In this session, you will learn to track the vehicles. But what is meant by tracking?
Here, we are trying to solve two problems – counting the number of vehicles passing through the lane and classifying them.
To solve the first problem, we have to count a vehicle only once while passing from the lane. The second problem is solved by identifying vehicles using hulls and classifying them using deep learning.
From what you have learned in the previous session, you can figure out which is a vehicle and which is not by validating the blob at every timestep. But in the next time step, you should be able to identify:
- Whether it’s the same vehicle that appeared in the previous timeframe.
- Whether it’s a different vehicle that’s appeared for the first time in the lane.
- Whether the vehicle being tracked has left the frame.
Let’s look at how these operations are carried out in the next segment.
Report an error