Inspiration
Our inspiration for this project was how many people across the world have car crashes and the amount of time for help to get to them is too much. Since there are already models to detect whether there was a car crash, we wanted to create a one-stop-shop that contained information about many things.
What it does
Our project uses CCTV footage to detect a crash on security cameras that will send a message that there is a crash. Taking this information, we also use geolocation to determine the current longitude and latitude of the crash and send this to a database that we created using firebase. This database contains all the information about all previous crashes imported into the database and all crashes that occur while our program starts and these are sent to the website. The website will show the location of crashes on a map using google maps API and police officers can input their location to find the route between their location and the crash to get there as fast as possible. Along with this, we will have a heatmap that uses a dataset found online with previous crashes and their latitude and longitude for the general public to see.
How we built it
We created a python file that scans a video and outputs a crash when a crash occurs. We then built a database using firebase that does all the backend for the website. We then used javascript to create a front-end website that uses the google maps API to see the google map with the locations. We also created a heatmap on javascript but didn't have enough time to display it on the website as connections between the two would have to be done.
Challenges we ran into
We ran into many challenges during our tenure with this hackathon. Our first challenge is we tried to get a machine learning model on GitHub as this wasn't the part of the system that we built, however, our laptops couldn't handle the software issues with TensorFlow and Keras. We met with many mentors but could never fix this issue. We created a program that acts as the ML model that outputs a crash at the correct time. Another issue we faced was working with javascript for the heatmap as we couldn't get it to show up on the website. We did get far enough to where we have a dataset that was split into coordinates and plotted using javascript, but we didn't integrate it into our website yet.
Accomplishments that we're proud of
We are proud of creating a database that stores all the data and can access both the python files side of our project and the javascript and HTML side of our project. Additionally, we are proud that we carried out an idea had a lot of features that can help people learn more about crashes and help police carry out their job more efficiently, saving lives. We believe that we did a very ambitious project that tested the limits of our understanding of concepts within programming.
What we learned
We learned many things about the google maps API including javascript. Also, we learned how to connect a database to both an HTML/javascript side of our project and a python/ML side of our project.
What's next for CrashNet
We have many things to look forward to for CrashNet. We would like to design our website to look more user-friendly. Also, we would like to get our updates fixed on our computers so that the ML side of our project that was taken from GitHub could work with our created website and database. Finally, we would like to improve on our heatmap as we created a heatmap but ran into an error that could be solved with a little more javascript expertise.
Log in or sign up for Devpost to join the conversation.