The Development of Recognizing Type of Vehicle System for Carbon Emission Monitoring

Abstract

A system that can recognize the type of vehicle by capturing and identifying the vehicles is developed. The image of each type of vehicle has been counted, and the carbon emission has been calculated based on the numbers of the vehicles detected, which then have been displayed on the dashboard to monitor low carbon emissions. This project's motivation is to support Malaysia's plan to reduce its carbon dioxide emissions and reach the goal of having a low-carbon city by 2030. In addition, most current systems face some difficulty recognizing, especially when using automatic number plate recognition, since the system requires data from the Road Transport Department of Malaysia (RTD), leading to user privacy issues. Therefore, the type of vehicle recognition is built to overcome the problems, especially on the main road, where the surveillance camera is placed. This system uses cameras and machine learning algorithms to identify vehicles by analyzing their shape and other visual characteristics. YOLOv7 and OpenCV-python have been applied to this system as such methods help to detect, track, count, and classify moving vehicles. Since Open-CV is Python's real-time computer vision library, the YOLO can be used directly with OpenCV. The result on functionality was tested based on the camera distance, which the detection can detect vehicles between 1 km and 3km. According to the analysis of the results, the number of carbon emissions measured every minute correlates with the total number of each type of vehicle. As a result, it shows that the number of carbon emissions increased along with the number of vehicles. Overall, this system can function well and raise awareness among road users to reduce the use of private vehicles as it can lead to higher carbon dioxide emissions to the environment.