shlokashah / SEGRO

An end to end solution for waste segregation, collection and disposal

SEGRO - say grow to our environment

Considering the current scenario in our country, we see that waste disposal, segregation and collection, collectively is one of the biggest challenges faced by the authorities. This application provides for an end to end solution for waste classification,collection and disposal.

Technologies Used

1. Django Rest Framework (API for backend)
2. Flutter (Mobile Application)
3. Image Classification with CNN and YOLOv3

Waste Classification

For the purpose of waste classification we have used Image Classification which classifies waste into 6 different categories: 1)metal 2)paper 3)cardboard 4)trash 5)plastic 6)glass

Results Using CNN with ResNet Architecture

Results Using YOLOv3

Waste Segregation

For the purpose of waste segregation we have simulated the working the working of Smart Bins.

Screenshots of the simulation of Bins

Route Optimization

For route optimization of the truck for waste collection we use the TSP algorithm.

Screenshots of optimized route generated

Code Structure

1. Master branch consists of the backend api.
2. Flutter branch consists of the mobile application.
3. Classifier branch consists of CNN classifier.
4. Yolo branch consists of classifier with YOLOv3.
5. Simulator branch consists of simulation of classifier over webcam.

References and credits

1. https://github.com/AntonMu/TrainYourOwnYOLO
2. https://github.com/Cartucho/mAP