Geospatial Tasks Research Articles

GeoEntity-type constrained knowledge graph embedding for predicting natural-language spatial relations

Natural-language spatial relations between geographic entities (geoentities) reflect diverse perceptions influenced by factors like location, culture, and linguistic conventions. These relations play a crucial role in supporting geospatial tasks, such as question answering and cognitive reasoning. While prior studies focused on a limited set of human-selected spatial terms and geometric attributes, they often overlooked essential semantic attributes. To overcome this limitation, we developed a Spatial Relation-based Knowledge Graph Embedding framework, SR-KGE, with new KG fusion functions to predict spatial relation terms among distinct geoentities. This method not only considers graph structures and the diversity of natural language expressions in the embedding and learning process, but also incorporates geoentity types as a constraint to capture spatial and semantic relations more accurately. Our experiments on two knowledge graph datasets, one small-scale and one large-scale, have both shown its superior performance in spatial relation inference compared to popular KGE models, including TransE, RotatE, and HAKE. We hope our research will advance the classic study of natural language described spatial relations in a more automated and intelligent way.

Geospatial foundation models for image analysis: evaluating and enhancing NASA-IBM Prithvi’s domain adaptability

Research on geospatial foundation models (GFMs) has become a trending topic in geospatial artificial intelligence (AI) research due to their potential for achieving high generalizability and domain adaptability, reducing model training costs for individual researchers. Unlike large language models, such as ChatGPT, constructing visual foundation models for image analysis, particularly in remote sensing, encountered significant challenges such as formulating diverse vision tasks into a general problem framework. This paper evaluates the recently released NASA-IBM GFM Prithvi for its predictive performance on high-level image analysis tasks across multiple benchmark datasets. Prithvi was selected because it is one of the first open-source GFMs trained on time-series of high-resolution remote sensing imagery. A series of experiments were designed to assess Prithvi’s performance as compared to other pre-trained task-specific AI models in geospatial image analysis. New strategies, including band adaptation, multi-scale feature generation, and fine-tuning techniques, are introduced and integrated into an image analysis pipeline to enhance Prithvi’s domain adaptation capability and improve model performance. In-depth analyses reveal Prithvi’s strengths and weaknesses, offering insights for both improving Prithvi and developing future visual foundation models for geospatial tasks.

On the Opportunities and Challenges of Foundation Models for GeoAI (Vision Paper)

Large pre-trained models, also known as foundation models (FMs), are trained in a task-agnostic manner on large-scale data and can be adapted to a wide range of downstream tasks by fine-tuning, few-shot, or even zero-shot learning. Despite their successes in language and vision tasks, we have not yet seen an attempt to develop foundation models for geospatial artificial intelligence (GeoAI). In this work, we explore the promises and challenges of developing multimodal foundation models for GeoAI. We first investigate the potential of many existing FMs by testing their performances on seven tasks across multiple geospatial domains, including Geospatial Semantics, Health Geography, Urban Geography, and Remote Sensing. Our results indicate that on several geospatial tasks that only involve text modality, such as toponym recognition, location description recognition, and US state-level/county-level dementia time series forecasting, the task-agnostic large learning models (LLMs) can outperform task-specific fully supervised models in a zero-shot or few-shot learning setting. However, on other geospatial tasks, especially tasks that involve multiple data modalities (e.g., POI-based urban function classification, street view image–based urban noise intensity classification, and remote sensing image scene classification), existing FMs still underperform task-specific models. Based on these observations, we propose that one of the major challenges of developing an FM for GeoAI is to address the multimodal nature of geospatial tasks. After discussing the distinct challenges of each geospatial data modality, we suggest the possibility of a multimodal FM that can reason over various types of geospatial data through geospatial alignments. We conclude this article by discussing the unique risks and challenges to developing such a model for GeoAI.

