检索结果-内蒙古大学图书馆

A computer vision enhanced IoT system for koala monitoring and recognition

INTERNET OF THINGS 2025年 29卷

作者： Trevathan, Jarrod Tan, Wee Lum Xing, Wangzhi Holzner, Daniela Kerlin, Douglas Zhou, Jun Castley, Guy Griffith Univ Inst Integrated & Intelligent Syst Gold Coast Australia Griffith Univ eRes Serv Gold Coast Australia Griffith Univ Ctr Planetary Hlth & Food Secur Gold Coast Australia

The significance of koalas to Australia extends beyond their iconic status, with the industry surrounding them valued at $3.2 billion and providing employment for 30,000 individuals. Ever increasing pressures from urbanisation, land clearing, disease, bushfires and the spread of nonnative vegetation and predators have resulted in the continual deterioration in natural koala populations. Effective monitoring and protection of koalas (and their habitat) is critical in reversing this decline. Of particular concern is how to manage koalas at the urban interface - especially with regard to koalas navigating structures such as roads and motorways. While there exist numerous types of infrastructure and strategies to theoretically allow safe passage of koalas between habitat that is divided by road networks, there is limited data on the known effectiveness of these approaches. Part of the problem lies in the ability to detect, monitor and report on the numbers of koalas in the area and their traversal behaviours. Unfortunately, strategies that rely on human observation are piecemeal, and using koala injury/fatalities statistics on roads is not an ideal approach. This paper presents a computer vision enhanced IoT koala monitoring and recognition system that can be used to detect koalas in their native surroundings non-intrusively. The cameras are deployed in places of interest near fauna road crossings. Motion sensing triggers the cameras to take several seconds of video footage that is relayed to the Cloud. Machine learning algorithms process the video footage to determine whether a koala has been spotted. Experimental results demonstrate that our best model on YOLO8 achieve 97.5 AP, 96.5 AR, 99.2 mAP@50, and 97.1 mAP@50-95 in our dataset which contains both daytime and night-time images. Relevant conservation groups and stakeholders can then use our outcomes to target their koala protection strategies accordingly. The system has now been used in multiple koala conservation

关键词： Koalas Internet of things (IoT) remote monitoring image processing Neural networks

来源：评论

学校读者我要写书评

暂无评论

Deep Learning Meets Object-Based image Analysis: Tasks, challenges, strategies, and perspectives

引用

IEEE GEOSCIENCE AND remote sensing MAGAZINE 2024年

作者： Ma, Lei Yan, Ziyun Li, Mengmeng Liu, Tao Tan, Liqin Wang, Xuan He, Weiqiang Wang, Ruikun He, Guangjun Lu, Heng Blaschke, Thomas Nanjing Univ Jiangsu Prov Key Lab Geog Informat Sci & Technol Key Lab Land Satellite Remote Sensing Applicat Sch Geog & Ocean SciMinist Nat Resources Nanjing 210023 Peoples R China Beijing Normal Univ State Key Lab Remote Sensing Sci Beijing 100875 Peoples R China Nanjing Univ Sch Geog & Ocean Sci Nanjing 210023 Peoples R China Fuzhou Univ Acad Digital China Fujian Key Lab Spatial Data Min & Informat Sharing Minist Educ Fuzhou 350108 Peoples R China Michigan Technol Univ Coll Forest Resources & Environm Sci Houghton MI 49931 USA Beijing Inst Satellite Informat Engn Beijing 100095 Peoples R China Sichuan Univ State Key Lab Hydraul & Mt River Engn Chengdu 610065 Peoples R China Univ Salzburg Dept Geoinformat A-5020 Salzburg Austria

Deep learning has gained significant attention in remote sensing, especially in pixel- or patch-level applications. Despite initial attempts to integrate deep learning into object-based image analysis (OBIA), its full potential remains largely unexplored. In this article, as OBIA usage becomes more widespread, we conducted a comprehensive review and expansion of its task subdomains, with or without the integration of deep learning. Furthermore, we have identified and summarized five prevailing strategies to address the challenge of deep learning's limitations in directly processing unstructured object data within OBIA, and this review also recommends some important future research directions. Our goal with these endeavors is to inspire more exploration in this fascinating yet overlooked area and facilitate the integration of deep learning into OBIA processing workflows.

