检索结果-内蒙古大学图书馆

How do human detect targets of remote sensing images with visual attention?

INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION 2024年 132卷

作者： He, Bing Qin, Tong Shi, Bowen Dong, Weihua Beijing Normal Univ State Key Lab Remote Sensing Sci Beijing Key Lab Remote Sensing Environm & Digital Beijing 100875 Peoples R China Beijing Normal Univ Fac Geog Sci Beijing 100875 Peoples R China Univ Ghent Dept Geog Res Grp CartoGIS B-9000 Ghent Belgium

Human visual attention is the basis of target recognition, change detection and classification in remote sensing images. However, the human visual attention of remote sensing images during target detection remains uninvestigated. In this study, we simultaneously collected eye tracking and electroencephalography (EEG) data of 40 experts during target detection in 1000 remote sensing images. We quantified their attention in different phases of target detection (i.e., target search, selection, and identification) from their eye movements and brain activities. The eye-tracking results indicated that hue and lightness were crucial visual features that guided visual search for remote sensing images. The size of the target also affected the allocation of human attention resources. Particularly, participants tended to miss the targets that were smaller than 3.9 % of the whole image area. The fixation event-related potentials (FRPs) in temporal and parietal brain regions further distinguished the attention of targets and distractors, which could be attributed to the process of memory and processing. Our findings offer new visual and neural evidence for human attention in target detection of remote sensing images, which not only contribute to the performance of attention-based algorithms in remote sensing interpretation, but also provide an empirical basis for guiding human-machine intelligent fusion.

关键词： remote sensing image Visual attention EEG Eye tracking Target detection

来源：评论

学校读者我要写书评

暂无评论

remote sensing Urban Green Space Layout and Site Selection Based on Lightweight Expansion Convolutional Method

引用

IEEE ACCESS 2023年 11卷 99889-99900页

作者： Fan, Ding Yu, Siwei Jin, Fengcheng Han, Xinyan Zhang, Guoqiang Univ Sains Malaysia Sch Housing Bldg & Planning Gelugor 11600 Malaysia Leshan Normal Univ Dept Fine Arts & Design Leshan 614000 Peoples R China

With the improvement of remote sensing image resolution, remote sensing image scene classification has become a major difficulty in the research of remote sensing Urban green space spatial layout and site selection. Complex data and network structure affect the processing effect of the traditional Convolutional neural network model, so it is particularly important to design a more efficient Convolutional neural network. This research will first expand the convolution design of the Convolutional neural network to improve the scope of model recognition, then select two methods of structural pruning and separable knowledge distillation for lightweight processing of the model, and finally introduce relevant models for comparative experiments to verify the lightweight effect of the model. The experimental results show that the global average accuracy of the lightweight model based on structural pruning is 95.5%, and the Kappa coefficient value is 0.947;The global classification accuracy of the lightweight model based on knowledge distillation reaches 95.60%, with a Kappa coefficient value of 0.939. It only uses 38.419MB of storage space to recognize a remote sensing image, 4698352 model parameters, and 1397527639 floating-point operations per second. The results show that the expansion convolution and network pruning methods improve the classification performance of the Convolutional neural network, improve the accuracy of the model, and the knowledge distillation method has a better effect in reducing the complexity and making up for the loss of the classification performance of the network model.

关键词： Dilated convolution structural pruning knowledge distillation convolutional neural network remote sensing image classification

来源：评论

学校读者我要写书评

暂无评论

Cloud removal using SAR and optical images via attention mechanism-based GAN

引用

pattern recognition LETTERS 2023年 175卷 8-15页

作者： Zhang, Shuai Li, Xiaodi Zhou, Xingyu Wang, Yuning Hu, Yue Harbin Inst Technol Sch Elect & Informat Engn Harbin 150001 Peoples R China

Clouds often appear in remote sensing images, which seriously affect the application of remote sensing images. Therefore, cloud removal is an important preprocessing process in remote sensing image applications. In this paper, we propose a generative adversarial network-based cloud removal method for optical remote sensing images with the assistance of synthetic aperture radar (SAR) images. Our model is an end-to-end model, which consists of a translation module, an attention module, a generator, and a discriminator. We introduce the attention mechanism to accurately locate the cloud regions. With the obtained attention maps as the prior information, the proposed method can remove the clouds while preserving the cloud-free regions. In addition, we include the structural similarity index (SSIM) and the attention penalty in the loss function to improve the performance of the proposed method. Numerical experiments show that the proposed model provides improved cloud removal performance compared with the state-of-the-art methods.

