Improving Passenger Count Accuracy in Crowded Public Transit Stops

In recent years, advancements in sensing, communication, and management technologies have enhance the traditional transportation systems globally. This trend is particularly significant in densely populated countries like India, where efficient transportation management is vital for enhancing overall quality of life. With Indian cities facing numerous challenges such as overcrowding, declining air quality, and escalating traffic accidents, there is a need to optimize public transportation systems. Accurate passenger counting at crowded bus stops plays a crucial role in this optimization, enabling authorities to allocate resources efficiently and improve service quality. Leveraging existing infrastructure, such as CCTV systems, presents a cost-effective solution for gathering passenger data. Additionally, advancements in camera sensor technology and computer vision algorithms, including YOLO (“You Only Look Once”), offer promising avenues for accurate passenger counting. By training models like YOLOv5 on datasets specific to Indian bus stops, satisfactory results can be achieved in terms of precision and recall. This paper explores the application of such detection approaches to passenger counting, demonstrating the feasibility of achieving accurate estimations without the need for additional infrastructure investments. Accurate passenger counting contributes to smoother, more efficient public transportation experiences, promoting sustainability and enhancing overall commuter satisfaction.