检索结果-内蒙古大学图书馆

A novel driving lane change intent prediction model based on image data mining approach and transformer

ICT Express 2025年第3期11卷 467-472页

作者： He, Junbo Guan, Wei Gou, Xuanyuan Zhang, Zhiqing Guangxi Science and Technology Development Institute Nanning 530001 China School of Aeronautical Engineering Civil Aviation University of China Tianjin 300300 China College of Mechanical Engineering Guangxi University Nanning 530004 China School of Mechanical and Automotive Engineering Guangxi University of Science and Technology Liuzhou 545000 China

Lane-changing represents not only a common driving behavior but also a potentially hazardous one. Accurately predicting lane change intentions plays a crucial role in enhancing road traffic safety and guiding autonomous vehicle planning. In this study, a Face-mesh model is used to extract salient features from complex driver behavior data. Subsequently, by using the Farneback optical flow algorithm in conjunction with the ResNet-50 neural network, important lane change cues were extracted from the vehicle surroundings. The Transformer model was optimized using the Teacher-forcing training strategy and the Scheduled-sampling method, fostering faster convergence and heightened prediction accuracy. Empirical tests had shown that this model had attained an impressive precision of 98.61%, recall of 98.24 %, and an F1 score of 98.42 % when forecasting lane change intentions 0.5 s ahead. © 2025 The Author(s)

关键词： Image data mining Lane change intention prediction Neural network optical flow algorithm

来源：评论

学校读者我要写书评

暂无评论

algorithm of Detection and Tracking for Laboratory Unsafe Behavior Based on Computer Vision

Algorithm of Detection and Tracking for Laboratory Unsafe Be...

引用

2023 International Conference on Evolutionary algorithms and Soft Computing Techniques, EASCT 2023

作者： Gao, Lu Bohai University College of Physical Science and Technology Jinzhou China

ISBN: (纸本)9798350313413

Computer vision gives computers the ability of visual perception similar to people, so that computers can feel the environment in the field of vision, understand the content of the feeling, and take corresponding actions. Object tracking is a problem of creating corresponding matching between successive image frames based on position, speed, shape, texture and color. Many security problems in laboratories are caused by unsafe behaviors, and most laboratories have built a complete video surveillance system based on computer vision, using these video surveillance data to identify unsafe behaviors. In this paper, image denoising based on morphological filter algorithm, moving target detection based on optical flow algorithm and moving target tracking based on Mean-Shift algorithm are used to provide key technical solutions for the study of unsafe behaviors in laboratories. The results show that the optical flow algorithm has good adaptability and is suitable for both static and dynamic background. If a more accurate similarity measurement function is used, the tracking time is shorter and the speed is faster. © 2023 IEEE.

关键词： computer vision detection algorithms Mean-Shift algorithm morphological filtering optical flow algorithm tracking algorithms unsafe behavior

来源：评论

学校读者我要写书评

暂无评论

Motion Compensated Compression for Event-Based Cameras

Motion Compensated Compression for Event-Based Cameras

引用

作者： Bairagi, Arnob Kumar University of Lethbridge (Canada)

学位级别：M.Sc., Master of Science

Event cameras detect significant changes at each pixel asynchronously and report these events in real-time. A changing scene can generate many events in a short time. Efficient storage and transmission are necessary for further processing of this event data. Inspired by this necessity, we propose a lossless Motion Compensated Compression algorithm based on optical flow (MCCOF) for event cameras. We analyzed our proposed algorithm performance compared with the lossless spike coding algorithm. We found that our MCCOF algorithm achieves a higher compression ratio on most datasets compared to the spike coding algorithm. Using a real-time event-based optical flow algorithm for motion compensation, our algorithm does not significantly increase the computational time for compression and decompression.

关键词： Event data processing Event-based cameras Log-luminance Lossless compression algorithm Motion compensated compression optical flow algorithm

来源：评论

学校读者我要写书评

暂无评论

Quantitative determination of SO2 flux from industrial chimney through machine vision with plume model verification

引用

MEASUREMENT 2024年 238卷

作者： He, Weiwei Zhang, Huiliang Wang, Rui Guo, Jianjun Zhang, Yikang Wu, Kuijun Yantai Univ Sch Phys & Elect Informat Yantai 264005 Peoples R China

Accurate emission rate data are essential for assessing the environmental impact of industrial emissions and calibrating the effectiveness of desulphurization equipment. However, accurate calculation of SO2 emissions is difficult to realize due to the environmental conditions, the time-varying and spatial non-uniformity of pollution source emissions. In this paper, a Dual TV Plus machine vision algorithm is proposed, calculated and validated with the conventional Farneback optical flow algorithm in terms of industrial stack emission rates. In addition, the effect of turbulence on the inversion of SO2 emission under complex environmental conditions is evaluated by combining simulations and outdoor experiments. The results show that the Dual TV Plus algorithm performs better than the Farneback algorithm and the Dual TV-L1 algorithm, with 50% and 30% improvement in accuracy and 5% and 8% improvement in immunity to interference, respectively. This work is important for promoting the engineering applications of remote sensing instruments for environmental monitoring.

