mở đầu chương và kết luận chương, có liệt kê tài liệu tham khảo và có trích dẫn đúng quy định 9 Kỹ năng viết xuất sắc (cấu trúc câu chuẩn, văn phong khoa học, lập luận 1 2 3 4 5 logic và có cơ sở, từ vựng sử dụng phù hợp v.) Thành tựu nghiên cứu khoa học (5) (chọn 1 trong 3 trường hợp) Có bài báo khoa học được đăng hoặc chấp nhận đăng/Đạt giải SVNCKH 10a giải 3 cấp Viện trở lên/Có giải thưởng khoa học (quốc tế hoặc trong 5 nước) từ giải 3 trở lên/Có đăng ký bằng phát minh, sáng chế Được báo cáo tại hội đồng cấp Viện trong hội nghị SVNCKH nhưng 10b không đạt giải từ giải 3 trở lên/Đạt giải khuyến khích trong các kỳ thi 2 quốc gia và quốc tế khác về chuyên ngành (VD: TI contest) 10c Không có thành tích về nghiên cứu khoa học 0 Điểm tổng /50 Điểm tổng quy đổi về thang 10 Nhận xét khác của cán bộ phản biện. Người nhận xét (Ký và ghi rõ họ tên) PREFACE Modern structures such as bridges, buildings, water dams, minerals mines, etc. have indeed played an important role in our daily life. They benefit us in many aspects and greatly contribute to our twenty-first-century society.
Owners and maintenance managers of these capital-intensive assets expect to increase durability of the structure. Therefore, Structural Health Monitoring (SHM) system is born in response to such a need. Acknowledged of structures’ defects, timely intervention shall reduce risks and maintenance costs. As a result, SHM system shall help to increase the service life of structures.
For example, The Canadian Highway Bridge considers a service life of 75 years for newly constructed bridges. Another example in this case is the New Cham- plain bridge in Montreal, Canada, which is designed for 125 years in service. Moreover, municipalities, provincial and federal governments are becoming interested in the con- cept of “Smart City”. This vision aims at implementing advanced technologies to create value-added services for citizens and the administration of the city [?].
In the context, proper inspection and maintenance have become more important than ever. Another crucial application of SHM system is to improve reliability of existing infrastructure. Existing infrastructures in North America are rapidly approaching the end of their design service life. According to ASCE 2017 infrastructure report card, nearly 10% of bridges (about 56,000 bridges) in the United States have structural defi- ciencies, which makes them vulnerable.
The conditions in Canada are not much better. There are about 75,000 highway bridges in Canada; According to the National Research Council of Canada (NRCC 2015), almost one-third of these bridges have structural or functional deficiencies. The collapse of Ponte Mirandi in Genova, Italy, in August 2018 clearly shows that existing infrastructure requires immediate attention. SHM system refers to an array of connected sensors, which collect and analyze data at every moment during the service life of the structure.
The goal is to identify and quantify any damage or deterioration state that might occur over the service life [?]. Vibrating Wire Sensors (VWS) are among such sensors. In Viet Nam, due to our rocket- developed economy in the past decade, constructions are seen everywhere. The tragic and catastrophic breakdown of Can Tho bridge in Vinh Long province, 26 September 2007 that causes the death of 55 people is evidence of how important structure monitor- ing is.
Thus, in Viet Nam, the need for VWS and SHM system in general is undeniable. According to surveys, most of the VWS sold in the market are GEOKON original, while Readout device compatible with these sensors is overpriced as well as hard to purchase. This study plays the part of introducing and developing a VWS Reader that is compatible with these VWS models, with the vision to develop it to a Data-logger for IoT implemen- tation. The product is designed to be affordable while still satisfy all standard industrial requirements.
To complete this graduation thesis, I would like to express my heartfelt gratitude to Mr. Nguyen Nam Phong (Dr.) for his unwavering support and guidance during my project writing process. I would love to give special thanks to the Hanoi University of Science and Technol- ogy lecturers who have enthusiastically passed on their experience over the years. The information I gained during the Hanoi University of Science and Technology lecture process is not only the cornerstone for completing the project but also a valuable asset that will enable me to confidently enter the labor market.
