Query：
学者姓名：陈锋
Refining：
Year
Type
Indexed by
Source
Complex
Co-
Language
Clean All
Abstract ：
文中介绍了基于边缘计算的拖轮智能安全预警系统设计原理与实施方案.通过边缘计算实时获取部署在拖轮上的摄像头数据,经视频取流解码后通过神经网络进行检测,实现驾驶员离岗预警、人员进入危险区域预警和人员定时巡查预警等.针对神经网络检测可能出现的误检问题,设计了一套推理均衡器算法和感兴趣区域检测算法,有效减少了因误检导致的系统误报问题.所设计的系统最终在福州港务集团的拖轮上得到实际部署、测试和验证.结果表明,设计的系统和算法运行可靠,可以有效实现拖轮智能安全预警功能.
Keyword ：
YOLOv5s YOLOv5s 区域检测 区域检测 射线法 射线法 嵌入式 嵌入式 拖轮安全 拖轮安全 推理均衡器 推理均衡器
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | 吴陈锋 , 郑华东 , 陈锋 . 基于边缘计算的拖轮智能安全预警系统 [J]. | 物联网技术 , 2025 , 15 (8) : 10-15 . |
| MLA | 吴陈锋 等. "基于边缘计算的拖轮智能安全预警系统" . | 物联网技术 15 . 8 (2025) : 10-15 . |
| APA | 吴陈锋 , 郑华东 , 陈锋 . 基于边缘计算的拖轮智能安全预警系统 . | 物联网技术 , 2025 , 15 (8) , 10-15 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
针对传统的单一链路传输无法满足流媒体业务高带宽、低延时需求的问题,基于流媒体传输协议SRT在OpenWRT系统上实现了多路并行传输,并提出了一种基于链路状态的数据分流算法.该算法根据服务器端反馈的链路信息确定链路的状态,按照每条链路的状态分配相应的传输任务量,并且在服务端实现分流合并,可以有效降低数据包的端到端时延,并提高系统传输的稳定性.实验结果表明:采用该多路并行传输系统进行传输时,数据包的端到端时延比单一链路更加稳定并且平均时延降低50%左右,系统的吞吐量也明显提高,能够较好地保证负载均衡,并且在OpenWRT系统下实现了更有利于实际业务的部署.
Keyword ：
OpenWRT系统 OpenWRT系统 SRT协议 SRT协议 分流合并 分流合并 多路并行传输 多路并行传输 数据调度 数据调度 链路状态 链路状态
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | 宋道斌 , 陈锋 , 余超群 . 基于链路状态的OpenWRT多路并行传输系统的设计 [J]. | 物联网技术 , 2025 , 15 (1) : 72-76,79 . |
| MLA | 宋道斌 等. "基于链路状态的OpenWRT多路并行传输系统的设计" . | 物联网技术 15 . 1 (2025) : 72-76,79 . |
| APA | 宋道斌 , 陈锋 , 余超群 . 基于链路状态的OpenWRT多路并行传输系统的设计 . | 物联网技术 , 2025 , 15 (1) , 72-76,79 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
针对港口矿山复杂工业环境下,带式输送机管理方式多为人工监控,传统料流检测方法存在精度差和难以落地问题,提出一种基于单目视觉与线结构光融合的输送带料流体积检测方法.首先,将线结构光投射到料流表面,用标定好的工业相机采集料流图像;其次,根据激光颜色特征,通过灰度重心法提取料流图像的激光条纹,为了防止皮带抖动带来的误差,基于负载时的激光条纹拟合出皮带空载时的激光条纹;最后,获取激光条纹上料流点的三维空间坐标,基于料流端点计算出沿皮带运行方向料流截面积与传输速度的积分得到料流体积.实验结果表明,所提出的方法具有较高的精度,平均误差最小可达到0.901%,满足工业环境要求,能够运用于实际工作现场.