关键词： SO 2 flux Machine vision optical flow algorithm flow field modeling UV camera

来源：评论

学校读者我要写书评

暂无评论

Monitoring Sea Surface Currents Based on China GF-4 Multispectral Data

引用

MARINE GEODESY 2021年第6期44卷 531-553页

作者： Sun, Hequan Li, Zeyu Shao, Ruoli Dalian Naval Acad Dept Mil Oceanog & Hydrog Dalian 116018 Peoples R China Dalian Univ Technol Design Inst Civil Engn & Architecture Dalian Peoples R China

The seawater tracing index that characterizes the sea surface feature can be calculated using the optimal band combination of China GF-4 high-resolution multispectral data, and the distribution of sea surface currents can be retrieved from the sequential band combination data. In this paper, the methods of monitoring sea surface currents are presented in detail by processing multispectral remote sensing images generated by China GF-4 Earth Observation satellite. The sea surface currents obtained from GF-4 imagery data are compared with the sea surface currents derived from FES2014 global ocean tide model, providing a validation of the proposed methods.

关键词： FES2014 tide model GF-4 satellite maximum cross-correlation optical flow algorithm remote sensing sea surface currents

来源：评论

学校读者我要写书评

暂无评论

A debris image tracking using optical flow algorithm

引用

ADVANCES IN SPACE RESEARCH 2012年第5期49卷 1007-1018页

作者： Fujita, K. Hanada, T. Kitazawa, Y. Kawabe, A. Kyushu Univ Dept Aeronaut & Astronaut Nishi Ku Fukuoka 8190395 Japan IHI Corp Koto Ku Tokyo 1358710 Japan

This study proposes a motion detection and object tracking technique for GEO debris in a sequence of images. A couple of techniques (called the "stacking method" and "line-identifying technique") were recently proposed to address the same problem. Although these techniques are effective at detecting the debris position and motion in the image sequences, there are some issues concerned with computational load and assumed debris motion. This study derives a method to estimate motion vectors of objects in image sequence and finally detect the debris locations by using a computer vision technique called an optical flow algorithm. The new method detects these parameters in low computational time in a serial manner, which implies that it has an advantage to track not only linear but also nonlinear motion of GEO debris more easily than the previous methods. The feasibility of the proposed methods is validated using real and synthesized image sequences which contain some typical debris motions. (C) 2011 COSPAR. Published by Elsevier Ltd. All rights reserved.

关键词： Debris observation Image processing optical flow algorithm

来源：评论

学校读者我要写书评

暂无评论

Locating object - of - interest & preventing Stampede using crowd management 3

Locating object - of - interest & preventing Stampede using ...

引用

3rd International Conference on Computing Methodologies and Communication (ICCMC)

作者： Suganya, P. Gaurav, Tannavi Rampuriya, Priyansh Dey, Subhadeep SRM Inst Sci & Technol Comp Sci Engn Chennai Tamil Nadu India

ISBN: (纸本)9781538678084

Congestion in urban areas is becoming a serious problem in modern society. Large scale events at locations such as movie theatres or sports stadiums, etc which attract thousands of spectators can be expected to cause more inconvenience in future as compared to the present day conditions. Crowd management is important as with the increase in number of individuals a number of problems can occur, like, the risks of people or objects getting lost and stampede in areas of high crowd density. It is essential for people to live safely and securely in such large cities. Real-time analysis and rapid response is necessary for the security and safety. Using Artificial Intelligence we will be able to track people and object of interest almost immediately. In cases of people or costly objects being lost it becomes imperative to find its location in real-time, and this is where this implementation has an upper-hand on all its predecessors. In places of high density of population it is very common for children to get separated from their parents. Even at places where CCTV cameras are installed it is very time seeking to search for a particular object in millions of objects.

