Spatial Metadata Research Articles

A framework for automatically generating composite keywords for geo-tagged street images

Due to the ubiquity of sensor-equipped cameras such as smartphones, images are associated with spatial metadata, including camera’s geographical location and viewing orientation, which can be used for automatically generating better semantic keywords about geo-tagged urban street images in addition to visual keywords extracted from image analysis. This study introduces a novel framework for auto-tagging images that integrates both spatial and visual properties to generate comprehensive and accurate tags. The framework operates through four phases: extraction, abstraction, composition, and assessment. Our research highlights the benefits of combining visual and spatial analyses, demonstrated through a case study using geo-tagged urban street images from Orlando, Pittsburgh, and Manhattan. Experimental results show that the proposed framework significantly enhances the accuracy of keyword-based searches compared to conventional methods. In particular, based on our experiments, image search using the tags generated by our proposed framework, referred to as descriptive tags, achieved an average precision improvement factor of 0.9 compared to conventional tags. Additionally, our proposed ranking algorithm, which extends the term frequency-inverse document frequency (TF-IDF) algorithm, resulted in improvement factors of 0.86 for mean average precision (MAP) and 0.57 for mean reciprocal rank (MRR). Moreover, our framework’s flexibility and robustness make it suitable for diverse applications, from smart cities to online shopping. The paper also includes a detailed evaluation and user study, confirming the precision and reliability of the generated tags.

Revisiting the Kahn collection: multimodal artificial intelligence and visual patterns of presence and absence in the Archives de la Planète, 1909–1931

Financed by the French banker Albert Kahn, the Archives of the Planet is one of the most important collections of early photography. Produced between 1909 and 1931, its 72,000 autochromes show a world we are accustomed to seeing in black and white, captured in colour. While Kahn focused on volume, scholars have mostly studied the collection by close reading a small number of pictures. In contrast, this article applies a distant viewing methodology – a combination of data analysis and multimodal AI – to examine, compare, and analyse all the thousands of pictures in the Archives of the Planet. Using spatial metadata, it reveals that the world represented by the collection is significantly smaller than previously assumed: only 22 countries appear more than 100 times in the archive. Applying multimodal AI reveals that a large proportion of the archive (∼25%) consists of the same or extremely similar images. Challenging previous claims about the collection's heterogeneity, algorithmic clustering of the images demonstrates that Kahn's photographers captured similar scenes, adhering closely to the scientific principles of the project. On a general level, the article demonstrates that a distant viewing methodology can help find patterns of presence as well as patterns of absence: the silent parts of the archive.

Mapping urban large‐area advertising structures using drone imagery and deep learning‐based spatial data analysis

AbstractThe problem of visual pollution is a growing concern in urban areas, characterized by intrusive visual elements that can lead to overstimulation and distraction, obstructing views and causing distractions for drivers. Large‐area advertising structures, such as billboards, while being effective advertisement mediums, are significant contributors to visual pollution. Illegally placed or huge billboards can also exacerbate those issues and pose safety hazards. Therefore, there is a pressing need for effective and efficient methods to identify and manage advertising structures in urban areas. This article proposes a deep‐learning‐based system for automatically detecting billboards using consumer‐grade unmanned aerial vehicles. Thanks to the geospatial information from the drone's sensors, the position of billboards can be estimated. Side by side with the system, we share the very first dataset for billboard detection from a drone view. It contains 1361 images supplemented with spatial metadata, together with 5210 annotations.

Enhancing Neurophysiological Analysis and Emotion Recognition with EEG Channel Spatial Metadata through Graph Neural Networks

Enhancing Neurophysiological Analysis and Emotion Recognition with EEG Channel Spatial Metadata through Graph Neural Networks

Leveraging Spatial Metadata in Machine Learning for Improved Objective Quantification of Geological Drill Core