关键词： image edge detection image segmentation Feature extraction Semantic segmentation remote sensing Reviews Classification algorithms Object recognition Accuracy Sensors

来源：评论

学校读者我要写书评

暂无评论

Improving Transformer-based image Compressed sensing via Filtering and Fusion 6

Improving Transformer-based Image Compressed Sensing via Fil...

引用

6th IEEE International Conference on pattern recognition and Artificial Intelligence, PRAI 2023

作者： Liu, Lingjun Chen, Yishan Zhong, Jiabao Xie, Zhonghua Huizhou University School of Computer Science and Engineering Huizhou China

ISBN: (纸本)9798350325485

Compressed sensing (CS) technology has a wide range of application prospects and research value in many fields, which will have a positive impact on the development of digital signal processing and communication technology. Considering the actual storage of the observation matrix, the observation of images is usually processed in blocks, which leads to the problem of block effect and discontinuous diffusion caused by zero filling operations. To improve the visual quality of the latest Transformer-based CS reconstruction, we propose a joint scheme based on Transformer reconstruction module, the filtering module constructed by the BM3D method and image fusion module. Given a reconstructed image with some singular pixels form the reconstruction module, nonlocal filtering is effective in removing irregular points by using self-similarity between image blocks, but it may lead to over-smoothed images. Then, fusion module is used to extract the advantages of the outputs of both modules. Experimental results show that the singular pixels can be significantly reduced while obtaining slightly higher objective evaluations. © 2023 IEEE.

关键词： Compressed sensing

来源：评论

学校读者我要写书评

暂无评论

DETECTION AND IDENTIFICATION OF SURFACE COVER IN COALBED METHANE ENRICHMENT AREA BASED ON SPECTRAL UNMIXING

DETECTION AND IDENTIFICATION OF SURFACE COVER IN COALBED MET...

引用

IEEE International Geoscience and remote sensing Symposium (IGARSS)

作者： Zhao, Shanshan Qin, Qiming Xian Jiaotong Liverpool Univ Sch AI & Adv Comp Suzhou Peoples R China Peking Univ Sch Earth & Space Sci Inst Remote Sensing & GIS Beijing Peoples R China

ISBN: (数字)9781665427920

ISBN: (纸本)9781665427920

Based on the existing geological data and a large amount of research data in the study area, the article compared and analyzed the difference in hyperspectral reflectance between the potential enrichment area and the reference area. The use of Sentinel-2 multispectral images has been investigated to construct the surface recognition model of coalbed methane enrichment areas from two aspects: vegetation covered area and bare areas. Firstly, semi-automatic endmember extraction is implemented by adopting a more pratical spectral unmixing method. And the vegetation research area and mineral research area are separated accordingly. In the vegetation research area, the rich red edge information of Sentinel-2 is used to construct an anomaly recognition model. While in bare soil areas, advanced information processing technology is used to extract weak mineral alteration anomalies. The results are validated with field data.

关键词： Multi-spectral image Endmember Extraction Spectral Un-mixing Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

TRANSFER LEARNING IN THE CLASSIFICATION OF SATELLITE imageS SHOWING AMAZON RAINFOREST 24

TRANSFER LEARNING IN THE CLASSIFICATION OF SATELLITE IMAGES ...

