Search

We present a benchmark for assessing the capability of Large Language Models (LLMs) to discern intercardinal directions between geographic locations and apply it to three prominent LLMs: GPT-3.5, GPT-4, and Llama-2. This benchmark specifically evaluates whether LLMs exhibit a hierarchical spatial bias similar to humans, where judgments about individual locations' spatial relationships are influenced by the perceived relationships of the larger groups that contain them. To investigate this, we formulated 14 questions focusing on well-known American cities. Seven questions were designed to challenge the LLMs with scenarios potentially influenced by the orientation of larger geographical units, such as states or countries, while the remaining seven targeted locations were less susceptible to such hierarchical categorization. Among the tested models, GPT-4 exhibited superior performance with 55 percent accuracy, followed by GPT-3.5 at 47 percent, and Llama-2 at 45 percent. The models showed significantly reduced accuracy on tasks with suspected hierarchical bias. For example, GPT-4's accuracy dropped to 33 percent on these tasks, compared to 86 percent on others. However, the models identified the nearest cardinal direction in most cases, reflecting their associative learning mechanism, thereby embodying human-like misconceptions. We discuss avenues for improving the spatial reasoning capabilities of LLMs., Comment: This manuscript has been accepted for publication in The Professional Geographer

Accurate building height estimation is key to the automatic derivation of 3D city models from emerging big geospatial data, including Volunteered Geographical Information (VGI). However, an automatic solution for large-scale building height estimation based on low-cost VGI data is currently missing. The fast development of VGI data platforms, especially OpenStreetMap (OSM) and crowdsourced street-view images (SVI), offers a stimulating opportunity to fill this research gap. In this work, we propose a semi-supervised learning (SSL) method of automatically estimating building height from Mapillary SVI and OSM data to generate low-cost and open-source 3D city modeling in LoD1. The proposed method consists of three parts: first, we propose an SSL schema with the option of setting a different ratio of "pseudo label" during the supervised regression; second, we extract multi-level morphometric features from OSM data (i.e., buildings and streets) for the purposed of inferring building height; last, we design a building floor estimation workflow with a pre-trained facade object detection network to generate "pseudo label" from SVI and assign it to the corresponding OSM building footprint. In a case study, we validate the proposed SSL method in the city of Heidelberg, Germany and evaluate the model performance against the reference data of building heights. Based on three different regression models, namely Random Forest (RF), Support Vector Machine (SVM), and Convolutional Neural Network (CNN), the SSL method leads to a clear performance boosting in estimating building heights with a Mean Absolute Error (MAE) around 2.1 meters, which is competitive to state-of-the-art approaches. The preliminary result is promising and motivates our future work in scaling up the proposed method based on low-cost VGI data, with possibilities in even regions and areas with diverse data quality and availability.

Artificial intelligence (AI) approaches nowadays have gained remarkable success in single-modality-dominated remote sensing (RS) applications, especially with an emphasis on individual urban environments (e.g., single cities or regions). Yet these AI models tend to meet the performance bottleneck in the case studies across cities or regions, due to the lack of diverse RS information and cutting-edge solutions with high generalization ability. To this end, we build a new set of multimodal remote sensing benchmark datasets (including hyperspectral, multispectral, SAR) for the study purpose of the cross-city semantic segmentation task (called C2Seg dataset), which consists of two cross-city scenes, i.e., Berlin-Augsburg (in Germany) and Beijing-Wuhan (in China). Beyond the single city, we propose a high-resolution domain adaptation network, HighDAN for short, to promote the AI model's generalization ability from the multi-city environments. HighDAN is capable of retaining the spatially topological structure of the studied urban scene well in a parallel high-to-low resolution fusion fashion but also closing the gap derived from enormous differences of RS image representations between different cities by means of adversarial learning. In addition, the Dice loss is considered in HighDAN to alleviate the class imbalance issue caused by factors across cities. Extensive experiments conducted on the C2Seg dataset show the superiority of our HighDAN in terms of segmentation performance and generalization ability, compared to state-of-the-art competitors. The C2Seg dataset and the semantic segmentation toolbox (involving the proposed HighDAN) will be available publicly at https://github.com/danfenghong.