AbstractHere we present a method for using the spatial x–y coordinate of an image cropped from the cylindrical surface of digital 3D drill core images and demonstrate how this spatial metadata can be used to improve unsupervised machine learning performance. This approach is applicable to any data set with known spatial context, however, here it is used to classify 400 m of drillcore imagery into 12 distinct classes reflecting the dominant rock types and alteration features in the core. We modified two unsupervised learning models to incorporate spatial metadata and an average improvement of 25% was achieved over equivalent models that did not utilize metadata. Our semi‐supervised workflow involves unsupervised network training followed by semi‐supervised clustering where a support vector machine uses a subset of M expert labeled images to assign a pseudolabel to the entire data set. Fine‐tuning of the best performing model showed an f1 (macro average) of 90%, and its classifications were used to estimate bulk fresh and altered rock abundance downhole. Validation against the same information gathered manually by experts when the core was recovered during the Oman Drilling Project revealed that our automatically generated data sets have a significant positive correlation (Pearson's r of 0.65–0.72) to the expert generated equivalent, demonstrating that valuable geological information can be generated automatically for 400 m of core with only ∼24 hr of domain expert effort.

Improving deep learning acoustic classifiers with contextual information for wildlife monitoring

Bioacoustics, the exploration of animal vocalizations and natural soundscapes, has emerged as a valuable tool for studying species within their habitats, particularly those that are challenging to observe. This approach has broadened the horizons of biodiversity assessment and ecological research. However, monitoring wildlife with acoustic recorders produces large volumes of data that can be labor-intensive to analyze. Deep learning has recently transformed many computational disciplines by enabling the automated processing of large and complex datasets and has gained attention within the bioacoustics community. Despite the revolutionary impact of deep learning on acoustic detection and classification, attaining both high detection accuracy and low false positive rates in bioacoustics remains a significant challenge. An intriguing yet unexplored avenue for enhancing deep learning in bioacoustics involves the utilization of contextual information, such as time and location, to discern animal vocalizations within acoustic recordings. As a first case study, a multi-branch Convolutional Neural Network (CNN) was developed to classify 22 different bird songs using spectrograms as a first input, and spatial metadata as a secondary input. A comparison was made to a baseline model with only spectrogram input. A geographical prior neural network was trained, separately, to estimate the probability of a species occurring at a given location. The output of this network was combined with the baseline CNN. As a second case study, temporal data and spectrograms were used as input to a multi-branch CNN for the detection of Hainan gibbon (Nomascus hainanus) calls, the world’s rarest primate. Our findings demonstrate that adding metadata to the bird song classifier significantly improves classification performance, with the highest improvement achieved using the geographical prior model (F1-score of 87.78% compared to 61.02% for the baseline model). The multi-branch CNNs also proved efficient (F1-scores of 76.87% and 78.77%) and simpler to use than the geographical prior. In the second case study, our findings revealed a decrease in false positives by 63% (94% of the calls were detected) when the metadata was used by the multi-branch CNN, and an increase of 19% in gibbon detection. This study has uncovered an exciting new avenue for improving classifier performance in bioacoustics. The methodology described in this study can assist ecologists, wildlife management teams, and researchers in reducing the amount of time spent analyzing large acoustic datasets obtained from passive acoustic monitoring studies. Our approach can be adapted and applied to other calling species, and thus tailored to other use cases.

Solar and Wind Quantity 24 h—Series Prediction Using PDE-Modular Models Gradually Developed according to Spatial Pattern Similarity