I’m also grateful to Ms. Bui Van Anh and Mr. Nguyen Tien Hoa for creating this thesis LaTeX template. Finally, I wish the teachers good health and continued achievement in their careers as educators.
PLEDGE I am Tran Thai Son, student ID 20212460M, my instructors are Dr. Nguyen Nam Phong. I assure that all of the content presented in the project is the result of my research; the data stated in the project is completely genuine, reflecting the actual measurement results; all cited information complies with intellectual property regulations; the refer- ences are clearly listed. I take full responsibility for the content written in this project.
Hanoi, 10 Sep, 2022 Student Tran Thai Son TABLE OF CONTENTS CHAPTER 1: INTRODUCTION AND SYSTEM’S OVERVIEW 1 1.1 Introduction to ITS problems and system’s overview .2 Violation detection flow on edge module. 13 CHAPTER 2: OBJECT DETECTION AND TRACKING APPROACHES 16 2.1 Convolutional Neural Networks and YOLO .2 Evolution of YOLO and selection of YOLOv5 model .3 Implementation of YOLO detector .4 Theory of BYTE Track .5 Implementation of BYTE tracker. 34 CHAPTER 3: VIOLATION DETECTION RESULTS AND DATA TRAIN- ING 37 3.1 Violation and license plate detection approaches .3 Data preparation and training .2 Exploratory data analysis .4 Model training and evaluation. 63 CONCLUSION 68 REFERENCES 71 ABBREVIATIONS AI Artificial Intelligence API Application Programming Interface CMOS Complementary Metal-Oxide-Semiconductor CNN Convolutional Neural Network CPU Central Processing Unit CTC Connectionist Temporal Classification CSP Cross Stage Partial DNN Deep Neural Network EDA Exploratory Data Analysis FLOP Floating-Point Operations per Second FPS Frames Per Second GPIO General-Purpose Input/Output GPU Graphics Processing Unit GUI Graphical User Interface HDMI High-Definition Multimedia Interface HSV Hue, Saturation, Value IDE Integrated Development Environment IFC Intersection Filtering Condition IoT Internet of Things IoU Intersection over Union ITS Intelligent Transportation System LGC Learned Gradient Compression LSTM Long Short-Term Memory mAP mean Average Precision MOT Multi-Object Tracking MOTA Multiple Object Tracking Accuracy MOTP Multiple Object Tracking Pricision NLP Natural Language Processing NMS Non-Maximum Suppression OCR Optical Character Recognition OGUI Operator Graphical User Interface ONNX Open Neural Network Exchange OOUI On-duty Officer User Interface OS Operating System RAFE Recovering Appearance Feature Extraction RAM Random Access Memory RCNN Region-based Convolutional Neural Network RGB Red, Green, Blue RNN Recurrent Neural Network ROI Region Of Interest SDK Software Development Kit SORT Simple Online and Realtime Tracking SOTA State Of The Art SQL Non - Structured Query Language TBD Tracking By Detection UART Universal Asynchronous Receiver/Transmitter USB Universal Serial Bus YOLO You Only Look Once LIST OF FIGURES 1.1 Overall system flow diagram .2 Configuration drawing tool for operators .3 Camera viewing access for mobile application .4 Image of Jetson Xavier .6 Operators utilise OGUI for set up configuration .7 Violation data are visualized to operators .8 Mobile application’s access to violation data .9 Violation data sample in Cloud Firestore .10 Data structure of camera’s configuration .11 Overview implementation of edge device .1 Bounding boxes with a car as object [1] .2 Bounding boxes illustration [1] .3 YOLO’s core idea [2] .4 YOLO’s loss function [1] .5 Speed comparison of YOLOv3 [3] .6 YOLOv5 different model sizes, where FP16 stands for the half floating- point precision, V100 is an inference time in milliseconds on the Nvidia V100 GPU, and mAP based on the original COCO dataset [4] .7 YOLOv5 network architecture [2] .8 Comparison of vanilla PyTorch and ONNX-TensorRT workflow .9 TensorRT with six optimizations approach [5] .10 TensorRT: Precision calibration, layer and tensor fusion, kernel auto- tuning and multi-stream execution [5] .11 TensorRT layer and tensor fusion optimization [5] .12 MOTA-IDF1-FPS comparisons of different trackers.
