vehicle-velocity-estimator

This projects aims to estimate the velocity of vehicles moving on road from the CCTV footage of the road.

The implementation uses Lucas Kanade algorithm along with Optical Flow for generating a set of good points and tracking then for estimating the velocity of the moving object. Background removal using KNN algorithm is used for instantiating the tracking process whenever a new object is detected.

Requirements

• OpenCV2

sudo apt-get install python-opencv

• Python2.x
• numpy

sudo pip install numpy

Instructions to run the code

• Inside global_variables.py replace video_path with the path of your video
• Replace top_left and bottom_right with the coordinates of the box inside which you want to track the object
• Run the file named automatic_track.py using python2 or simply python automatic_track.py
• Use ESC key to exit from the video