Research and develop solutions to estimate traffic density from traffic cameras at main intersections

VIETNAM NATIONAL UNIVERSITY HO CHI MINH CITY HO CHI MINH CITY UNIVERSITY OF TECHNOLOGY BUI MINH HIEU RESEARCH AND DEVELOP SOLUTIONS TO ESTIMATE TRAFFIC DENSITY FROM TRAFFIC CAMERAS AT MAIN INTERSECTIONS Major: COMPUTER SCIENCE Major code: 8480101 MASTER’S THESIS HO CHI MINH CITY, month 07 year 2023 THIS THESIS IS COMPLETED AT HO CHI MINH CITY UNIVERSITY OF TECHNOLOGY – VNU-HCM Supervisor: Assoc Prof Ph D Tran Minh Quang Examiner 1: Assoc Prof Ph D Nguyen Van Vu Examiner 2: Assoc Prof Ph D Nguyen Tuan Dang This master’s thesis is defended at HCM City University of Technology, VNU- HCM City on July 11th 2023 Master’s Thesis Committee: (Please write down full name and academic rank of each member of the Master’s Thesis Committee) Chairman: Assoc Prof Ph D Le Hong Trang Secretary: Ph D Phan Trong Nhan Review 1: Assoc Prof Ph D Nguyen Van Vu Review 2: Assoc Prof Ph D Nguyen Tuan Dang Member: Assoc Prof Ph D Tran Minh Quang Approval of the Chairman of Master’s Thesis Committee and Dean of Faculty of Computer Science and Engineering after the thesis being corrected (If any) CHAIRMAN OF THESIS COMMITTEE HEAD OF FACULTY OF COMPUTER SCIENCE AND ENGINEERING i VIETNAM NATIONAL UNIVERSITY - HO CHI MINH CITY HO CHI MINH CITY UNIVERSITY OF TECHNOLOGY SOCIALIST REPUBLIC OF VIETNAM Independence – Freedom - Happiness THE TASK SHEET OF MASTER’S THESIS Full name : Bui Minh Hieu Student ID : 2170461 Date of birth : 17/02/1997 Place of birth : HCM Major : Computer science Major ID : 8480101 I THESIS TITLE : RESEARCH AND DEVELOP SOLUTIONS TO ESTIMATE TRAFFIC DENSITY FROM TRAFFIC CAMERAS AT MAIN INTERSECTIONS (NGHIÊN CỨU, XÂY DỰNG CÁC PHÉP ƯỚC LƯỢNG MẬT ĐỘ GIAO THÔNG DỰA VÀO DỮ LIỆU CAMERA Ở NHỮNG NÚT GIAO THÔNG QUAN TRỌNG) II TASKS AND CONTENTS : The objective of this thesis is to research and develop a method for estimating traffic density at intersection areas using images or videos Accordingly, the tasks involved in this work include determining the calculation method and evaluating traffic density for a given area, comparing existing papers and systems with the proposed method to identify any differences, proposing solutions to address challenges, and advancing the calculation method of the thesis Additionally, designing a prototype system to demonstrate the functionality will be undertaken III THESIS START DAY : (According to the decision on assignment of Master’s thesis) 06/09/2022 IV THESIS COMPLETION DAY : (According to the decision on assignment of Master’s thesis) 08/06/2023 V SUPERVISOR : (Please fill in the supervisor’s full name and academic rank) Assoc Prof Ph.D Tran Minh Quang HCM City, June 8th 2023 SUPERVISOR (Full name and signature) CHAIR OF PROGRAM COMMITTEE (Full name and signature) HEAD OF FACULTY OF COMPUTER SCIENCE AND ENGINEERING (Full name and signature) ii ACKNOWLEDGEMENT First and foremost, I would want to express my gratitude and extend my heartfelt gratitude to Assoc Prof Ph D Tran Minh Quang, the instructor who has completely guided and helped me to accomplish my thesis I'd want to thank the Computer Science and Engineering instructors, as well as the teachers who generously shared their knowledge with me throughout my time at the HCM city University of Technology Finally, I'd want to thank my family and friends, who have always supported and encouraged me throughout the process of doing this research HCM City, June 8th 2023 Bui Minh Hieu iii ABSTRACT Traffic congestion has become a pressing issue, not only for the government but also for the general public, as it directly impacts the quality of life It is necessary to research and develop solutions to manage traffic conditions and minimize the impact of congestion Understanding this, this study provides readers with a method to estimate traffic conditions, specifically traffic density