Triangulated Irregular Network Research Articles

Hydrological drought and floodplain disconnectivity: quantifying flow thresholds in a large coastal plain river

ABSTRACT Hydrological drought and river disconnectivity worsen from dams and human activities. This research investigates how dam constructions exacerbate hydrological drought and river disconnectivity in Apalachicola River system. Data from two hydrological stations (1928-2022), a hydrographic survey, high-resolution LiDAR, and a triangular irregular network (TIN) were used to identify three connectivity flow (141.5 m3/s, 169.8 m3/s, and 226.4 m3/s) categories, each impacting floodplain differently. A relative elevation model (REM) was also developed for floodplain connectivity assessment. Findings show a significant increase in the frequency, duration, and magnitude of connectivity low flow events from pre-dam to post-dam periods, especially at the 169.8 m³/s connectivity flow threshold. A 20th percentile variable threshold shows increasing hydrological drought. Drought increases have led to decreased connectivity between tributaries and the main river system, and an increase in the moderately disconnected floodplain area from 37.1% to 46.0%. Adaptive management strategies are needed to address drought-induced floodplain disconnectivity

An Improved Adaptive Grid-Based Progressive Triangulated Irregular Network Densification Algorithm for Filtering Airborne LiDAR Data

Ground filtering is crucial for airborne Light Detection and Ranging (LiDAR) data post-processing. The progressive triangulated irregular network densification (PTD) algorithm and its variants outperform others in accuracy, stability, and robustness, using grid-based seed point selection, TIN construction, and iterative rules for ground point identification. However, these methods still face limitations in removing low points and accurately preserving terrain details, primarily due to their sensitivity to grid size. To overcome this issue, a novel PTD filtering algorithm based on an adaptive grid (AGPTD) was proposed. The main contributions of the proposed method include an outlier removal method using a radius outlier removal algorithm and Kd-tree, a method for establishing an adaptive two-level grid based on point cloud density and terrain slope, and an adaptive selection method for angle and distance thresholds in the iterative densification processing. The performance of the AGPTD algorithm was assessed based on widely used benchmark datasets. Results show that the AGPTD algorithm outperforms the classical PTD algorithm in retaining ground feature points, especially in reducing Type I error and average total error significantly. In comparison with other advanced algorithms developed in recent years, the novel algorithm showed the lowest average Type I error, the minimal average total error, and the greatest average Kappa coefficient, which were 1.11%, 2.28%, and 90.86%, respectively. Additionally, the average accuracy, precision, and recall of AGPTD were 97.69%, 97.52%, and 98.98%, respectively.

Intelligent processing of UAV remote sensing data for building high-precision DEMs in complex terrain: A case study of Loess Plateau in China

The Loess Plateau in China is renowned for its dense gullies and complex terrain, with drastic changes primarily due to soil erosion and human activities, significantly affecting the evolution of the ecological environment. The complex terrains and dense vegetation make precise terrain measurement and modeling challenging. Although the development of Unmanned Aerial Vehicle (UAV) light detection and ranging (LiDAR) scanning and photogrammetry technologies has improved data acquisition precision, relying solely on one remote sensing technology struggles with accurately extracting bare earth information. This study adopted a method that fuses UAV lidar scanning with aerial photogrammetric imagery, generating detailed lidar point cloud data that includes coordinate, reflectance, true color, and texture information to enhance data classifiability and interpretability. Subsequently, a point cloud classification model based on the Transformer architecture (Stratified Transformer) is introduced to intelligently complete the initial ground point cloud extraction in complex gully terrains. Further, to address residual non-ground noise in the initial ground point clouds, a new point cloud classification optimization algorithm (MDD, Multi-scale C2M Distance Difference) is proposed. This algorithm, based on the characteristics of discrete and non-continuous with the ground surface of the noisy point clouds, effectively eliminates the discrete noisy point clouds by analyzing the distances between the point clouds and TINs (Triangular Irregular Networks) of different scales and their differences. This study effectively addresses the technical challenges of ground point cloud extraction in the mixed environment of complex terrain and vegetation, solving the problem of precise terrain measurement and intelligent data processing in complex gully terrains, and offering new technical pathways for detecting geomorphological changes.