Accurate building height estimation is key to the automatic derivation of 3D city models from emerging big geospatial data, including Volunteered Geographical Information (VGI). However, an automatic solution for large-scale building height estimation based on low-cost VGI data is currently missing. The fast development of VGI data platforms, especially OpenStreetMap (OSM) and crowdsourced street-view images (SVI), offers a stimulating opportunity to fill this research gap. In this work, we propose a semi-supervised learning (SSL) method of automatically estimating building height from Mapillary SVI and OSM data to generate low-cost and open-source 3D city modeling in LoD1. The proposed method consists of three parts: first, we propose an SSL schema with the option of setting a different ratio of "pseudo label" during the supervised regression; second, we extract multi-level morphometric features from OSM data (i.e., buildings and streets) for the purposed of inferring building height; last, we design a building floor estimation workflow with a pre-trained facade object detection network to generate "pseudo label" from SVI and assign it to the corresponding OSM building footprint. In a case study, we validate the proposed SSL method in the city of Heidelberg, Germany and evaluate the model performance against the reference data of building heights. Based on three different regression models, namely Random Forest (RF), Support Vector Machine (SVM), and Convolutional Neural Network (CNN), the SSL method leads to a clear performance boosting in estimating building heights with a Mean Absolute Error (MAE) around 2.1 meters, which is competitive to state-of-the-art approaches. The preliminary result is promising and motivates our future work in scaling up the proposed method based on low-cost VGI data, with possibilities in even regions and areas with diverse data quality and availability., Comment: Accepted for GIScience 2023

Routes with a high share of greenery are attractive for cyclist and pedestrians. We analyze how strongly such green routes differ from the respective fast routes using the openrouteservice. Greenness of streets was estimated based on OpenStreetMap data in combination with Sentinel-II imagery, 3d laser scan data and administrative information on trees on public ground. We assess the effect both at the level of the individual route and at the urban level for two German cities: Dresden and Heidelberg. For individual routes, we study how strongly green routes differ from the respective fast routes. In addition, we identify parts of the road network which represent important green corridors as well as unattractive parts which can or cannot be avoided at the cost of reasonable detours. In both cities, our results show the importance of urban green spaces for the provision of attractive green routes and provide new insights for urban planning by identifying unvegetated bottlenecks in the street network for which no green alternatives exist at this point.

A worldwide increase in the number of people and areas affected by disasters has led to more and more approaches that focus on the integration of local knowledge into disaster risk reduction processes. The research at hand shows a method for formalizing this local knowledge via sketch maps in the context of flooding. The Sketch Map Tool enables not only the visualization of this local knowledge and analyses of OpenStreetMap data quality but also the communication of the results of these analyses in an understandable way. Since the tool will be open-source and several analyses are made automatically, the tool also offers a method for local governments in areas where historic data or financial means for flood mitigation are limited. Example analyses for two cities in Brazil show the functionalities of the tool and allow the evaluation of its applicability. Results depict that the fitness-for-purpose analysis of the OpenStreetMap data reveals promising results to identify whether the sketch map approach can be used in a certain area or if citizens might have problems with marking their flood experiences. In this way, an intrinsic quality analysis is incorporated into a participatory mapping approach. Additionally, different paper formats offered for printing enable not only individual mapping but also group mapping. Future work will focus on advancing the automation of all steps of the tool to allow members of local governments without specific technical knowledge to apply the Sketch Map Tool for their own study areas.

Routes with a high share of greenery are attractive for cyclist and pedestrians. We analyze how strongly such green routes differ from the respective fast routes using the openrouteservice. Greenness of streets was estimated based on OpenStreetMap data in combination with Sentinel-II imagery, 3d laser scan data and administrative information on trees on public ground. We assess the effect both at the level of the individual route and at the urban level for two German cities: Dresden and Heidelberg. For individual routes, we study how strongly green routes differ from the respective fast routes. In addition, we identify parts of the road network which represent important green corridors as well as unattractive parts which can or cannot be avoided at the cost of reasonable detours. In both cities, our results show the importance of urban green spaces for the provision of attractive green routes and provide new insights for urban planning by identifying unvegetated bottlenecks in the street network for which no green alternatives exist at this point.

Current indoor mapping approaches can detect accurate geometric information but are incapable of detecting the room type or dismiss this issue. This work investigates the feasibility of inferring the room type by using grammars based on geometric maps. Specifically, we take the research buildings at universities as examples and create a constrained attribute grammar to represent the spatial distribution characteristics of different room types as well as the topological relations among them. Based on the grammar, we propose a bottom-up approach to construct a parse forest and to infer the room type. During this process, Bayesian inference method is used to calculate the initial probability of belonging an enclosed room to a certain type given its geometric properties (e.g., area, length, and width) that are extracted from the geometric map. The approach was tested on 15 maps with 408 rooms. In 84% of cases, room types were defined correctly. It, to a certain degree, proves that grammars can benefit semantic enrichment (in particular, room type tagging).