关键词： remote sensing Cloud removal SAR Generative adversarial network

来源：评论

学校读者我要写书评

暂无评论

Dense small object detection algorithm for unmanned aerial vehicle remote sensing images in complex backgrounds

引用

DIGITAL SIGNAL processing 2025年 158卷

作者： Ji, Jingyu Zhao, Yuefei Li, Aihua Ma, Xiaolin Wang, Changlong Lin, Zhilong Shijiazhuang Campus Army Engn Univ Shijiazhuang Hebei Peoples R China

The high altitude, high-resolution imaging, complex background, and skewed shooting angles of unmanned aerial vehicle (UAV) remote sensing images present significant challenges to the detection and identification of dense and small objects. To overcome the above issues, a novel algorithm for the detection of dense small objects in remote sensing images under complex backgrounds is presented. Firstly, we introduce an innovative attention mechanism, Efficient Multiscale Attention Module (EMA), and apply it to the backbone network with the objective of enhancing the multi-scale feature extraction capability. Secondly, an innovative Object Reinforcement Mechanism (ORM) is proposed. This utilizes a multi-branch atrous convolution technique to broaden the receptive field in the network, thereby enhancing the ability to capture feature information in local regions. Furthermore, a novel Context Region Aware Fusion Mechanism (CRAFM) has been devised to further enhance the accuracy of object recognition. The objective of this mechanism is to construct global contextual relationships, with particular attention paid to the correlation between small objects and the overall scene. Additionally, it is able to differentiate between objects and backgrounds by fusing contextual information. Finally, Distribution Shifting Convolution (DSConv) is also introduced to reduce the computational complexity and the number of parameters of the network. Experimental results show that our proposed algorithm significantly outperforms eight state-of-the-art methods in all metrics on three publicly available remote sensing image datasets. And this can lay the foundation for airborne applications.

关键词： remote sensing images Dense small objects Attention mechanism Context Region Aware Fusion Mechanism Object Reinforcement Mechanism

来源：评论

学校读者我要写书评

暂无评论

Detecting Out-Of-Distribution Earth Observation images with Diffusion Models

Detecting Out-Of-Distribution Earth Observation Images with ...

引用

IEEE/CVF Conference on Computer Vision and pattern recognition (CVPR)

作者： Le Bellier, Georges Audebert, Nicolas Cnam CEDRIC EA4629 F-75141 Paris France Univ Gustave Eiffel ENSG IGN LASTIG F-94160 St Mande France

ISBN: (纸本)9798350365474

Earth Observation imagery can capture rare and unusual events, such as disasters and major landscape changes, whose visual appearance contrasts with the usual observations. Deep models trained on common remote sensing data will output drastically different features for these out-of-distribution samples, compared to those closer to their training dataset. Detecting them could therefore help anticipate changes in the observations, either geographical or environmental. In this work, we show that the reconstruction error of diffusion models can effectively serve as unsupervised out-of-distribution detectors for remote sensing images., using as a plausibility score. Moreover, we introduce ODEED, a novel reconstruction-based scorer using the probability-flow ODE of diffusion models. We validate it experimentally on SpaceNet 8 with various scenarios, such as classical OOD detection with geographical shift and near-OOD setups: pre/post-flood and non-flooded/flooded image recognition. We show that our ODEED scorer significantly outperforms other diffusion-based and discriminative baselines on the more challenging near-OOD scenarios of flood image detection, where OOD images are close to the distribution tail. We aim to pave the way towards better use of generative models for anomaly detection in remote sensing.

关键词： deep learning diffusion models out-of-distribution detection remote sensing

来源：评论

学校读者我要写书评

暂无评论

FarmSeg_VLM: A farmland remote sensing image segmentation method considering vision-language alignment

引用

ISPRS JOURNAL OF PHOTOGRAMMETRY AND remote sensing 2025年 225卷 423-439页

作者： Wu, Haiyang Mu, Weiliang Zhong, Dandan Du, Zhuofei Li, Haifeng Tao, Chao Cent South Univ Sch Geosci & Info Phys Changsha 410083 Peoples R China