Keyword ：
单目视觉 单目视觉 工业环境 工业环境 料流检测 料流检测 生产安全 生产安全 线结构光 线结构光
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | 林灿辉 , 俞佳宝 , 陈锋 et al. 基于单目视觉的输送带料流体积检测方法研究 [J]. | 电视技术 , 2025 , 49 (1) : 37-41 . |
| MLA | 林灿辉 et al. "基于单目视觉的输送带料流体积检测方法研究" . | 电视技术 49 . 1 (2025) : 37-41 . |
| APA | 林灿辉 , 俞佳宝 , 陈锋 , 郭恩特 , 黄锦楠 , 陈晨炜 . 基于单目视觉的输送带料流体积检测方法研究 . | 电视技术 , 2025 , 49 (1) , 37-41 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
With the rapid development of mobile networks and devices, real-time video transmission has become increasingly important worldwide. Constrained by the bandwidth limitations of single networks, extensive research has shifted towards video transmission in multi-network environments. However, differences in bandwidth and latency in heterogeneous networks (such as LTE and Wi-Fi) lead to high latency and packet loss issues, severely affecting video quality and user experience. This paper proposes a Forward Error Correction (FEC)-based Low-Delay Multipath Scheduling algorithm (LDMP-FEC). This algorithm combines the Gilbert model with a continuous Markov chain to adaptively adjust FEC redundancy, thereby enhancing data integrity. Through the FEC Recovery Priority Scheduling (FEC-RPS) algorithm, it dynamically optimizes the transmission order of data packets, reducing the number of out-of-order packets (OFO-packets) and end-to-end latency. Experimental results show that LDMP-FEC significantly reduces the number of out-of-order packets in heterogeneous network environments, improving performance by 50% compared to the round-robin and MinRtt algorithms, while maintaining end-to-end latency within 150 ms. Under various packet loss conditions, LDMP-FEC sustains a playable frame rate (PFR) above 90% and a Peak Signal-to-Noise Ratio (PSNR) exceeding 35 dB, providing an efficient and reliable solution for real-time video and other low-latency applications.
Keyword ：
FEC FEC heterogeneous networks heterogeneous networks multipath multipath OFO-packets OFO-packets round robin round robin video transmission video transmission
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Gao, Tingjin , Chen, Feng , Chen, Pingping . LDMP-FEC: A Real-Time Low-Latency Scheduling Algorithm for Video Transmission in Heterogeneous Networks [J]. | ELECTRONICS , 2025 , 14 (3) . |
| MLA | Gao, Tingjin et al. "LDMP-FEC: A Real-Time Low-Latency Scheduling Algorithm for Video Transmission in Heterogeneous Networks" . | ELECTRONICS 14 . 3 (2025) . |
| APA | Gao, Tingjin , Chen, Feng , Chen, Pingping . LDMP-FEC: A Real-Time Low-Latency Scheduling Algorithm for Video Transmission in Heterogeneous Networks . | ELECTRONICS , 2025 , 14 (3) . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Mobile edge caching (MEC) has emerged as a promising and economical solution to complement conventional infrastructure caching. Nonetheless, the vulnerability of vehicle-to-vehicle (V2V) links causes connection loss and limits data exchange. On the other hand, the traditional content popularity-based request model would lead to lower cache hit ratios and increased content retrieval times. To tackle this issue, a novel scheme of MEC-assisted public vehicular network is proposed in this study, where the random linear network coding (RLNC) based caching strategy is applied to allow public vehicles to simultaneously obtain coded blocks from multiple vehicles and infrastructures on the move. Besides, a content request model that considers content popularity, historical interest, and social attributes is explored. A cost minimization problem is formulated under the proposed scheme, which is a highly non-trivial stochastic problem. To this end, the data volume of V2V offloading is obtained by a divide and conquer (DC) algorithm, and then the caching strategy is derived by a heuristic-based algorithm. Finally, extensive simulations show that the proposed content model can achieve a higher local offloading ratio and the proposed scheme has a much lower cost compared to the baselines.