A status digital twin approach for physically monitoring over-and-under excavation in large tunnels

Accurate assessments of over-and-under-excavation volumes during tunnel construction are critical for stability and structural safety. Over-and-under-excavation assessment is typically time-consuming as it is undertaken manually and is prone to inaccuracies that adversely impact quality, safety, and costs. To overcome this problem, a status digital twin (to monitor physical conditions) is developed to assess over-and-under excavation during tunneling, enabling a robust and automated assessment of volumes of earth to be determined. The development of the status digital twin comprises three stages: (1) creating an as-designed building information model and as-built point cloud model; (2) producing an automated status update with newly added data; and (3) calculating the of over-and-under excavation volume by building a Triangulated Irregular Network (TIN). The status digital twin is tested in a large-scale tunnel project to evaluate its feasibility and effectiveness. The average over-and-under excavation thickness is 0.05 m, and the mean error of volume calculation is 4.54 %. The results demonstrate that the status digital twin can accurately assess over-and-under excavation during the construction of large-scale tunnel projects by providing instant feedback concerning over/under excavation, thus enabling quality to be effectively managed and costs to be controlled.

Interactive effects of elevation difference, slope variation, and terrain formation on road traffic crashes occurrences using triangular irregular network

AbstractNumerous factors contribute to traffic crashes, which result in fatalities and injuries. Road traffic crashes and their severity level are likely caused by variations in elevation, slope, and terrain formation. As far as our knowledge is concerned, the triangular irregular network (TIN) was not used to analyze dangerous zones of the road network with respect to traffic crashes' location. The main purpose of this study was to investigate a dangerous road network that was affected by traffic crashes using TIN. To address the goal, this study used elevation differences slope variation, and terrain between road traffic crash locations. The study used the intersection zone road network coordinates of traffic crashes in the city of Budapest that were recorded from 2017 to 2021. Using tools like ArcGIS and MS Excel, the study organized, modeled, and analyzed the data. Geometrical interval classification was used to classify the data in accordance with the characteristics of elevation deference and slope variation. Both inferential and descriptive statistics were engaged to explain, analyze, and summarize the findings. The results of this study demonstrated that the majority of traffic crashes occurred in locations with lower elevations and their differences, steeper/cliff slopes between points, and flat terrain. The finding was in contrast to what was observed in other studies. This study advocates that TIN is the optimal method and substitute for analyzing the risky zone of the road network at macroscopic level.

Preparing Cadastral Map Databases for Mosul District

Investigates the development and implementation of digital cadastral map databases for Mosul District, Iraq, addressing the transition from traditional paper-based systems to modern geospatial technologies. The research explores the integration of Geographic Information Systems (GIS), remote sensing, and Global Positioning System (GPS) technologies to enhance cadastral mapping processes. Encompass data collection from diverse sources, including high-resolution satellite imagery, aerial photography, and existing cadastral maps. A comprehensive analysis of data types— vector, raster, Triangulated Irregular Network (TIN), and tabular—is conducted to establish an optimal database structure. The study delineates the technical capabilities of GIS in data processing, spatial analysis, and model preparation, emphasizing the importance of data accuracy and consistency. A systematic approach to data entry, including map scanning, georeferencing, and digitization, is presented. The research outlines the creation of a multi-layered geographic database, incorporating both spatial and descriptive data. Protocols for database updating using satellite imagery and field surveys are developed, ensuring the currency and reliability of cadastral information. The results demonstrate significant improvements in data management, spatial analysis capabilities, and information accessibility compared to traditional methods. The study concludes that digital cadastral databases offer enhanced decision-making support for land administration and urban planning in Mosul District. Contributes to the body of knowledge on modernizing cadastral systems in developing regions, providing insights into technological integration, data standardization, and the challenges associated with transitioning to digital cadastral management systems. The findings have implications for policy makers and land administrators seeking to implement similar systems in comparable urban contexts.

