Skip to content

A-Chaudhary/lexington-traffic

Repository files navigation

lexington-traffic

Downloads real-time traffic camera video from any specific intersection in the Lexington-Fayette Urban County from https://trafficvid.lexingtonky.gov/publicmap/

YOLO_v4 on the Darknet framework is run to detect cars in the video feed.

Object Detection Output for yolov4.weights

Object Tracking with DeepSORT on Daytime and Nighttime Footage

Daytime

Nighttime

Custom Allowed Classes & Count Per Camera Segmentation

Daytime

Nighttime

Movement Based Labeling

Daytime

Nighttime

Calculated Total Time Spent Wasted Sitting at Trafficless Stop Light

Daytime

Nighttime

TODO

  • Extract Source of Lexington Traffic Cameras
  • Save Video File from Parsed chunklist
  • Find Traffic Camera with Optimal Positioning
  • Prevent Repeat Video Files with Different Source
  • Setup CUDA, OpenCV, Darknet, and Yolo v4
  • Run Object Detection with Darknet on Video
  • Convert yolov4.weights to Tensorflow
  • Apply DeepSORT for Object Tracking
  • Determine Object Movement Boolean
  • Calculate Benefit from Reduced Car Idling
  • Extrapolate to City Wide Scale

References

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages