High-Level Project Summary
Marine debris is one of the most pervasive threats to the health of coastal areas, oceans, and waterways. Challenge is to leverage Artificial Intelligence/Machine Learning to monitor, detect.The flow of mismanaged plastic waste into our waters is threatening critical natural resources, economies, and industries around the world.Marine debris is preventable through increased public awareness, changing individuals’ behaviors, and movement towards an economy of reduced pollution and waste. Educating the AI system through constant feeding of data making it more intelligent. Modifying the existing algorithm for using less computation power and to bring productive results in less time
Link to Project "Demo"
Link to Final Project
Detailed Project Description
By power of cloud computing and big data analytics, use of algorithm like convolution neural network in artificial Intelligence. We were able to create model which can serve for governments and the private sector to identify strategic methods for building capacity in systems to detect waste and litter. Some of these may include improving infrastructure, collection systems, government coordination, and public education and engagement, incentivizing the global recycling market in partnership with the private sector; advancing research into, and the development of innovative solutions and technology; and promoting the removal of marine litter.
We have used python , numpy, pandas, matplotlib, tensorflow, keras, jupyter notebook, google collab.
Intel 8th Generation machine ,
core i5, 8gb ram, 4gb gpu memory
MS Windows operating System.
we are trying to design model which can detect marine plastic waste with lesser computational power and give productive results in less time. By creating alerts of newly found plastic waste on a global scale.
Space Agency Data
https://lpdaac.usgs.gov/products/gweldmov031/
https://earthdata.nasa.gov/eosdis/eosdis-data-news-archive/eosdis-data-news-august-2021
This was the dataset we used to train our Deep learning Aritficial Intelligendce model for various green landscape.
It helped our model to study the difference between landscape and marine debris.
Hackathon Journey
The Journey to Space App challenge is very exciting. There is lots of thing to learn from. I learned about the nasa, open source data set. Various information of upcoming nasa project. The most motivating factor was video that I saw on youtube :https://youtu.be/4wH878t78bw
A research team led by Christine Figgener (Texas A&M University) found a male olive ridley sea turtle during an in-water research trip in Costa Rica. He had a 10-12 cm PLASTIC STRAW lodged in his nostril and they removed it. This video shows graphically why plastic waste is detrimental to marine life, especially single-use plastics (such as straws, which are one of the most redundant items). This turtle suffers from an item that is human-made and used by most of us frequently.
Plastics and other debris such as bottle caps, balloons, and lighters are also ingested directly by wildlife, such as sea turtles, seabirds, and marine mammals. This made work on the project.
My approach using the power of Artificial Intelligence, latest algorithm such CNN can we design a model to detect plastic waste. which can futhur develop to send alerts such as messages, photos, videos with location details to near by authorities for people or mammals in distress. so a rescue team can be arranged immediately.
Biggest challenge was to find right dataset to train out model so that we can test it on fresh dataset. We had to gather data from nasa open source website and then merge it with other various dataset from other website mentioned below to create one big large dataset. Then we had to preprocess the dataset. Fill with the missing values. remove the outliers. To train the dataset it took around 30min to 45 minutes minutes time. We had to use jupyter notebook and then google collab to higher computation power.
I would like to thank:
1, A research team led by Christine Figgener for making us release that lots of thing can be done if we have inner desire to .
2,Nasa and Pravishya tech for presenting an oppurtunity to participate in this hacathon.
3, All my professor who help me in developing this project.
References
https://www.doi.gov/ocl/marine-debris-impacts
https://www.epa.gov/sites/default/files/2020-10/documents/marinelitter_booklet_10.16.20_v10epa.pdf
https://github.com/Chaetll/NASA-Space-app-challenge.git
https://www.nasa.gov/open/data.html
Kaggle.com
github
Tags
# AI/ML #plastic marine debris #plastic #marine debris

