In order to detect a car on the image, we need to identify feature(s) which uniquely represent a car. We could try using simple template matching or relaying on color features but these methods are not robust enough when it comes to changing perspectives and shapes of the object.
In order to have a robust feature set and increase our accuracy rate we will be using Histogram of Oriented Gradients (HOG). This feature descriptor is much more resilient to the dynamics of the traffic. In essence, you should think of features as thumbprints of the objects you are interested in.