Real-time detection of vehicles with transition superiority using artificial intelligence methods

Object tracking and detection are among the most researched topics today due to their critical importance in areas like defense, security, medicine, robotics, autonomous vehicles, and license plate recognition. The increase in automated and card-based entry systems has reduced the need for human intervention at vehicle checkpoints, allowing only authorized vehicles to pass. Unauthorized vehicles must either contact an authority or turn back. In institutions, vehicles need authorization to pass through barriers into restricted areas. Civilian vehicles are not allowed on routes used by red-striped emergency ambulances and blue-striped patient transport ambulances, particularly in hospitals, to prevent delays caused by traffic congestion. Arm barriers are commonly used to address this issue. The aim of this thesis is to utilize artificial intelligence methods to detect vehicles with priority access at unmanned checkpoints and quickly grant them passage. The YOLO algorithm, one of the object detection algorithms, will be used to detect priority vehicles as quickly and accurately as possible. Once the vehicle is detected, an IoT system will signal the barrier to the main driver to open it, allowing the vehicle to pass through. For model training, 7216 JPG format datasets were used. The dataset was divided into 80% for training and 20% for validation. The training resulted in a mAP50 score of 98.3%.