The design and implementation of efficient photovoltaic (PV) plants and wind farms require a precise analysis and definition of specifics in the region of interest. Reliable Artificial Intelligence (AI) models can recognize long-term spatial and temporal variability, including anomalies in solar and wind patterns, which are necessary to estimate the generation capacity and configuration parameters of PV panels and wind turbines. The proposed 24 h planning of renewable energy (RE) production involves an initial reassessment of the optimal day data records based on the spatial pattern similarity in the latest hours and their follow-up statistical AI learning. Conventional measurements comprise a larger territory to allow the development of robust models representing unsettled meteorological situations and their significant changes from a comprehensive aspect, which becomes essential in middle-term time horizons. Differential learning is a new unconventionally designed neurocomputing strategy that combines differentiated modules composed of selected binomial network nodes as the output sum. This approach, based on solutions of partial differential equations (PDEs) defined in selected nodes, enables us to comprise high uncertainty in nonlinear chaotic patterns, contingent upon RE local potential, without an undesirable reduction in data dimensionality. The form of back-produced modular compounds in PDE models is directly related to the complexity of large-scale data patterns used in training to avoid problem simplification. The preidentified day-sample series are reassessed secondary to the training applicability, one by one, to better characterize pattern progress. Applicable phase or frequency parameters (e.g., azimuth, temperature, radiation, etc.) are related to the amplitudes at each time to determine and solve particular node PDEs in a complex form of the periodic sine/cosine components. The proposed improvements contribute to better performance of the AI modular concept of PDE models, a cable to represent the dynamics of complex systems. The results are compared with the recent deep learning strategy. Both methods show a high approximation ability in radiation ramping events, often in PV power supply; moreover, differential learning provides more stable wind gust predictions without undesirable alterations in day errors, namely in over-break frontal fluctuations. Their day average percentage approximation of similarity correlation on real data is 87.8 and 88.1% in global radiation day-cycles and 46.7 and 36.3% in wind speed 24 h. series. A parametric C++ executable program with complete spatial metadata records for one month is available for free to enable another comparative evaluation of the conducted experiments.

Optimization technique for increasing resolution in computed tomography imaging

Starting from the importance of conforming to biological reality in medicine, in this paper we propose an optimization technique for increasing resolution of computed tomography (CT) images acquired using various existing scanners. Considering a three-dimensional Hounsfield Units (HU) array, together with the corresponding spatial metadata of interest (pixel sizes and slice thickness), the procedure is based on halving each voxel along the directions of the device's Cartesian frame of reference and find those values which are both satisfying the X-Rays attenuation coefficient average requirement and minimizing the HU distance to classical interpolation points. The discussed method was tested by implementing a C# .Net 6, cross-platform library containing two algorithm flavors that could be independently applied: “Z” for doubling the number of slices, and “XY” for doubling the resolution of individual slices. This design allows also chaining (e.g. one could apply the “Z,XY,Z” sequence in order to reduce four times slice thickness). In the context of existing unavoidable limitations, the first results are suggesting the “CT compatible” interpolation technique could provide a reasonable approximation of reality. However, the main advantage comes from satisfying mass conservation, which is of high importance in medical diagnosis and treatment.•The Hounsfield Units scale is defined as a linear transformation of the X-Rays attenuation coefficients. Thus, splitting a computed tomography voxel into two congruent volumes must satisfy the HU average requirement (the initial value must equal the average of the two output HU values).•Existing interpolation methods (linear, spline, etc.) are not compatible with the computed tomography HU average requirement. This could also result in mass estimate anomalies with significant impact in medical diagnosis.•The proposed “CT compatible” interpolation method is based on finding those values which are both satisfying the X-Rays attenuation coefficient average requirement and minimizing the Hounsfield Units distance to classical interpolation points.

The Problem of Reference Rot in Spatial Metadata Catalogues

The content at the end of any hyperlink is subject to two phenomena: the link may break (Link Rot) or the content at the end of the link may no longer be the same as it was when it was created (Content Drift). Reference Rot denotes the combination of both effects. Spatial metadata records rely on hyperlinks for indicating the location of the resources they describe. Therefore, they are also subject to Reference Rot. This paper evaluates the presence of Reference Rot and its impact on the 22,738 distribution URIs of 18,054 metadata records from 26 European INSPIRE spatial data catalogues. Our Link Rot checking method detects broken links while considering the specific requirements of spatial data services. Our Content Drift checking method uses the data format as an indicator. It compares the data formats declared in the metadata with the actual data types returned by the hyperlinks. Findings show that 10.41% of the distribution URIs suffer from Link Rot and at least 6.21% of records suffer from Content Drift (do not declare its distribution types correctly). Additionally, 14.94% of metadata records only contain intermediate HTML web pages as distribution URIs and 31.37% contain at least one HTML web page; thus, they cannot be accessed or checked directly.

