Description of Invention
The growing number of registered vehicles in Malaysia emphasizes the importance of vehicle detection for effective traffic management. Advanced Driver Assistance Systems (ADAS) have become common in premium cars, aiming to prevent accidents. However, the ADAS system requires high resolution cameras and sensors but most Malaysian cars are only equipped with dashboard cameras with mid- or low-range resolution. Hence, computer vision plays a crucial role in processing image data efficiently, aiding in object detection and identification. This project presents a lightweight approach using deep learning for vehicle specifications, including number plates, vehicle types, and brands detection under poor visibility.