Our Software

The entire software has multiple components, including trash detection software and the physical drone itself. We utilize autonomous flight software (waypoint) and obstacle avoidance software to ensure that video footage is taken autonomously and safely. We set waypoints around park trails and send the drone to fly through those points autonomously, which it does via geological positioning systems (GPS) embedded in the drone. The drone is set to automatically begin videoing the path and after this flight is completed, the video is stopped and sent to the trash detection software. 

​

In the trash detection software, we train the software to recognize trash. We manually highlight the trash, which is done by the EccoDrone team, and we feed that video footage into the software. We then allow the software another test video and we determine the success rate of the trash detection software. We have inputted over 50 different types of trash from over 300 different angles so that our software can correctly identify trash. We then send these videos to our partners for them to use in planning and preparing park and beach cleanups. 

​

We are currently in the testing phase where we are manually inputting trash into the software to train the software. From current tests, the software can identify around 80% of the trashes correctly and identifies around 5% of the trashes incorrectly. With continuous training and help from our sponsors and partners, as well as efforts from the EccoDrone team, we hope to increase the reliability and effectiveness of the drone software to help organizations more efficiently clean beaches and parks.