A better understanding of ecosystem processes requires accurate estimates of forest biomass and structure on global scales. Recently, there have been demonstrations of the ability of remote sensing instruments, such as radar and lidar, for the estimation of forest parameters from spaceborne platforms in a consistent manner. These advances can be exploited for global forest biomass accounting and structure characterization, leading to a better understanding of the global carbon cycle. The popular techniques for estimation of forest parameters from radar instruments in particular, use backscatter intensity, interferometry and polarimetric interferometry. This dissertation analyzes the accuracy of biomass and structure estimates over temperate forests of the North-Eastern United States. An empirical approach is adopted, relying on ground truth data collected during field campaigns over the Harvard and Howland Forests in 2009. The accuracy of field biomass estimates, including the impact of the diameter-biomass allometry is characterized for the field sites. Full waveform lidar data from two LVIS field campaigns of 2009 over the Harvard and Howland forests is analyzed to assess the accuracy of various lidar-biomass relationships. Radar data from NASA JPL's UAVSAR is analyzed to assess the accuracy of the backscatter-biomass relationships with a theoretical radar error model. The relationship between field biomass and InSAR heights is explored using SRTM elevation and LVIS derived ground topography. Temporal decorrelation, a major factor affecting the accuracy of repeat-pass InSAR observations of forests is analyzed using the SIR-C single-day repeat data from 1994. Finally, PolInSAR inversion of heights over the Harvard and Howland forests is explored using UAVSAR repeat-pass data from the 2009 campaign. These heights are compared with LVIS height estimates and the impact of temporal decorrelation is assessed.

The prospect of sea level rise due to melting ice sheets affirms the urgency of gaining knowledge on ice sheet dynamics during deglaciation. The Fennoscandian Ice Sheet serves as an analogue, whose retreat can be reconstructed from the geomorphological record. The recent development of a high-resolution LiDAR-derived elevation model can reveal new relationships between landforms, even for well-studied areas such as the Torneträsk region in northwestern Sweden. Therefore, this study aims to refine the reconstruction of the deglaciation in this region based on an updated glacial geomorphological map. A range of glacial landforms were mapped, which by means of an inversion model were utilized to form swarms representing spatially and temporally coherent ice sheet flow systems. Additionally, glacial lake traces allowed for the identification of ice margins that dammed lakes in Torneträsk, Rautasjaure, and other (former) lake basins. Eight glacial lake stages were identified for the Torneträsk basin, where final drainage occurred through Tornedalen. Over 20 glacial lake stages were identified for the Rautasjaure basin, where drainage occurred along the margins of a thinning ice lobe. The disparity between the glacial lake systems results from different damming mechanisms in relation to the contrasting topography of the basins. A strong topographic control on the retreat pattern is evident, as the ice sheet retreated southward in an orderly fashion in the premontane region, but disintegrated into ice lobes in the montane region. The temporal resolution of current dating techniques is insufficient to constrain the timing of ice retreat at the spatial scale of this study. Precise dating of the Pärvie fault would pinpoint the age of the ice margin which at the time of rupture was located between two glacial lake stages of Torneträsk. Collectively, this study provides data for better understanding the final retreat of the ice sheet and associated processes, such as interactions between glacial lakes and ice dynamics.

No description available.

La representación del entorno juega un papel crucial en la navegación autónoma. A esta representación, se le suele denominar en la literatura como mapa, y su construcción suele realizarse mediante vehículos de mapeado dedicados. Sin embargo, aunque este tipo de mapas son muy precisos a nivel local, presentan el inconveniente de ser globalmente inconsistentes debido a la acumulación de pequeños errores que se hacen relevantes cuando los mapas crecen en tamaño. En la presente tesis doctoral, se propone como alternativa, la navegación autónoma basada en mapas públicos geo-referenciados. Este tipo de mapas, a diferencia de los construidos mediante vehículos de mapeado, son por naturaleza globalmente consistentes debido a que proceden de imágenes aéreas que se encuentran geo-referenciadas. Esto, además de ser una ventaja en sí misma, conlleva otro tipo de beneficios, como la no dependencia de un proceso de mapeado, o la posibilidad de navegar sin restricciones en el tamaño del entorno. No obstante, la integración de los mapas públicos geo-referenciados introduce algunas particularidades en la implementación de los algoritmos de navegación autónoma. Durante la investigación que se expone en esta memoria, se han desarrollado diferentes métodos para abordar dichas particularidades. En el módulo de localización, la representación de mapas geo-referenciados introduce la dependencia de un tipo de marcas que deben ser observables, tanto desde imágenes por satélite, como desde los sensores locales del vehículo. Esta restricción genera representaciones escasas que, a menudo, resultan en zonas ambiguas para la asociación de datos (efecto aliasing). Para abordar estos temas, se han desarrollado diferentes métodos robustos, como Delta-Angle Lane Markings Representation, una estrategia de representación para el proceso de asociación de datos, y Distance-Compatible SAmple Consensus, un método de asociación de datos. Para mitigar el efecto del aliasing en el módulo de localización, también se han empleado capacidades de autoajuste, que modifican de manera dinámica la configuración del método de asociación de datos en función de la pseudo-entropía medida en las observaciones. Por otra parte, para el módulo de planificación de rutas, se ha desarrollado un método llamado Naive-Valley-Path que corrige las imprecisiones locales intrínsecas en los mapas públicos. Todos estos métodos han sido comparados con sus homólogos en el estado del arte, demostrando en todos los casos mejoras que han resultado en contribuciones de gran impacto para la comunidad científica. / La presente tesis doctoral ha sido financiada por la Conselleria d’Innovació, Universitats, Ciència i Societat Digital de la Generalitat Valenciana y el Fondo Social Europeo de la Unión Europea a través de las subvenciones ACIF/2019/088 y BEFPI/2021/069.

