Urban Remote Sensing with Spatial Big Data: A Review and Renewed Perspective of Urban Studies in Recent Decades

Remote Sensing, Journal Year: 2023, Volume and Issue: 15(5), P. 1307 - 1307

Published: Feb. 26, 2023

During the past decades, multiple remote sensing data sources, including nighttime light images, high spatial resolution multispectral satellite unmanned drone and hyperspectral among many others, have provided fresh opportunities to examine dynamics of urban landscapes. In meantime, rapid development telecommunications mobile technology, alongside emergence online search engines social media platforms with geotagging has fundamentally changed how human activities landscape are recorded depicted. The combination these two types sources results in explosive mind-blowing discoveries contemporary studies, especially for purposes sustainable planning development. Urban scholars now equipped abundant theoretical arguments that often result from limited indirect observations less-than-ideal controlled experiments. For first time, can model, simulate, predict changes using real-time produce most realistic results, providing invaluable information planners governments aim a healthy future. This current study reviews development, status, future trajectory studies facilitated by advancement big analytical technologies. review attempts serve as bridge between growing “big data” modern communities.

Language: Английский

---

Deep learning for urban land use category classification: A review and experimental assessment

Remote Sensing of Environment, Journal Year: 2024, Volume and Issue: 311, P. 114290 - 114290

Published: July 14, 2024

Mapping the distribution, pattern, and composition of urban land use categories plays a valuable role in understanding environmental dynamics facilitating sustainable development. Decades effort mapping have accumulated series approaches products. New trends characterized by open big data advanced artificial intelligence, especially deep learning, offer unprecedented opportunities for patterns from regional to global scales. Combined with large amounts geospatial data, learning has potential promote higher levels scale, accuracy, efficiency, automation. Here, we comprehensively review advances based research practices aspects sources, classification units, approaches. More specifically, delving into different settings on learning-based mapping, design eight experiments Shenzhen, China investigate their impacts performance terms sample, model. For each investigated setting, provide quantitative evaluations discussed inform more convincing comparisons. Based historical retrospection experimental evaluation, identify prevailing limitations challenges suggest prospective directions that could further facilitate exploitation techniques using remote sensing other spatial across various

Language: Английский

---

The spatiotemporal evolution and prediction of vegetation NPP in the Huangshui River Basin of Qilian Mountains

Frontiers in Environmental Science, Journal Year: 2025, Volume and Issue: 12

Published: Jan. 13, 2025

The Qilian Mountains and Huangshui River Basin (HRB) represent significant ecological functional areas carbon reservoirs within China. estimation prediction of vegetation net primary productivity (NPP) in this area is beneficial for the management China’s terrestrial ecosystems. Nevertheless, existing methods NPP at local scale are characterised by considerable uncertainty error, have not accounted influence multi-factor interactions. Accordingly, study initially sought to quantify data HRB from 2000 2019 through implementation an improved Carnegie-Ames-Stanford Approach (CASA) model. Subsequently, it endeavoured elucidate spatiotemporal evolution patterns influencing factors over years. ConvGRU model was employed investigate prospective trajectory HRB. findings revealed a notable upward annual variation between 2019. majority regions demonstrated increase NPP, although few exhibited decline. Furthermore, correlation PRE, TEMP, SR, NDVI exhibits regional disparities. spatial characteristics future also demonstrate overall increasing trend. Additionally, distribution characteristics, with evident trends hot spot contraction or cold expansion. This provides pivotal theoretical support assessment sequestration status analogous regions.

Language: Английский

---

A Scalable Framework for Sensor Data Ingestion and Real-Time Processing in Cloud Manufacturing

Algorithms, Journal Year: 2025, Volume and Issue: 18(1), P. 22 - 22

Published: Jan. 4, 2025

Cloud Manufacturing enables the integration of geographically distributed manufacturing resources through advanced Computing and IoT technologies. This paradigm promotes development scalable adaptable production systems. However, existing frameworks face challenges related to scalability, resource orchestration, data security, particularly in rapidly evolving decentralized settings. study presents a novel nine-layer architecture designed specifically address these issues. Central this framework is use Apache Kafka for robust, high-throughput ingestion, Spark Streaming enhance real-time processing. underpinned by microservice-based that ensures high scalability reduced latency. Experimental validation using sensor from UCI Machine Learning Repository demonstrated substantial improvements processing efficiency throughput compared with conventional frameworks. Key components, such as RabbitMQ, contribute low-latency performance, whereas durability supports application. Additionally, in-memory rapid dynamic analysis, yielding actionable insights. The experimental results highlight potential operational efficiency, utilization, offering resilient solution suited demands modern industrial applications. underscores contribution advancing providing detailed insights into its applicability contemporary ecosystems.