Keyword ：
Content request model Content request model Mobile edge caching Mobile edge caching Public vehicular networks Public vehicular networks V2V offloading V2V offloading
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Chen, Xiaopei , Lin, Zhijian , Wu, Wenhao et al. Cost-Efficient and Preference-Aware Mobile Edge Caching in Public Vehicular Networks [J]. | MOBILE NETWORKS & APPLICATIONS , 2025 . |
| MLA | Chen, Xiaopei et al. "Cost-Efficient and Preference-Aware Mobile Edge Caching in Public Vehicular Networks" . | MOBILE NETWORKS & APPLICATIONS (2025) . |
| APA | Chen, Xiaopei , Lin, Zhijian , Wu, Wenhao , Chen, Feng , Chen, Pingping . Cost-Efficient and Preference-Aware Mobile Edge Caching in Public Vehicular Networks . | MOBILE NETWORKS & APPLICATIONS , 2025 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Estimating a full 3D human shape from a Single RGB Image is an important and challenging inverse task in virtual reality. Recent approaches apply deep learning to achieve significant success by learning geometric features from given images and corresponding 3D annotations of human datasets. However, humans in images are often occluded or self-occluded due to complex outdoor environments. Additionally, the difficulty in obtaining accurate 3D data means that most 3D annotations are inaccurate pseudo-labels generated by simulation software. These issues limit the effective learning of human geometric features, resulting in reduced performance of existing deep learning-based 3D human shape estimation methods. In this paper, to fully leverage the prior knowledge from such low-quality data (i.e.,), we propose a novel geometric prior learning approach with multimodal representation compensation for 3D human shape estimation. First, we design a multimodal pre-training task that reconstructs the human image and mesh from masked input data in a self-supervised manner, facilitating geometric prior learning for our network. Then, we use the pre-trained image and mesh network to guide and fine-tune the end-to-end human shape estimation framework. Evaluations across multiple public datasets show that our method clearly improves upon the baseline of previous work with more accurate human shapes, particularly in outdoor scenes with occlusion and inaccurate labeling.
Keyword ：
3D human shape estimation 3D human shape estimation Monocular image Monocular image Multimodal pre-training Multimodal pre-training Pseudo-3d annotations Pseudo-3d annotations Self-occlusion Self-occlusion
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Chen, Feng , Huang, Jilin , Jiang, Biao et al. GLMP: Geometric prior learning with multimodal pre-training representation compensation for 3D human shape estimation [J]. | KNOWLEDGE-BASED SYSTEMS , 2025 , 326 . |
| MLA | Chen, Feng et al. "GLMP: Geometric prior learning with multimodal pre-training representation compensation for 3D human shape estimation" . | KNOWLEDGE-BASED SYSTEMS 326 (2025) . |
| APA | Chen, Feng , Huang, Jilin , Jiang, Biao , Chen, Pingping , Jiang, Mengxi . GLMP: Geometric prior learning with multimodal pre-training representation compensation for 3D human shape estimation . | KNOWLEDGE-BASED SYSTEMS , 2025 , 326 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Detection-based tracking is one of the main methods of multi-object tracking. It can achieve good tracking performance when using excellent detectors but it may associate wrong targets when facing overlapping and low-confidence detections. To address this issue, this article proposes a novel multi-object tracker (SCTracker) by exploiting shape constraint and confidence. In the data association stage, an intersection of union (IoU) distance with shape constraints is developed to calculate the cost matrix between tracks and detections, which can reduce the track of the wrong target with the similar position but inconsistent shape. Moreover, the detection confidence is calculated in the update stage of the Kalman filter to improve the track performance with the inaccurate detection result. Experimental results on the MOT 17 dataset show that the proposed SCTracker can improve the tracking performance of multi-object tracking when compared with the state-of-the-art methods.
Keyword ：
Deep learning Deep learning distance measurement distance measurement motion estimation motion estimation object tracking object tracking
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Mao, Huan , Chen, Yulin , Li, Zongtan et al. SCTracker: Multi-Object Tracking With Shape and Confidence Constraints [J]. | IEEE SENSORS JOURNAL , 2024 , 24 (3) : 3123-3130 . |
| MLA | Mao, Huan et al. "SCTracker: Multi-Object Tracking With Shape and Confidence Constraints" . | IEEE SENSORS JOURNAL 24 . 3 (2024) : 3123-3130 . |
| APA | Mao, Huan , Chen, Yulin , Li, Zongtan , Chen, Pingping , Chen, Feng . SCTracker: Multi-Object Tracking With Shape and Confidence Constraints . | IEEE SENSORS JOURNAL , 2024 , 24 (3) , 3123-3130 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Anomaly detection is represented as an unsupervised learning to identify deviated images from normal images. In general, there are two main challenges of anomaly detection tasks, i.e., the class imbalance and the unexpectedness of anomalies. In this paper, we propose a multiresolution feature guidance method based on Transformer named GTrans for unsupervised anomaly detection and localization. In GTrans, an Anomaly Guided Network (AGN) pre-trained on ImageNet is developed to provide surrogate labels for features and tokens. Under the tacit knowledge guidance of the AGN, the anomaly detection network named Trans utilizes Transformer to effectively establish a relationship between features with multiresolution, enhancing the ability of the Trans in fitting the normal data manifold. Due to the strong generalization ability of AGN, GTrans locates anomalies by comparing the differences in spatial distance and direction of multi-scale features extracted from the AGN and the Trans. Our experiments demonstrate that the proposed GTrans achieves state-of-the-art performance in both detection and localization on the MVTec AD dataset. GTrans achieves image-level and pixel-level anomaly detection AUROC scores of 99.0% and 97.9% on the MVTec AD dataset, respectively.