TorchIO: A Python library for efficient loading, preprocessing, augmentation and patch-based sampling of medical images in deep learning

Background and objectiveProcessing of medical images such as MRI or CT presents different challenges compared to RGB images typically used in computer vision. These include a lack of labels for large datasets, high computational costs, and the need of metadata to describe the physical properties of voxels. Data augmentation is used to artificially increase the size of the training datasets. Training with image subvolumes or patches decreases the need for computational power. Spatial metadata needs to be carefully taken into account in order to ensure a correct alignment and orientation of volumes. MethodsWe present TorchIO, an open-source Python library to enable efficient loading, preprocessing, augmentation and patch-based sampling of medical images for deep learning. TorchIO follows the style of PyTorch and integrates standard medical image processing libraries to efficiently process images during training of neural networks. TorchIO transforms can be easily composed, reproduced, traced and extended. Most transforms can be inverted, making the library suitable for test-time augmentation and estimation of aleatoric uncertainty in the context of segmentation. We provide multiple generic preprocessing and augmentation operations as well as simulation of MRI-specific artifacts. ResultsSource code, comprehensive tutorials and extensive documentation for TorchIO can be found at http://torchio.rtfd.io/. The package can be installed from the Python Package Index (PyPI) running pip install torchio. It includes a command-line interface which allows users to apply transforms to image files without using Python. Additionally, we provide a graphical user interface within a TorchIO extension in 3D Slicer to visualize the effects of transforms. ConclusionTorchIO was developed to help researchers standardize medical image processing pipelines and allow them to focus on the deep learning experiments. It encourages good open-science practices, as it supports experiment reproducibility and is version-controlled so that the software can be cited precisely. Due to its modularity, the library is compatible with other frameworks for deep learning with medical images.

Synchronising Spatial Metadata Records and Interfaces to Improve the Usability of Metadata Systems

The spatial data infrastructure literature reveals significant gaps in metadata systems concerning their efficiency and effectiveness for end-users. The literature proposes improvements to make the metadata systems more user-friendly. These improvements include new metadata elements and user interfaces that are in concert with each other. In this paper, we implement the proposed improvements in a prototype system and engage with end-users to assess if the proposals help users’ expectations. The prototype is evaluated by conducting think-aloud protocol (TAP) usability testing and semi-structured interviews with end-users. The result demonstrates an increased level of satisfaction about existing systems and some more areas to improve. We conclude that a synchronised development approach for the spatial metadata and the user interface will increase the usability of the metadata for data discovery and selection.

A Proposal for a User-Oriented Spatial Metadata Profile

Spatial metadata profiles have been designed and evolved by data custodians to manage, share, discover, and use spatial data. The end-users of spatial data often do not have much input in designing the profiles. The spatial data infrastructure literature reveals that they question the usability of spatial metadata. This paper analyzes the usability of metadata profiles by engaging end-users and clarifying their requirements in response to this problem. Over 60 users from 18 countries were engaged using an online survey based on a purposive sampling method. The results show that the most widely used metadata standard, ISO 19115, provides metadata elements to accommodate most user requirements for searches. However, an extension to the standard is necessary to assist users in discovery and selection. Two new metadata elements are proposed as part of the extension. The extension also involves changing the obligation type of existing elements to improve data discovery.

Developing Lesson Plans for Teaching Spatial Data Management in Academic Libraries through a Lens of Popular Culture

ABSTRACT Understanding data management in spatial data science is fundamentally important when ensuring a project’s data is appropriately collected, analyzed, stored, and communicated. However, these concepts are often introduced to learners through compliance (i.e., we do this because we are required to or it is the right thing to do), and as a result can mean learning experiences around spatial data management are viewed less than enthusiastically. Academic libraries are places where engaging learning experiences around spatial data management can be explored and implemented through unique pedagogical techniques. For example, activities teaching spatial data management in academic libraries through a popular culture lens can make these topics more approachable and reduce learning barriers. In this paper, I present three lesson plans for academic librarians which approach teaching spatial data management through popular culture. The following concepts are represented: (1) understanding spatial metadata through Dungeons & Dragons; (2) learning to write good documentation through drawing Pokémon; and (3) developing good file naming habits with 1980s album covers. Alongside the lesson plans, I provide a narrative outlining why popular culture pedagogy is a useful technique to spatial data management education to increase the inclusivity and engagement around geospatial learning experiences in academic libraries.

