Vehicles Detection On Expressway Via Deep Learning: Single Shot Multibox Object Detector

Abstract

Nowadays, the growing number of vehicles in Taiwan, traffic jams on the expressway, especially on holidays, and during rush hours are increasing. To solve this problem, the first requirement is to detect the vehicles on the expressway. However, vehicle detection is a challenging problem due to among other things, small size of the vehicles, low resolution videos, high speeds of vehicles, and the long distance between the camera and the vehicles. Recently, deep learning-based approaches have shown extraordinary performance in many computer vision tasks. That is, the deep learning model: SSD (Single Shot MultiBox Detector) is applied and fine-turnedto detect vehicles in expressway. From the experimental results described herein, it is demonstrated that the method presented can detect more vehicles than the pre-trained SSD model.