We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
A learning approach for garbage and debris identification
Summary
This study developed machine learning approaches for detecting garbage and plastic debris in river and ocean environments from imagery. The authors present detection models and discuss the challenges of building training datasets for environmental plastic monitoring systems.
Plastic pollution in the sea is causing several environmental and economic problems all over the world. It is estimated that around 8 million tonnes of plastic every year end up in the sea, the vast majority coming from rivers. In the recent years some interesting approaches to detect these plastics have been proposed using satellite images. However, the difficulty in obtaining datasets makes it a difficult task for classifiers using modern deep learning approaches. The purpose of this project is to develop new tools for the estimation of garbage patches in seas and rivers. Machine learning algorithms that can interpret satellite hyperspectral imaging show viability for identification and monitoring of the plastic debris. In this thesis we will create a dataset with suspected regions of garbage patches and test different deep learning approaches to classify the dataset created. Moreover, we propose a modification for the U-Net architecture and feature reduction techniques for this task.