Search Results (23)

Unmanned Aerial Vehicle (UAV) orthophoto generation in complex environments remains challenging because weak textures, reflective surfaces, occlusions, and large scene extents can cause incomplete reconstruction, ghosting, and seam artifacts. Although 3D Gaussian Splatting (3DGS) offers an efficient explicit scene representation, its use in large-scale UAV orthophoto generation is limited by high memory consumption, unstable densification, and insufficient support for mapping-oriented orthographic rendering. This paper proposes a single-GPU 3DGS framework for UAV orthophoto generation by integrating adaptive spatial block partitioning, deterministic structure-sensitive adaptive density control, and core–buffer tiled orthographic rendering with weighted blending. The proposed framework decomposes large scenes into resource-bounded subregions, guides Gaussian densification using fixed multi-view neighborhoods and edge-enhanced dynamic consistency, and generates large-format orthophotos with reduced boundary and seam artifacts. Experiments on MatrixCity-S and multiple UAV photogrammetric datasets show that the method achieves competitive reconstruction quality and improved resource efficiency. On MatrixCity-S, it reaches 29.01 dB PSNR and 0.901 SSIM, while completing training in 1 h 49 min on a single NVIDIA RTX 3090 GPU. Compared with BlockGS, peak VRAM consumption is reduced by more than 38% across datasets. Under geo-aligned comparison conditions, line-measurement comparisons with MetaShape and Pix4DMapper yield RMSE values of 0.099 m and 0.087 m, respectively. These results demonstrate the potential of the proposed framework for memory-efficient 3DGS-based UAV orthophoto generation under constrained hardware resources, while further control-point-based validation is still needed for rigorous surveying-grade applications. Full article

Exploring the relationship between park green spaces and urban functional spaces provides valuable insights into the production of organically integrated urban spaces that combine production, living, and ecological functions. It also offers guidance for urban spatial structure adjustments and supports the development of park-centered cities. Recent studies have demonstrated that park green spaces offer significant ecological and social benefits; however, evaluations have mostly focused on specific indicators of park green spaces, lacking a detailed and comprehensive assessment. Therefore, this study aims to combine multi-source data and various indicators using methods such as spatial profile analysis and geographical detectors to assess the effectiveness of park green spaces in influencing urban clustering. Firstly, it was determined that both park green spaces and urban single and integrated functional spaces in Hangzhou exhibit clustering distribution. Secondly, by measuring the impact of 12 park green spaces on the clustering of urban functional spaces, specific results were obtained. It was found that there are significant differences in the impact effectiveness across different park green spaces. Thirdly, exploring the factors influencing the agglomeration effect of park green spaces on urban functional spaces reveals that transportation, public services and administration, and residential, commercial, and industrial production functions around parks all influence this effect, albeit with diminishing strength in that order. Interaction between any of these functions further enhances the influence, and the introduction of vitality factors helps eliminate potential misjudgments caused by “ghost city” phenomena. Additionally, park characteristics, such as area, service range, and accessibility, all significantly impact the agglomeration effectiveness of urban functional spaces, with the influence further amplified by the interactions between these characteristics. Finally, directions for future research and planning insights are summarized. Full article

This paper analyzes changes in sacrifice by burning and the space inhabited by ghosts in ancient China from philological and linguistic perspectives. During the Shang and Zhou dynasties, rulers believed that they could convey their offerings and reverence to their ancestors in heaven by burning firewood and sacrifices (尞). From the Spring and Autumn Period to the Han dynasty, the ancient Chinese metaphors for naming the underground space inhabited by ghosts experienced a transformation from a natural space (Yellow Spring (黄泉)) to human settlements (li (里), big cities (都)) and then to government institutions for criminal penalty (government (府), prisons (狱)), which symbolized the gradual establishment of a living order in the space inhabited by ghosts based on the human society. When the new living order of the space inhabited by ghosts was established, the ancient Chinese began to reconstruct sacrifice by burning during the Wei and Jin dynasties, and the objects burnt were represented by joss paper. The use of the term “transforming (化)” to refer to sacrifice by burning suggests that people believed that burning with fire was a way to transfer objects from the real world to the world of ghosts. The act of burning joss paper not only embodied the Chinese concept of ancestor worship to “treat the dead as if they were alive” but also gave “fire (火)” rich religious connotations while greatly simplifying the process and cost of sacrificial rituals, thus gradually becoming popular. Full article

