Light Lens

Inspiration

What it does

We have a software that uses a camera to recognize how many cars are on the street (i.e North-South and East-West). If there are no cars/pedestrians in one of those streets and many cars in the other while the light is red, the software switches the lights to green in the direction that has the queue of cars. The software sends a constant stream of data to the computer which then analyzes the data and sends it to the cloud server. The cloud server then outputs a binary value to the lights where a 0 means that there are a lot of cars in the N-S directions so the red light should switch to green and a 1 means that there are lot of cars going in the E-W direction so their lights should be switched to green and other direction red.

How we built it

We had a software created using tensorflow to find if there are any cars or pedestrians in the field of vision. Then the information goes to the particle board (electron) which analyzes the frames sent to it and outputs a value to the google cloud which finally outputs a value to the circuit telling it what lights to switch or to stay the same.

Challenges we ran into

We had some issues with the wireless configuration of the raspberry pi, downloading python onto the Qualcomm board, and accessing the SQL database on Azure. When trying to configure the raspberry pi, we had connected it to an Ethernet cable to a laptop so that it would get faster speeds but it would not connect to the networks and after spending a couple hours we had given up on it. The Qualcomm dragonboard was hard to configure as whenever we downloaded python 3.6 it would only show up as 3.4.2 which caused many problems and we could not continue with the rest of the necessary steps to finish the problem. Finally we had trouble accessing the SQL in Azure

Accomplishments that we're proud of

We believe our diorama depiction of how it will be used is very well executed given the time and we are extremely happy with he results that it is outputting. We are also proud of ourselves for making this more challenging by applying concepts we have just learned as a basis of our whole project.

What we learned

We learned that by adding too many different hardware it may not work well together and may over-complicate the whole solution by "biting off more than we can chew". We also learned that under certain conditions, we are capable of doing things that we would never have thought possible since we started the competition. It is also safe to say that we are also very capable in handling new concepts thrown at us with minimal stress.

What's next for Light Lens

If given the opportunity, we would like to continuously update this and try to optimize its functions so that it can be useful for cities around the world. Our goal is to limit traffic times and make it more sustainable/cheaper for cities to use.