Awards & Nominations
Project AMPHITRITE: Ocean4Good has received the following awards and nominations. Way to go!
Global Nominee
Project AMPHITRITE
High-Level Project Summary
Team Ocean4Good aims to build a web app for monitoring and classifying the type of plastic marine debris using crowdsourced photos of aquatic waste and satellite data for garbage patches, presenting figures of environmental impact.AMPHITRITE tackles SDG 14: Living Below Water, by addressing the plastic pollution in Manila Bay, which was found to be contributing about 0.28-0.75 million tonnes of plastic waste per year (SEA Circular, 2020) - ranking third among the top contributors of plastic pollution in the global marine ecosystem. The project is expected to serve as tool to help our policymakers in formulating strategic actions in marine pollution management and help preserve marine life.
Link to Project "Demo"
Link to Final Project
Detailed Project Description
PROJECT OVERVIEW
AMPHITRITE is a comprehensive web dashboard that analyzes information from satellite spectral imagery, and crowdsourced images of plastic waste to quantify the amount, classify the types, and monitor hotspots of plastic debris, and measure its environmental and economic impact. The project is initially implemented in the proximity of Manila Bay, covering the province of Bulacan, and the west portion of Metro Manila.
The application uses two sources: (a) crowdsourced pictures taken from inland/ marine areas, and (b) spectral satellite data from Copernicus Sentinel-2.
OPERATION
Figure 1. Crowdsourced Photo Obtained in the Angat River, one of the primary water channels connected to Manila Bay.
Crowdsourced Pictures. The photos obtained from the local community within the coastal area of Manila Bay, and the inland river networks in Pampanga and Bulacan will undergo object detection using the AutoML Vision of the Google Cloud Platform according to the type of waste. As part of the training, the team used pre-annotated images from Taco Dataset. The information obtained from the plastic waste, alongside with the location on where the photo was taken will be retrieved by AMPITHRITE for visualization (under quantitative dashboard).
Figure 2. (left) Satellite Optical Imagery of Manila Bay, (top-right) Magnified Image within the Connection of Pasig River and Manila Bay, (bottom-right) Magnified and Enhanced Image of Garbage Patches within the Manila Bay
Satellite Images. The raster file acquired by Sentinel covers the area of intersection between the Pasig River and Manila Bay in September 2019. Garbage patches, which will be identified through the difference of water moisture levels in comparison with surface of sea/ocean (most commonly known as Normalized Difference Water Index or NDWI) with QGIS, an open-source Geographic Information Systems platform. The information will be acquired by the AMPITHRITE dashboard as well, under the geographic dashboard.
IMPACT
The developed web dashboard is expected to provide primary understanding for local communities about the value and types of waste in their locations, and provide LGUs a comprehensive tool to explore the environmental and economic impact for policymaking, as well as pave a way for strategic and efficient actions in managing the present marine garbage in Manila Bay. The team believes that achieving a safe haven for marine life as stated in SDG 14 will be made possible with the collaborative efforts of the government institutions, local community, and the advanced technologies of space agencies
ADVANTAGES OF THE PROJECT
Team Ocean4Good believes that the foundation of an effective policy is accessible and comprehensive data. With the advent of technology and the availability of resources particularly from the satellite data, researchers are able to remotely monitor plastic waste that do not require excessive consumption of resources.
LIMITATIONS OF THE PROJECT
-The team experiences constraints in the angles of plastic wastes (e.g. bottles), restricting an accurate reading for the plastic type.
-Obtained data are limited to floating plastics.
-The learning model is not capable of classifying recyclable items and nonrecyclable ones.
FUTURE RECOMMENDATIONS
-More enhanced object detection model
-Use ocean roughness to identify the location of other plastic debris, particularly microplastics
-Understand the capacity/volume of a plastic debris per square meter
-Utilize satellite data over a period of time to forecast movement of wastes
Space Agency Data
Satellite Imagery
Our team utilized raster files obtained by the Copernicus Sentinel 2 satellites in Manila Bay from September 2019 to September 2020. The acquired file included the bands necessary for determining the Normalized Difference Water Index (NDWI), which will be used to identify the patches of marine wastes over time.
Hackathon Journey
Working on a hackathon has been a roller coaster experience for our team.
References
[1] https://www.sea-circular.org/wp-content/uploads/2020/04/SEA-circular-Country-Briefing_THE-PHILIPPINES.pdf
[2] https://www.uow.edu.au/media/2020/marine-debris-costs-asia-pacific-economies-us108b-annually-report-.php#:~:text=At%20current%20estimated%20leakage%20rates,then%20this%20number%20would%20increase
[3] https://news.mongabay.com/2018/10/plastic-trash-from-the-sachet-economy-chokes-the-philippines-seas/
[4] https://www.esa.int/Enabling_Support/Preparing_for_the_Future/Discovery_and_Preparation/A_step_forward_in_detecting_plastic_marine_litter_from_space
[5]
Data Used
https://github.com/AgaMiko/waste-datasets-review
http://tacodataset.org/
https://earthexplorer.usgs.gov/
https://drive.google.com/drive/u/0/folders/1m3drr2HjBOAHBOTEv-5ob_bOkdtLdcmr
Tools Used
- Canva
- Google Cloud Platform
- React and Node.js
- MongoDB
Tags
#ProtectTheOceans#PlasticDebris
Global Judging
This project has been submitted for consideration during the Judging process.