The Internet of Things (IoT) constitutes the foundation of a deeply interconnected society in which objects communicate through the Internet. This innovation, coupled with 5G and artificial intelligence (AI), finds application in diverse sectors like smart cities and advanced manufacturing. With increasing IoT adoption comes heightened vulnerabilities, prompting research into identifying IoT malware. While existing models excel at spotting known malicious code, detecting new and modified malware presents challenges. This paper presents a novel six-step framework. It begins with eight malware attack datasets as input, followed by insights from Exploratory Data Analysis (EDA). Feature engineering includes scaling, One-Hot Encoding, target variable analysis, feature importance using MDI and XGBoost, and clustering with K-Means and PCA. Our GhostNet ensemble, combined with the Gated Recurrent Unit Ensembler (GNGRUE), is trained on these datasets and fine-tuned using the Jaya Algorithm (JA) to identify and categorize malware. The tuned GNGRUE-JA is tested on malware datasets. A comprehensive comparison with existing models encompasses performance, evaluation criteria, time complexity, and statistical analysis. Our proposed model demonstrates superior performance through extensive simulations, outperforming existing methods by around 15% across metrics like AUC, accuracy, recall, and hamming loss, with a 10% reduction in time complexity. These results emphasize the significance of our study’s outcomes, particularly in achieving cost-effective solutions for detecting eight malware strains. Full article

Road instance segmentation is vital for autonomous driving, yet the current algorithms struggle in complex city environments, with issues like poor small object segmentation, low-quality mask edge contours, slow processing, and limited model adaptability. This paper introduces an enhanced instance segmentation method based on SOLOv2. It integrates the Bottleneck Transformer (BoT) module into VoVNetV2, replacing the standard convolutions with ghost convolutions. Additionally, it replaces ResNet with an improved VoVNetV2 backbone to enhance the feature extraction and segmentation speed. Furthermore, the algorithm employs Feature Pyramid Grids (FPGs) instead of Feature Pyramid Networks (FPNs) to introduce multi-directional lateral connections for better feature fusion. Lastly, it incorporates a convolutional Block Attention Module (CBAM) into the detection head for refined features by considering the attention weight coefficients in both the channel and spatial dimensions. The experimental results demonstrate the algorithm’s effectiveness, achieving a 27.6% mAP on Cityscapes, a 4.2% improvement over SOLOv2. It also attains a segmentation speed of 8.9 FPS, a 1.7 FPS increase over SOLOv2, confirming its practicality for real-world engineering applications. Full article

The layout of a city is complex, and indoor spaces have thousands of aspects that make them susceptible to fire. If a fire breaks out, it is difficult to quell, so a fire in the city will cause great harm. However, the traditional fire detection algorithm has a low detection efficiency and high detection rate of small targets, and disasters have occurred during detection. Therefore, this paper proposes a fire safety detection algorithm based on CAGSA-YOLO and constructs a fire safety dataset to integrate common fire safety tools into fire detection, which has a preventive detection effect before a fire occurs. In the improved algorithm, the upsampling in the original YOLOv5 is replaced with the CARAFE module. By adjusting its internal Parameter contrast, the algorithm pays more attention to local regional information and obtains stronger feature maps. Secondly, a new scale detection layer is added to detect objects larger than 4 × 4. Furthermore, the sampling Ghost lightweight design replaces C3 with the C3Ghost module without reducing the mAP. Finally, a lighter SA mechanism is introduced to optimize visual information processing resources. Using the fire safety dataset, the precision, recall, and mAP of the improved model increase to 89.7%, 80.1%, and 85.1%, respectively. At the same time, the size of the improved model is reduced by 0.6 M to 13.8 M, and the Param is reduced from 7.1 M to 6.6 M. Full article