The concept of place recently gains momentum in GIScience. In some fields like human geography, spatial cognition or information theory, this topic already has a longer scholarly tradition. This is however not yet completely the case with statistical spatial analysis and cartography. Despite that, taking full advantage of the plethora of user-generated information that we have available these days requires mature place-based statistical and visualization concepts. This paper contributes to these developments: We integrate existing place definitions into an understanding of places as a system of interlinked, constituent characteristics. Based on this, challenges and first promising conceptual ideas are discussed from statistical and visualization viewpoints.

Tailored routing and navigation services utilized by wheelchair users require certain information about sidewalk geometries and their attributes to execute efficiently. Except some minor regions/cities, such detailed information is not present in current versions of crowdsourced mapping databases including OpenStreetMap. CAP4Access European project aimed to use (and enrich) OpenStreetMap for making it fit to the purpose of wheelchair routing. In this respect, this study presents a modified methodology based on data mining techniques for constructing sidewalk geometries using multiple GPS traces collected by wheelchair users during an urban travel experiment. The derived sidewalk geometries can be used to enrich OpenStreetMap to support wheelchair routing. The proposed method was applied to a case study in Heidelberg, Germany. The constructed sidewalk geometries were compared to an official reference dataset (“ground truth dataset”). The case study shows that the constructed sidewalk network overlays with 96% of the official reference dataset. Furthermore, in terms of positional accuracy, a low Root Mean Square Error (RMSE) value (0.93 m) is achieved. The article presents our discussion on the results as well as the conclusion and future research directions.

The concept of place recently gains momentum in GIScience. In some fields like human geography, spatial cognition or information theory, this topic already has a longer scholarly tradition. This is however not yet completely the case with statistical spatial analysis and cartography. Despite that, taking full advantage of the plethora of user-generated information that we have available these days requires mature place-based statistical and visualization concepts. This paper contributes to these developments: We integrate existing place definitions into an understanding of places as a system of interlinked, constituent characteristics. Based on this, challenges and first promising conceptual ideas are discussed from statistical and visualization viewpoints.

The study presents a conceptual model to build a high level geo knowledge based system to manage real world processes. A generic event oriented perspective was used to implicitly represent causal relationships among different components of a Spatio-Temporal Geographic Information system. To utilize the collected event information, a Volunteered Location Based Service was developed. In addition a tool was signed to increase the general awareness of the volunteers in spatial analysis and shift their power and influences from a mere data collector to a knowledge producer., Diese Studie stellt ein konzeptionelles Modell vor, um ein geo-wissensbasiertes System zum Managen der realen Weltprozessen zu erstellen. Eine generische ereignisorientierte Perspektive wurde verwendet, um implizit die kausalen Zusammenhänge zwischen den verschiedenen Komponenten eines räumlich-zeitlichen Systems darzustellen. Um die gesammelten Ereignisinformationen zu nutzen, wurde ein “Volunteered Location Based Service” entwickelt. Darüber hinaus wurde ein Tool entwickelt, um das allgemeine Bewusstsein der Freiwilligen in räumlichen Analyse zu erhöhen.

In this work, we propose a framework to assess the quality of OpenStreetMap (OSM) building footprints data without using any reference data. More specifically, the OSM history data will be examined regarding the development of attributes, geometries and positions of building footprints. In total seven quality indicators are defined for the intrinsic quality assessment. For our case study in the federal state of Baden-Württemberg (BW), Germany, a PostgreSQL database is established based on a spatiotemporal data model which can track both individual objects and editing events on OSM. The preliminary experiments show that the quality of building footprints in BW is relatively high. And the quality in terms of semantics, geometries and positions are getting increasingly high over the time thanks to the considerable contribution of OSM volunteers., In dieser Arbeit stellen wir ein Konzept zur Bewertung von der Qualität von Gebäudegrundrissen aus OpenStreetMap (OSM) ohne Verwendung von Referenzdaten vor. Insbesondere wird der Verlauf der Bearbeitung von Stützpunkten und Attributen der Objekte untersucht. Sieben Indikatoren Bewertung der intrinsischen Datenqualität wurden definiert. Für die vorliegende Studie ist am Beispiel von Baden-Württemberg eine PostgreSQL-Datenbank erstellt worden, um ein räumlich-zeitliches Datenmodell zu implementieren, welches sowohl einzelne Objekte als auch Bearbeitungsereignisse (Events) verfolgen kann. Vorläufige Ergebnisse zeigen eine relativ hohe Qualität der OSM-Gebäudedaten, wobei eine Steigerung der Qualität hinsichtlich Semantik, Geometrie und Positionsgenauigkeit als Beitrag der freiwilligen OSM-Bearbeiter zu beobachten ist.