Compared to remote sensing image (RSI) segmentation methods that rely solely on label-based learning, visionlanguage models (VLMs) can leverage language guidance to model the inherent relationships between farmland and its surrounding environment. However, unlike natural images, where the objects are typically singular and concentrated, farmland remote sensing images often contain a variety of complex and dispersed ground objects. Additionally, due to the limitations of VLMs in processing long texts, the limited textual descriptions are insufficient for accurately mapping the spatial distribution of ground objects in farmland remote sensing images, resulting in insufficient alignment granularity between the image and text spaces. To address this issue, a FarmSeg_VLM considering vision-language alignment is proposed. First, to enhance the image-text alignment granularity in farmland remote sensing images, an image-text spatial alignment strategy under multi-label background priors (ITSA_MLBP) is designed. This strategy provides category labels for various ground objects along with corresponding textual descriptions. FarmSeg_VLM is trained through backpropagation by combining the respective text prompts and masks for each ground object, allowing it to better fit the spatial correspondence between the visual and linguistic representations of different ground objects, thereby enhancing the image-text spatial alignment granularity in farmland remote sensing images. Second, to further correct the spatial alignment mapping between the language descriptions and the corresponding visual features of ground objects, an imagetext alignment adapter (ITAA) is designed. This adapter combines the image embeddings and masks from the segmentation model as dense visual prompts and aligns them with sparse textual prompts from a multimodal large language model (MLLM) using cross-attention, thereby correcting the text mapping. Finally, this paper compares existing farmland recognition me

关键词： Vision-language models Farmland segmentation Vision-language alignment

来源：评论

学校读者我要写书评

暂无评论

GeoChat : Grounded Large Vision-Language Model for remote sensing

GeoChat : Grounded Large Vision-Language Model for Remote Se...

引用

IEEE/CVF Conference on Computer Vision and pattern recognition (CVPR)

作者： Kuckreja, Kartik Danish, Muhammad Sohail Naseer, Muzammal Das, Abhijit Khan, Salman Khan, Fahad Shahbaz Mohamed bin Zayed Univ AI Abu Dhabi U Arab Emirates Birla Inst Technol & Sci Hyderabad India Australian Natl Univ Canberra ACT Australia Linkoping Univ Linkoping Sweden

ISBN: (纸本)9798350353006

Recent advancements in Large Vision-Language Models (VLMs) have shown great promise in natural image domains, allowing users to hold a dialogue about given visual content. However, such general-domain VLMs perform poorly for remote sensing (RS) scenarios, leading to inaccurate or fabricated information when presented with RS domain-specific queries. Such a behavior emerges due to the unique challenges introduced by RS imagery. For example, to handle high-resolution RS imagery with diverse scale changes across categories and many small objects, region-level reasoning is necessary alongside holistic scene interpretation. Furthermore, the lack of domain-specific multimodal instruction following data as well as strong backbone models for RS make it hard for the models to align their behavior with user queries. To address these limitations, we propose GeoChat - the first versatile remote sensing VLM that offers multitask conversational capabilities with high-resolution RS images. Specifically, GeoChat can not only answer image-level queries but also accepts region inputs to hold region-specific dialogue. Furthermore, it can visually ground objects in its responses by referring to their spatial coordinates. To address the lack of domain-specific datasets, we generate a novel RS multi-modal instruction-following dataset by extending image-text pairs from existing diverse RS datasets. We establish a comprehensive benchmark for RS multitask conversations and compare with a number of baseline methods. GeoChat demonstrates robust zero-shot performance on various RS tasks, e.g., image and region captioning, visual question answering, scene classification, visually grounded conversations and referring detection. Our code is available here.

关键词：

来源：评论

学校读者我要写书评

暂无评论

D-CANet: Diverse class-aware coding and decoding structure network for semantic segmentation of high-resolution remote sensing images

D-CANet: Diverse class-aware coding and decoding structure n...

引用

2024 International Conference on image, Signal processing, and pattern recognition, ISPP 2024

作者： Yuan, Zhengwu Shao, Wen Chen, Qiang Ke, Yingqi College of Computer Science and Technology Chongqing University of Posts and Telecommunication Chongqing China

ISBN: (纸本)9781510680425

The substantial scale variation and intra-class diversity within remote sensing imagery pose significant challenges for semantic segmentation, rendering methods developed for natural images inapplicable. These challenges, we introduce a novel semantic segmentation model named D-CANet, which primarily comprises three modules: the Global Class Center Awareness (GCCA), the Local Class Awareness Module (LCAM), and the Global Class Generation Module (GCG). Specifically, the GCCA module is dedicated to modeling the global representation of class context to mitigate the interference from image backgrounds;the LCAM module generates a local class representation, serving as an intermediary perceptual element that facilitates an implicit linkage between pixels and global class representations, minimizing the variance within classes;following the processing by the LCAM module, the GCG module enhances the global class representation. This encoder-decoder structure equipped with GCCA, LCAM, and GCG modules achieves precise segmentation of objects of varying scales within remote sensing imagery through the interactive perception and fusion of global and local features. Experimental assessments conducted on the Potsdam dataset and the Vaihingen dataset illustrate that D-CANet surpasses the current state-of-the-art semantic segmentation techniques in terms of efficacy. © 2024 SPIE.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Diffusion Models in Low-Level Vision: A Survey