The serious mismatch between industrialization and urbanization has led to the emergence of ghost cities. Industry-and-city integration aims to agglomerate industries and the population simultaneously by coordinating the planning and construction, and by mixing different functional areas including industry, office, living, and commercial functions. Based on the population spatial vector database of Jimei District in Xiamen in 2020, this paper empirically analyzes the effects of spatial patterns between industry and city, in terms of residential location and job accessibility, on the housing occupancy rate in new towns and cities. The findings demonstrate that: (1) The attraction of residential location to population varies among three different urban expansion models. The housing occupancy rate of residential areas that meet the concentric circle model is the highest, followed by the sector model, and the multiple nuclei model is the lowest; (2) The jobs–housing relationship has a stable and positive impact on the occupancy rate of commercial housing in the new town, which verifies that job accessibility is the basic demand for families’ residential location choice; (3) There is a significant pattern difference in the influence of job accessibility on the occupancy rate. The occupancy rate of the sector model residential area is highly dependent on job accessibility: the higher the job accessibility, the lower the occupancy rate of the concentric residential area, while job accessibility has a weak impact on the occupancy rate of the multiple nuclei residential area. The conclusions suggest that the spatial planning of new towns should include a clear population absorbing strategy, and the residential location should follow the expansion law of the urban residential functional area, balance the relationship between industrial agglomeration and the job–housing relationship, and allocate life factors in a targeted manner according to the actual impact of job accessibility. Full article

China has undergone rapid urbanization in the past few decades, and it has been accompanied by overdevelopment. Residential vacancies caused by overdevelopment result in a waste of resources and generate greenhouse gases associated with land surface changes. Due to the poor spatial resolution and limited availability of data, previous studies performed analyses at low resolutions at the county scale, thus lacking spatial detail. In addition, they used complicated subjective indicators difficult to apply to cities of various sizes across China. To understand the detailed spatial pattern of residential vacancies in megacities, we designed a more generally applicable approach with multisource high-resolution spatiotemporal data and tested it in Beijing, the capital of China. At first, a statistical regression with features derived from multisource data was used. Then, the predicted values of the regression function were used as standard heat values, and the observed heat value in each unit was divided by the corresponding standard heat value. Next, residential vacancies were estimated by calculating the quantiles of these division results in all analysis units. This approach requires no prior knowledge or complicated indicators and can be easily applied across cities in China, which is beneficial for development planning at the provincial and national levels. Full article

The paper establishes a connection between the becoming-writer of Burroughs, who found his calling and style during the 1950s and his signature characteristic of becoming-animal. This can first be observed in Queer, where Burroughs develops his so-called routine; a short sketch-like text that often involves instances of metamorphosis or transformation. The theoretical background for this short form and the term becoming-animal is taken from Deleuze’s and Guattari’s book on Kafka, who also worked best in short texts and frequently wrote about animals. “The Composite City” may be the central text to understanding Burroughs’ work. It is the text where Burroughs found his style and his identity as a writer. Becoming-animal is a logical consequence that further develops Burroughs’ aesthetic ideal. Over the following decades, he experimented with it in different forms, and toward the end of his career, it became part of an environmental turn. In Ghost of Chance, one can find the same aesthetic ideal that starts Burroughs’ writing in 1953, but the political implications have turned toward saving the lemurs of Madagascar. Full article

This paper has three main objectives. It traces the “closed” urban model of city development, critiques it at length, showing how it has led to an unsustainable dead-end, represented in post-Covid-19 “ghost town” status for many central cities, and proposes a new “open” model of city design. This is avowedly an unsegregated and non-segmented utilisation of now often abandoned city-centre space in “open” forms favouring urban prairie, or more formalised urban parklands, interspersed with so-called “agritecture” in redundant high-rise buildings, shopping malls and parking lots. It favours sustainable theme-park models of family entertainment “experiences” all supported by sustainable hospitality, integrated mixed land uses and sustainable transportation. Consideration is given to likely financial resource issues but the dearth of current commercial investment opportunities from the old carbonised urban model, alongside public policy and consumer support for urban greening, are concluded to form a propitious post-coronavirus context for furthering the vision. Full article