at intersections To achieve this, the research team divided the task into two main components: vehicle counting and area measurement Vehicle counting is relatively simple with the assistance of modern technologies and techniques In contrast, and notably, this study offers readers methods to calculate the area of a region from images, without relying on camera specifications, based on the concept of object reference Additionally, through this research, we also present the design of the testing system, the challenges faced, and the accompanying solutions during the design process, as well as the results achieved when applied in a real-world environment Finally, alongside the expected outcomes, some limitations need to be discussed and addressed in the future Keyword: Traffic density, Area of a region, Object reference iv TÓM TẮT LUẬN VĂN THẠC SĨ Tắc đường trở thành vấn đề cấp bách, khơng phủ mà cịn với người dân, ảnh hưởng trực tiếp đến chất lượng sống Việc nghiên cứu phát triển giải pháp quản lý tình trạng giao thông để giảm thiểu tác động ùn tắc trở nên cần thiết Nhằm để hiểu rõ vấn đề nêu, nghiên cứu cung cấp cho độc giả phương pháp để ước tính, đánh giá điều kiện giao thơng, cụ thể tính mật độ giao thơng ngã tư Để đạt mục tiêu này, nhóm nghiên cứu chia cơng việc thành hai thành mục chính: đếm số phương tiện đo diện tích khu vực Việc đếm số phương tiện đơn giản với hỗ trợ công nghệ kỹ thuật đại Ngược lại, đo diện tích khu vực lại khó khăn thử thách hơn, nghiên cứu đề xuất phương pháp tính tốn diện tích khu vực từ hình ảnh, mà khơng dựa vào thơng số máy ảnh, dựa khái niệm tham chiếu đối tượng Bên cạnh đó, thơng qua nghiên cứu này, giới thiệu kiến trúc hệ thống mà thiết kế, thách thức gặp phải trình chạy hệ thống giải pháp kèm, kết đạt áp dụng môi trường thực tế Cuối cùng, bên cạnh kết dự kiến, số hạn chế cần thảo luận giải tương lai Từ khóa: Mật độ giao thơng, Diện tích khu vực, Tham chiếu đối tượng v THE COMMITMENT I confirm that this is my research The data utilized in the thesis's complete analytic process has a clear and transparent provenance, and it was released in compliance with scientific research standards and ethics In this thesis, I have presented the results of my study openly and fairly The thesis results are presented in this report for the first time and have not been published in any earlier thesis HCM City, June 8th 2023 Bui Minh Hieu vi TABLE OF CONTENTS I INTRODUCTION 1.1 Research problem 1.2 Objectives of the topic 1.3 Scope of study 1.4 Scientific and practical significances 1.4.1 Practical significance 1.4.2 Academy significance II THEORETICAL BASIC 2.1 Definition of traffic density 2.2 Definition of level of service (LOS) 2.3 Definition of computer vision - machine learning 2.3.1 Haar-cascade 2.3.1.1 Calculating Haar Features 2.3.1.2 Creating Integral Images 2.3.1.3 Adaboost Training 10 2.2.1.4 Implementing Cascading Classifiers 10 2.3.2 Convolutional neural network models 11 2.3.2.1 Convolution layer 12 2.3.2.2 Pooling layer 12 2.3.2.3 Fully connected layer 13 2.3.3 YOLO 13 2.3.3.1 Residual blocks 14 2.3.3.2 Bounding box regression 14 2.3.3.3 Intersection over union (IOU) 15 2.3 Definition of pixel per meter 16 III RELATED WORKS 18 3.1 Traffic situation in Ho Chi Minh City 18 3.1.1 Overview of traffic situation in Ho Chi Minh City 18 3.1.2 Statistics of damage 19 3.2 Vehicles detect and count approaches 20 3.2.1 Foreground extraction 20 3.2.2 Haar-cascade 21 3.2.3 Convolutional neural network 23 vii 3.2.4 YOLO 24 3.3 Object size measuring methods 25 3.3.1 Math-based calculating method 26 3.3.2 Object reference method 29 3.4 Traffic density calculating 30 IV PROPOSED SOLUTIONS 32 4.1 Calculate traffic density 32 4.2 Vehicle counting and categorizing 33 4.2.1 Vehicle counting 33 4.2.2 Vehicle classifying and converting 34 4.3 Calculate intersection area 36 4.3.1 Distance-based method 38 4.3.2 Mean-based method 43 V EXPERIMENTAL