In remotely sensed images, high intraclass variance and interclass similarity are ubiquitous due to complex scenes objects with multivariate features, making semantic segmentation a challenging task. Deep convolutional neural networks can solve this problem by modeling the context of features improving their discriminability. However, current learning paradigms model feature affinity in spatial dimension channel separately then fuse them sequential or parallel manner, leading suboptimal...

Since DCNNs (deep convolutional neural networks) have been successfully applied to various academic and industrial fields, semantic segmentation methods, based on DCNNs, are increasingly explored for remote-sensing image interpreting information extracting. It is still highly challenging due the presence of irregular target shapes, similarities inter – intra-class objects in large-scale high-resolution satellite images. A majority existing methods fuse multi-scale features that always fail...

High spatial resolution remote sensing images (HRRSIs) contain intricate details and varied spectral distributions, making their semantic segmentation a challenging task. To address this problem, it is crucial to adequately capture both local global contexts reduce ambiguity. While self-attention modules in vision transformers long-range context, they tend sacrifice details. In article, we propose geometric prior-guided interactive network (GPINet), hybrid that refines features across...

The rapid advancements in remote sensing technology have enabled the widespread availability of fine-resolution images (RSIs), offering rich spatial details and semantics. Despite applicability scalability transformers semantic segmentation RSIs by learning pairwise contextual affinity, they inevitably introduce irrelevant context, hindering accurate inference patch To address this, we propose a novel multi-head attention-attended module (AAM) that refines self-attention mechanism. AAM...

Semantic segmentation plays a pivotal role in the intelligent interpretation of remote sensing images (RSIs). However, conventional methods predominantly focus on learning representations within spatial domain, often resulting suboptimal discriminative capabilities. Given intrinsic spectral characteristics RSIs, it becomes imperative to enhance potential these by integrating context alongside information. In this paper, we introduce spectrum-space collaborative network (SSCNet), which is...

Semantic segmentation of remote sensing images (RSIs) is pivotal for numerous applications in urban planning, agricultural monitoring, and environmental conservation. However, traditional approaches have primarily emphasized learning within the spatial domain, which frequently leads to less than optimal discrimination features. Considering inherent spectral qualities RSIs, it essential bolster these representations by incorporating context conjunction with information improve discriminative...

Semantic segmentation of Remote Sensing Images (RSIs) entails assigning semantic labels to each pixel accurately. RSIs are rich in spatial and spectral data, revealing diverse material object characteristics. Yet, current RSI-focused computer vision models struggle with significant intra-class variation inter-class resemblance due limited data usage. We propose the Frequency Domain Feature-Guided Network (FFGNet) for RSI segmentation, influenced by digital signal processing theories. FFGNet...

Contextual information plays a pivotal role in the semantic segmentation of remote sensing imagery (RSI) due to imbalanced distributions and ubiquitous intra-class variants. The emergence transformer intrigues revolution vision tasks with its impressive scalability establishing long-range dependencies. However, local patterns, such as inherent structures spatial details, are broken tokenization transformer. Therefore, ICTNet is devised confront deficiencies mentioned above. Principally,...

Deep learning has excelled in image classification, but noisy labels large datasets pose a significant challenge, impacting performance and generalization. To tackle this, we propose novel co-training method using cyclic rates. This trains two networks simultaneously, each selecting clean samples based on loss values to optimize the other's parameters, reducing overfitting confirmation bias. The rate allows oscillate between underfitting overfitting, enhancing distinction samples. Our...

Semantic segmentation of high-resolution remote sensing images (HRRSIs) presents unique challenges due to the intricate spatial and spectral characteristics these images. Traditional methods often prioritize information while underutilizing rich context, leading limited feature discrimination capabilities. To address issues, we propose a novel frequency attention-enhanced network (FAENet), which incorporates attention model (FreqA) jointly contexts. FreqA leverages discrete wavelet...

Pan-sharpening is a significant task that aims to generate high spectral- and spatial- resolution remote-sensing image by fusing multi-spectral (MS) panchromatic (PAN) image. The conventional approaches are insufficient protect the fidelity both in spectral spatial domains. Inspired robust capability outstanding performance of convolutional neural networks (CNN) natural super-resolution tasks, CNN-based pan-sharpening methods worthy further exploration. In this paper, novel method proposed...

