*Background

The marine animals have died from the affect of trash that people throwing to the ocean. We consider this as a problem with people's attitudes. So, how can we solve it? Our idea is we have to decrease the number of existed trash as well as making people realize about it and will not throw it to the ocean again. By our system and web game, people can interact with the real problem at home. We hope that our system can reduce the number of trash and make people realizing how much does the trash in the ocean today.

*Resources

garbage dataset for CNN classification model training:

https://github.com/garythung/trashnet

Tools we use:

MATLAB, Simulink

Tensorflow python

React javascript

Powerdirector

Source code and Demo video:

1. about Machine Learning's trash classification network

https://github.com/yihliang831209/NASAapps2019_WhatOnEarth

2. about website

https://github.com/r06942072/Great-Pacific-Garbage-Patch

3. Demo video

https://drive.google.com/drive/folders/1w-VknpFo3B...

*Challenges faced

1. For conveying meaningful and educational information to people, we want to use natural language processing(NLP) technique from some keywords like "plastic" plus "ocean" among websites, which is difficult to start implementing.

2. Connecting three works into one complete project is challenging:

For boat moving analysis, we use matlab.

For image recognition and classification, we use python tensorflow framework.

For interactive website, we use react javascript frontend library.

3. The real data is hard to implement so we use the basic data to demonstrate.

4. How to get the real dynamics of the boat real time, localizeation+sensor?

5. Connection between the real world and the game.

*final presentation, PPT

https://drive.google.com/file/d/1KHxcCxGh4USO06PJY...