The horizontal axis is FPS (running speed), the vertical axis is MOTA, and the radius of circle is IDF1 [6] .13 Architectural model one method MOT [6] .14 BYTE Track Algorithms [6] .15 Simulate the working principle of BYTE Track [6] .16 Tracking flow of Byte Track .17 Recovering Appearance Feature Extraction method. For four-wheel ve- hicles, the features in the area of the red box should be extracted. While these two-wheel vehicles are featured in the green and blue boxes.1 Violation detection algorithm .2 Extracting bounding boxes center from YOLO format .3 Traffic light recognition flow .4 Intersection algorithm filtering .5 The long, single-row license plate type in Vietnam .6 License plate with different colors indicates different categories of vehi- cles in Vietnam .7 Pipeline in license plate detection .8 Original images of vehicles .9 Result of grayscale process in license plate flow .10 Result of Filter Bilateral in license plate flow .11 Edge detection with Canny Edge method in license plate flow .12 Finding contours process in license plate flow .13 Extracted ROI in license plate flow .14 Cropped short, two-line license plate as input of character segmentation and separation process .15 Cropped long, one-line license plate .16 CRNN architecture in Paddle OCR approach .17 License plate before detection (left) and after detection(right) .18 Determine characters’ position on the license plate .19 Image of violation event, test run on recorded video at an intersection .20 Image of violation event, test run on stream video at an intersection .21 Image of violation event with a violated motorcycle vehicle .22 Image of violation event with a violated car vehicle .23 Recognizing the license plate of a violated vehicle .24 Automated annotation using YOLOv7 model .25 Result of automated annotation tool .26 Flipping transformation result in our dataset .27 Filtering transformation approach result in our dataset .28 Scaling transformation approach result in our dataset .29 Translation transformation approach result in our dataset .30 Shear transformation by five degree in our dataset .31 Shear transformation by minus five degree in our dataset .32 Rotation transformation approach result in our dataset .33 MixUp augmentation approach in our dataset .34 Two common approaches to tuning hyper-parameters .35 Data in VOC 2007 format .36 Model training process on server .37 Training results on dataset. 66 LIST OF TABLES Table 1.1 Specfications of Jetson Xavier .1 Maximum batch size for each network architecture and workflow 28 Table 2.2 Utilized frameworks with corresponding version included in Jet- Pack 4.3 Other frameworks with corresponding version .4 Comparison of different data association methods on the MOT17 validation set [6] .1 Range of components in HSV color space of signal lights.2 Comparison of improved RAFE-BYTE Track with others approaches 52 Table 3.3 Comparision of data analysis metrics between our dataset and others 58 ABSTRACT An intelligent transportation system (ITS) [7] plays an essential role in public trans- port management, security, and other issues.
The ITS requires traffic flow detection and tracking to be a vital component. Based on the real-time acquisition of urban road traffic flow information, an ITS provides intelligent guidance for relieving traffic jams and re- ducing environmental pollution. The traffic flow detection in an ITS usually adopts the cloud computing mode. The edge node will transmit all the captured video to the cloud computing center.
However, the increasing traffic monitoring has brought enormous challenges to the storage, communication, and processing of traditional transportation systems based on cloud computing. To address this issue, a traffic flow detection and tracking scheme based on deep learning on the edge node is presented in this article. Apart from the edge node, the proposed ITS also contains a data-storing server and client-based user interface, which makes this ITS more practical and user-oriented than others.