Submerged Surface Area and Water Storage Volume of Hamrin Dam Lake Using Contour Lines

The predictions and warnings from drying up lakes and rivers are increasing nowadays, especially in Iraq, due to climatic changes and water flow reduction from neighboring countries. Consequently, this paper deals with monitoring the changes in submerged surface areas and Hamrin dam lake storage volume from December 2011 to the end of 2022. Hamrin Lake supplies Diyala province, Iraq, with irrigation water. Multitemporal satellite images were acquired from google earth pro and Earth Observation (EO) browsers. Triangular Irregular Network (TIN) layers generated contour lines through the ArcGIS environment. Volumes were calculated using the end area method. The results show that the water storage volume reached 1.8 billion cubic meters (bcm) in December 2016 and decreased to 12 million cubic meters in December 2021. Hamrin Lake lost 99% and 98% of its capacity relative to the normal water storage volume in 2021 and 2022, respectively. At normal storage, the maximum water volume of Hamrin Lake is about 3 bcm. The scarcity of irrigation water from Hamrin Lake causes decreasing agricultural areas, desertification, and people migrating to other regions.

Triangulated Irregular Network based Seamline Determination for Fast Image Stitching of Multiple UAV Images

Abstract. To maintain the efficiency of UAV (unmanned aerial vehicle) remote sensing, rapid image stitching is essential to make multiple UAV images into a seamless mosaic image. Relief displacement introduces variations in object appearance for each image, causing mismatch errors at mosaic seamlines. Traditional approaches involve orthorectifying images using DSMs (digital surface models). While these approaches allow for accurate image stitching, they do not cope with the advantages of UAVs due to their time consumption. In contrast, fast image stitching techniques that do not use orthorectification are well suited for UAV image processing. Related researches have attempted to optimize seamlines to eliminate the errors caused by relief displacement without the use of DSMs. We propose to utilize a TIN (triangular irregular network) of tiepoints to effectively eliminate errors caused by relief displacement while maintaining the fast speed of image stitching. In this study, a TIN is constructed based image tiepoints whose ground coordinates have been obtained through bundle adjustment. The edges of the TIN are used to generate seamlines for image stitching, and the facets of the TIN are used to select minimal images for image stitching and to optimize seamlines. Image stitching results of our proposed method had small error of 1–2 pixels and the processing time of less than 10 minutes for 97 UAV images. This study showed that the proposed method could stitch multiple images while maintaining stable quality using only geometric clues of a TIN of tiepoints.

Impact of Hurricane Irma on coral reef sediment redistribution at Looe Key Reef, Florida, USA

Abstract. Understanding event-driven sediment transport in coral reef environments is essential to assessing impacts on reef species, habitats, restoration, and mitigation, yet a global knowledge gap remains due to limited quantitative studies. Hurricane Irma made landfall in the Lower Florida Keys with sustained 209 km h−1 winds and waves greater than 8 m on 10 September 2017, directly impacting the Florida Reef Tract (FRT) and providing an opportunity to perform a unique comprehensive, quantitative assessment of its impact on coral reef structure and sediment redistribution. We used lidar and multibeam derived digital elevation models (DEMs) collected before and after the passing of Hurricane Irma over a 15.98 km2 area along the lower FRT including Looe Key Reef to quantify changes in seafloor elevation, volume, and structure due to storm impacts. Elevation change was calculated at over 4 million point locations across 10 habitat types within this study area for two time periods using data collected (1) approximately 1 year before the passing of Irma and 3 to 6 months following the storm's impact as well as (2) 3 to 6 months after and up to 16.5 months after the storm. Elevation change data were then used to generate triangulated irregular network (TIN) models in ArcMap to calculate changes in seafloor volume during each time period. Our results indicate that Hurricane Irma was primarily a depositional event that increased mean seafloor elevation and volume at this study site by 0.34 m and up to 5.4 Mm3, respectively. Sediment was transported primarily west-southwest (WSW) and downslope, modifying geomorphic seafloor features including the migration of sand waves and rubble fields, formation of scour marks in shallow seagrass habitats, and burial of seagrass and coral-dominated habitats. Approximately 16.5 months after Hurricane Irma (during a 13-month period between 2017 and 2019), net erosion was observed across all habitats with mean elevation change of −0.15 m and net volume change up to −2.46 Mm3. Rates of elevation change during this post-storm period were 1 to 2 orders of magnitude greater than decadal and multi-decadal rates of change in the same location, and changes showed erosion of approximately 50 % of sediment deposited during the storm event as seafloor sediment distribution began to re-equilibrate to non-storm sea-state conditions. Our results suggest that higher-resolution elevation change data collected over seasonal and annual time periods could enhance characterization and understanding of short-term and long-term rates and processes of seafloor change.