RESULT 46 5.1 Experiment setup environment 46 5.2 Experimental system architecture 47 5.2.1 Data collection module 47 5.2.2 Training server module 48 5.2.3 Data analysis 50 5.2.4 Diagnosis 51 5.2.4.1 Obtain real area approach 51 5.2.4.2 Calculate error rate 52 5.3 Result from training and detecting 53 5.4 Result from inferring intersection area 58 5.5 Result from evaluating traffic density 61 VI DISCUSSION AND FURTHER RESEARCH 63 6.1 Achievement 63 6.2 Limitation of the study 64 6.3 Recommendations for further research 65 PUBLICATIONS 66 REFERENCES 67 I INTRODUCTION 1.1 Research problem Urbanization is understood as the process of urban expansion expressed as a percentage of the urban area or population over the total area or population of an area or region Moreover, urbanization is also considered a huge development process, improving quality of life, maintaining a balanced population, controlling population density, etc By the end of June 2021, the coverage rate of urban zoning planning compared to construction land area in urban areas across the country will reach about 53%, in which special urban areas (Hanoi and Ho Chi Minh City) and 19 grade I cities reach about 80–90%, and in urban areas of grades II, III, and IV, about 40–50% The detailed coverage rate of urban planning is about 39% compared to the area of construction land [1] According to several recent reports, the urbanization of Vietnam is at a gradually increasing pace, with the percentage of the country’s coverage reaching 40% in 2019 [2] This process of urbanization brings many benefits to a country, such as accelerating economic growth, shifting labor and economic structures, and changing population distribution Cities are not only big consumers of goods but also places to create job opportunities and income for workers [3] Consequently, the necessity for travel has led to an increase in the number of means of transportation, which has increased traffic congestion as a result of the large cities' rapid population growth Traffic jams have always been a nuisance for every citizen in urbanized cities to cope with since it is uncomfortable to travel, and for Vietnam's governments to deal with since it not only costs a lot of money and consideration to establish an effective plan to solve the problem but is also very dangerous if left as is, as transport within Vietnam will be delayed and the economy will be affected due to such circumstances [3] It is estimated that traffic jams in one of the most urbanized cities, Ho Chi Minh City, can damage the government budget 58 5.4 Result from inferring intersection area According to Fig 36, after detecting the intersection region from Ba Huyen Thanh Quan – Vo Thi Sau, since only two motorcycles are detected, the distancebased method is used to calculate the ratio based on object reference (the motorcycle with “RefObj” text is the reference) As a result, the intersection area inferred from the calculation is returned as 68.7 m2, whereas the real one is 73 m2 Therefore, the error rate in this sample is 5.89%, with a traffic density of 0.029 Meanwhile, Fig 37 illustrates the results when applying the mean-based method In the figure, the blue color represents cars, and the green color represents motorcycles The obtained result for the calculated area is 78.11 m2, with an error rate of 7% and a traffic density of 0.166 Fig 36 Distance-based method is applied Fig 37 Mean-based method is applied 59 In addition to calculating the error rate, we also conducted experiments by running the two methods in section 4.3 multiple times in real-time to verify their effectiveness Firstly, to assess the effectiveness of the distance-based method, we ran 1000 samples Since this method requires an environment with less traffic, we conducted the experiments during off-peak hours (from 11:00 to 14:00) From Fig 38, it can be observed that choosing the center of