Investigations of Quality Aspects in UAV-Based Laser Scanning for Agriculture and Deformation Monitoring

This dissertation investigates the quality of UAV-based laser scanning and its applications in agriculture and deformation monitoring. UAV-based laser scanning is a growing technology that integrates sensors such as GNSS (Global Navigation Satellite System), IMU (Inertial Measurement Unit) and a laser scanner to produce high-resolution georeferenced 3D point clouds. This technology is increasingly used across various fields, including mining, surveying, forestry, and agriculture, due to its flexibility, detailed data acquisition, and the benefit of direct georeferencing, eliminating the need for ground control points in many cases. <br /> The quality of these point clouds is crucial, as errors in trajectory estimation, system calibration, and sensor performance can significantly impact the quality and, therefore, the potential use of data. This dissertation addresses the use in precision agriculture and deformation monitoring with different quality aspects. <br /> The study is divided into four main objectives: <br /> - Evaluation of Direct Georeferencing: This objective assesses the quality of point clouds produced by UAV-based laser scanning with a focus on direct georeferencing. Key aspects evaluated include point cloud noise, accuracy, precision of repeated measurements, and the impact of GNSS master stations. By analyzing various flight patterns, such as cross-flight and corridor mapping, this research aims to understand and improve the performance of direct georeferencing, providing insights into how these systems and processing can be improved for different applications.<br /> - Quality Aspects of a Full-Waveform 2D Laser Scanner: The second objective examines the quality of the 2D laser scanner (RIEGL miniVUX-2UAV) used in UAV-based laser scanning systems. This part of the research evaluates important parameters such as range precision, rangefinder offset, resolution capability, and multi-target capability. The goal is to understand the impact of the laser scanner on the overall error budget of the scanning system, providing a detailed assessment of how the sensor performs with a focus on multi-target detection.<br /> - Estimation of Structural Plant Traits: This part applies UAV-based laser scanning to precision agriculture, focusing on the estimation of wheat crop height and plant area index (PAI). A novel ground classification algorithm based on k-means clustering is introduced to differentiate between ground and crop points. This approach investigates the quality in terms of the accuracy of structural plant trait estimation, addressing challenges related to the small object size and spatial resolution.<br /> - Deformation Analysis of Water Dams: The final objective investigates the use of UAV-based laser scanning for deformation monitoring of water dams. This addresses challenges in trajectory estimation, flight planning, and the unique measurement environment. The study highlights both the potential and limitations in terms of the quality of UAV-based scanning, particularly in GNSS-denied environments. A following instance segmentation algorithm for rubble masonry is developed to enhance deformation analysis, enabling comparisons between point clouds from different epochs. <br /> Overall, it provides a comprehensive discussion of UAV-based laser scanning technology and its applications, offering valuable insights into the understanding of point cloud quality and its use in different fields.