Evaluating Efficiency: Speed Violator Detection Before and After Image Recognition System Installation

Road traffic accidents are a major public issue, with around 1.2 millions fatalities, mostly due to driver negligence, most commonly speeding. This paper shows the importance of detecting speed violations with Haar Cascade to track vehicle. In testing, the model detected 66 vehicles compared to 56 by humans, with actual count of 58. For violations, model identified 16, while human detects 7, with an actual count of 12. Performance was assessed with MAE, revealing human having smaller MAE for vehicle detection, while the model performing slightly better in detecting speed violations. While the model struggle with normal sized vehicle, it shows promise in detecting speed violators.