In this work, we propose a framework to assess the quality of OpenStreetMap (OSM) building footprints data without using any reference data. More specifically, the OSM history data will be examined regarding the development of attributes, geometries and positions of building footprints. In total seven quality indicators are defined for the intrinsic quality assessment. For our case study in the federal state of Baden-Württemberg (BW), Germany, a PostgreSQL database is established based on a spatiotemporal data model which can track both individual objects and editing events on OSM. The preliminary experiments show that the quality of building footprints in BW is relatively high. And the quality in terms of semantics, geometries and positions are getting increasingly high over the time thanks to the considerable contribution of OSM volunteers., In dieser Arbeit stellen wir ein Konzept zur Bewertung von der Qualität von Gebäudegrundrissen aus OpenStreetMap (OSM) ohne Verwendung von Referenzdaten vor. Insbesondere wird der Verlauf der Bearbeitung von Stützpunkten und Attributen der Objekte untersucht. Sieben Indikatoren Bewertung der intrinsischen Datenqualität wurden definiert. Für die vorliegende Studie ist am Beispiel von Baden-Württemberg eine PostgreSQL-Datenbank erstellt worden, um ein räumlich-zeitliches Datenmodell zu implementieren, welches sowohl einzelne Objekte als auch Bearbeitungsereignisse (Events) verfolgen kann. Vorläufige Ergebnisse zeigen eine relativ hohe Qualität der OSM-Gebäudedaten, wobei eine Steigerung der Qualität hinsichtlich Semantik, Geometrie und Positionsgenauigkeit als Beitrag der freiwilligen OSM-Bearbeiter zu beobachten ist.

The study presents a conceptual model to build a high level geo knowledge based system to manage real world processes. A generic event oriented perspective was used to implicitly represent causal relationships among different components of a Spatio-Temporal Geographic Information system. To utilize the collected event information, a Volunteered Location Based Service was developed. In addition a tool was signed to increase the general awareness of the volunteers in spatial analysis and shift their power and influences from a mere data collector to a knowledge producer., Diese Studie stellt ein konzeptionelles Modell vor, um ein geo-wissensbasiertes System zum Managen der realen Weltprozessen zu erstellen. Eine generische ereignisorientierte Perspektive wurde verwendet, um implizit die kausalen Zusammenhänge zwischen den verschiedenen Komponenten eines räumlich-zeitlichen Systems darzustellen. Um die gesammelten Ereignisinformationen zu nutzen, wurde ein “Volunteered Location Based Service” entwickelt. Darüber hinaus wurde ein Tool entwickelt, um das allgemeine Bewusstsein der Freiwilligen in räumlichen Analyse zu erhöhen.

In this work, we propose a framework to assess the quality of OpenStreetMap (OSM) building footprints data without using any reference data. More specifically, the OSM history data will be examined regarding the development of attributes, geometries and positions of building footprints. In total seven quality indicators are defined for the intrinsic quality assessment. For our case study in the federal state of Baden-Württemberg (BW), Germany, a PostgreSQL database is established based on a spatiotemporal data model which can track both individual objects and editing events on OSM. The preliminary experiments show that the quality of building footprints in BW is relatively high. And the quality in terms of semantics, geometries and positions are getting increasingly high over the time thanks to the considerable contribution of OSM volunteers., In dieser Arbeit stellen wir ein Konzept zur Bewertung von der Qualität von Gebäudegrundrissen aus OpenStreetMap (OSM) ohne Verwendung von Referenzdaten vor. Insbesondere wird der Verlauf der Bearbeitung von Stützpunkten und Attributen der Objekte untersucht. Sieben Indikatoren Bewertung der intrinsischen Datenqualität wurden definiert. Für die vorliegende Studie ist am Beispiel von Baden-Württemberg eine PostgreSQL-Datenbank erstellt worden, um ein räumlich-zeitliches Datenmodell zu implementieren, welches sowohl einzelne Objekte als auch Bearbeitungsereignisse (Events) verfolgen kann. Vorläufige Ergebnisse zeigen eine relativ hohe Qualität der OSM-Gebäudedaten, wobei eine Steigerung der Qualität hinsichtlich Semantik, Geometrie und Positionsgenauigkeit als Beitrag der freiwilligen OSM-Bearbeiter zu beobachten ist.

The assessment of the quality of volunteered geographic information (VGI) is cornerstone to understand the fitness for purpose of datasets in many application domains. While most analyses focus on geometric and positional quality, only sporadic attention has been devoted to the interpretation of the data, i.e., the communication process through which consumers try to reconstruct the meaning of information intended by its producers. Interpretability is a notoriously ephemeral, culturally rooted, and context-dependent property of the data that concerns the conceptual quality of the vocabularies, schemas, ontologies, and documentation used to describe and annotate the geographic features of interest. To operationalize conceptual quality in VGI, we propose a multi-faceted framework that includes accuracy, granularity, completeness, consistency, compliance, and richness, proposing proxy measures for each dimension. The application of the framework is illustrated in a case study on a European sample of OpenStreetMap, focused specifically on conceptual compliance.