Language: Английский

---

A first attempt to model global hydrology at hyper-resolution

Earth System Dynamics, Journal Year: 2025, Volume and Issue: 16(1), P. 29 - 54

Published: Jan. 7, 2025

Abstract. Global hydrological models are one of the key tools that can help meet needs stakeholders and policy makers when water management strategies policies developed. The primary objective this paper is therefore to establish a first-of-its-kind, truly global hyper-resolution model spans multiple-decade period (1985–2019). To achieve this, two limitations addressed, namely lack high-resolution meteorological data insufficient representation lateral movement snow ice. Thus, novel downscaling procedure better incorporates fine-scale topographic climate drivers incorporated, module capable frozen resembling glaciers, avalanches, wind included. We compare 30 arcsec version PCR-GLOBWB (PCR – Water Balance) previously published 5 arcmin versions by evaluating simulated river discharge, cover, soil moisture, land surface evaporation, total storage against observations. show provides more accurate simulation in particular for smaller catchments. highlight modeling possible with current computational resources results realistic representations cycle. However, our also suggest still incorporate cover heterogeneity relevant processes at sub-kilometer scale provide estimates moisture evaporation fluxes.

Language: Английский

---

Trends in crop yield estimation via data assimilation based on multi-interdisciplinary analysis

Field Crops Research, Journal Year: 2025, Volume and Issue: 322, P. 109745 - 109745

Published: Jan. 10, 2025

Language: Английский

---

FastVSDF: An Efficient Spatiotemporal Data Fusion Method for Seamless Data Cube

IEEE Transactions on Geoscience and Remote Sensing, Journal Year: 2024, Volume and Issue: 62, P. 1 - 22

Published: Jan. 1, 2024

Spatiotemporal data fusion provides an efficacious strategy for addressing gaps within time series datasets. This approach significantly enhances the feasibility of large-scale remote sensing applications by, example, enabling creation seamless Data Cubes (SDC). Nevertheless, strict input requirements and low computational efficiency current methods severely limit practicality SDC production. In this study, we propose efficient spatiotemporal method, Fast Variation-based Fusion (FastVSDF) method. FastVSDF consists 3 steps, i.e., unmixing, distributing global residuals, local residuals. unmixing process, introduces fast abundant variation classification (FAVC) to mitigate sample imbalance expedite unsupervised classification. Then, in-class Gaussian weight function is introduced accelerate distribution residuals by considering introduce information on spectral similarity. Besides, employs Guided Filter combat "block artifacts" efficiently. Results show that demonstrated superior performance over Fit-FC, STARFM, RASDF, FSDAF. More importantly, yields a remarkable improvement in efficiency, reducing predicting 43 573 times. As practical application, generated Sentinel-2 Yangtze River Basin, China. The process single period's Basin dataset was accomplished 20 minutes, with average 3.85 seconds each scene. Comprehensively accuracy, feasibility, universality, demonstrates potential constructing long-term SDC. Our code will be publicly available at https://github.com/ChenXuAxel/FastVSDF.

Language: Английский

---

Promoting forest landscape dynamic prediction with an online collaborative strategy

Journal of Environmental Management, Journal Year: 2024, Volume and Issue: 352, P. 120083 - 120083

Published: Jan. 18, 2024

Language: Английский

---

ERKT-Net: Implementing Efficient and Robust Knowledge Distillation for Remote Sensing Image Classification

EAI Endorsed Transactions on Industrial Networks and Intelligent Systems, Journal Year: 2024, Volume and Issue: 11(3)

Published: July 3, 2024

The classification of Remote Sensing Images (RSIs) poses a significant challenge due to the presence clustered ground objects and noisy backgrounds. While many approaches rely on scaling models enhance accuracy, deployment RSI classifiers often requires substantial computational storage resources, thus necessitating use lightweight algorithms. In this paper, we present an efficient robust knowledge transfer network named ERKT-Net, which is designed provide yet accurate Convolutional Neural Network (CNN) classifier. This method utilizes innovative simple concepts better accommodate inherent nature RSIs, thereby significantly improving efficiency robustness traditional Knowledge Distillation (KD) techniques developed ImageNet-1K. We evaluated ERKT-Net three benchmark datasets found that it demonstrated superior accuracy very compact volume compared 40 other advanced methods published between 2020 2023. On most challenging NWPU45 dataset, outperformed KD-based with maximum Overall Accuracy (OA) value 22.4%. Using same criterion, also surpassed first-ranked multi-model minimum OA 0.7 but presented at least 82% reduction in parameters. Furthermore, ablation experiments indicated our training approach has improved classic DA techniques. Notably, can reduce time expenditure distillation phase by 80%, slight sacrifice accuracy. study confirmed logit-based KD technique be more effective developing classifiers, especially when tailored characteristics RSIs.

Language: Английский

---

Mapping built infrastructure in semi-arid systems using data integration and open-source approaches for image classification

Remote Sensing Applications Society and Environment, Journal Year: 2025, Volume and Issue: unknown, P. 101472 - 101472

Published: Jan. 1, 2025

Language: Английский

---