Parallel Topology-aware Mesh Simplification on Terrain Trees

We address the problem of performing a topology-aware simplification algorithm on a compact and distributed data structure for triangle meshes, the Terrain trees. Topology-aware operators have been defined to coarsen a Triangulated Irregular Network (TIN) without affecting the topology of its underlying terrain, i.e., without modifying critical features of the terrain, such as pits, saddles, peaks, and their connectivity. However, their scalability is limited for large-scale meshes. Our proposed algorithm uses a batched processing strategy to reduce both the memory and time requirements of the simplification process, and thanks to the spatial decomposition on the basis of Terrain trees, it can be easily parallelized. Also, since a Terrain tree after the simplification process becomes less compact and efficient, we propose an efficient post-processing step for updating hierarchical spatial decomposition. Our experiments on real-world TINs, derived from topographic and bathymetric LiDAR data, demonstrate the scalability and efficiency of our approach. Specifically, topology-aware simplification on Terrain trees uses 40% less memory and half the time compared to the most compact and efficient connectivity-based data structure for TINs. Furthermore, the parallel simplification algorithm on the Terrain trees exhibits a 12× speedup with an OpenMP implementation. The quality of the output mesh is not significantly affected by the distributed and parallel simplification strategy of Terrain trees, and we obtain similar quality levels compared to the global baseline method.

Evaluating the performance of various interpolation techniques on digital elevation models in highly dense forest vegetation environment

A Digital Elevation Model (DEM) is a numerical representation of the Earth's topography using an array of cells or pixels. DEMs are important for various purposes such as hydrological studies and environmental monitoring in situations where there is a need to assess land use changes, deforestation, and disaster response. In this study, various spatial interpolation methods, such as the Inverse distance weighting (IDW) technique, Natural Neighbor, and Triangulated Irregular Network (TIN) methods, were compared to evaluate the performance in generating accurate and reliable DEMs. We also considered factors such as computational efficiency and ease of implementation to determine the most suitable method. The precision of the generated DEMs was assessed through metrics such as root mean square error (RMSE) and mean absolute bias (MAB). The predicted elevation models were compared with elevation values obtained through ground surveys. Furthermore, a raster agreement analysis was used to compare the elevation values recorded by the two data sources, the USGS 3D Elevation Program (USGS 3DEP) and Unmanned Aerial Vehicle (UAV). The analysis revealed that TIN method demonstrated the highest level of agreement at 92%, while the binning technique showed inconsistency in predicting the elevation values based on the data source. In regions with high forest cover, the results were less accurate due to the limited number of last returns available for developing the DEM. This study also found that TIN recorded the fastest processing time per square acre. The presence of vegetation significantly influences the quantity of Light Detection and Ranging (LiDAR) pulses capable of reaching the forest floor, subsequently impacting the quantity of points available for the creation of a DEM.

A Study of Precipitation Forecasting for the Pre-Summer Rainy Season in South China Based on a Back-Propagation Neural Network

In South China, the large quantity of rainfall in the pre-summer rainy season can easily lead to natural disasters, which emphasizes the importance of improving the accuracy of precipitation forecasting during this period for the social and economic development of the region. In this paper, the back-propagation neural network (BPNN) is used to establish the model for precipitation forecasting. Three schemes are applied to improve the model performance: (1) predictors are selected based on individual meteorological stations within the region rather than the region as a whole; (2) the triangular irregular network (TIN) is proposed to preprocess the observed precipitation data for input of the BPNN model, while simulated/forecast precipitation is the expected output; and (3) a genetic algorithm is used for the hyperparameter optimization of the BPNN. The first scheme reduces the mean absolute percentage error (MAPE) and the root mean square error (RMSE) of the simulation by roughly 5% and more than 15 mm; the second reduces the MAPE and RMSE by more than 15% and 15 mm, respectively, while the third improves the simulation inapparently. Obviously, the second scheme raises the upper limit of the model simulation capability greatly by preprocessing the precipitation data. During the training and validation periods, the MAPE of the improved model can be controlled at approximately 35%. For precipitation hindcasting in the test period, the anomaly rate is less than 50% in only one season, and the highest is 64.5%. According to the anomaly correlation coefficient and Ps score of the hindcast precipitation, the improved model performance is slightly better than the FGOALS-f2 model. Although global climate change makes the predictors more variable, the trend of simulation is almost identical to that of the observed values over the whole period, suggesting that the model is able to capture the general characteristics of climate change.