Keyword ：
Anomaly detection Anomaly detection Deep learning Deep learning Knowledge distillation Knowledge distillation Transformer Transformer
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Yan, Shuting , Chen, Pingping , Chen, Honghui et al. Multiresolution feature guidance based transformer for anomaly detection [J]. | APPLIED INTELLIGENCE , 2024 , 54 (2) : 1831-1846 . |
| MLA | Yan, Shuting et al. "Multiresolution feature guidance based transformer for anomaly detection" . | APPLIED INTELLIGENCE 54 . 2 (2024) : 1831-1846 . |
| APA | Yan, Shuting , Chen, Pingping , Chen, Honghui , Mao, Huan , Chen, Feng , Lin, Zhijian . Multiresolution feature guidance based transformer for anomaly detection . | APPLIED INTELLIGENCE , 2024 , 54 (2) , 1831-1846 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
随着远程监控和人工智能的融合发展,传统的码率优化算法并不适用于现阶段的移动监控网络场景.在机器视觉应用场景中,相对于传统码率优化算法只关注视频的质量,机器更关注于视频所表达的语义信息.以5G路侧摄像头远程智能检测为应用场景,提出一种基于视频语义的码率优化算法,在有限的码率传输范围内最大化目标检测准确率.具体地,该算法引入视频语义任务模型,将目标检测作为语义任务.分析目标比特与语义之间的特征关系,建立复杂度与运动区域结合的新权重来分配目标比特,使目标检测准确率达到最大化.实验结果表明,相较于HM16.23所使用的帧级树编码单元(Coding Tree Unit,CTU)层码率控制算法,所提算法不仅能够节省码率而且更符合无线远程监控的目标检测需求.在测试环境下平均提升了 1.4％的目标检测准确率,最高能够提升2.5％的目标检测准确率.
Keyword ：
人工智能 人工智能 机器视觉 机器视觉 目标检测 目标检测 视频语义 视频语义
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | 黄发仁 , 柯捷铭 , 郑楚飞 et al. 基于视频语义的码率控制算法 [J]. | 无线电工程 , 2024 , 54 (8) : 1890-1899 . |
| MLA | 黄发仁 et al. "基于视频语义的码率控制算法" . | 无线电工程 54 . 8 (2024) : 1890-1899 . |
| APA | 黄发仁 , 柯捷铭 , 郑楚飞 , 周简心 , 张森林 , 陈锋 . 基于视频语义的码率控制算法 . | 无线电工程 , 2024 , 54 (8) , 1890-1899 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
针对多模态信息在三维空间融合过程中数据一致性和有效性的问题,提出鸟瞰视角(BEV)下图像与雷达融合的编码模块BEVIRF.与传统的透视视角下深度信息缺失的方案相比,本方法利用可变注意力的改进方案聚合图像和雷达信息,解决不同模态信息的统一表示问题,生成语义丰富且包含空间位置信息BEV特征图.同时在基于Transformer的网络结构中提出动态位置编码,旨在通过感知物体的空间信息来生成对应的位置编码,让模型专注于目标的回归,减少查询与匹配的不稳定性.所提出的方案在nuScenes数据集上取得了优秀结果.
Keyword ：
BEV特征 BEV特征 三维目标检测 三维目标检测 卷积神经网络 卷积神经网络 注意力机制 注意力机制
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | 蔡甘霖 , 陈锋 , 张森林 . 图像-雷达融合的三维目标检测算法 [J]. | 福州大学学报（自然科学版） , 2024 , 52 (6) : 659-666 . |
| MLA | 蔡甘霖 et al. "图像-雷达融合的三维目标检测算法" . | 福州大学学报（自然科学版） 52 . 6 (2024) : 659-666 . |
| APA | 蔡甘霖 , 陈锋 , 张森林 . 图像-雷达融合的三维目标检测算法 . | 福州大学学报（自然科学版） , 2024 , 52 (6) , 659-666 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Export
| Results： |
Selected to |
| Format： |
