Back to publications

Scalable object detection pipeline for traffic cameras: Application to Tfl JamCams

Gan, Huan Min and Fernando, Senaka and Molina-Solana, Miguel
Expert Systems with Applications (under review) (SUBMITTED)


With CCTV systems being installed in the transport infrastructures of many cities, there is an abundance of data to be extracted from the footage. This paper explores the application of the YOLOv3 object detection algorithm on the Transport for London’s (TfL) JamCam feed. The result is a series of easy to deploy Docker pipelines to create, store and serve (through a REST API) data on identified objects on that feed. The pipelines can be deployed to any Linux machine with a NVIDIA GPU to support accelerated computation. We fine-tuned the system to detect car, truck and pedestrian objects in London JamCam scenes by studying how different confidence thresholds affect detections of each object. By running the system continuously for 3 weeks, we built a dataset of more than 2200 detection datapoints for each camera (~6 datapoints an hour). The detections, visualised on an animated geospatial map, prove their effectiveness in identifying traffic patterns typical of an urban city like London, portraying the variation on different object population levels throughout the day.




  author = {Gan, Huan~Min and Fernando, Senaka and Molina-Solana, Miguel},
  title = {Scalable object detection pipeline for traffic cameras: Application to Tfl JamCams},
  journal = {Expert Systems with Applications},
  year = {SUBMITTED},
  volume = {(under review)},
  doi = {},
  comment = {},
  timestamp = {31}