The assessment of the quality of volunteered geographic information (VGI) is cornerstone to understand the fitness for purpose of datasets in many application domains. While most analyses focus on geometric and positional quality, only sporadic attention has been devoted to the interpretation of the data, i.e., the communication process through which consumers try to reconstruct the meaning of information intended by its producers. Interpretability is a notoriously ephemeral, culturally rooted, and context-dependent property of the data that concerns the conceptual quality of the vocabularies, schemas, ontologies, and documentation used to describe and annotate the geographic features of interest. To operationalize conceptual quality in VGI, we propose a multi-faceted framework that includes accuracy, granularity, completeness, consistency, compliance, and richness, proposing proxy measures for each dimension. The application of the framework is illustrated in a case study on a European sample of OpenStreetMap, focused specifically on conceptual compliance.

The purpose of multimodal route planning in urban transportation system is to provide a user with an optimal route between the source and the target of a trip where the route may utilize different transportation modes or mode combinations. This problem can be abstracted in a general form by decoupling the domain-specific details from the underlying data model and algorithms. This research work is dedicated to a general approach of modeling multimodal network data and performing optimal path queries on it. The multimodal transportation networks are modeled as Plug-and-Play Multimodal Graph Set based on the proposed concept of Switch Point. The path-finding problem is formalized as two types of multimodal shortest path problem. The solutions to both the types of problem are elaborated. Performance evaluations and the implementation of the approach in a prototype system verified the effectiveness and feasibility of the methodology., Diese Arbeit befasst sich mit einer intelligenten multimodalen Navigation, um den Menschen die bestmögliche Mobilität in einer multimodalen Verkehrsinfrastruktur anzubieten. Das Problem kann in allgemeiner Form durch die Entkopplung der Domain-spezifischen Details aus dem zugrunde liegenden Datenmodell und Algorithmen abstrahiert werden. Diese Forschungsarbeit widmet sich dem allgemeinen Ansatz der Modellierung eines multimodalen Datenmodells, den Abfragen von optimalen Routen in diesem Modell. Dafür werden die multmodalen Verkehrsnetze nach Konzept des Switch-Points als Plug-and-Play Multimodalen Graphenset modelliert. Dann wird die Routenplanungsaufgabe in zwei Typen kategorisiert und als multimodales kürzester Pfad Problem formalisiert. Für die beiden Arten von Problem werden Lösungsansatze vorgeschlagen und implementiert. Die Web-basiertes Prototyp-System demonstriert den gesamten Workflow der multimodalen Routenplanungsfunktion.

The purpose of multimodal route planning in urban transportation system is to provide a user with an optimal route between the source and the target of a trip where the route may utilize different transportation modes or mode combinations. This problem can be abstracted in a general form by decoupling the domain-specific details from the underlying data model and algorithms. This research work is dedicated to a general approach of modeling multimodal network data and performing optimal path queries on it. The multimodal transportation networks are modeled as Plug-and-Play Multimodal Graph Set based on the proposed concept of Switch Point. The path-finding problem is formalized as two types of multimodal shortest path problem. The solutions to both the types of problem are elaborated. Performance evaluations and the implementation of the approach in a prototype system verified the effectiveness and feasibility of the methodology., Diese Arbeit befasst sich mit einer intelligenten multimodalen Navigation, um den Menschen die bestmögliche Mobilität in einer multimodalen Verkehrsinfrastruktur anzubieten. Das Problem kann in allgemeiner Form durch die Entkopplung der Domain-spezifischen Details aus dem zugrunde liegenden Datenmodell und Algorithmen abstrahiert werden. Diese Forschungsarbeit widmet sich dem allgemeinen Ansatz der Modellierung eines multimodalen Datenmodells, den Abfragen von optimalen Routen in diesem Modell. Dafür werden die multmodalen Verkehrsnetze nach Konzept des Switch-Points als Plug-and-Play Multimodalen Graphenset modelliert. Dann wird die Routenplanungsaufgabe in zwei Typen kategorisiert und als multimodales kürzester Pfad Problem formalisiert. Für die beiden Arten von Problem werden Lösungsansatze vorgeschlagen und implementiert. Die Web-basiertes Prototyp-System demonstriert den gesamten Workflow der multimodalen Routenplanungsfunktion.