A5040960573- Traffic Prediction and Management Techniques
- Time Series Analysis and Forecasting
- Transportation Planning and Optimization
- Traffic control and management
- Data Management and Algorithms
- Internet Traffic Analysis and Secure E-voting
- Music and Audio Processing
- Vehicle License Plate Recognition
- Network Security and Intrusion Detection
- Advanced Manufacturing and Logistics Optimization
- Anomaly Detection Techniques and Applications
- Traffic and Road Safety
- Advanced Welding Techniques Analysis
- Aluminum Alloys Composites Properties
- Urban and Freight Transport Logistics
- Data Mining Algorithms and Applications
- Spam and Phishing Detection
- Vehicle Routing Optimization Methods
- Currency Recognition and Detection
- Advanced Malware Detection Techniques
- Advanced SAR Imaging Techniques
- Advanced Measurement and Detection Methods
- Neural Networks and Applications
- Manufacturing Process and Optimization
- Magnesium Alloys: Properties and Applications
University of Electronic Science and Technology of China
2023-2025
Fujian University of Technology
2019-2022
Harbin Institute of Technology
2020-2022
The targets in Synthetic Aperture Radar (SAR) images are often tiny, irregular, and difficult to detect against complex backgrounds, leading a high probability of missed or incorrect detections by object detection algorithms. To address this issue improve the recall rate, we introduce an improved version YOLOv8 (You Only Look Once), named MDD-YOLOv8. This model is not only fast but also highly accurate, with fewer instances detections. Our proposed outperforms baseline SAR image utilizing...
Abstract Despite the significant advancements made by deep visual networks in detecting surface defects at a regional level, challenge of achieving high-quality pixel-wise defect detection persists due to varied appearances and limited availability data. To address over-reliance on appearance enhance accuracy segmentation, we proposed Transformer-based Siamese network with change awareness, which formulates segmentation under complex background as mimic human inspection process....
As one of the largest Internet Things systems in world, China’s expressway electronic toll collection (ETC) generates nearly billion pieces transaction data every day, recording traffic trajectories almost all vehicles on expressway, which has great potential application value. However, there are inevitable missed transactions and false ETC system, leads to certain missing rates data. In this work, a dynamic search step SegrDTW algorithm based an improved DTW is proposed according...
The travel time prediction of vehicles is an important part intelligent expressways. It can not only provide the vehicle distribution trend each section for expressway management department to assist fine expressway, but it also owners with dynamic and accurate services formulate more reasonable plans. However, there are still some problems in current research (e.g., different types processed separately, proximity road network considered, capture information spatial-temporal perspective...
Abstract With the rapid development of expressway Electronic Toll Collection (ETC) technology in China, management system is becoming digital and intelligent, which provides a solid foundation for vehicle infrastructure cooperation autonomous driving. The gantry position key part ETC system. However, there are still some problems (e.g. missing or false), can seriously affect intelligent expressways. To address these two issues, an positioning method proposed. First, transaction data GPS on...
Despite the eye-catching breakthroughs achieved by deep visual networks in detecting region-level surface defects, challenge of high-quality pixel-wise defect detection remains due to diverse appearances and data scarcity. To avoid over-reliance on appearance achieve accurate segmentation, we proposed a change-aware Siamese network that solves segmentation change framework. A novel multi-class balanced contrastive loss is introduced guide Transformer-based encoder, which enables encoding...
To overcome the drawbacks of maximum speed limit information expressways (i.e., long update cycle and great complexity recognition), in this work, an Electronic Toll Collection (ETC) gantry data-based method for dynamically identifying is proposed. Firstly, characteristics ETC data are analyzed, then cleaned reconstructed, after which algorithm proposed constructing a vehicle travel set. Secondly, feature vector model road section established by taking relationship among distribution...
The development of the Industrial Internet has promoted progress social productivity, but it also faces attacks from abnormal network traffic. Network intrusion detection systems (NIDSs) ensure safe and reliable operation networks by monitoring traffic status detecting in a timely manner. To detect intrusions real time efficiently, we propose hierarchical model CNNTransformer NIDS with spatio-temporal feature fusion, combined soft selection based on attention mechanism. is used for...
With the rapid development of industrial Internet, security control network has become a vital topic. The intrusion detection system (NIDS), as an important device in contemporary networks, is importance detecting traffic and ensuring security. In this paper, two-stage training Deep Neural Network (DNN) based on cross-entropy focal loss, called CE-FL-NIDS, proposed to handle with problems caused NIDS due uneven data distribution dataset. effectiveness model verified by performing 10-class...
Recognizing speed limit information is crucial for advanced driver assistance systems (ADAS) as it directly affects the safety planning and decision-making process of intelligent driving systems. However, traditional image recognition-based solutions confront inherent restrictions precision issues due to uncontrolled external factors. This paper endeavors present a novel, data-driven solution recognition that leverages stability maturity technologies, overcoming these challenges. We...
The similarity measurement of time series is a significant approach to mine the rich and valuable law information hidden in massive data. As most advantageous measuring similarities series, Dynamic Time Warping (DTW) has become one hottest researches field data mining. However, DTW algorithm does not satisfy trigonometric inequality, its space complexity are extremely high, how efficiently realize retrieval similar sequences large-scale sequential remains challenge. This paper first...
Accurate modeling of travel speeds is crucial for optimizing roadway management, yet traditional methods overlook a key factor the influence vehicle dwell times in service areas.This oversight introduces bias into speed measurements, impairing their utility fine-grained traffic monitoring.To address this problem, we propose an innovative framework that integrates machine learning prediction area calculation.We focus on 9.3 km segment major highway Fujian Province, China includes Qingyunshan...
Monocular vehicle perception has been a vital problem in autonomous driving. Anchor-based detectors are always used to solve the but bounding box's center calibrate is actually not corresponded with vehicle's 3D center. Therefore, we designed network by point detector U-Net and propose calibration method using 3D2D reprojection constraints prior plane imaging system. We evaluate our on ApolloCar3D dataset, experimental results prove that proposed feasible effective. On testing mean average...