One Way

My house, located within the suburbs, faces onto a one-way street. Ever since we moved in, we have always had trouble with drivers coming down the one way illegally. Now, there are cases where people do not know the area and do this without knowing. However, the rate at which people come up and down the one way, with the increasing population and traffic problems are making the situation worse. There is a stop street at the one end of the one way which people often do not stop at, never mind look up the one way to see if someone is coming down the one way. This situation is illustrated by the following image. The dotted line over here indicates the part where the road transitions from a one-way to a normal road.

Naturally, as you can see, my house is relatively near the transition. So, technically, I could simply go down the one way. I don’t. I go up the road and around the triangular block to go in some directions. However, as you can see, by the time the cars get to this point on the road, they have picked up some significant speed; if they know they are in the wrong, then it is highly likely that they are trying to get out of that section as fast as possible as well. This, has not yet (to my knowledge), led to any accidents.

I am going to attempt to create a system to try and stop this. Or at least have some fun tracking these people going down the one way.

The advent of networked cameras has made this a possibility. All that I need to do is be able to run some cable down to the gate, set up a camera, and set up some sort of tracking system for these cars.

At the moment, I have created a repository to make use of for the process.

An added benefit to installing cameras down by the gate is that we will then actually have a camera down by the gate to see when visitors are outside (which is really convenient as they currently hoot at the gate).

There are also currently a number of machine learning algorithms (and many datasets out there for vehicles) that I will attempt to make use of in conjunction with OpenCV (I am going to try Python before I attempt to use Matlab) to perform this.