A stream-aligned mixed polyhedral meshing strategy for integrated surface-subsurface hydrological models

Mesh refinement near river corridors is commonly deployed to resolve topographic details important for accurately representing riverine flow dynamics in watershed-scale spatially distributed integrated hydrology models. However, this increased accuracy comes with higher computational costs due to larger mesh and smaller time steps resulting from small mesh cells. We present a novel stream-aligned mixed-polygonal meshing approach that resolves river-corridor topography efficiently by placing long quadrilateral cells along the mapped center lines of the rivers and standard Triangulated Irregular Network (TIN) mesh in the rest of the domain. Our integrated hydrology simulations using mixed-polyhedral mesh generated from the mixed-polygonal approach closely match those using a highly refined extruded-TIN mesh but with a 96.43% reduction in the number of mesh cells and a 99.75% reduction in CPU hours. Conversely, coarser extruded-TIN meshes, due to poor water conveyance relative to the refined mesh simulations, yielded lower peak flows, longer recession curves, and higher inundation fractions.The creation of this mixed-polyhedral mesh is implemented within the Watershed Workflow tool, a Python-based workflow library for watershed simulation. We also implemented a layered process of hydrologic conditioning specific to the cells in the river corridor, which are explicitly meshed within the mixed-polyhedral mesh. The hydrologic conditioning enforces monotonic topographic gradients, removes spurious obstructions, and allows for burning-in depths of the river cells based on ancillary information. This approach paves the way for efficiently modeling narrow engineered channels and specifying river-specific hydrodynamic details and biogeochemical processes.

PERFORMANCE ASSESSMENT OF SPATIAL INTERPOLATIONS FOR TRAFFIC NOISE MAPPING ON UNDULATING AND LEVEL TERRAIN

Traffic noise mapping frequently employs Kriging, Inverse Distance Weighted (IDW), and Triangular Irregular Networks (TIN) spatial interpolations. This study uses the Henk de Kluijver noise model to evaluate the performance of spatial interpolations. Effective traffic noise mapping requires that noise observation points (Nops) be designed as 2 m grids. The upper and lower slopes function as noise barriers to reduce sound levels. Therefore, assessment of accuracy is essential for visualising noise levels in undulating and level terrain. In addition, this work gives an accurate comparison of traffic noise interpolation in undulating areas. The elements of spatial interpolations, such as the weighting factor, variogram, radius, and number of points influence the interpolation accuracy. The Kriging with a Gaussian variogram, where the radius is 5 m and the number of points is 12 demonstrates the highest level of precision. However, there is no direct relationship between accuracy validation and cross-validation. In cross-validation, however, the accuracy of the Gaussian variogram with a 7 m radius and 18 points is more accurate. In addition, this study demonstrates that Kriging is superior for extrapolating noise levels in undulating regions. Accurate visualising traffic noise levels requires a prior understanding of spatial interpolations.

Визначення об’єму насипів методом GNSS та лазерного сканування

In the modern period, the calculation of volumes of embankments plays an important role in a number of industries, starting from construction and ending with mining. This calls for the improvement of methods and technologies to ensure accurate and operational calculations of embankment volumes. One of the promising directions in this context is the use of Triangulated Irregular Networks (TIN) and Mesh models based on data obtained during geodetic measurements. Surveying is a large amount of accurate spatial measurement data that is necessary to create detailed geometric models of objects. The obtained geodetic data allow creating TIN and Mesh models that reflect the relief of the land surface with the necessary accuracy and detail. The use of these models to calculate the volumes of embankments becomes a key element in improving the processes of design, construction and management of natural resources. The purpose of this work is to compare the accuracy of the calculation of embankment volumes based on the data obtained by the hand-held laser scanner Stonex X120GO and the GNSS receiver Stonex S700A. Method. The method of comparing equal-precision measurements was used. Since there were no reference volumes of embankments at the site of the works, the difference between the volumes calculated on the basis of the data obtained with the hand-held laser scanner Stonex X120GO and the GNSS receiver Stonex S700A was used. Also, comparative and visual analysis was used to compare the difference between Mesh models obtained on the basis of different data sources. The results. The accuracy of determining the volumes of embankments obtained by a handheld laser scanner Stonex X120GO and a GNSS receiver Stonex S700A was studied. The average error of the differences between the determined volumes was 4.62%. Point clouds and Mesh models themselves were also analyzed (number of points, triangles, surface shape). The advantages and disadvantages of using a GNSS receiver and a hand-held laser scanner to collect data for calculating the volumes of material mounds are identified. Scientific novelty and practical significance. A technique for checking the accuracy of determining the volume of embankments without known reference values is proposed. The influence of the embankment area error on the accuracy of volume determination was studied. According to the results of the study, it is possible to assert the advantages in the detail of the received data and the speed of shooting when using a manual laser scanner, and about some difficulties associated with the desired presence of a powerful computer, a large array of data, an increase in the volume of camera work in comparison with the data obtained under the time of application of the GNSS receiver and the RTK method for determining the volumes of embankments