the image as the anchor point for blocking conditions is relatively accurate As the distance approaches zero, the error rate decreases gradually However, it can be observed that in some cases, the distance is very low but the error rate is still high To explain this phenomenon, we examined and realized that the numerator (area_pixelintersection) in Eq (14) is the cause When everything is stable, YOLO v7 plays a major role in controlling this error Specifically, if YOLO v7 returns incorrect results for detecting the intersection area, then Eq (15) will produce incorrect area results The consequence of this is a high error rate To address this issue, enhancing the training for the model will be necessary in the future Based on this reason, we recognize that the problem has shifted to another problem, which is improving the performance of the model Fig 38 1000 times running distance-based method 60 Fig 39 1000 times running mean-based method For the mean-based method, the essence of this method is that the number of vehicles will affect the calculation process Therefore, we don't need to be concerned about choosing a specific time with fewer vehicles Instead, we also conducted experiments with 1000 samples throughout the morning (from 7:00 to 11:00) In this method, we expect that as the number of vehicles increases, the error rate will gradually decrease due to the influence of Eq (18) Through Fig 39, it can be seen that our expectation is not wrong when the number of vehicles falls between and 10, and the error rate does not skyrocket to 100% However, the stability of this method is not high yet The reason is derived from the analysis in section 4.3.2 Additionally, similar to the distance-based method, this method is also affected by the area_pixelintersection in Eq (14), which is determined by YOLO v7 61 5.5 Result from evaluating traffic density In this section, we focus on analyzing the results of the traffic density calculation as mentioned above Since the research goal is to draw conclusions about the traffic situation based on the calculated results, we conducted experiments to collect 3000 real-time data at the intersection of Ba Huyen Thanh Quan - Vo Thi Sau Each sample required 16 seconds for the transportation department's server to generate a new image, resulting in a total execution time of 13.3 hours An important point to note here is that the "number of motorcycles" refers to the total number of motorcycles after the conversion step in Eq (11) Table Survey between traffic density and traffic status Number of motorcycles Traffic density Level 0~5 ~ 0.055 Free ~ 13 0.075 ~ 0.21 Normal 14 ~ 0.22 ~ Busy The reason we experimented with the intersection of Ba Huyen Thanh Quan Vo Thi Sau is that the camera system there always operates without being turned off, unlike other cameras Unfortunately, the camera system at the Ly Chinh Thang Truong Dinh intersection experienced that issue Additionally, we experimented continuously for 13.3 hours to examine the changes in traffic density throughout the day and draw conclusions from it Based on Table 3, we propose to divide the traffic conditions at the intersection into three levels, following the idea mentioned in section 3.4: fast-flowing, normalflowing, and slow-flowing To classify these levels, we conducted observations and self-evaluations based on vehicle counts For example, at the intersection area, if there are only to motorcycles, it will be classified as fast-flowing Similarly, if there are to 13 motorcycles, the area will be classified as normal-flowing, and finally, it will be classified as slow-flowing when the number of motorcycles exceeds 14 In 62 addition to vehicle counting, we also recorded the corresponding values of traffic density To explain why we had to use the observational method to evaluate the level in Table 3, the definition of level of service needs to be reconsidered