Parking Detection

Parking Detector is a cloud based application. Through Alexa, users are able to ask and learn if parking is available in a location. To help users not be inconvenienced when parking.

Comment

Inspiration

Much of our teams inspiration for this project came from the frustration of driving around in the library parking lot looking for parking. In Blacksburg, in any place on campus it takes approximately 5 minutes, however, finding parking anywhere on campus takes at least 10 to 20 minutes. Our inspiration came from the JB Hunt prompt, where they suggested using Alexa to get information about where to park within a parking lot.

What it does

Parking Detector using a camera and Google Vision, takes a picture of a parking lot. With that picture and Google Vision's object detection API we are able to determine whether there are parking spots available for the user.

How we built it

We built this application by first developing our understanding of how Google Vision API and Alexa API works. From there we got our credentials for the Google Cloud and began to develop the backend code in Node.js. From there we used ngrok and our webcam to start testing how to take the camera pictures and move them to a location where we could also run the Google Vision API over it. Once we managed to do that, we began to start integrating Alexa into the process. Alexa interaction model was developed from the developer console through Amazon. However, due to compatibility issues we weren't able to fully connect the interaction model with our Google Cloud backend code.

Challenges we ran into

The main challenge that we ran into when developing this project was connect the Google Cloud Service to the Alexa device. The main problem was converting the request that Alexa sent to our backend into a format that Alexa could process when our program sends back response. Much of our struggle came from understanding what the Alexa API was saying as well as how to use the service with HTTP rather than their built-in service Lambda.

Accomplishments that we're proud of

One of the main accomplishments that our team is proud of is developing the connection between our camera and Google Vision and using the API to detect objects. Although we hit many bumps in the road, we were very pleased to see at the end of the project that the we were able to detect multiple objects from the pictures we took.

What we learned

Our biggest discovery was the ngrok platform. It's a platform that exposes local servers behind firewalls to the public internet over secure tunnels. Which made the process of creating our application and attempting to connect it to Alexa much easier. As a group we can also see a lot more applications of this software in future classroom and personal projects.

What's next for Parking Detection

The next iteration of parking detection will be able to work seamlessly with the Alexa platform as well as have a developed machine learning model that will be able to better identify cars and their locations in the parking lot. This would give us the capabilities to not only know how many cars are there, but also run analytics on the data we collect and give users data visualizations of when the parking lot is busy and when's the best time to park.

Share this project:

Updates