Development of Vehicle Detection Technique Using Sliding Window Tracking

This project develops a vehicle detection system to detect, count and record vehicles autonomously. Vehicle detection system is used to perform traffic measurement such as tracking, detecting and monitoring vehicles. In Malaysia, Malaysian Highway Authority (MHA) or Lembaga Lebuhraya Malaysa (LLM) is a body responsible for conducting the traffic measurement to measure the congestion level of the highway. The congestion data collected can be further used by other official bodies such as traffic police in controlling traffic and Jabatan Kerja Raya (JKR) in route planning and management. However, the vehicle detection systems in Malaysia are not fully automated and many are relying on human to manually monitor the congestion. Motivated by the need of automated vehicle detection system, this project implements a vehicle detection system that consists of a video camera for capturing the vehicles on the road and a sliding window technique developed in Visual C++ to record, detect and count the vehicles. The performance is measured in term of accuracy by measuring the percentage of the detections.