Marine plastic waste is a large environmental problem whose extent isn’t always easy to quantify. This project attempted to approximate the amount of plastic waste in the oceans by measuring how much plastic is washed ashore on beaches and extrapolate the overall amount of plastic waste based on that.
To do so, I developed a deep learning based object detection framework, based on the popular YOLO framework to compute the location and type of plastic waste on images gathered by drones. Additionally, using a SLAM approach based on OKVIS I localized the plastic waste in 3D space and for larger waste estimated the volume of each piece.
This work was part of The Plastic Tide project.