유튜브 라이브 영상과 YOLOv5 모형을 이용한 군중 집계와 실시간 모니터링 시스템 - 리조트 영상을 중심으로 - = Crowd Counting and Real-time Monitoring System Using YouTube Live Camera and YOLOv5 Model - Focusing on Resort Images -

다국어 초록 (Multilingual Abstract)

Large gatherings pose serious problems for crowd safety and raise security concerns for the organizers. Estimating the crowd counting in a single image is very important for crowd control and crowd safety. In this study, we try to figure out the density level of people in real-time by counting the number of person with resort image of YouTube live camera. YOLOv5 was used to detect the person. As a result of comparing the YOLOv5x trained with the resort image and the YOLOv5x6 model trained with the COCO dataset, the proposed YOLOv5x custom model well detected the pattern for the change in the number of person. The results of counted by real-time are summarized in graphs and tables to detect the density level of people. The person detection and summary results were applied to a monitoring system that observes the change in the number of person over time. The presented dashboard can check real-time images, detection result images and changes in crowding. In addition, notification messages and images are sent to the mobile device, so administrators will be able to quickly check the situation.

참고문헌 (Reference)

1 최수진, "유튜브 관광콘텐츠 품질과 유튜버의 신뢰성이 지각된 즐거움, 사용자 만족 및 관광행동에 미치는 영향" 한국관광학회 44 (44): 123-145, 2020

2 김진호, "노상 주차 차량 탐지를 위한 YOLOv4 그리드 셀 조정 알고리즘" (사)디지털산업정보학회 18 (18): 31-40, 2022

3 YouTube, "[4K] Live Camera - Niseko Hanazono Resort"

4 Redmon, J., "You only look once: Unified, real-time object detection" 8 : 779-788, 2016

5 Dang, F., "YOLOWeeds : A novel benchmark of YOLO object detectors for multi-class weed detection in cotton production systems" 205 : 107655-, 2023

6 Anushkannan, N. K., "YOLO Algorithm for Helmet Detection in Industries for Safety Purpose" 225-230, 2022

7 Roy, A. M., "WilDect-YOLO : An efficient and robust computer vision-based accurate object localization model for automated endangered wildlife detection" 75 : 101919-, 2023

8 Bhuiyan, M. R., "Video analytics using deep learning for crowd analysis : a review" 81 (81): 27895-27922, 2022

9 Zheng, Z., "Toward real-time congestion measurement of passenger flow on platform screen doors based on surveillance videos analysis" 128474-, 2023

10 Ahmad, M., "Overhead view person detection using YOLO" IEEE 0627-0633, 2019

11 Punn, N. S., "Monitoring COVID-19 social distancing with person detection and tracking via fine-tuned YOLO v3 and Deepsort techniques" 2020

12 Lin, T. Y., "Microsoft coco: Common objects in context. Computer Vision–ECCV 2014" Springer International Publishing 740-755, 2014

13 GitHub, "GitHub - ultralytics/yolov5: YOLOv5"

14 Hsu, Y. W., "Estimation of the number of passengers in a bus using deep learning" 20 (20): 2178-, 2020

15 Fitwi, A., "Estimating interpersonal distance and crowd density with a single-edge camera" 10 (10): 143-, 2021

16 Al-Smadi, Y., "Early Wildfire Smoke Detection Using Different YOLO Models" 11 (11): 246-, 2023

17 He, G., "Dynamic region division for adaptive learning pedestrian counting" IEEE 1120-1125, 2019

18 Rahman, R., "Densely-Populated Traffic Detection Using YOLOv5 and Non-maximum Suppression Ensembling" Springer 567-578, 2021

19 Teoh, S. K., "Computer vision and machine learning approaches on crowd density estimation : A review" 2654 (2654): 030009-, 2023

20 COCO, "COCO - Common Objects in Context"

21 Purwar, R. K., "Analytical Study of YOLO and Its Various Versions in Crowd Counting" Springer Nature Singapore 975-989, 2021

22 Nagpal, R., "An Application of Deep Learning for Sweet Cherry Phenotyping using YOLO Object Detection" 2023

23 Pham, D. L., "A YOLO-based Real-time Packaging Defect Detection System" 217 : 886-894, 2023

24 Chandra, N., "A Human Intruder Detection System for Restricted Sensitive Areas" 1-4, 2021