关键词： Artificial Intelligence Open-CV optical flow algorithm Real-time tracking Crowd Behaviour Analysis

来源：评论

学校读者我要写书评

暂无评论

Evaluating a potential technique with local optical flow vectors for automatic organ-at-risk (OAR) intrusion detection and avoidance during radiotherapy

引用

PHYSICS IN MEDICINE AND BIOLOGY 2019年第14期64卷 145008-145008页

作者： Teo, P. Troy Guo, Kaiming Ahmed, Bilal Alayoubi, Nadia Kehler, Katherine Fontaine, Gabrielle Sasaki, David Pistorius, Stephen Univ Manitoba Dept Phys & Astron Winnipeg MB R3T 2N2 Canada CancerCare Manitoba Med Phys 675 McDermot Ave Tuti Peg MB R3E 0V9 Canada Univ Manitoba Dept Radiol Winnipeg MB R3T 2N2 Canada Univ Manitoba Biomed Engn Grad Program Winnipeg MB R3T 2N2 Canada

Various techniques of deep inspiration breath hold (DIBH) have been used to mitigate the likelihood and risk of exposing the heart, an organ-at-risk (OAR) for unintended radiation during left breast radiotherapy. However, issues of reproducibility of these techniques warrant further investigation into the feasibility of detecting the intrusion of an OAR into the treatment field during intrafractional treatment delivery. The increase of high-dose, low-fraction radiotherapy treatments makes it important to immediately adapt treatment once an OAR is detected in the treatment field. This proof-of-concept implementation includes an algorithm that detects and tracks the motion at the edges of a treatment field and a control algorithm that adapts the treatment aperture according to the motion detected. In accordance to the AAPM Task-Group (TG-132) report, image registration techniques should be verified with virtual and physical phantoms prior to clinical application. Since most OARs move as a result of respiration-induced motion, we have used a lung phantom to generate images of a generic OAR intruding into a treatment field with known velocity. The phantom was programmed to move with sinusoidal and lung patient tumor motion patterns and the accuracy of intrusion tracking and MLC adaptation were benchmarked with the ground truth-programmed motion of the OAR. The motions were recorded with an electronic portal imaging device (EPID). An optimal cluster size of 9 x 9 motion vectors was found to provide the smallest average absolute position error of 0.3 mm. A strong linear correlation between the adapted MLC leaves and the actual OAR position was observed. The algorithm had a mean position tracking error of -0.4 +/- 0.3 mm and a precision of 1.1 mm. It is possible to adapt MLC leaves based on the motion detected at the edges of the irradiated field, and it would be feasible to shield an unplanned intrusion of an OAR into the treatment field.

关键词： deep inspiration breath hold (DIBH) breast cancer treatment cardiac block image registration technique optical flow algorithm

来源：评论

学校读者我要写书评

暂无评论

optical flow Based Violence Detection in Video Surveillance 1

Optical Flow Based Violence Detection in Video Surveillance

引用

1st International Conference on Advances in Communication and Computing Technology (ICACCT)

作者： Garje, Prasad. D. Nagmode, M. S. Davakhar, Kiran. C. Zeal Coll Engn & Res Pune Maharashtra India Govt Coll Engn & Res E&TC Dept Pune Maharashtra India Smt Kashibai Navale Coll Engn E&TC Dept Pune Maharashtra India

ISBN: (纸本)9781538609262

Video Surveillance has become an issue of utmost importance due to rising crime and violence rate in the world. Many algorithms exist to analyze the behavior of the moving objects. In this paper, an approach to detect any violent behavior in videos using optical flow algorithms. The SDHA 2000 dataset is used to analyze foreground behavior. The main task in violence detection is to detect moving objects and classify their behavior using mean of intensity as violent or non-violent.

关键词： RGB to HSV Conversion optical flow algorithm Gunner Farneback algorithm Violent Activity

来源：评论

学校读者我要写书评

暂无评论

Foreground Segmentation Using Motion Vector for Camouflaged Surveillance Scenario 2

Foreground Segmentation Using Motion Vector for Camouflaged ...

引用

2nd IEEE International Conference on Wireless Communications, Signal Processing and Networking (WiSPNET)

作者： Ratthi, K. Iyshwarya Nithya, V. Yogameena, B. Menaka, K. Thiagarajar Coll Engn Elect & Commun Engn Madurai Tamil Nadu India

ISBN: (纸本)9781509044429

Detection of moving objects is the first relevant step of information extraction in many computer vision applications. It gains focus of attention in applications such as abnormal event detection, human gait characterization, congestion analysis, person identification and fall detection for elderly people. Foreground segmentation fails due to various reasons such as presence of camouflage, shadows, slow moving objects, illumination changes etc. Detecting moving objects from similarly colored background (known as camouflage problem), has been a long-standing open question in this field. Thus, camouflage is one of the major issues where foreground and background gets occluded or merged when they are of similar color. In case of terrorist attacks, terrorist disguise themselves, by dressing similar to the environment so that they can hide themselves. Foreground segmentation by eliminating camouflage in such cases play a vital role which is one of the major issues being faced nowadays. In order to overcome this issue, the proposed work extracts motion information using optical flow algorithm to detect the moving object and Motion Assisted Matrix Restoration (MAMR) for foreground segmentation in fully camouflaged surveillance scenario. Experimental results have been shown emphasizing that the proposed methodology outperforms other state-of-the-art algorithms.

关键词： Foreground segmentation motion vector optical flow algorithm camouflage removal MAMR

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：