Nanoparticles in terrestrial sediments and the behavior of the spectral optics of Sentinel-3B OLCI Satellite images in a river basin of UNESCO World Cultural and Natural Heritage.

The dangerous chemical elements associated with nanoparticles (NPs) and ultra-fine sediment particles in hydrological bays have the capacity to move contaminants to large oceanic regions. The general objective of this study is to quantify the major chemical elements present in NPs and ultra-fine particles in aquatic sediments sampled from Guanabara Bay and compare these data to values determined through spectral optics using the Sentinel-3B Satellite OLCI (Ocean Land Color Instrument) during the winter and summer seasons of 2018, 2019, 2020, 2021, and 2022. This is done to highlight the impacts anthropogenic environmental hazards have on the marine ecosystem and human beings. Ten aquatic sediment field collection points were selected by triangulated irregular network (TIN). Samples were subjected to analysis by X-ray diffraction (XRD), scanning electron microscopy (SEM), electron dispersion spectroscopy (EDS), and transmission electron microscopy (TEM), which enabled a detailed analysis using scanning transmission electron microscopy (STEM). Geospatial analyses using Sentinel-3B OLCI Satellite images considered Water Full Resolution (WFR) at 300m resolution, in neural network (NN), normalized at 0.83µg/mg. A maximum average spectral error of 6.62% was utilized for the identification of the levels of Absorption Coefficient of Detritus and Gelbstoff (ADG443_NN) at 443m-1, Chlorophyll-a (CHL_NN) (m-3), and Total Suspended Matter (TSM_NN) (g m-3) at 581 sample points. The results showed high levels of ADG443_NN, with average values as high as of 4444m-1 (summer 2021). When related to the analyses of nanoparticulate sediments and ultrafine particles collected in the field, they showed the presence of major chemical elements such as Ge, As, Cr, and others, highly toxic to human health and the aquatic environment. The application of satellite and terrestrial surveys proved to be efficient, in addition to the possibility of this study being applied to other hydrological systems on a global scale.

Proximity Queries on Point Clouds using Rapid Construction Path Oracle

The prevalence of computer graphics technology boosts the developments of point clouds in recent years, which offer advantages over terrain surfaces (represented by Triangular Irregular Networks, i.e., TINs) in proximity queries, including the shortest path query, the k-Nearest Neighbor (kNN) query and the range query. Since (1) all existing on-the-fly and oracle-based shortest path query algorithms on a TIN are very expensive, (2) all existing on-the-fly shortest path query algorithms on a point cloud are still not efficient, and (3) there are no oracle-based shortest path query algorithms on a point cloud, we propose an efficient (1+ε)-approximate shortest path oracle that answers the shortest path query for a set of Points-Of-Interests (POIs) on the point cloud, which has a good performance (in terms of the oracle construction time, oracle size and shortest path query time) due to the concise information about the pairwise shortest paths between any pair of POIs stored in the oracle. Our oracle can be easily adapted to answering the shortest path query for any points on the point cloud if POIs are not given as input, and also achieve a good performance. Then, we propose efficient algorithms for answering the (1+ε)-approximate kNN and range query with the assistance of our oracle. Our experimental results show that when POIs are given (resp. not given) as input, our oracle is up to 390 times, 30 times and 6 times (resp. 500 times, 140 times and 50 times) better than the best-known oracle on a TIN in terms of the oracle construction time, oracle size and shortest path query time, respectively. Our algorithms for the other two proximity queries are both up to 100 times faster than the best-known algorithms.

