Object Counting on Low Quality Images: A Case Study of Near Real-Time Traffic Monitoring

The installation and management of traffic monitoring devices can be costly from both a financial and human resource point of view. It is therefore important to take advantage of available infrastructures to maximize the information extraction for each technology. Here we show how low-quality urban road traffic images from cameras, already installed in many cities such as Montreal, Vancouver and Toronto can be used as a non-intrusive traffic monitoring. To this end, we use a pre-trained object detection neural network to count vehicles within images. We then compare the results with human annotations gathered through crowdsourcing campaigns. We use this comparison to assess performance and calibrate the neural network annotations. The performance of our system allows us to consider applications which can monitor the traffic conditions in near real-time, making the counting usable for traffic-related services. Furthermore, the resulting annotations pave the way for building a historical vehicle counting dataset to be used for analysing the impact of road traffic on many city-related issues such as urban planning, security, and pollution.