constraints, anchored by the theoretical underpinnings of sensor fusion theory, network adjustment theory, and urban geometry effects on GNSS visibility. Practical contributions include actionable guidelines for system configuration, data processing pipelines, and performance benchmarks for scale-up in urban cadastral agencies. The primary conclusion anticipates that an adaptive UAV LiDAR and GNSS network substantially improves high-precision cadastral surveying in urban environments, enabling reliable boundary determination with formal uncertainty quantification and reduced field operational burden. Recommendations encompass standardization of adaptive fusion parameters across building typologies, investment in regional GNSS augmentation infrastructure, development of a modular software toolkit for real-time data fusion and adjustment, and further research into real-time change detection for dynamic urban cadastral data.