We were given this task as our final year project and would like to share our project in this competition.
What it does
We modified a drone called Blacklion-068 from the Aerolion Technologies to perform monitoring & inspection functions. By implementing various functions of autonomous flight using lidar and range sensor, smart detection using software implementation, and intelligent payload dispensing.
How we built it
For Smart detection, we used TensorFlow and OpenCV to train data and code to achieve crack, color, shape, and number detections. We used ROS to achieve Autonomous flight by programming the waypoints and coordinates into the Operating system. In this case Ubuntu.
Challenges we ran into
We faced issues like being given a weak router, which gave us problems in giving commands from our laptop to the drone. We solved it by changing the router to a newer one, which required us to change the codes in the drone and commands to allow everything to run.
Accomplishments that we're proud of
Our drone is able to successfully fly autonomously. Our detections are able to detect accurately. Cracks. Colors. Numbers.
What we learned
Learned how to work with AI, work with robotics, programming in python.
What's next for Smart detection with autonomous UAV
We are hoping this project can be taken outdoors to do crack detection on the outside and also be able to scan the whole building with just information about the building's height, width, and breadth.