GeoGPT: An assistant for understanding and processing geospatial tasks

Decision-makers in GIS often need to combine multiple spatial algorithms and operations to solve geospatial tasks. While professionals can understand and solve these tasks by using GIS tools sequentially, developing workflows for various tasks can be inefficient, as even slight differences in tasks require corresponding adjustments in the workflow. Recently, large language models (e.g., ChatGPT) presented a strong performance in semantic understanding and reasoning. Especially, AutoGPT can further extend the capabilities of large language models (LLMs) by automatically reasoning and calling externally defined tools. Inspired by these studies, we attempt to increase the efficiency of developing workflows for handling geoprocessing tasks by integrating the semantic understanding ability inherent in LLMs with mature tools within the GIS community. Specifically, we develop a new framework called GeoGPT that can conduct geospatial data collection, processing, and analysis in an autonomous manner. In this framework, a LLM is used to understand the demands of users, and then think, plan, and execute defined GIS tools sequentially to output final effective results. In this process, our framework is user-friendly, accepting natural language instructions as input and adapting to various geospatial tasks, which can serve as an assistant for GIS professionals. A systemic evaluation and several cases, including geospatial data crawling, spatial query, facility siting, and mapping, validate the effectiveness of our framework. Though limited cases are presented in this paper, GeoGPT can be further extended to various tasks by equipping with more GIS tools, and we think the paradigm of “foundational plus professional” implied in GeoGPT provides an effective way to develop next-generation GIS in this era of large foundation models.

Advancing Urban Life: A Systematic Review of Emerging Technologies and Artificial Intelligence in Urban Design and Planning

The advancement of cutting-edge technologies significantly transforms urban lifestyles and is indispensable in sustainable urban design and planning. This systematic review focuses on the critical role of innovative technologies and digitalization, particularly artificial intelligence (AI), in urban planning through geo-design, aiming to enhance urban life. It begins with exploring the importance of AI and digital tools in revolutionizing contemporary urban planning practices. Through the methodology based on the Systematic Reviews and Meta-Analyses (PRISMA) protocol, this review sifts through relevant literature over the past two decades by categorizing artificial intelligence technologies based on their functionalities. These technologies are examined for their utility in urban planning, environmental modeling, and infrastructure development, highlighting how they contribute to creating smarter and more livable cities. For instance, machine learning techniques like supervised learning excel in forecasting urban trends, whereas artificial neural networks and deep learning are superior in pattern recognition and vital for environmental modeling. This analysis, which refers to the comprehensive evaluation conducted in this Systematic Review, encompasses studies based on diverse data inputs and domains of application, revealing a trend toward leveraging AI for predictive analytics, decision-making improvements, and the automation of complex geospatial tasks in urban areas. The paper also addresses the challenges encountered, including data privacy, ethical issues, and the demand for cross-disciplinary knowledge. The concluding remarks emphasize the transformative potential of innovative technologies and digitalization in urban planning, advocating for their role in fostering better urban life. It also identifies future research avenues and development opportunities. In light of our review findings, this study concludes that AI technologies indeed hold transformative promise for the field of geo-design and urban planning. They have proven instrumental in advancing predictive analytics, refining decision-making, and streamlining complex geospatial tasks. The AI’s capacity to process expansive datasets and improve urban planning accuracy has facilitated more sustainable urban development and enhanced the resilience of urban environments.

Integration of generative artificial intelligence and Google Earth Engine for mangrove land cover mapping

Mangrove ecosystems, crucial for coastal sustainability, are threatened by human activities, underscoring the need for accurate mapping for effective conservation. This research explores the novel integration of generative artificial intelligence, specifically Microsoft Copilot, with Google Earth Engine (GEE) for mapping mangrove land cover in Kembung River, Bengkalis Island, Indonesia. The methodology leverages Copilot’s natural language processing capabilities to generate GEE JavaScript code, streamlining the process of Sentinel-2 imagery processing and land cover classification using the Random Forest algorithm. Copilot assists in automating complex coding tasks, reducing development time and potential human errors. However, challenges emerge in hyperparameter tuning within GEE’s computational constraints. The results demonstrate an overall accuracy of 84.4% (Kappa = 0.794) in identifying nine land cover classes, with mangroves covering 46.6% of the study area. This innovative approach enhances mangrove mapping efficiency and accuracy, paving the way for improved monitoring and conservation. The study also highlights the potential of AI in environmental science applications, particularly in conservation. Future research should optimize Copilot’s performance for advanced geospatial tasks, address spectral variability challenges, and explore its applicability across diverse ecosystems. This study contributes to mangrove conservation efforts and demonstrates the potential of AI-assisted coding in environmental science applications.

Geo-Storylines: Integrating Maps into Storyline Visualizations.