as mentioned in section 2.2 Although traffic density is a parameter that helps assess the level of service, there is no specific formula to represent this relationship Through [46] and [47], we realize that the relationship between these two parameters is relative Specifically, as traffic density increases, it indicates that the area or road segment is more congested, resulting in a lower level of service due to lower vehicle speeds and excessive traffic volume Conversely, if traffic density is low, it indicates that the area or road segment is less crowded, resulting in a higher level of service However, no specific relationship has been established, so we cannot directly connect the value of traffic density with the level of service Instead, we present the level category based on subjective evaluations 63 VI DISCUSSION AND FURTHER RESEARCH In this section, we would first like to present the results that we have obtained during the research process, including both scientific and practical outcomes Alongside the expected results, there are also remaining limitations that this study has yet to address Finally, we would like to propose future development suggestions for the system, as well as ideas for future research 6.1 Achievement Firstly, related to scientific result, this thesis offers scientific methods to evaluate congestion status at intersections in Ho Chi Minh City The outcomes of this research have the potential to make the following contributions:  Introducing an approach within Python scripts to create new instances that automatically access web databases, enabling more convenient data collection through command execution  Conducting experiments that demonstrate how the position of an object impacts pixel intensity, providing a means to describe the object  Presenting methods to calculate parameters like pixel-per-meter, which are independent of physical specifications These methods utilize the concept of object reference to convert area from pixels to square meters  Offering alternative approaches, distinct from the level of service (LOS), to predict traffic status at intersections by calculating traffic density  Developing a prototype system for automated data collection and processing, addressing the issue of staffing limitations  Providing the necessary sample size for training models, along with techniques for labeling, segmenting, and managing data As evident from section 5.4, the error rate remains acceptable, at approximately 5%, for the distance-based method However, the mean-based method still requires improvement, as it exhibits an error rate of about 20% Regarding practical result, the thesis also contributes a way to provide information about traffic 64 status, especially at intersection regions, for citizens living in Ho Chi Minh City As a result, people can identify and adjust their travel routes to avoid congestion Not only does it bring direct benefits in terms of avoiding and minimizing traffic jams, but the results of this thesis also contribute indirectly to improving people's lives by reducing environmental pollution, improving fuel economy, shortening travel time, etc Moreover, this thesis supports the government in terms of managing traffic status by providing a report and evaluating traffic density at intersection regions in Ho Chi Minh City Furthermore, the thesis provides another way to use CCTV cameras to exploit their full potential beyond security purposes 6.2 Limitation of the study Despite achieving expected results, this research still has some remaining weaknesses that need to be examined and improved in the future: - Collecting 50 samples for each application of YOLO v7 at a new intersection still does not meet the goal of "generalizing the problem." Therefore, it is necessary to gather a sufficient amount of data for the system to achieve better results - For the distance-based method, more samples