The use of Light Detection and Ranging (lidar) technology is revolutionizing Maya archaeology, as it penetrates through thick vegetation prevalent in Maya environments, uncovering the structures and features below. At the site of El Mirador in the Petén Department of Guatemala, lidar data has been analyzed using Geographic Information Systems (GIS) to map features, such as residential buildings and quarries, that other technologies like satellite imagery have missed. El Mirador is a large site dating to the Preclassic through Post Classic periods (1000 BCE to 1500 BCE) and is argued to have the largest monumental architecture built by the Lowland Maya, but the nature of socioeconomic and political coordination at the site is poorly understood. Through analysis of quarry and residential structure volumes outlying areas of El Mirador at various distances from the city center, this research seeks to understand more about the nature of coordination at the site in terms of limestone production. Buffer zones of 150m and 300m were created around a central residence group in each selected area. This research shows that zones closer to the city center produced a greater volume of limestone than those further away; however, the quarries within each buffer zone did not produce enough stone even for the structures within their immediate zone. The total quarry volumes in the 150m buffer zones are greater than the combined volumes in the area between the 150m and 300m buffers, indicating a measure of coordination from each central structure group. Further research of quarrying at residential groups could help uncover the nature of supra-household coordination at Preclassic sites where the exact nature of elite involvement in quarrying is still not completely understood.

Landslides are a natural hazard that is expected to increase in the future, due to climate change. In order to keep risk management plans up to date, an efficient inventory method is needed. In previous studies, multi-temporal high-resolution digital elevation models (DEM) produced with LiDAR technology have been used successfully for landslide inventory and monitoring in different parts of the world. The aim of this study has been to discover an inventory method for landslides in Sweden, using two generations of elevation data produced with LiDAR. The analysis was performed in GIS with the creation of a DEM of difference (DoD) and visual comparison as key components. The sites were also verified using Google Earth satellite imagery and aerial photos. The result of the study shows that a functional, efficient method was developed and several potential landslides were found in the three different study areas. The soil characteristics, slope gradient and distance to areas affected by forestry were recorded for all potential landslide sites. Using multi-temporal DEM for landslide inventory is time- and cost efficient, and the results are more accurate compared to traditional inventory techniques. Hopefully the method developed in this study can be used on a larger scale and lead to updated risk management and prevention plans throughout all risk areas for landslides in Sweden. In future studies field work is recommended to verify the potential landslide sites.

As more data sources have become abundantly available, an increased interest in 3D reconstruction has emerged in the image processing academic community. Applications for 3D reconstruction of urban and residential buildings consist of urban planning, network planning for mobile communication, tourism information systems, spatial analysis of air pollution and noise nuisance, microclimate investigations, and Geographical Information Systems (GISs). Previous, classical, 3D reconstruction algorithms solely utilized aerial photography. With the advent of LIDAR systems, current algorithms explore using captured LIDAR data as an additional feasible source of information for 3D reconstruction. Preprocessing techniques are proposed for the development of an autonomous 3D Reconstruction algorithm. The algorithm is designed for autonomously deriving three dimensional models of urban and residential buildings from raw LIDAR data. First, a greedy insertion triangulation algorithm, modified with a proposed noise filtering technique, triangulates the raw LIDAR data. The normal vectors of those triangles are then passed to an unsupervised clustering algorithm – Fuzzy Simplified Adaptive Resonance Theory (Fuzzy SART). Fuzzy SART returns a rough grouping of coplanar triangles. A proposed multiple regression algorithm then further refines the coplanar grouping by further removing outliers and deriving an improved planar segmentation of the raw LIDAR data. Finally, further refinement is achieved by calculating the intersection of the best fit roof planes and moving nearby points close to that intersection to exist at the intersection, resulting in straight roof ridges. The end result of the aforementioned techniques culminates in a well defined model approximating the considered building depicted by the LIDAR data.

Autonomous vehicles are complex robotic and artificial intelligence systems working together to achieve safe operation in unstructured environments. The objective of this work is to provide a foundation to develop more advanced algorithms for off-road autonomy. The project explores the sensors used for off-road autonomy and the data capture process. Additionally, the point cloud data captured from lidar sensors is processed to restore some of the geometric information lost during sensor sampling. Because ground truth values are needed for quantitative comparison, the MAVS was leveraged to generate a large off-road dataset in a variety of ecosystems. The results demonstrate data capture from the sensor suite and successful reconstruction of the selected geometric information. Using this geometric information, the point cloud data is more accurately segmented using the SqueezeSeg network.

Obstacle detection and avoidance plays a crucial role in the autonomous navigation of unmanned ground vehicles (UGV). Information about the obstacles decreases as the distance between the UGV and obstacles increases. However, this information decreases much more rapidly for negative obstacles than for positive obstacles. UGV navigation becomes more challenging in off-road environments due to the higher probability of finding negative obstacles (e.g., potholes, ditches, trenches, etc.) compared with on-road environments. One approach to solve this problem is to avoid the candidate path with a negative obstacle, but in off-road environments avoiding negative obstacles in all situations is not possible. In such cases, the local path planner may need to choose a candidate path with a negative obstacle that causes the least amount of damage to the vehicle. To deal better with these types of scenarios, this research introduces a novel approach to perform 3D shape estimation of negative obstacles using LiDAR point cloud data. The dimensions (width, diameter, and depth), location (center), and curvature of negative obstacles were calculated based on an estimated shape. The presented approach can estimate the shape of different kinds of negative obstacles such as holes, trenches, in addition to large and complicated negative obstacles. This approach was tested on different terrain types using the Mississippi Autonomous Vehicle Simulation (MAVS).

No description available.