Australian author Michelle de Kretser addresses in her literary work ideas of home and belonging. In Springtime. A Ghost Story (2014) the author gives voice to an ambiguous and variable subject who coexists with her past, present and future, inhabiting a fluid trans-space where love has a principal role. Frances, the main character in Springtime, sees ghosts who unconsciously allow her to voice her insecurities and doubts concerning her life existence. These phantoms contribute to the formation of Frances’ alternative conceptualization of subjectivity. At the same time, de Kretser offers in this dystopic novella a much-needed escape from binary definitions of inclusion/exclusion, offering palimpsests of the spaces that Frances inhabits—Melbourne, Sydney and Paris. This main character is a fluid flâneuse who tries to adjust to her glocality constituted and reconstituted by a discursive imaginary. In this article, I analyze how de Kretser subverts ghost story patterns, destabilizes binary thinking, and decentralizes the human subject offering the reader an alternative haunting love story with an open ending, where cities, ghosts, humans, dogs and nature become active characters who are-in-this-together-but-who-are-not-one-and-the-same. Full article

Artificial intelligence (AI), like many revolutionary technologies in human history, will have a profound impact on societies. From this viewpoint, we analyze the combined effects of AI to raise important questions about the future form and function of cities. Combining knowledge from computer science, urban planning, and economics while reflecting on academic and business perspectives, we propose that the future of cities is far from being a determined one and cities may evolve into ghost towns if the deployment of AI is not carefully controlled. This viewpoint presents a fundamentally different argument, because it expresses a real concern over the future of cities in contrast to the many publications who exclusively assume city populations will increase predicated on the neoliberal urban growth paradigm that has for centuries attracted humans to cities in search of work. Full article

Rapid urbanization in China has been accompanied by spatial inefficiency in patterns of human activity, of which ‘ghost towns’ are the most visible result. In this study, we measure the density and diversity of human activity in the built environment and relate this to various explanatory factors. Using the Pearl River Delta (PRD) as an empirical case, our research demonstrates the distribution of human activity by multi-source data and then explores its dynamics within these areas. This empirical study is comprised of two parts. The first part explores location information regarding human activity in urbanized areas and shows density and diversity. Regression models are applied to explore how density and diversity are affected by urban scale, morphology and by a city’s administrative level. Results indicate that: 1) cities with smaller populations are more likely to be faced with lower density and diversity, but they derive greater marginal benefits from improving land use efficiency; 2) the compactness of the layout of urban land, an index reflecting the plane shapes of the built environment, is highly correlated with density and diversity in built-up areas; and 3) the administrative importance of a city has a significant and positive impact on the density of human activity, but no obvious influence on its diversity. Full article

Short-term property rentals are perhaps one of the most common traits of present day shared economy. Moreover, they are acknowledged as a major driving force behind changes in urban landscapes, ranging from established metropolises to developing townships, as well as a facilitator of geographical mobility. A geolocation ontology is a high level inference tool, typically represented as a labeled graph, for discovering latent patterns from a plethora of unstructured and multimodal data. In this work, a two-step methodological framework is proposed, where the results of various geolocation analyses, important in their own respect, such as ghost hotel discovery, form intermediate building blocks towards an enriched knowledge graph. The outlined methodology is validated upon data crawled from the Airbnb website and more specifically, on keywords extracted from comments made by users of the said platform. A rather solid case-study, based on the aforementioned type of data regarding Athens, Greece, is addressed in detail, studying the different degrees of expansion & prevalence of the phenomenon among the city’s various neighborhoods. Full article

Although many efforts have been made on building shadow detection from aerial images, little research on simultaneous shadows detection on both building roofs and grounds has been presented. Hence, this paper proposes a new method for simultaneous shadow detection on ghost image. In the proposed method, a corner point on shadow boundary is selected and its 3D approximate coordinate is calculated through photogrammetric collinear equation on the basis of assumption of average elevation within the aerial image. The 3D coordinates of the shadow corner point on shadow boundary is used to calculate the solar zenith angle and the solar altitude angle. The shadow areas on the ground, at the moment of aerial photograph shooting are determined by the solar zenith angle and the solar altitude angle with the prior information of the digital building model (DBM). Using the relationship between the shadows of each building and the height difference of buildings, whether there exists a shadow on the building roof is determined, and the shadow area on the building roof on the ghost image is detected on the basis of the DBM. High-resolution aerial images located in the City of Denver, Colorado, USA are used to verify the proposed method. The experimental results demonstrate that the shadows of the 120 buildings in the study area are completely detected, and the success rate is 15% higher than the traditional shadow detection method based on shadow features. Especially, when the shadows occur on the ground and on the buildings roofs, the successful rate of shadow detection can be improved by 9.42% and 33.33% respectively. Full article