引用

IEEE TRANSACTIONS ON pattern ANALYSIS AND MACHINE INTELLIGENCE 2025年第6期47卷 4630-4651页

作者： He, Chunming Shen, Yuqi Fang, Chengyu Xiao, Fengyang Tang, Longxiang Zhang, Yulun Zuo, Wangmeng Guo, Zhenhua Li, Xiu Tsinghua Univ Tsinghua Shenzhen Int Grad Sch Shenzhen 518055 Peoples R China Duke Univ Dept Biomed Engn Durham NC 27708 USA Shanghai Jiao Tong Univ AI Inst MoE Key Lab Artificial Intelligence Shanghai 200240 Peoples R China Harbin Inst Technol Sch Comp Sci & Technol Harbin 150006 Peoples R China Tianyijiaotong Technol Ltd Suzhou 215131 Peoples R China

Deep generative models have gained considerable attention in low-level vision tasks due to their powerful generative capabilities. Among these, diffusion model-based approaches, which employ a forward diffusion process to degrade an image and a reverse denoising process for image generation, have become particularly prominent for producing high-quality, diverse samples with intricate texture details. Despite their widespread success in low-level vision, there remains a lack of a comprehensive, insightful survey that synthesizes and organizes the advances in diffusion model-based techniques. To address this gap, this paper presents the first comprehensive review focused on denoising diffusion models applied to low-level vision tasks, covering both theoretical and practical contributions. We outline three general diffusion modeling frameworks and explore their connections with other popular deep generative models, establishing a solid theoretical foundation for subsequent analysis. We then categorize diffusion models used in low-level vision tasks from multiple perspectives, considering both the underlying framework and the target application. Beyond natural image processing, we also summarize diffusion models applied to other low-level vision domains, including medical imaging, remote sensing, and video processing. Additionally, we provide an overview of widely used benchmarks and evaluation metrics in low-level vision tasks. Our review includes an extensive evaluation of diffusion model-based techniques across six representative tasks, with both quantitative and qualitative analysis. Finally, we highlight the limitations of current diffusion models and propose four promising directions for future research. This comprehensive review aims to foster a deeper understanding of the role of denoising diffusion models in low-level vision.

关键词： Diffusion models Noise Surveys Training remote sensing Superresolution Reviews Noise reduction Optimization Mathematical models score-based stochastic differential equations low-level vision tasks medical image processing remote sensing data processing video processing

来源：评论

学校读者我要写书评

暂无评论

Adaptive Cross-Spatial sensing Network for Change Detection 7th

Adaptive Cross-Spatial Sensing Network for Change Detection

引用

7th Chinese Conference on pattern recognition and Computer Vision

作者： Jin, Liyuan Zhang, Yanduo Lu, Tao Wang, Jiaming Wuhan Inst Technol Hubei Key Lab Intelligent Robot Wuhan 430205 Peoples R China Hubei Univ Arts & Sci Xiangyang 441053 Peoples R China

ISBN: (纸本)9789819784929;9789819784936

The problem of unbalanced category distribution of foreground and background in remote sensing change detection (CD) tasks has become a key problem in improving the performance of CD models when processing complex remote sensing image data. To address this issue, we develop a novel adaptive cross-spatial sensing network (ACS-Net), aiming to dynamically adjust the degree of attention of the foreground and background to achieve accurate and efficient CD. The proposed ACS-Net combines the global context module (GCM) and the cross-spatial learning module (CSLM) by introducing the adaptive weight adjustment mechanism (AWAM) to adaptively adjust the weight coefficients of the GCM and CSLM, thereby enhancing the ability to extract foreground features. GCM captures the global context information of the image to generate global perceptual features, thereby helping to distinguish the foreground and background. While CSLM focuses on learning the information association between different spatial locations, which can enhance the model's sensitivity to spatially changing details, and then focus on local feature extraction of foreground. The experimental results show that the F1-Score index of ACS-Net on two public datasets (Levir-CD and Lebedev-CD) reaches 91.53% and 96.25% respectively, validating its effectiveness and superiority.

关键词： Change detection Deep learning remote sensing High-resolution images

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：