Vision-based Vehicle Detection and Distance Estimation

Abstract

Real-time vehicle detection is one of the most important topics under the Autonomous Vehicles (AVs) research paradigm and traffic surveillance. Detecting vehicles and estimating their distances are essential to ensure that the vehicles can keep a safe distance and run safely on the roads. The technology can also be utilized to determine traffic flow and estimate vehicle speed. In this paper, we apply two different deep learning models and compare their performances in detecting vehicles such as cars and trucks for deployment on the self-driving cars to ensure road safety. Our models are based on YOLOv4 and Faster R-CNN which are efficient and accurate in object detection within a given distance. We also propose a vision-based distance estimation algorithm to estimate other vehicles’ distances. In detecting vehicles within 100 meters, the two variations of our models, YOLOv4 and Faster R-CNN, achieved 99.16% and 95.47% mean precision, and 79.36% and 85.54% Fl-measure respectively on a two-way road. The detection speed is 68 fps and 14 fps for YOLOv4 and Faster R-CNN respectively.