引用

24th ISPRS Congress on Imaging Today, Foreseeing Tomorrow

作者： Karwowska, K. Mil Univ Technol Fac Civil Engn & Geodesy Dept Imagery Intelligence PL-00908 Warsaw Poland

In recent years, we have been dealing with the dynamic technological progress of the space sector, which allows for the observation of the Earth with better temporal, spatial and spectral resolution. The increasing availability of satellite data has contributed to the development of data processing algorithms Thanks to the use of digital image processing methods and deep neural networks, it is possible to perform automatic image classification, segmentation or detection and recognition of objects on the images. This article presents the methodology that allows to accelerate the classification process of satellite images representing the Amazon rainforest based on the Transfer Learning method. Additionally, the influence of the choice of optimization, i.e. the network weight estimation strategy, on the classification of objects was checked. In order to verify the method, an additional raster image classifier was created on the basis of Lidar data. Research shows that the transfer learning method allows the preparation of an image classifier based on a small database (less than 100 images representing one class). The network training process can be shortened to a few minutes.

关键词： convolutional neural networks scene classification urban land cover classification high-resolution satellite imagery image analysis rainforest

来源：评论

学校读者我要写书评

暂无评论

remote sensing image Instance Segmentation Based on Attention Balanced Feature Pyramid

引用

INTERNATIONAL JOURNAL OF pattern recognition AND ARTIFICIAL INTELLIGENCE 2023年第1期37卷 2254020-2254020页

作者： Nie, Xuan Wang, Hailin Chai, Bosong Duan, Mengyang Northwestern Polytech Univ Sch Software Xian 710072 Shaanxi Peoples R China

In recent years, with the development of remote sensing technology and the enhancement of the value of remote sensing images in military and civil fields, remote sensing image object segmentation has also received more and more attention. This paper mainly studies the application of instance segmentation based on deep convolutional neural network in the remote sensing image. This paper proposes an attention balanced feature pyramid module, which strengthens multi-level features and uses the attention module to suppress the interference features of noise in the complex background. In addiction, Soft-NMS is introduced to improve the performance of the network, and GIoU loss is introduced to improve the effect of object detection. The proposed network improves the average detection and segmentation accuracy (mAP) values from 41.75% and 35.34% to 43.05% and 36.02%, respectively.

关键词： remote sensing object detection instance segmentation

来源：评论

学校读者我要写书评

暂无评论

Detection and Classification of Satellite remote sensing images Using Hybrid Segmentation and Feature Extraction with Effective Algorithms

Detection and Classification of Satellite Remote Sensing Ima...

引用

Distributed Computing and Optimization Techniques (ICDCOT), International Conference on

作者： G. Vinuja N. Bharatha Devi Department of CSE Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Saveetha University Chennai India

The remote sensing image analysis, classification, and pattern recognition processes all depend on image segmentation. In this research, a search-based convolutional neural network (SBCNN) is used to identification method for remote sensing images. Prior to applying the image data to the SKFCM with PeSOA segmentation step, the image data must first undergo pre-processing. When pre-processing satellite images for road networks, noise is removed using an improved median filtering technique. The image is then segmented using the SKFCM with PeSOA Segmentation technique to have inverse shape determination with lowest energy usage. Using an intensity constraint, it is possible to identify the segments of a building and vegetation, a road, and a barren area of land. Following segmentation, MLBP with DWT feature extraction is performed on the road satellite images, and the SBCNN is then used to categorize the images. After associated with obtainable methods, the findings of the suggested technique display excellent precision of 98.6%.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Lightweight Synthetic Aperture Radar image Transmission Facility Detection Method Based on YOLOv5 5

Lightweight Synthetic Aperture Radar Image Transmission Faci...

引用

5th International Conference on Computer Vision, image and Deep Learning, CVIDL 2024

作者： Wen, Gang Zhou, Fangrong Yu, Hui Ma, Yi Wang, Guofang Wang, Yifan Geng, Hao Cao, Jun Kunming650217 China