Storyline visualizations are a powerful way to compactly visualize how the relationships between people evolve over time. Real-world relationships often also involve space, for example the cities that two political rivals visited together or alone over the years. By default, Storyline visualizations only show implicitly geospatial co-occurrence between people (drawn as lines), by bringing their lines together. Even the few designs that do explicitly show geographic locations only do so in abstract ways (e.g., annotations) and do not communicate geospatial information, such as the direction or extent of their political campains. We introduce Geo-Storylines, a collection of visualisation designs that integrate geospatial context into Storyline visualizations, using different strategies for compositing time and space. Our contribution is twofold. First, we present the results of a sketching workshop with 11 participants, that we used to derive a design space for integrating maps into Storylines. Second, by analyzing the strengths and weaknesses of the potential designs of the design space in terms of legibility and ability to scale to multiple relationships, we extract the three most promising: Time Glyphs, Coordinated Views, and Map Glyphs. We compare these three techniques first in a controlled study with 18 participants, under five different geospatial tasks and two maps of different complexity. We additionally collected informal feedback about their usefulness from domain experts in data journalism. Our results indicate that, as expected, detailed performance depends on the task. Nevertheless, Coordinated Views remain a highly effective and preferred technique across the board.

A Comparison of Monoscopic and Stereoscopic 3D Visualizations: Effect on Spatial Planning in Digital Twins

From the user perspective, 3D geospatial data visualizations are one of the bridges between the physical and the digital world. As such, the potential of 3D geospatial data visualizations is frequently discussed within and beyond the digital twins. The effects on human cognitive processes in complex spatial tasks is rather poorly known. No uniform standards exist for the 3D technologies used in these tasks. Although stereoscopic geovisualizations presented using 3D technologies enhance depth perception, it has been suggested that the visual discomfort experienced when using 3D technology outweighs its benefits and results in lower efficiency and errors. In the present study, sixty participants using 3D technologies were tested in terms of their ability to make informed decisions in selecting the correct position of a virtual transmitter in a digital twin and a digital terrain model, respectively. Participants (n = 60) were randomly assigned into two groups, one using 3D technology engaging stereoscopic shutter glasses and the second working with standard computer screen-based visualizations. The results indicated that the participants who used shutter glasses performed significantly worse in terms of response time (W = 175.0; p < 0.001, r = −0.524). This finding verifies previous conclusions concerning the unsuitability of stereoscopic visualization technology for complex decision-making in geospatial tasks.

Collaborating Underwater Vehicles Conducting Large-Scale Geospatial Tasks

Consider large groups of unmanned underwater vehicles (UUVs) conducting large-scale geospatial tasks, such as information gathering or area sensing. Major costs of long duration missions include expensive underwater positioning systems and propulsion, which consumes energy. Exploiting the ocean currents can increase endurance, but requires accounting for forecast uncertainty, which lies beyond the scope of this article. State-of-the-art techniques, such as Monte Carlo tree search or cross entropy method that coordinate underwater vehicles for path-dependent rewards, do not scale well to such large groups. Furthermore, solving the mentioned tasks requires accounting for overlaps in the areas each vehicle searches, increasing the complexity of the problem. We therefore investigate planning techniques that can evaluate path-dependent rewards, account for the ocean forecast, and efficiently coordinate plans for many agents. Two formulations are investigated, which either search the space of action sequences or the space of feedback policies to find dynamically feasible trajectories. We present what we believe to be the first application of the cross entropy method to create joint plans for large groups of 8-128 UUVs. We also develop a novel iterative greedy method that further refines the best discovered constant action sequences to improve other greedy techniques. The iterative greedy method gathers the most information on average, scales well to deploying large groups of agents, and gathers 3%-8% more reward than other techniques.

Impact Mapping for Geospatial Reasoning and Decision Making.

The reported study evaluated a novel approach to aiding geospatial reasoning and decision making. Impact mapping aims to alleviate the cognitive demands of geospatial tasks in part by externalizing data in the form of an integrated decision surface. This is achieved by aggregating data across multiple sources of information and visualizing their combined utility rather than objective measurements or individual utility. Previous research has shown that geospatial decisions improve when aided in this manner, but it remains unknown if dynamic decision making, often plagued by fatigue and anchoring bias, would benefit similarly. The experiment implemented a systematic manipulation of the presence of a composite impact map and the number of attributes present in a two-stage disaster relief, resource allocation task to investigate when and how impact mapping is beneficial or deleterious to decision makers. The presence of the composite impact map increased the utility of selected sites, increased re-planning decisions, reduced information display views, and reduced workload. Generally, the effect of the composite impact map was greater when participants were asked to evaluate more attributes. Composite impact maps appear to improve repeated geospatial reasoning and minimize anchoring bias because they alleviate the cognitive demands otherwise necessary to interpret and maintain information from multiple attributes. Data visualization techniques, such as impact mapping, can improve repeated geospatial decision making in environments that include high cognitive demand.