still need to be collected to demonstrate the relationship mentioned in Fig 18 - Further research is needed for the mean-based method, as its mathematical formula remains undiscovered Additionally, it lacks a stopping condition to halt the conversion ratio update process, resulting in area calculation reliance on two parameters: the region area in pixel units from YOLO and the conversion ratio - Currently, this research has not yet addressed the issue of multiple cameras representing and collecting data for an intersection The current system still lacks a "connection" mechanism for the cameras 65 - Lastly, in terms of the system aspect, there are still existing issues such as NULL data It is necessary to consider and further develop "Diagnosis" mechanisms to detect and eliminate such data at the input stage 6.3 Recommendations for further research In terms of technicality, it is necessary to find and improve the algorithms, together with approaches, for the object reference concept Currently, the error rate from calculating is still high because the result depends heavily on the YOLO v7 detection result If there is a mechanism to support the model’s result, it could reduce the impact of the YOLO v7 detection task on the density calculation In addition, techniques related to image processing should be considered to be added before training and after performing running predictions to obtain a better result instead of depending only on the detection from the model Besides, applying image processing can also help the model and system release whenever the camera view is changed and perform diagnosis and re-calibration Finally, the system is proposed to be integrated into the Intelligent Transportation System (ITS), especially UTraffic, to have additional input sources for performing traffic status analysis and updating the map state In terms of management, it is expected that the Department of Transportation can improve the way data is accessed Currently, performing analysis and evaluating the time cost of computation for real-time cases is very hard since one can obtain only low-quality images It is suggested that the government improve the system to apply more advanced techniques related to real-time situations Finally, regarding the mean-based method, to address the issue of too many vehicles in the intersection area leading to fragmented intersection images, one possible solution to consider is using images of the intersection area with fewer vehicles to determine the intersection region On the other hand, vehicle counting and traffic density calculation will still be performed using real-time images This solution demonstrates great potential, although it still needs careful consideration and thorough experimentation 66 PUBLICATIONS During the research process, I and the research team wrote papers and had them published in conferences as listed below International Conferences H M Bui and Q T Minh, “Traffic Density Estimation at Intersections via Image-Based Object Reference Method,” in The 2nd International Conference on Intelligence of Things 2023 - Internet of Things Applications (ICIT – 2023), Ho Chi Minh, Vietnam, 2023 67 REFERENCES [1] S Nguyen “Tỷ lệ thị hóa tồn quốc đạt 40,4%.” Internet: https://baotainguyenmoitruong.vn/ty-le-do-thi-hoa-toan-quoc-dat-40-4328230.html, July 26, 2021 [2] Anonymous “Đơ thị hố gì? Q trình ảnh hưởng thị hóa.” Internet: https://online.vinhomes.vn/do-thi-hoa-la-gi, Nov 4, 2022 [3] Anonymous “Xu hướng thị hóa Việt Nam giai đoạn 2021-2030 số hệ lụy.” Internet: https://atschool.vn/blogs/mon-dia-li/xu-huong-do-thihoa-o-viet-nam-giai-doan-2021-2030-va-mot-so-he-luy, Dec 1, 2021 [4] D Thu “TP.HCM thiệt hại tỉ USD năm ùn tắc.” Internet: https://tuoitre.vn/tp-hcm-co-the-thiet-hai-6-ti-usd-moi-nam-do-un-tac20220929133058583.htm, Sep 29, 2022 [5] Anonymous “Ùn tắc giao thông Hà Nội gây thiệt hại nặng nề kinh tế.” Internet: