High-Level Project Summary
I have made a Trash Plastic Detection system. It comes with both CLI and web versions. Embedded computers with Satellites, drones, submarines, etc. detect and send pictures of trash plastic to a database. It can detect plastics from a video, and send them to a server. A web interface also available where we can upload video, and trash pastics of the videos shown on website.
Link to Project "Demo"
Link to Final Project
Detailed Project Description
This project detects trash plastics in the ocean.
How does it work?
Method 1. A Computer integrated on Satellites or Drones or Submarine run ocean.py which takes a photograph of the ocean and whenever plastic is detected it will send it to a server. Go to https://oceanplastic.herokuapp.com/ to see images
Method 2. Satellites or Drones, Submarine collect video of ocean. Pass this video through ocean.py. All images of plastics will send to a database. Go to https://oceanplastic.herokuapp.com/ to see images
Go to the following Link for Details :
https://github.com/ujjwalkar0/Garuda/tree/master/Plastic%20Detector
Method 3: Collect a video of the ocean and upload it to https://oceanplastic.herokuapp.com/upload. Go to oceanplastic.heroku.app to see images
What benefits does it have?
If we make a similar system that detects ocean life and sends images to the same server, and show these in an app, people interested in ocean life download the app. People see how we waste our earth, and they reduce the use of plastic. This is the benefit.
What do you hope to achieve?
My hope was to complete the project, which I have done already.
Technology Uses : Python, Opencv, Django, Bootstrap, and Most important thing Yolo object detection.
Future Plan: Use Reinforcement Learning
Space Agency Data
Get started with https://towardsdatascience.com/identifying-and-removing-marine-plastic-a-deep-learning-approach-6ad5e41bec41 this article. From there I go to https://github.com/gautamtata/DeepPlastic Github repo. Studying notebook I find some difficulties and https://youtu.be/yGMZOD44GrI this video help me a lot. Then I find the dataset https://drive.google.com/drive/folders/1tk4HDftOVzmROHpkWRUIwL0Gu_p2E-qa , by which I generate .cfg and continue
Hackathon Journey
Nice Experience
References
Get started with: https://github.com/gautamtata/DeepPlastic
Dataset: https://drive.google.com/drive/folders/1tk4HDftOVzmROHpkWRUIwL0Gu_p2E-qa
Youtube: https://youtu.be/yGMZOD44GrI and others...
Tags
#yolo4, #object_detection, #computer_vision, #opencv
Global Judging
This project has been submitted for consideration during the Judging process.