A Geospatial Thinking Multiyear Study

In the field of environmental sustainability and landscape management, geospatial thinking is necessary. A good level of geospatial thinking is related to academic success in engineering degrees. It is relevant, therefore, to detect the possible deficiencies that university students may have in tasks related to geospatial thinking. This research presents the results of a 2014‒2019 multiyear study with agricultural engineering students, in which seven geospatial tasks were analyzed. The statistical analysis shows that geospatial tasks related to slope, stream/water flow, visibility, and relief interpretation are the best at predicting the final course mark. The present research provides quantitative data on the efficiency that four technologies have to reinforce geospatial thinking focused on each task. Augmented Reality is an appropriate 3D technology for geospatial tasks related to route search, stream/water flow, and elevation points. SketchUp Make 2017 and Autodesk 123D Make showed their potential to solve tasks related to terrain slope and visibility analysis. Spatial Data Infrastructure has given the best results in geospatial tasks related to the photointerpretation of the relief and with topographic profiles of the terrain. Our findings will help teachers to select the most appropriate geospatial tasks to include in their courses.

Analytical Study of 3D Transformation Parameters Between WGS84 and Adindan Datum Systems in Sudan

Coordinate transformation between various reference datums or systems is an essential tool in geospatial tasks such as surveying, geodesy and photogrammetry. Transformation of coordinates is a mathematical process that converts coordinates of a point in one reference datum into coordinates of the same point in the other reference datum. This issue is still critical for the earth sciences scholars in Sudan because the topographic maps were carried out based on the local datum (Adindan) and the recent surveying activities are performed using Global Positioning System, i.e., based on WGS84 reference datum. To resolve the transformation problem between the WGS 84 and Adindan datum systems, the current study explores two transformation methods, namely Bursa–Wolf and Molodensky–Badekas. The least squares observation equations and the combined case versions of Bursa–Wolf and Molodensky–Badekas are employed to estimate different transformation parameters. These parameters are tested and validated using coordinates of 32 first-order points known in both Adindan and WGS84 datum systems. The results are further verified by a statistical test. Through the residuals estimated, the suitability of the transformation methods for application in the study area is thoroughly discussed. To make survey data usable and accessible to all, it is essential to establish reliable transformation parameters where data collected in one system can easily be transformed into the other. Therefore, if the relationship between the WGS84 and Adindan (local) reference systems is accurately established and verified by higher quality control indices, we can promote mapping and other survey activities in Sudan. Again, Sudan has a large and extensive local network. This network can be easily reduced to the worldwide platform if the transformation parameters between the two systems are precisely estimated.

Integrating Free and Open Source Solutions into Geospatial Science Education

While free and open source software becomes increasingly important in geospatial research and industry, open science perspectives are generally less reflected in universities’ educational programs. We present an example of how free and open source software can be incorporated into geospatial education to promote open and reproducible science. Since 2008 graduate students at North Carolina State University have the opportunity to take a course on geospatial modeling and analysis that is taught with both proprietary and free and open source software. In this course, students perform geospatial tasks simultaneously in the proprietary package ArcGIS and the free and open source package GRASS GIS. By ensuring that students learn to distinguish between geospatial concepts and software specifics, students become more flexible and stronger spatial thinkers when choosing solutions for their independent work in the future. We also discuss ways to continually update and improve our publicly available teaching materials for reuse by teachers, self-learners and other members of the GIS community. Only when free and open source software is fully integrated into geospatial education, we will be able to encourage a culture of openness and, thus, enable greater reproducibility in research and development applications.

Cloud Computing Enabled Web Processing Service for Earth Observation Data Processing

The OpenGIS Web Processing Service (WPS) can process both simple and complex geospatial tasks including Earth Observation tasks. As the requirements of Earth Observation data, algorithms, calculation models, and daily life become increasingly complicated; WPS needs to provide high-performance service-oriented computing capability. This paper proposes a cloud computing enabled WPS framework for Earth Observation data processing. It consists of a client layer and a WPS layer, which further consists of a WPS server layer and a cloud computing layer. The cloud computing environment is based on the open-source software Apache Hadoop. The three layers of the proposed cloud computing enabled WPS are outlined, followed by a workflow that processes a user's task using these three layers. Then technological implementation details are explained. An experiment processing Moderate Resolution Imaging Spectroradiometer (MODIS) data shows that WPS can be enabled in a cloud computing environment.