ISBN: (纸本)9798350373820

Transmission facilities play a crucial role in ensuring electrical safety. Implementing remote sensing monitoring for transmission facilities in challenging terrains offers robust support for daily maintenance inspections and early accident warnings. In view of the small scale of power facility targets relative to surrounding terrain features in high-resolution Synthetic Aperture Radar (SAR) images, this paper proposes a lightweight transmission facility target recognition method based on an improved YOLOv5 network. The paper mainly focuses on network improvement and lightweight processing. In the YOLOv5 network, fused dilated convolutions and Convolutional Block Attention Module (CBAM) attention mechanisms are introduced to adjust the weights of small target receptive fields, enhancing the feature extraction of small target power facilities. The Filter Pruning via Geometric Median (FPGM) pruning is used to adjust the network architecture, remove redundant feature modules, and achieve lightweight models. The Varifocal loss is used as the loss function to balance positive and negative classes, highlight the contribution of positive examples, and improve model accuracy. The results show that after optimization, the accuracy can be improved to a maximum of 9 8. 2%. After pruning, the model size significantly decreases, and the inference time can be reduced to below 3 ms. Compared with current mainstream algorithms, the performance is greatly improved, which can meet the needs of transmission facility target detection in remote sensing SAR images. © 2024 IEEE.

关键词： Synthetic aperture radar

来源：评论

学校读者我要写书评

暂无评论

DEFECTIVE PIXEL DETECTION AND CORRECTION IN PRISMA HYPERSPECTRAL DATA 12

DEFECTIVE PIXEL DETECTION AND CORRECTION IN PRISMA HYPERSPEC...

引用

12th Workshop on Hyperspectral Imaging and Signal processing: Evolution in remote sensing (WHISPERS)

作者： Diani, M. Acito, N. Alibani, M. Corsini, G. Univ Pisa Dipartimento Ingn Informaz Pisa Italy Accademia Navale Livorno Italy

ISBN: (数字)9781665470698

ISBN: (纸本)9781665470698

We present a new strategy to identify bad pixels in hyperspectral pushbroom sensors and to replace the inaccurate radiance values with estimates derived from spectral and spatial analysis. The proposed method is quite effective to correct spaceborne hyperspectral data where the regular calibration of the instrument is more complex than in airborne applications. In this paper we discuss the results obtained on images acquired by the PRISMA hyperspectral instrument operated by the Italian Space Agency (ASI). These preliminary results show the effectiveness of the proposed strategy both in detecting even subtle sources of fixed pattern noise, otherwise undetectable using visual inspection of a single spectral band, and in accurately reconstructing the missing radiance values.

关键词： hyperspectral image bad pixel detection residual nonuniformity PRISMA

来源：评论

学校读者我要写书评

暂无评论

remote sensing target tracking in satellite videos based on a variable-angle-adaptive Siamese network

引用

IET image processing 2021年第9期15卷 1987-1997页

作者： Bi, Fukun Sun, Jiayi Han, Jianhong Wang, Yanping Bian, Mingming North China Univ Technol Dept Sch Informat Sci & Technol Beijing Peoples R China Beijing Inst Spacecraft Syst Engn Qian Xuesen Lab Beijing Peoples R China

remote sensing target tracking in satellite videos plays a key role in various fields. However, due to the complex backgrounds of satellite video sequences and many rotation changes of highly dynamic targets, typical target tracking methods for natural scenes cannot be used directly for such tasks, and their robustness and accuracy are difficult to guarantee. To address these problems, an algorithm is proposed for remote sensing target tracking in satellite videos based on a variable-angle-adaptive Siamese network (VAASN). Specifically, the method is based on the fully convolutional Siamese network (Siamese-FC). First, for the feature extraction stage, to reduce the impact of complex backgrounds, we present a new multifrequency feature representation method and introduce the octave convolution (OctConv) into the AlexNet architecture to adapt to the new feature representation. Then, for the tracking stage, to adapt to changes in target rotation, a variable-angle-adaptive module that uses a fast text detector with a single deep neural network (TextBoxes++) is introduced to extract angle information from the template frame and detection frames and performs angle consistency update operations on the detection frames. Finally, qualitative and quantitative experiments using satellite datasets show that the proposed method can improve tracking accuracy while achieving high efficiency.

关键词： Optical, image and video signal processing image recognition Computer vision and image processing techniques Video signal processing

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：