Moving Object Activator in Background Subtraction Algorithm for Automatic Passenger Counter System in Public Transportation

Abstract

Buses are the most used transportation by people in Indonesia when traveling between cities. However, to improve passenger comfort, a tool is required to determine the number of passengers on the bus. This research presents an automated system to count the number of passengers based on a background subtraction algorithm and moving object activator.  This research aims to provide the number of passengers based on video images taken from the entranceway. The system is built with a camera, Raspberry Pi, and LCD. The APC system starts the counting process by removing the video background image from the captured object image. The entry and exit direction of the object is determined using the concept of moving object activator. The experiments were applied in several scenarios to determine the robustness of the system. The best APC performance was achieved when the system is positioned perpendicularly above the entranceway at a height of 230 cm and a light intensity of 800-1000 lux. Meanwhile, the moving object activator is effective in supporting the system's performance to determine the passenger's direction. In this scenario, the results stated that the accuracy of APC system performance reached 93.8%.