Building Dataset Research Articles

Building for tomorrow: Analyzing ideal thermal transmittances in the face of climate change in Brazil

The climate will become hotter, and buildings will perform differently as outdoor conditions evolve. If the lowest energy demand is desired, it is crucial to determine the ideal thermophysical properties of the envelope over the buildings' life span. However, the scientific literature is still scarce in providing a compelling answer. Therefore, this study (i) determines ideal thermal transmittance values (U-values) for present-day and future climates, (ii) determines to what extent the thermophysical properties will need to change to remain ideal, (iii) identifies different trends of U-values over time, (iv) establishes a relationship between outdoor air temperatures, cooling and heating setpoints, and ideal U-values, and (v) proposes a set of design strategies according to each trend. The EPSAP generative design method was used to create a large dataset of residential buildings with random geometries and U-values to evaluate their energy demand for heating and cooling in EnergyPlus. The thermal performance of each building was evaluated for 30 locations in Brazil for the current period and two future timeframes (2050 and 2080). The Future Weather Generator tool was used to morph today's typical meteorological weather to match the EC-Earth3 data for the SSP5–8.5 scenario. Although climate change has a similar relative impact, its consequences differ over time in each location. The ideal U-values have different trends in different regions: (a) remaining unchanged in the future, (b) changing from being the highest possible to the lowest of the analyzed range in 2050 or 2080, and (c) being mid-range values in the present and with similar or lower values in the future climate. The impact on the thermal loads of maintaining the present-day ideal U-values also varies significantly in the future timeframes, from being nil to representing an increase reaching 30 % in 2050 (∆ 2.94 MW·h ± 0.06 MW·h) and 57 % in 2080 (∆ 6.05 MW·h ± 0.09 MW·h). Therefore, building design professionals need to use different strategies according to each region and consider how climate evolves during the lifetime of the building.

Comparing multiple machine learning models to investigate the relationship between urban morphology and PM2.5 based on mobile monitoring

The development of compact cities has deteriorated the air quality and threatened public health. To improve air quality, it is necessary to systematically analyze the complex relationship between urban morphology and air pollution to propose spatial optimization strategies. This study selected Nanshan District in Shenzhen, China as a case study, and used the street-level exposure to PM2.5 obtained from mobile monitoring as the dependent variable. Nine urban morphology parameters extracted from 3D building dataset and street view images, as well as other environmental parameters, were used as independent variables. A multiple linear regression model and six machine learning models were constructed for investigating the optimal one. For the study area, the results indicated that the RF and the XGBoost were the best-fit models with the R2 of 0.95 and 0.96 respectively. The frontal area index at 20 m height, the standard deviation of volume obstruction ratio, the building coverage ratio at 20 m height and the standard deviation of building height are the key urban form factors influencing PM2.5. In urban renewal and design practices, it is recommended to maintain the frontal area index within the range of 0.10–0.35, the standard deviation of volume obstruction ratio within the range of 4 × 10−5–8 × 10−5, and the standard deviation of building height below 48 m or above 64 m within the 1000 m buffer zone. These findings provide a generalizable analytical framework for understanding PM2.5 dispersion, and the differentiated applicability of models and parameters is recommended to be considered for different cities.

A large‐scale dataset of buildings and construction sites

AbstractWith the rapid development of deep learning and machine automation technology, as well as workforce aging, increasing labor costs, and other issues, an increasing number of scholars have paid attention to the use of these techniques to solve problems in civil engineering. Although progress has been made in applying deep learning to damage detection, many subfields in civil engineering are still in the initial stage, and a large amount of data has not been used. Moreover, the rapid development of a field cannot be separated from large open‐source datasets and many researchers. Therefore, this study attempts to construct a dataset named the BCS dataset of nearly 212,000 photos of buildings and construction sites using multi‐threaded parallel crawler technology and offline collection. The dataset will be expanded regularly. As a practical demonstration, the StyleGAN3 and StyleGAN2 generative adversarial networks were utilized on the dataset to create faked safety hat images and high‐resolution architectural images. Subsequently, four classic classification models were employed to validate the dataset, achieving a Top‐1 accuracy of up to 0.947. These results underscore the dataset's excellent potential for practical applications.

