Multiple Camera Set-up for Detection of Road Obstruction Violation

In this paper, the use of multi-camera system for contactless obstruction violation apprehension was investigated. The researchers designed a two-camera surveillance system and employed Python to process the outputs from both cameras concurrently. The first camera was placed at an elevated location to facilitate the detection of obstructions, focusing on vehicles obstructing the pedestrian lane during the red-light phase of the traffic light. The second camera was positioned on the shoulder lane dedicated to capturing the license plates for optical character recognition. For obstruction detection, the researchers utilized background subtraction algorithms using a static background element to identify obstructing vehicles. The results obtained from this dual-camera system were analyzed to assess the effectiveness of the obstruction detection process, the accuracy of the license plate detection, and the reliability of the OCR in correctly identifying the license plate numbers of vehicles violating the pedestrian lane rules. The system's performance in detecting obstructions using the background subtraction algorithm achieved an accuracy rate of 99.5%, license plate detection accuracy of 74.33% and license plate character recognition of 85.99%. Overall, the combined performance metrics of the system indicate a high level of efficiency and accuracy in monitoring traffic obstructions, detecting license plates, and correctly recognizing the characters on those plates.