The application of convolutional neural networks has been shown to significantly improve the accuracy building extraction from very high-resolution (VHR) remote sensing images. However, there exist so-called semantic gaps among different kinds buildings due large intraclass variance buildings, and most present-day methods are ineffective in extracting various areas that cover scenes, for example, urban villages high-rise because existing strategies same scenes. With improvement resolution...

Benefiting from the developments in deep learning technology, learning-based algorithms employing automatic feature extraction have achieved remarkable performance on change detection (CD) task. However, of existing CD methods is hindered by imbalance between changed and unchanged pixels. To tackle this problem, a progressive foreground-balanced sampling (PFBS) strategy basis not adding information proposed to help model accurately learn features pixels during early training process thereby...

Building, as an integral aspect of human life, is vital in the domains urban management and analysis. To facilitate large-scale planning applications, acquisition complete reliable building data becomes imperative. There are a few publicly available products that provide lot data, such Microsoft Open Street Map. However, East Asia, due to more complex distribution buildings scarcity auxiliary there lack these regions, hindering application Asia. Some studies attempt simulate information...

The rapid advancement of automated artificial intelligence algorithms and remote sensing instruments has benefited change detection (CD) tasks. However, there is still a lot space to study for precise detection, especially the edge integrity internal holes phenomenon features. In order solve these problems, we design Change Guiding Network (CGNet), tackle insufficient expression problem features in conventional U-Net structure adopted previous methods, which causes inaccurate holes. maps...

Foundation models have reshaped the landscape of Remote Sensing (RS) by enhancing various image interpretation tasks. Pretraining is an active research topic, encompassing supervised and self-supervised learning methods to initialize model weights effectively. However, transferring pretrained downstream tasks may encounter task discrepancy due their formulation pretraining as classification or object discrimination In this study, we explore Multi-Task (MTP) paradigm for RS foundation address...

China's Earth Observation(EO) System has undergone significant development since the 1970s, as China dedicated substantial efforts to advancing remote sensing technology. With fifty years of development, successfully narrowed technology gap with foreign countries through collaborative endeavors government and enterprises. At present, constructed a comprehensive EO system that been proven indispensable for driving economic growth facilitating sustainable development. This paper provides an...

Change detection (CD) is a fundamental and important task for monitoring the land surface dynamics in earth observation field. Existing deep learning-based CD methods typically extract bi-temporal image features using weight-sharing Siamese encoder network identify change regions decoder network. These methods, however, still perform far from satisfactorily as we observe that 1) layers focus on irrelevant background 2) models' confidence inconsistent at different stages. The first problem...

Very high-resolution (VHR) earth observation systems provide an ideal data source for man-made structure detection such as building footprint extraction. Manually delineating footprints from the remotely sensed VHR images, however, is laborious and time-intensive; thus, automation needed in extraction process to increase productivity. Recently, many researchers have focused on developing algorithms based encoder-decoder architecture of convolutional neural networks. However, we observe that...

Visual impairments significantly impact individuals' ability to perceive their surroundings, affecting everyday tasks and spatial navigation. This study explores SEEK VR,s a multi-modal virtual reality game designed foster empathy raise awareness about the challenges faced by visually impaired individuals. By integrating visual feedback, 3D audio, haptic provides an immersive experience that helps participants understand physical emotional struggles of impairment. The paper includes review...

Buildings and roads are the two most basic man-made environments that carry interconnect human society. Building road information has important application value in frontier fields of regional coordinated development, disaster prevention, auto-driving, etc. Mapping buildings from very high-resolution (VHR) remote sensing images become a hot research topic. However, existing methods often extract with separate models, ignoring their strong spatial correlation. To fully utilize complementary...

Building footprint information is one of the key factors for sustainable urban planning and environmental monitoring.Mapping building footprints from remote sensing images an important challenging task in earth observation field.Over years, convolutional neural networks have shown outstanding improvements extraction field due to their ability automatically extract hierarchical features make predictions.However, as buildings are various different sizes, scenes, roofing materials, it hard...

In order to mitigate the spread of COVID-19, Wuhan was first city implement strict lockdown policy in 2020. Even though numerous researches have discussed travel restriction between cities and provinces, few studies focus on effect transportation control inside due lack measurement available data Wuhan. Since public transports been shut down beginning lockdown, change traffic density is a good indicator reflect intracity population flow. Therefore, this paper, we collected time-series...

Recently, the flourishing large language models(LLM), especially ChatGPT, have shown exceptional performance in understanding, reasoning, and interaction, attracting users researchers from multiple fields domains. Although LLMs great capacity to perform human-like task accomplishment natural image, their potential handling remote sensing interpretation tasks has not yet been fully explored. Moreover, lack of automation planning hinders accessibility techniques, non-remote experts research...

Sudden natural disasters and man-made pose a threat to human life property safety, real-time semantic segmentation of high-resolution remote sensing images is crucial for disaster damage assessment applications. In recent years, with the wide application high spatial resolution (HSR) change detection methods based on deep learning (DL), acquisition information damaged areas has become more convenient accurate. However, due black box characteristics existing methods, lack interpretability...

In recent years, using deep learning for large area building change detection has proven to be very efficient. However, the current methods pixel-wise still have some limitations, such as a lack of robustness false-positive changes and confusion about boundary dense buildings. To address these problems, novel method called multiscale attention edge-aware Siamese network (MAEANet) is proposed. The principal idea integrate both discriminative edge structure information improve quality...

Foundation models have reshaped the landscape of Remote Sensing (RS) by enhancing various image interpretation tasks. Pretraining is an active research topic, encompassing supervised and self-supervised learning methods to initialize model weights effectively. However, transferring pretrained downstream tasks may encounter task discrepancy due their formulation pretraining as classification or object discrimination In this study, we explore Multi-Task (MTP) paradigm for RS foundation address...

The merging zones of freeways often witness frequent traffic crashes due to their complex organization, phenomena, and dispersed speed distribution. Quantifying crash risk characteristics predicting evolution are crucial for proactive control on freeways. Previous studies freeway mainly focused the relationship between flow factors, with a few considering influence dangerous driving behaviors risk, but they all neglected superposition impact risk. To address this issue, study utilizes Naive...