Visual Brick model authoring tool for building metadata standardization

The Brick ontology is a unified semantic metadata standard for building assets and their relationships, serving as a key enabler for effective interoperability and automation of building systems and analytics. However, creating a Brick model, in other words, standard semantic metadata based on the Brick ontology for a building dataset, can be a complex task. This paper presents two case studies of the creation of Brick models for real-world residential and commercial building datasets, highlighting the challenges during the Brick model creation process. Additionally, the paper introduces VizBrick, an interactive authoring tool for creating semantic building metadata. VizBrick facilitates the creation of Brick models by providing an intuitive visual interface and interactive capabilities, such as keyword search, automatic mapping suggestions, and recommendations. The use of VizBrick is shown to significantly reduce the time and effort required during the Brick model creation process.

Beyond plain toxic: building datasets for detection of flammable topics and inappropriate statements

Beyond plain toxic: building datasets for detection of flammable topics and inappropriate statements

Extraction of building from remote sensing imagery base on multi-attention L-CAFSFM and MFFM

Building extraction from high-resolution remote sensing images is widely used in urban planning, land resource management, and other fields. However, the significant differences between categories in high-resolution images and the impact of imaging, such as atmospheric interference and lighting changes, make it difficult for high-resolution images to identify buildings. Therefore, detecting buildings from high-resolution remote sensing images is still challenging. In order to improve the accuracy of building extraction in high-resolution images, this paper proposes a building extraction method combining a bidirectional feature pyramid, location-channel attention feature serial fusion module (L-CAFSFM), and meticulous feature fusion module (MFFM). Firstly, richer and finer building features are extracted using the ResNeXt101 network and deformable convolution. L-CAFSFM combines feature maps from two adjacent levels and iteratively calculates them from high to low level, and from low to high level, to enhance the model’s feature extraction ability at different scales and levels. Then, MFFM fuses the outputs from the two directions to obtain building features with different orientations and semantics. Finally, a dense conditional random field (Dense CRF) improves the correlation between pixels in the output map. Our method’s precision, F-score, Recall, and IoU(Intersection over Union) on WHU Building datasets are 95.17%、94.83%、94.51% and 90.18%. Experimental results demonstrate that our proposed method has a more accurate effect in extracting building features from high-resolution image.

PS-Insar point cloud densification using Sentinel-1 and TerraSAR-X data

ABSTRACT The persistent scatterer interferometric SAR (PS-InSAR) was presented to overcome the temporal and spatial circumspection in the studied areas. The interferometric phase contains different phase components, with many researchers being particularly interested in the displacement component. In this article, contrary to the conventional use of this method in evaluating the displacement phase, we try to estimate the residual phase component of the topography contribution by changing master image selection criteria and spatial baseline maximization. By calculating the height error of the topography model utilized in PS-InSAR processing, also known as the height phase, it becomes possible to determine the height correction for each permanent scatterer. There is a challenge with generating 3D models of urban buildings due to the absence of acceptable point clouds. To overcome this obstacle, we used shape-completion learning methods. The training process directly transforms from discrete point cloud space to dense space by extracting the general geometric structure of the desired shape while preserving the details. The loss criterion of this network is based on the Chamfer Distance, which measures the distance between the non-dense input and dense output point clouds. We generated 30,000 individual building datasets for network training. We utilized two SAR image datasets consisting of 27 Sentinel-1 and 20 TerraSAR-X images. After evaluating the test dataset of 3D building models and comparing them with aerial LiDAR reference data, we calculated the Root Mean Square Error (RMSE) for the PS height point clouds. For Sentinel-1 VH and VV channels, the RMSE values were 6.369 m and 5.478 m, respectively. As for TerraSAR-X HH channel, the RMSE was 4.317 m.

