Traffic Counting using YOLO Version-5 (A case study of Jakarta-Cikampek Toll Road)

The Jakarta-Cikampek toll road is the main access to the Tanjung Priok port, which is connected directly via the Cilincing-Tanjung Priuk Port toll road as a development of the North Jakarta reclamation coastal area. YOLO (You Only Look Once) is a common object detection model that offers faster and more accurate results.. The purpose of this article is to use advancements in information technology to automate the process of manually recording traffic counts on the highway. The method utilized in this study was to record a video of traffic movements with a smartphone camera and save it in MP4 format. Calculations are performed at the office after receiving recorded video and utilizing a program written by the author that makes use of Python, OpenCV, Pytorch, and YOLO version 5 software. When passing through a counter box, the traffic volume is counted and saved in Excel format (.xls). The video records footage near the Tambun area of the Jakarta-Cikampek toll road. According to the measurement accuracy of 95% for cars, 96% for buses, and 89% for trucks respectively, it can be stated that using YOLO version 5 for detecting vehicle volume and categorization is fairly satisfactory.


Introduction
The Jakarta-Cikampek toll road is the main access to the Tanjung Priok port which is connected directly via the Cilincing-Tanjung Priuk Port toll road as a development of the North Jakarta coastal area, reclamation of North Jakarta coastal area, and West Jakarta coastal area included Indah Kapuk coastal area developments.
Information and computing technology is currently developing very quickly according to Industrial Era 4.0.The use of technological advances also has an impact on archiving important documents, which have changed from paper archives to digital archives, especially with a good deal of free storage media facilities or the presence of local computer hard disks, computer server hard disks, and online network hard disks such as Google-Drive, Mediafire, OneDrive, Datacomm Cloud Service, and others.
The development of this technology is also rapidly growing in use in the fields of Artificial Intelligence [12], Machine Learning (ML)[3], and Computer Vision, which boosts the development of artificial human intelligence that can respond by talking to the respondent.Even developments have penetrated the computer assembly industry, car assembly, and television assembly so they affect human resources.Also, in the world of the hotel industry, some have used human robots as hotel services.In the future developments will penetrate the world of the automotive industry, especially in the case of autonomous vehicles, which have been actively developed lately, even though older generations of vehicles already exist, especially luxury cars.
After introducing artificial intelligence and machine learning, which include deep learning techniques, Python, Deepsort, TensorFlow, and YOLO) version 5 [4] (You Only Look One) and others, digital traffic volume calculations will be easier and more accurate.YOLO is a technology of computer vision for detecting objects that is very important for various applications these days.YOLO is a model for identifying objects quickly and most precisely and is a free resource.The capacity of the model to detect in the YOLO new version has been improved both in providing detection accuracy and increased detection speed as well as higher image resolution detection capabilities compared to versions of YOLO v5 and below.
YOLO works to imitate the work of the human brain or artificial intelligence [5], by carrying out detection using an artificial neural network method or Convolution Neural Network (CNN).Therefore, YOLO uses data used for recognition or training (trained), the next process uses additional data as data for the validation process, which will produce an object classification that has currently achieved 80 detections.object.YOLOv5 is a much-loved program at the vision level of AI, developed by Ultralytics as open-source.Because the process is simple, in the future, likely, AI vision methods will likely always be used in the fields of college learning, automotive industry development, and so on.
Yolo-V5 has developed object detection with a total group of 80 (eighty) object classifications obtained from identification and validation using Microsoft Common Objects in Context (COCO) data in 2015 [14].It is used as a single neural network [6] to process the whole image.The image is divided into several boxes and an algorithm predicts the probability and bounding box for each box.In this study, the utilization of YOLO-V5 will be studied, to detect objects, especially vehicles which are only detected in 5 (five) groups of vehicles, namely motorcycles, cars, buses, and trucks.Traffic volume calculations are important in overcoming congestion at the road section [7]

Methods
The study began with the collection of primary data, specifically video data, in the field using a smartphone with a resolution of 1200 pixels and a memory card with a capacity of 128 GB for storing recordings.Video samples were obtained by filming them on the scene and then processing them with a computer at the office.
This research was conducted on the Jakarta -Cikampek Toll road, specifically at Tambun, Bekasi City, West Java, in July 2023.The location of the survey can be seen in Figure 1.The placement of the location of the camera and the equipment used can be seen in Figure 2.

Result and Discussion
Data analysis was carried out by software application using OpenCV  The survey results of video capture of traffic conditions on the Jakarta-Cikampek Toll Road, near Tambun, Bekasi City, West Java acquired video data with a duration of 2 (two) hours.The video data obtained is traffic data collected during the morning rush hour for two hours, from 07.00 to 09.00 (Western Indonesia Time).The initial step before continuing to count the number of vehicles passing through a road lane is to mark a box at the beginning of the frame, which will be used to mark the lane through which the vehicle is passing, as shown in Figure 4. Manual calculations are also performed to compare the level of accuracy acquired from software versus manual calculations on the Jakarta-Cikampek toll road at Tambun between 7:00 a.m. and 09:00 a.m.The accuracy results of calculating traffic volume are provided in Table 1.

Conclusion
According to the measurement accuracy of 95% for cars, 96% for buses, and 89% for trucks respectively, it can be stated that using YOLO version 5 for detecting vehicle volume and categorization is fairly satisfactory.In the future, this paper suggests that vehicles be classified according to the Bina Marga 7 (seven) standard traffic classifications: motorcycle, automobile, small bus, large bus, small truck, medium truck, and large truck is suggested to be improved.

Acknowledgement
The author is thankful to Jayabaya University for helping me with research funding assistance.
[8],Python, OpenXL, and YOLO  version 5 [9].This study used the YOLO version 5 software application to analyze data from videos recorded from the Jakarta-Cikampek toll road to quantify the traffic volume on the roadway.Software algorithms such as Python, YOLO, OpenCV, NumPy, and OpenXL are used to do calculations of Traffic volume.The calculation results will be saved in an Excel file using the OpenXL function.The goal of this study was to compare the results to manual data traffic counts and volume traffic counts by YOLO version 5 in a different lane.

Figure 4 .
Figure 4. Traffic counts by using open-access software/applications.

Table 1 .
Comparison of Manual Traffic count and Software YOLO version 5.
Gündüz M Ş and Işık G 2023 A new YOLO-based method for real-time crowd detection from video and performance analysis of YOLO models Journal of Real-Time Image Processing 20(1) [5] Lee, H, Yu, D, Kim T, H Kim and S H Hwang 2022 Predictive Intelligent Driver Model using Deep Learning-based Prediction of Surrounding Vehicle Trajectory In IJCAI AI4AD Workshop [6] Paszke ,A, S Gross F, Massa, Lerer, A, J Bradbury, G Chanan,... and Chintala S 2019 PyTorch: An imperative style, high-performance deep learning library Advances in Neural Information Processing Systems 32 [7] R Mudiyono 2018 Flyover Model to Overcome Congestion at The Junction of Jatingaleh Road Semarang-Indonesia, International Journal Of Engineering Sciences & Research Technology 7(12) pp 2277-9655 [8] J Redmon, S Divvala, R Girshick and A Farhadi 2016 You only look once: Unified, real-time object detection Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition pp 779-788 [9] J Redmon and A Farhadi 2018 YOLO v.3 Tech Report pp 1-6