Drones are already fitted with so many sensors whose data can be used to solve so many problems, we wanted to hack it and access this data to solve some problems in the world.
What it does
1) Uses infrared vision to capture IR images of houses that have been subject to natural disasters and process them via IBM Watson Studio to recognize if someone is trapped inside the house (especially disabled people). 2) Uses IMU data to predict wind speed and altitude to give a wind speed distribution gradient. Can eventually be fitted with pressure sensors, temperature probes and Relative humidity sensors to more accurately predict weather phenomena.
How we built it
1) Use IMB watson to train it to recognize IR images of houses and people. 2) Hacked a drone to download log files and and processed it using MATLAB to predict wind speed and direction.
Challenges we ran into
1) How to train IBM Watson, developing a suitable database, figuring out how cloud computing and real time data could be processed. 2) How to obtain log files and turn them into readable data for MATLAB, make sense of the numbers, equations to translate data into wind speed, reducing noise and integrating live processing of data.
Accomplishments that we're proud of
We got the image recognition software to work and our data base was accurate enough. We also managed to predict the live weather data quite accurately.
What we learned
Theres so much backend data that can be used to solve problems and so many processing tools available to us to create solutions.
What's next for Droniverse
Develop a website that enables users to deploy a drone fitting their needs, integrate cloud computing and real time processing.