Semantic enrichment of building functions through geospatial data integration and ontological inference

The comprehensive definition of buildings in urban spatial databases (SDBs) and geographic information systems (GISs) is crucial for city management, considering their essential role in cities. Modern urban geospatial applications such as smart cities need a rich-information infrastructure, relying upon urban SDBs and GISs, powered by multiple data sources. In this respect, many geospatial techniques are available for acquiring geometric data of buildings, but their semantic definitions require extra effort. Today, volunteered geographic information (VGI) platforms are assumed to be alternative geospatial data sources and their integration with official datasets provides a new opportunity for the enrichment of urban geospatial datasets. In this context, geospatial semantic web technologies can contribute to the semantic enrichment process. This study presents a methodology for enriching various building datasets using GIS and geospatial semantic web technologies, enabling an enhanced definition of building functions. The proposed method makes use of an OpenStreetMap (OSM) dataset to obtain missing semantic information about building features. The enrichment process was implemented by transferring OSM point of interest (POI) values to associated OSM building features through the application of semantic web rule language (SWRL) rules. Furthermore, the same process was also applied to two additional official datasets at different levels of detail (TOPO1K and TOPO25K), which were subsequently incorporated into the geospatial ontology. The resulting geospatial ontology offers new opportunities for functional definition and logical rule-making for buildings, with the potential to uncover new classes and insights. In addition, it allows for the definition of terms and explanations in multiple languages, covering most of the OSM values used worldwide for selected tags to define building functions.

Hybrid Task Cascade-Based Building Extraction Method in Remote Sensing Imagery

Instance segmentation has been widely applied in building extraction from remote sensing imagery in recent years, and accurate instance segmentation results are crucial for urban planning, construction and management. However, existing methods for building instance segmentation (BSI) still have room for improvement. To achieve better detection accuracy and superior performance, we introduce a Hybrid Task Cascade (HTC)-based building extraction method, which is more tailored to the characteristics of buildings. As opposed to a cascaded improvement that performs the bounding box and mask branch refinement separately, HTC intertwines them in a joint multilevel process. The experimental results also validate its effectiveness. Our approach achieves better detection accuracy compared to mainstream instance segmentation methods on three different building datasets, yielding outcomes that are more in line with the distinctive characteristics of buildings. Furthermore, we evaluate the effectiveness of each module of the HTC for building extraction and analyze the impact of the detection threshold on the model’s detection accuracy. Finally, we investigate the generalization ability of the proposed model.

Merging databases for CNN image recognition, increasing bias or improving results?

Automated microscopy, image processing, and recognition using artificial intelligence is getting a growing interest from the scientific community, as more and more research centres are actively working on building datasets of images for training convolutional neural networks (CNNs) to identify microscopic objects. However, images acquired between institutes might show differences in light and contrast intensity leading to potential bias in identification when using datasets or CNNs from another institute.One might then question if combining datasets acquired in different conditions is likely to improve the efficiency of the resulting CNN by increasing the number of images and integrating lighting variability into the learning process, or on the contrary introduce bias in the CNN training by adding a recurrent noise, common to all classes, through a substantial light and contrast variability.In order to ease collaboration between laboratories, two datasets of middle Eocene radiolarian images, acquired separately at GNS Science (NZ) and the University of Lille (France), were generated to assess the accuracy of CNNs trained on both datasets individually, and also when combined into a third dataset. The performance of the three resulting CNNs was then evaluated on new images acquired at both institutions.Finally, the new radiolarian dataset generated at GNS allowed to easily detect unknown taxa, which are otherwise abundant in the studied material. Seven new species are described: Ceratospyris metroid n. sp., Ceratospyris okazakii n. sp., Desmospyris biloba n. sp., Botryostrobus lagena n. sp., Buryella apiculata n. sp., Lophocyrtis cortesei n. sp., and Cromyosphaera fulgurans n. sp.