The AirSensor open-source R-package and DataViewer web application for interpreting community data collected by low-cost sensor networks

While large-scale low-cost sensor networks are now recording air pollutant concentrations at finer spatial and temporal scales than previously measured, the large environmental data sets generated by these sensor networks can become overwhelming when considering the scientific skills required to analyze the data and generate interpretable results. This paper summarizes the development of an open-source R package (AirSensor) and interactive web application (DataViewer) designed to address the environmental data science challenges of visualizing and understanding local air quality conditions with community networks of low-cost air quality sensors. AirSensor allows users to access historical data, add spatial metadata, and create maps and plots for viewing community monitoring data. The DataViewer application was developed to incorporate the functionality and plotting functions of the R package into a user-friendly web experience that would serve as the primary source for data communication for community-based organizations and citizen scientists.

Reducing Consumer Uncertainty: Towards an Ontology for Geospatial User-Centric Metadata

With the increased use of geospatial datasets across heterogeneous user groups and domains, assessing fitness-for-use is emerging as an essential task. Users are presented with an increasing choice of data from various portals, repositories, and clearinghouses. Consequently, comparing the quality and evaluating fitness-for-use of different datasets presents major challenges for spatial data users. While standardization efforts have significantly improved metadata interoperability, the increasing choice of metadata standards and their focus on data production rather than potential data use and application, renders typical metadata documents insufficient for effectively communicating fitness-for-use. Thus, research has focused on the challenge of communicating fitness-for-use of geospatial data, proposing a more “user-centric” approach to geospatial metadata. We present the Geospatial User-Centric Metadata ontology (GUCM) for communicating fitness-for-use of spatial datasets to users in the spatial and other domains, to enable them to make informed data source selection decisions. GUCM enables metadata description for various components of a dataset in the context of different application domains. It captures producer-supplied and user-described metadata in structured format using concepts from domain-independent ontologies. This facilitates interoperability between spatial and nonspatial metadata on open data platforms and provides the means for searching/discovering spatial data based on user-specified quality and fitness-for-use criteria.

Spatial Metadata Usability Evaluation

Spatial metadata is a critical part of any spatial data infrastructure, which enables the organising, sharing, discovery and use of spatial data. This paper highlights a knowledge gap in the usability of the metadata systems for the end–users. It then addresses the gap by applying the User Centred Design approach to investigate the usability of metadata records. The research engages with end–users concerning efficiency and effectiveness of metadata systems, and end–users’ satisfaction and expectations. The results indicate significant gaps with the effectiveness and efficiency of metadata systems for spatial data discovery and selection. Inconsistency and irrelevant information in the metadata records were found in the title, keywords, abstracts, data quality and other elements of the metadata. Additionally, essential improvements were identified for user interfaces. Discouraging presentation of the metadata is a prominent problem found in the interface of the metadata systems.

Sistem Informasi Pelaporan dan Pencatatan Aktifitas Kegiatan Hutan di Taman Nasional Gunung Ciremai Berbasis Image Geotagging

Current advances in information technology, especially cellular phones, have become tools that can facilitate the dissemination and delivery of information from one party to another. The concept of geotagging is very similar to marking points on GPS devices, geotagging not only adds waypoints but also adds spatial metadata information to digital media or photos. The process of reporting and recording activities is still using manual processes, both recorded and unrecorded and there is no system for save the reporting process. The system design in this study is an information system that can be accessed by the public in the form of reporting geotagging images on an android application (.apk) and integrated management by Balai TNGC. Achievements or research targets to be achieved are the ease of delivery of information from the community or TNGC officers in terms of reporting and recording of forest activities and the wishes of the authors are expected to be a pilot or reference that can be used by all National Parks in Indonesia. The system development in this study uses the Extreme Programing (XP) method, where this method is suitable for the existing problems. For modeling design using UML (Unified Modeling Language. Keywords : TNGC; Geotagging; Android; Extreme Programing; UML

HBIM (heritage Building Information Modell) of the Wang Stave Church in Karpacz – Case Study

ABSTRACT In recent years, Building Information Model (BIM) has become a leading topic for construction-related fields. This technology may be viewed as a process, a type of database, a software or even a 3D model, but in fact, it integrates all these roles and many more. Heritage (Historic) Building Information Model (HBIM) is a standard for heritage objects. It not only allows for the storing of spatial information and metadata but also provides the means of documenting changes that such structures undergo. The scope of application varies from simple documentation repository, through conservation planning tools, to construction and renovation simulation instruments. In this paper, we focus on the example of the so-called Wang Temple in Karpacz, Poland. This object was chosen due to its unique history. This project was done in cooperation with BIMPoint Company (Kraków). It was aimed at showing how effective HBIM can be in accurate spatial documentation of small-scale heritage sites.