UAV-MAPPING – A USER REPORT

Abstract. This paper reports on first hand experiences in operating an unmanned airborne system (UAS) for mapping purposes in the environment of a mapping company. Recently, a multitude of activities in UAVs is visible, and there is growing interest in the commercial, industrial, and academic mapping user communities and not only in those. As an introduction, the major components of an UAS are identified. The paper focuses on a 1.1kg UAV which is integrated and gets applied on a day-to-day basis as part of an UAS in standard aerial imaging tasks for more than two years already. We present the unmanned airborne vehicle in some detail as well as the overall system components such as autopilot, ground station, flight mission planning and control, and first level image processing. The paper continues with reporting on experiences gained in setting up constraints such a system needs to fulfill. Further on, operational aspects with emphasis on unattended flight mission mode are presented. Various examples show the applicability of UAS in geospatial tasks, proofing that UAS are capable delivering reliably e.g. orthomosaics, digital surface models and more. Some remarks on achieved accuracies give an idea on obtainable qualities. A discussion about safety features puts some light on important matters when entering unmanned flying activities and rounds up this paper. Conclusions summarize the state of the art of an operational UAS from the point of the view of the author.

Distributed Fusion Process for Remotely Sensed Images Based on GIS Services Chain

In this paper,we propose a framework of GIS service chain for geodata and geoprocessing web services to handle complex geospatial tasks in web environment using XML web service,OGC standard and workflow techniques.An online image fusion web service is developed as an example based on the framework.This web service allows different images can be fused on-line.The prototype system integrated industry standards,including Web services,OGC WCS,OGC WPS,WSDL and BPEL4WS,allows invoke geoprocessing functionality in a platform and language independent manner on the Internet.

Optimizing grid computing configuration and scheduling for geospatial analysis: An example with interpolating DEM

Many geographic analyses are very time-consuming and do not scale well when large datasets are involved. For example, the interpolation of DEMs (digital evaluation model) for large geographic areas could become a problem in practical application, especially for web applications such as terrain visualization, where a fast response is required and computational demands exceed the capacity of a traditional single processing unit conducting serial processing. Therefore, high performance and parallel computing approaches, such as grid computing, were investigated to speed up the geographic analysis algorithms, such as DEM interpolation. The key for grid computing is to configure an optimized grid computing platform for the geospatial analysis and optimally schedule the geospatial tasks within a grid platform. However, there is no research focused on this. Using DEM interoperation as an example, we report our systematic research on configuring and scheduling a high performance grid computing platform to improve the performance of geographic analyses through a systematic study on how the number of cores, processors, grid nodes, different network connections and concurrent request impact the speedup of geospatial analyses. Condor, a grid middleware, is used to schedule the DEM interpolation tasks for different grid configurations. A Kansas raster-based DEM is used for a case study and an inverse distance weighting (IDW) algorithm is used in interpolation experiments.

Geospatial Concept Understanding and Recognition in G6–College Students: A Preliminary Argument for Minimal GIS

As geographic information systems (GIS) are increasingly implemented in K–12 classrooms, the risk becomes one of teaching “buttonlogy” or simply how to point and click to complete certain functions. Through the development of a geospatial concept lexicon and corresponding geospatial task ontology along with simple concept-based tasks completed by students in different grade levels, this research has illuminated grade-related differences in geospatial concept recognition and understanding. In these experiments, simple paper and pencil tasks were given to 6th grade, high school, and undergraduate students to provide insight into different levels of concept understanding, specifically in terms of grade-related abilities to comprehend descriptions of spatial relationships. Results indicate significant differences in geospatial concept recognition, understanding, and use among the grade-based participants tested during the course of the project. These results can be used to inform the development of a “Minimal GIS” in which a pedagogic goal of grade-appropriate concept understanding becomes the driving force behind the GIS, suggesting the structure of an effective support system for spatial thinking.

Why Some Students Have Trouble with Maps and Other Spatial Representations

Geoscience learning requires mastery of various spatially demanding tasks, and the learning-science literature offers research findings that illuminate the mental processes underlying such geospatial tasks. Research on spatial abilities shows that there are large individual differences in performance on spatial tasks, that spatial skills can improve with appropriate training but that the improvement may not transfer to related tasks, and that the form of effective training may vary with the student's spatial ability. Research on use of maps in real-world settings shows that the map-reading task involves three constituent understandings: representational correspondence, configurational correspondence and directional correspondence between a map and the real world. Research on topographic-map use has uncovered consistent, teachable strategies used by successful map users. These include grouping features into configurations rather than focusing on individual features separately, and evaluating multiple hypotheses about one's viewpoint. Research on how people comprehend 2-D representations of 3-D structures aids in diagnosing the nature of students' errors on such tasks. Students making non-penetrative errors, in which they only use information visible on the exterior of the 3-D volume, may need different interventions than students who make penetrative errors, in which they try to envision the unseen portions of the 3-D volume.