Fast single-shot multi-frame aerial image traffic density analyzer Based on deep learning

Abstract

This paper studies a real-time traffic density analyzer based on target detection. Based on the development of deep learning technology in recent years and taking full advantage of big data strength, the intelligent and real-time traffic density analyzer is constructed. It collects the road condition in the current status in real time, and estimates the current number of driving vehicles and the specific positions of the corresponding vehicles through the real-time calculation and analysis of efficient target detection algorithms such as faster rcnn [1], SSD [2] and so on. It can easily be carried on unmanned aerial vehicles or fixed road stations, and communicate with other digital devices in intelligent cities in real time to reflect traffic conditions.