https://antv.gov.vn/tin-tuc/xa-hoi/un-tac-giao-thong-tai-ha-noi-gaythiet-hai-nang-ne-ve-kinh-te-474690.html, Jan 3, 2023 [6] H M Tan et al., "Traffic Condition Estimation Based on Historical Data Analysis," in 2020 IEEE Eighth International Conference on Communications and Electronics (ICCE), Phu Quoc Island, Vietnam, 2021, pp 256-261 [7] Anonymous “Levels of service.” Internet: https://en.wikipedia.org/wiki/Levels_of_service, Nov 2, 2022 [8] H M Tan “Xây dựng hệ thống dự đốn tình trạng giao thông TP.HCM,” Master's thesis, HCMC University of Technology - Vietnam National University, Vietnam, 2020 [9] N N Ngoc “Nghiên cứu mơ hình khuyến khích chia sẻ liệu từ cộng đồng, áp dụng vào tốn phịng chống ùn tắc giao thông Tp HCM,” Master's thesis, HCMC University of Technology - Vietnam National University, Vietnam, 2018 [10] Anonymous “Machine Learning in Computer Vision.” Internet: https://fullscale.io/blog/machine-learning-computer-vision/, May 4, 2019 68 [11] P Viola and M Jones, "Rapid object detection using a boosted cascade of simple features," in Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition CVPR 2001, Kauai, HI, USA, 2001, pp I-I [12] A Vidhya “Haar Cascades, Explained.” Internet: https://medium.com/analytics-vidhya/haar-cascades-explained38210e57970d, Dec 20, 2020 [13] R Yamashita et al., “Convolutional neural networks: an overview and application in radiology,” Insights into Imaging, vol 09, pp 611–629, June 2018 [Online] Available: SrpingerOpen, https://insightsimaging.springeropen.com [14] “Convolutional DataBricks Layer.” Internet: https://www.databricks.com/glossary/convolutional-layer, May 14, 2020 [15] J Redmon et al., "You Only Look Once: Unified, Real-Time Object Detection," 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, 2016, pp 779-788 [16] A Vidhya “IOU (Intersection over Union).” Internet: medium.com/analyticsvidhya/iou-intersection-over-union-705a39e7acef, Jan 17, 2021 [17] W Chien-Yao et al., “YOLOv7: Trainable bag-of-freebies sets new state-ofthe-art for real-time object detectors,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2023, June 20-22, 2023, Canada [Online] Available: IEEE Xplore, http://www.ieee.org [18] “Traffic Shanmukha Internet: density.” https://knowledge4civil.wordpress.com/2017/03/05/traffic-density/, Mar 5, 2017 [19] M Rouse “Pixel.” Internet: https://www.techopedia.com/definition/24012/pixel, Aug 31, 2020 [20] Anonymous “What is PPM (Pixels Per Meter)?” www.infinitioptics.com/glossary/ppm-pixels-meter, Apr 1, 2023 Internet: 69 [21] H Minh “Giao Thông TPHCM Gần Chạm Mức Cảnh Báo Cao Nhất.” Internet: https://thesaigontimes.vn/giao-thong-tphcm-da-gan-cham-muc- canh-bao-cao-nhat/, Oct 5, 2022 [22] T Dao “TP.HCM Phát Sinh Điểm Ùn Tắc Giao Thông Mới.” Internet: https://plo.vn/tphcm-phat-sinh-6-diem-un-tac-giao-thong-moipost725183.html, Mar 23, 2023 [23] H Nguyen “Ùn Tắc Giao Thông Khiến TPHCM Thiệt Hại Tỉ USD/Năm.” Internet: https://laodong.vn/giao-thong/un-tac-giao-thong-khien-tphcm-thiethai-6-ti-usdnam-1067354.ldo, July 12, 2022 [24] K Srivathsa D S and Kamalraj R., “Vehicle detection and counting of a vehicle using Opencv,” in International Research Journal of Modernization in Engineering Technology and Science, vol 03, no 05, p 04, 2021 [25] A P Gulati “Vehicle Detection and Counting System using OpenCV.” Internet: www.analyticsvidhya.com/blog/2021/12/vehicle-detection-and- counting-system-using-opencv/, Dec 31, 2021 [26] T K N Nguyen “Luật an ninh mạng 2018.” Internet: https://thuvienphapluat.vn/van-ban/Cong-nghe-thong-tin/Luat-an-ninhmang-2018-351416.aspx, June 12, 2018 [27] H Tahir et al., “Performance Analysis and Comparison of Faster R-CNN, Mask R-CNN and ResNet50 for the Detection and Counting of Vehicles,” in 2021 International Conference on Computing, Communication, and Intelligent System, ICCCIS, Greater Noida, India, 2021, pp 587-594 [28] C S Asha and A V Narasimhadhan, “Vehicle