Neural Network Based Controller for Demand Response Management for a College Building

Abstract: Energy management refers to the optimization of the energy system, one of the most intricate and significant technological inventions that we are aware of. While optimising has a lot of experience, the demand side is what drives the production and distribution of energy. Research and business are paying more attention to Demand Side Management (DSM) portfolio includes actions to enhance the energy infrastructure next to the consumer. It varies from enhancing energy efficiency by the use of superior materials over clever incentives in energy prices for specific use habits, increasing advanced real-time management of dispersed energy resources. In this paper, Neural Network based contreiler has been designed to perform the load balancing using a load profile of commercial college building dataset. The results obtained show the regression fir of 99.9% and error of 0.0005%.

SCA-Net: Multiscale Contextual Information Network for Building Extraction Based on High-Resolution Remote Sensing Images

Accurately extracting buildings is essential for urbanization rate statistics, urban planning, resource allocation, etc. The high-resolution remote sensing images contain rich building information, which provides an important data source for building extraction. However, the extreme abundance of building types with large differences in size, as well as the extreme complexity of the background environment, result in the accurate extraction of spatial details of multi-scale buildings, which remains a difficult problem worth studying. To this end, this study selects the representative Xinjiang Tumxuk urban area as the study area. A building extraction network (SCA-Net) with feature highlighting, multi-scale sensing, and multi-level feature fusion is proposed, which includes Selective kernel spatial Feature Extraction (SFE), Contextual Information Aggregation (CIA), and Attentional Feature Fusion (AFF) modules. First, Selective kernel spatial Feature Extraction modules are used for cascading composition, highlighting information representation of features, and improving the feature extraction capability. Adding a Contextual Information Aggregation module enables the acquisition of multi-scale contextual information. The Attentional Feature Fusion module bridges the semantic gap between high-level and low-level features to achieve effective fusion between cross-level features. The classical U-Net, Segnet, Deeplab v3+, and HRNet v2 semantic segmentation models are compared on the self-built Tmsk and WHU building datasets. The experimental results show that the algorithm proposed in this paper can effectively extract multi-scale buildings in complex backgrounds with IoUs of 85.98% and 89.90% on the two datasets, respectively. SCA-Net is a suitable method for building extraction from high-resolution remote sensing images with good usability and generalization.

A Weak Sample Optimisation Method for Building Classification in a Semi-Supervised Deep Learning Framework

Deep learning has gained widespread interest in the task of building semantic segmentation modelling using remote sensing images; however, neural network models require a large number of training samples to achieve better classification performance, and the models are more sensitive to error patches in the training samples. The training samples obtained in semi-supervised classification methods need less reliable weakly labelled samples, but current semi-supervised classification research puts the generated weak samples directly into the model for applications, with less consideration of the impact of the accuracy and quality improvement of the weak samples on the subsequent model classification. Therefore, to address the problem of generating and optimising the quality of weak samples from training data in deep learning, this paper proposes a semi-supervised building classification framework. Firstly, based on the test results of the remote sensing image segmentation model and the unsupervised classification results of LiDAR point cloud data, this paper quickly generates weak image samples of buildings. Secondly, in order to improve the quality of the spots of the weak samples, an iterative optimisation strategy of the weak samples is proposed to compare and analyse the weak samples with the real samples and extract the accurate samples from the weak samples. Finally, the real samples, the weak samples, and the optimised weak samples are input into the semantic segmentation model of buildings for accuracy evaluation and analysis. The effectiveness of this paper’s approach was experimentally verified on two different building datasets, and the optimised weak samples improved by 1.9% and 0.6%, respectively, in the test accuracy mIoU compared to the initial weak samples. The results demonstrate that the semi-supervised classification framework proposed in this paper can be used to alleviate the model’s demand for a large number of real-labelled samples while improving the ability to utilise weak samples, and it can be used as an alternative to fully supervised classification methods in deep learning model applications that require a large number of training samples.