A Single Data Extraction Algorithm for Oblique Photographic Data Based on the U-Net

In the automated modeling generated by oblique photography, various terrains cannot be physically distinguished individually within the triangulated irregular network (TIN). To utilize the data representing individual features, such as a single building, a process of building monomer construction is required to identify and extract these distinct parts. This approach aids subsequent analyses by focusing on specific entities, mitigating interference from complex scenes. A deep convolutional neural network is constructed, combining U-Net and ResNeXt architectures. The network takes as input both digital orthophoto map (DOM) and oblique photography data, effectively extracting the polygonal footprints of buildings. Extraction accuracy among different algorithms is compared, with results indicating that the ResNeXt-based network achieves the highest intersection over union (IOU) for building segmentation, reaching 0.8255. The proposed “dynamic virtual monomer” technique binds the extracted vector footprints dynamically to the original oblique photography surface through rendering. This enables the selective representation and querying of individual buildings. Empirical evidence demonstrates the effectiveness of this technique in interactive queries and spatial analysis. The high level of automation and excellent accuracy of this method can further advance the application of oblique photography data in 3D urban modeling and geographic information system (GIS) analysis.

EXPERIMENTING TIN DATA STRUCTURE FOR REPRESENTING PLANE SURFACE AND SUBSURFACE SPATIAL OBJECTS - PRELIMINARY WORK

Abstract. Understanding and analyzing complex spatial objects involves the integration of surface, terrain, and subsurface data into 3D spatial models. However, describing and efficiently analyzing the connections between these data model is particularly challenging. This study aims to generate TIN data structure that enable the 3D representation of surface terrain subsurface integration. Furthermore, to ensure accurate representation within a unified 3D model, spatial relationships between the TIN data model could be obtained. Consequently, three – dimensional Triangular Irregular Networks (3D TIN) mesh of spatial objects representations and indexing structures were created. The integration of spatial objects, encompassing surface, terrain, and subsurface elements, is realized through a unified model. This synergy is achieved by combining the distinct 2D and 3D topology structures of each component. The resulting comprehensive model captures the intricate relationships and interactions within the spatial environment. This study provides additional insight into the representation and analysis of surface, terrain, and subsurface management in 3D spatial models. Hence, better decision making, resource management, underground utility installations, and evaluation of spatial objects will be made possible in our future work.

GeoPDNN 1.0: a semi-supervised deep learning neural network using pseudo-labels for three-dimensional shallow strata modelling and uncertainty analysis in urban areas from borehole data

Abstract. Borehole data are essential for conducting precise urban geological surveys and large-scale geological investigations. Traditionally, explicit modelling and implicit modelling have been the primary methods for visualizing borehole data and constructing 3D geological models. However, explicit modelling requires substantial manual labour, while implicit modelling faces problems related to uncertainty analysis. Recently, machine learning approaches have emerged as effective solutions for addressing these issues in 3D geological modelling. Nevertheless, the use of machine learning methods for constructing 3D geological models is often limited by insufficient training data. In this paper, we propose the semi-supervised deep learning using pseudo-labels (SDLP) algorithm to overcome the issue of insufficient training data. Specifically, we construct the pseudo-labels in the training dataset using the triangular irregular network (TIN) method. A 3D geological model is constructed using borehole data obtained from a real building engineering project in Shenyang, Liaoning Province, NE China. Then, we compare the results of the 3D geological model constructed based on SDLP with those constructed by a support vector machine (SVM) method and an implicit Hermite radial basis function (HRBF) modelling method. Compared to the 3D geological models constructed using the HRBF algorithm and the SVM algorithm, the 3D geological model constructed based on the SDLP algorithm better conforms to the sedimentation patterns of the region. The findings demonstrate that our proposed method effectively resolves the issues of insufficient training data when using machine learning methods and the inability to perform uncertainty analysis when using the implicit method. In conclusion, the semi-supervised deep learning method with pseudo-labelling proposed in this paper provides a solution for 3D geological modelling in engineering project areas with borehole data.