High-Level Project Summary
Floating marine debris is a global pollution problem which leads to the loss of marine and terrestrial biodiversity. Large swaths of marine debris are also navigational hazards to ocean vessels. The use of Earth observation data and artificial intelligence techniques can revolutionize the detection of floating marine debris on satellite imagery and pave the way to a global monitoring system for controlling and preventing the accumulation of marine debris in oceans. This project seeks to reduce marine debris by detecting marine debris through satellite data.AQUA AI then sends geo-location of the debris, and also tracks the debris
Link to Final Project
Link to Project "Demo"
Detailed Project Description
Floating marine debris is a global pollution problem which leads to the loss of marine and terrestrial biodiversity. Large swaths of marine debris are also navigational hazards to ocean vessels. The use of Earth observation data and artificial intelligence techniques can revolutionize the detection of floating marine debris on satellite imagery and pave the way to a global monitoring system for controlling and preventing the accumulation of marine debris in oceans. This project seeks to reduce marine debris by detecting marine debris through satellite data.
AQUA AI then sends geo-location of the debris, and also tracks the debris
A platform to detect ocean hotspots / plastic sources and a machine / interceptor to pick up ocean waste
Our solution is an AI object detection software that detects plastic pollution hotspots in real time using satellite data. This would help create detailed maps of plastic densities in remote ocean locations to aid ocean cleanup efforts. Current datasets are built using conventional methods (trawls) which are very labor-intensive, or less conventional methods (airplane and drones) which are very costly and complex to organize. Our team is developing a more intelligent and effective manner using Artificial Intelligence (AI) specifically deep learning to detect and monitor plastic waste in real time as well as predict plastic pollution sources before they enter the ocean.
We do this by using data from the satellites in the Copernicus program and apply floating debris index (FDI), Normalized Difference Vegetation Index (NDVI) and spectral signatures to detect patches of floating plastics on the ocean surface.
We also trained the Bayesian algorithm to classify the mixed floating material and provide details on the kind of debris found. Our solution uses the eo-learn and scikit-Learn python libraries.
We accompany our solution with a web application that provides information on plastics pollution on certain remote locations and allow users to input preferred locations and get data on the extent of plastic pollution over there.
The solution targets mainly governments, environmental agencies, ocean cleanup organizations and researchers.
Our solution which is a web application integrated with an AI model allows the user to select the location of interest. The images of the selected location are sent to our AI model for analysis.
Images with a higher confident prediction score indicate marine debris in it.
Our model then uses openCV to visualize the debris by producing bounding contours around the debris.
A feedback which contains the the location of the debris, its densities and size is displayed to the user
Space Agency Data
The images were obtained from Planet Scope optical imagery which has a spatial resolution of approximately 3 meters. In this dataset, marine debris consists of floating objects on the ocean surface which can belong to one or more classes namely plastics, algae, sargassum, wood, and other artificial items. Several studies were used for data collection and validation. While a small percentage of the dataset represents the coastlines of Ghana and Greece, most of the observations surround the Bay Islands in Honduras. This data was downloaded from Radiant ML.com
NASA IMPACT is an interdisciplinary team that works to further ESDS’s goal of overseeing the lifecycle of Earth science data to maximize the scientific return of NASA's missions and experiments for research and applied scientists, decision makers, and the society at large. IMPACT’s three focus areas are interagency collaboration, assessment and evaluation, and advanced concepts.
Hackathon Journey
We are currently writing exams in school, but we decided to try our hands on something impactful. We did face challenges getting this data.
This project is a desktop app which uses machine learning. TensorFlow framework from google was used because it is ocean source and very fast
References
Shah, A., Thomas, L., & Maskey, M. (2021) "Marine Debris Dataset for Object Detection in Planetscope Imagery,RADIANT ML
Tags
#ocean_clean_up #resilient_ocean #tensorflow #keras #NASA #Copernicus