BiGTA-Net: A Hybrid Deep Learning-Based Electrical Energy Forecasting Model for Building Energy Management Systems

The growth of urban areas and the management of energy resources highlight the need for precise short-term load forecasting (STLF) in energy management systems to improve economic gains and reduce peak energy usage. Traditional deep learning models for STLF present challenges in addressing these demands efficiently due to their limitations in modeling complex temporal dependencies and processing large amounts of data. This study presents a groundbreaking hybrid deep learning model, BiGTA-net, which integrates a bi-directional gated recurrent unit (Bi-GRU), a temporal convolutional network (TCN), and an attention mechanism. Designed explicitly for day-ahead 24-point multistep-ahead building electricity consumption forecasting, BiGTA-net undergoes rigorous testing against diverse neural networks and activation functions. Its performance is marked by the lowest mean absolute percentage error (MAPE) of 5.37 and a root mean squared error (RMSE) of 171.3 on an educational building dataset. Furthermore, it exhibits flexibility and competitive accuracy on the Appliances Energy Prediction (AEP) dataset. Compared to traditional deep learning models, BiGTA-net reports a remarkable average improvement of approximately 36.9% in MAPE. This advancement emphasizes the model’s significant contribution to energy management and load forecasting, accentuating the efficacy of the proposed hybrid approach in power system optimizations and smart city energy enhancements.

Exploring Relationships between Boltzmann Entropy of Images and Building Classification Accuracy in Land Cover Mapping.

Remote sensing images are important data sources for land cover mapping. As one of the most important artificial features in remote sensing images, buildings play a critical role in many applications, such as population estimation and urban planning. Classifying buildings quickly and accurately ensures the reliability of the above applications. It is known that the classification accuracy of buildings (usually indicated by a comprehensive index called F1) is greatly affected by image quality. However, how image quality affects building classification accuracy is still unclear. In this study, Boltzmann entropy (an index considering both compositional and configurational information, simply called BE) is employed to describe image quality, and the potential relationships between BE and F1 are explored based on images from two open-source building datasets (i.e., the WHU and Inria datasets) in three cities (i.e., Christchurch, Chicago and Austin). Experimental results show that (1) F1 fluctuates greatly in images where building proportions are small (especially in images with building proportions smaller than 1%) and (2) BE has a negative relationship with F1 (i.e., when BE becomes larger, F1 tends to become smaller). The negative relationships are confirmed using Spearman correlation coefficients (SCCs) and various confidence intervals via bootstrapping (i.e., a nonparametric statistical method). Such discoveries are helpful in deepening our understanding of how image quality affects building classification accuracy.

Double U-Net (W-Net): A change detection network with two heads for remote sensing imagery

Recently, the deep learning algorithms have been increasingly utilized in remote sensing change detection. However, incomplete buildings and the blurred edges caused by the complex scenes in change detection applications make the detection results fail to describe the real land cover changes. Superpixels can be used to alleviate edge blurring, but the existing superpixel methods cannot be trained jointly with the models in change detection. In this work, we investigated an innovative double-head method using deep learning, called double U-Net (W-Net), which consists of a superpixel module and a change detection module. Due to the superpixel module, W-Net can handle building edges very well. In order to solve problem that multiple subtasks fail to achieve the optimal results, a two-branch multi-task coupling framework of change detection and superpixels is designed for W-Net, which enables the model to achieve a globally optimal detection performance. The advancement of the W-Net was demonstrated using three public datasets. The F1score on LEVIR-CD dataset was 0.9031 and kappa coefficient was 0.8969. The F1-score on WHU building dataset was 0.9172 and kappa coefficient was 0.9142. The F1-score on SYSU-CD dataset was 0.8167and and kappa coefficient was 0.7724. The experiments confirmed that the W-Net is capable to detect the edges of changed area better and outperforms the other advanced change detection methods.