The locations and users' information can be shared interacted in the IoV (Internet of Vehicles), which provides sufficient data for traffic deployment behavior pattern analysis. However, privacy issues had become more severe since personal or sensitive is inclined to revealed a big environment. In this work, novel differential privacy-based algorithm named DPTD (Differentially Private Trajectory Database) proposed trajectory database releasing. Firstly, 3-dimensional generalized dataset...

Water body extraction is a typical task in the semantic segmentation of remote sensing images (RSIs). Deep convolutional neural networks (DCNNs) outperform traditional methods mining visual features; however, due to inherent mechanism network, spatial details and abstract representations at different levels are difficult capture accurately same time, then results decline become suboptimal, especially on narrow areas boundaries. To address above-mentioned problem, multiscale successive...

Automatic detection of workers wearing safety helmets at the construction site is essential for safe production. Aiming problem low recognition rate caused by factors such as background and light in automatic using traditional machine learning methods, this paper proposes an object framework that combines Online Hard Example Mining (OHEM) multi-part combination. In our framework, we first use multi-scale training increasing anchors strategies to enhance robustness original Faster RCNN...

Semantic segmentation of remote sensing imagery is a fundamental task in intelligent interpretation. Since deep convolutional neural networks (DCNNs) performed considerable insight learning implicit representations from data, numerous works recent years have transferred the DCNN-based model to data analysis. However, wide-range observation areas, complex and diverse objects illumination imaging angle influence pixels easily confused, leading undesirable results. Therefore, semantic network,...

Verticillium wilt (VW) is a common soilborne disease of cotton. It occurs mainly in the seedling and boll-opening stages severely impairs yield quality fiber. Rapid accurate identification evaluation VW severity (VWS) forms basis field cotton control, which has great significance to production. Cotton VWS values are conventionally measured using in-field observations laboratory test diagnoses, require abundant time professional expertise. Remote proximal sensing imagery spectrometry have...

Attention mechanisms have revolutionized the semantic segmentation network in interpreting remotely sensed images (RSIs) due to their amazing ability establishing contextual dependencies.Nevertheless, complex scenes and diverse objects RSIs, a variety of details correlations are not available Euclidean space.Therefore, similarityhybrid attention module (SHAM) is devised attentively learn hyperbolic maps between any two positions, followed by weighted element-wise summation.The hybrid posses...

The accurate extraction of rivers is closely related to agriculture, socio-economic, environment, and ecology. It helps us pre-warn serious natural disasters such as floods, which leads massive losses life property. With the development popularization remote-sensing information technologies, a great number river-extraction methods have been proposed. However, most them are vulnerable noise interference perform inefficient in big data environment. To address these problems, river method...

Transformers have emerged as a transformative tool in various computer vision tasks, excelling at capturing long-range dependencies. Their potential applicability and scalability the interpretation of high-resolution remote sensing images (HRRSIs) thus garnered substantial interest. However, unlike natural images, HRRSIs present intricate scenes characterized by scale variations diverse appearances. These challenges underscore importance enabling networks to effectively assimilate both local...

Land cover classification (LCC) is of paramount importance for assessing environmental changes in remote sensing images (RSIs) as it involves assigning categorical labels to ground objects. The growing availability multi-source RSIs presents an opportunity intelligent LCC through semantic segmentation, offering a comprehensive understanding Nonetheless, the heterogeneous appearances terrains and objects contribute significant intra-class variance inter-class similarity at various scales,...

Extracting water bodies is an important task in remote sensing imagery (RSI) interpretation. Deep convolution neural networks (DCNNs) show great potential feature learning; they are widely used the body interpretation of RSI. However, accuracy DCNNs still unsatisfactory due to differences many hetero-features bodies, such as spectrum, geometry, and spatial size. To address problem mentioned above, this paper proposes a multiscale normalization attention network (MSNANet) which can accurately...

Removing duplicate proposals is a critical process in pedestrian detection, and usually performed via Non-Maximum Suppression (NMS); however, crowded scenes, the detection of occluded pedestrians are hard to distinguish from proposals, making results inaccurate. In order address above-mentioned problem, authors this paper propose Multi-Attribute NMS (MA-NMS) algorithm, which combines density count attributes adaptively adjust suppression, effectively preserving while removing proposals....