Exploring the sensitivity of coastal inundation modelling to DEM vertical error

ABSTRACTAs sea level is projected to rise throughout the twenty-first century due to climate change, there is a need to ensure that sea level rise (SLR) models accurately and defensibly represent future flood inundation levels to allow for effective coastal zone management. Digital elevation models (DEMs) are integral to SLR modelling, but are subject to error, including in their vertical resolution. Error in DEMs leads to uncertainty in the output of SLR inundation models, which if not considered, may result in poor coastal management decisions. However, DEM error is not usually described in detail by DEM suppliers; commonly only the RMSE is reported. This research explores the impact of stated vertical error in delineating zones of inundation in two locations along the Devon, United Kingdom, coastline (Exe and Otter Estuaries). We explore the consequences of needing to make assumptions about the distribution of error in the absence of detailed error data using a 1 m, publically available composite DEM with a maximum RMSE of 0.15 m, typical of recent LiDAR-derived DEMs. We compare uncertainty using two methods (i) the NOAA inundation uncertainty mapping method which assumes a normal distribution of error and (ii) a hydrologically correct bathtub method where the DEM is uniformly perturbed between the upper and lower bounds of a 95% linear error in 500 Monte Carlo Simulations (HBM+MCS). The NOAA method produced a broader zone of uncertainty (an increase of 134.9% on the HBM+MCS method), which is particularly evident in the flatter topography of the upper estuaries. The HBM+MCS method generates a narrower band of uncertainty for these flatter areas, but very similar extents where shorelines are steeper. The differences in inundation extents produced by the methods relate to a number of underpinning assumptions, and particularly, how the stated RMSE is interpreted and used to represent error in a practical sense. Unlike the NOAA method, the HBM+MCS model is computationally intensive, depending on the areas under consideration and the number of iterations. We therefore used the HBM+ MCS method to derive a regression relationship between elevation and inundation probability for the Exe Estuary. We then apply this to the adjacent Otter Estuary and show that it can defensibly reproduce zones of inundation uncertainty, avoiding the computationally intensive step of the HBM+MCS. The equation-derived zone of uncertainty was 112.1% larger than the HBM+MCS method, compared to the NOAA method which produced an uncertain area 423.9% larger. Each approach has advantages and disadvantages and requires value judgements to be made. Their use underscores the need for transparency in assumptions and communications of outputs. We urge DEM publishers to move beyond provision of a generalised RMSE and provide more detailed estimates of spatial error and complete metadata, including locations of ground control points and associated land cover.

A new non-linear framework for localization of acoustic sources

Source localization is the process of assigning spatial metadata to a set of unlabeled time of arrival for acoustic signals that may be interlaced with extreme outliers caused by detection and material-induced errors. Exploiting the fact that location error is attributed to high residuals, a new framework for non-linear source localization is proposed in this article. The detected time of arrival was employed to calculate the intermediate source and event time. The framework then quarantines sensors with high residual based on the intermediate event time and relative distance from the intermediate source. Then, the time of arrivals for quarantined sensors were reviewed and adjusted. Finally, with new arrival time, the framework iteratively optimizes the spatial location and event time, along with the updating of velocity field through a feedback loop. The new framework was used to localize acoustic emission sources with extreme outliers in an anisotropic medium. The results indicate that the proposed framework can accurately quarantine the bad sensor and localize sources with an accuracy better than existing source localization methods. Finally, the framework was employed to study the progressive damage in a steel fiber–reinforced beam under four-point bending. The source localization result is consistent with that from visual inspection of the member at failure.