Counting for Traffic Management System using YOLO and Correlation Filter,” in 2018 IEEE International Conference on Electronics, Computing and Communication Technologies, CONECCT, Bangalore, India, 2018, pp 1-6 [29] V X Can et al., “Vehicle detection and counting under mixed traffic conditions in vietnam using yolov4,” in International Journal of Advanced Research in Engineering and Technology, IJARET, vol 12, no 2, pp 722-730, Feb 2021 70 [30] J Vipin et al., “Traffic density estimation from highly noisy image sources,” in Transportation Research Board 91st Annual Meeting, Washington, DC United States, 2012, no 12–1849 [31] A Rosebrock “Measuring Size of Objects in an Image with OpenCV.” Internet: https://pyimagesearch.com/2016/03/28/measuring-size-of-objects- in-an-image-with-opencv/, Mar 28, 2016 [32] B Szyk and A “Traffic Benn Density Calculator,” Internet: www.omnicalculator.com/everyday-life/traffic-density, Jun 05, 2023 [33] D Prasad et al., “HOG, LBP and SVM based Traffic Density Estimation at Intersection,” in 2019 IEEE Pune Section International Conference, PuneCon, Pune, India, 2019, pp 1-5 [34] K Mahesh and J Lingaiah, “Density based traffic signal system,” in International Journal of Scientific Engineering and Technology Research, IJSETR, India, vol 5, no 38, 2016, pp 7845-7853 [35] Lahinta, F et al., “Vehicle Detection and Counting to Identify Traffic Density in The Intersection of Road Using Image Processing,” in 1st International Conference on Science and Technology, ICOST, Makassar, Indonesia, 2019 [36] B Dwyer “Train YOLOv7 on a custom dataset.” Internet: https://blog.roboflow.com/yolov7-custom-dataset-training-tutorial/, Jul 13, 2022 [37] “Image Anonymous Moments.” Internet: https://theailearner.com/2020/10/16/image-moments/, Jan 15, 2021 [38] Anonymous “GPS Fields Area Measure.” Internet: https://apps.apple.com/us/app/gps-fields-area-measure/id1123033235, Jun 1, 2023 [39] R Peters “Using Python and wget to Download Web Pages and Files.” Internet: https://www.scrapingbee.com/blog/python-wget/, Feb 2, 2022 [40] P S Udimudi “YOLO Algorithm.” Internet: https://www.linkedin.com/pulse/yolo-algorithm-prasanna-swethaudimudi?trk=pulse-article_more-articles_related-content-card, Apr 29, 2022 71 [41] Anonymous “Kích thước xe máy.” Internet: https://bilparking.com.vn/article/kich-thuoc-bai-do-xe-may-tieu-chuan, Aug 10, 2020 [42] Anonymous “Kích thước xe hơi.” Internet: https://nghiencar.com/kich-thuocxe-hoi/, Dec 7, 2022 [43] Anonymous “Kích thước xe tải.” Internet: https://vanchuyenachau.com.vn/tin-tuc/kich-thuoc-cac-loai-xe-trong-van-tai/, Dec 1, 2019 [44] Anonymous “Kích thước xe buýt.” Internet: https://shac.vn/kich-thuoc-xebuyt-cac-loai, Sep 21, 2022 [45] M Traore “Roboflow Train: Understanding Training Graphs.” Internet: https://help.roboflow.com/faqs/roboflow-train-understanding-traininggraphs, Oct 8, 2022 [46] S Penny “Level of Service: Defining Scores for Different Transportation Facilities.” Internet: https://www.smatstraffic.com/2021/07/26/level-of- service/, Jul 26, 2021 [47] Gustavo de Andrade et al (2022, Nov.) Highway Capacity Manual (6th edition) [Online] Available: https://mctranswordpresssa.blob.core.windows.net/uploads/2023/01/D4-TO22-137-HCM7RG-FINAL-HCM7-Reference-Guide.pdf 72 VITA Full name: Bui Minh Hieu Date of birth: 17/02/1997 Place of birth: TP.HCM Address: 92 Nguyễn Hữu Cảnh, phường 22, quận Bình Thạnh TRAINING PROCESS Educational Place Degree Field Year of graduation Ho Chi Master Minh City University of Computer Science 2023 Automation 2021 Technology (HCMUT) Ho Chi Bachelor Minh City University of Technology (HCMUT) WORKING PROCESS Time Position Organizational Work Company Design software 2022 ~ on-going Embedded architecture and Engineer developing software for Forvia lightning system 2020 ~ 2022 Developing and Robert Embedded introducing safety Bosch Engineer system for chassis Engineering system Vietnam