I recently saw a video of a man who had used the thermal camera on his drone to help track down and rescue a small boy who had wandered off and was lost. This showed me the potential of object detection/classification from the aerial standpoint. I wanted to create a project that could take this advance in technology and help others in need, such as survivors or victims of natural disasters.
What it does
QuakeFind is a two part application. The first part is a mobile app that can potentially be put on the bottom of a drone or any UAV. This mobile app uses machine learning object detection to determine if there is a human on the floor of a wreckage. If there happens to be a person that is injured in that location, the mobile apps records its current location and sends it out to a database. The second part of the application is the web portion. A website that has access to the database of locations that the aerial camera has recorded is used to display these place on a map. The time, latitude, longitude, and full Google Maps links are also provided on the website for a certain location. Together, both parts of the project can help streamline the search and rescue of individuals.
How I built it
The app was built using Android Studio and Java. Google Firebase and ML Kit APIs were used to store the locations and detect humans respectively. For the website, I used Flask along with Python, HTML, and CSS. Google Firebase and Google Maps APIs were used to retrieve locations and display maps.
Challenges I ran into
The biggest challenge that I ran into was getting the locations to store in the Google Firebase Realtime Database. This was because the object detection and API requests needed to line up in order to properly write out the location. Also, dealing with streams in Python for the website was difficult.
Accomplishments that I'm proud of
I am proud of the way the UI came out. I think it looks visually pleasing.
What I learned
I learned how to work with Google Firebase and ML Kit as I had never worked with them before. Also, I learned a lot about about Android layouts because getting the camera to function properly played a major part.
What's next for QuakeFind
I would like to test the app and website out on a real UAV outdoors to see how it performs under real conditions.