EMailMe: A Method to Build Datasets of Corporate Emails in Portuguese

One of the areas in which knowledge management has application is in companies that are concerned with maintaining and disseminating their practices among their members. However, studies involving these two domains may end up suffering from the issue of data confidentiality. Furthermore, it is difficult to find data regarding organizations processes and associated knowledge. Therefore, this paper presents a method to support the generation of a labeled dataset composed of texts that simulate corporate emails containing sensitive information regarding disclosure, written in Portuguese. The method begins with the definition of the dataset’s size and content distribution; the structure of its emails’ texts; and the guidelines for specialists to build the emails’ texts. It aims to create datasets that can be used in the validation of a tacit knowledge extraction process considering the 5W1H approach for the resulting base. The method was applied to create a dataset with content related to several domains, such as Federal Court and Registry Office and Marketing, giving it diversity and realism, while simulating real-world situations in the specialists’ professional life. The dataset generated is available in an open-access repository so that it can be downloaded and, eventually, expanded.

Consumers profiling based federated learning approach for energy load forecasting

Energy load estimation is critical for the smooth functioning of several activities, such as reliable supply, reduced wastage, decision making and generation planning tasks. So far, deep learning and hybrid models have demonstrated great capabilities in capturing non-linear energy load variations patterns. However, these models are data-intensive, requiring a large amount of data to be centrally available for achieving the desired accuracy. Sharing electricity consumption profiles hampers users’ privacy and is computationally very expensive in terms of communication cost. In this regard, the current work proposes a novel federated learning-based approach mitigating the shortcomings of benchmark approaches. The proposed approach combines time-series characteristics patterns and clustering with federated learning to achieve desired privacy, cost benefits and accuracy. To validate these benefits, the performance comparison of the proposed approach is performed with local learning and benchmark prediction approaches. The experimental results on a real-world dataset of residential household buildings demonstrate that the proposed approach performs as good as the local learning approach while ensuring users’ data privacy and reducing communication costs.

OPENSTREETMAP AS AN INPUT SOURCE FOR PRODUCING GOVERNMENTAL DATASETS: THE CASE OF THE ITALIAN MILITARY GEOGRAPHIC INSTITUTE

Abstract. The role of Volunteered Geographic Information (VGI) to integrate, update or complement authoritative datasets released by governments has become increasingly important. This work analyses the contribution of OpenStreetMap (OSM), the most popular VGI project, as one of the input sources that the Military Geographic Institute (IGM), one of the Italian governmental mapping agencies, has used for producing the National Summary Database (DBSN). This database, which was recently released for 12 out of the 20 Italian regions, has a schema organised into a hierarchical structure composed of 10 layers, 30 themes and 93 classes, where each geospatial object carries information on the specific data source it was derived from. For each DBSN layer and theme, we first calculated the fraction of objects derived from OSM in all the Italian regions and related provinces. We found a heterogeneous picture with OSM contribution generally being limited, with the exception of few regions and layers/themes where the DBSN was almost exclusively derived from OSM. An in-depth comparison between the DBSN and OSM building datasets showed that OSM building completeness is varying across Italian regions and provinces, but in all regions there are buildings in OSM that are not included in the DBSN. The work shed light on the opportunities and obstacles for OSM to become a primary input source for the production of governmental datasets.

An Enhanced Algorithm for Active Fire Detection in Croplands Using Landsat-8 OLI Data

Burning biomass exacerbates or directly causes severe air pollution. The traditional active fire detection (AFD) methods are limited by the thresholds of the algorithms and the spatial resolution of remote sensing images, which misclassify some small-scale fires. AFD for burning straw is interfered with by highly reflective buildings around urban and rural areas, resulting in high commission error (CE). To solve these problems, we developed a multicriteria threshold AFD for burning straw (SAFD) based on Landsat-8 imagery in the context of croplands. In solving the problem of the high CE of highly reflective buildings around urban and rural areas, the SAFD algorithm, which was based on the LightGBM machine learning method (SAFD-LightGBM), was proposed to differentiate active fires from highly reflective buildings with a sample dataset of buildings and active fires and an optimal feature combining spectral features and texture features using the ReliefF feature selection method. The results revealed that the SAFD-LightGBM method performed better than the traditional threshold method, with CE and omission error (OE) of 13.2% and 11.5%, respectively. The proposed method could effectively reduce the interference of highly reflective buildings for active fire detection, and it has general applicability and stability for detecting discrete, small-scale fires in urban and rural areas.