Smart Dustbin on BrainyPi Using EdgeAI

Edge AI is the deployment of AI applications in devices throughout the physical world. It’s called “edge AI” because the AI computation is done near the user at the edge of the network, close to where the data is located, rather than centrally in a cloud computing facility or private data center.

In this project we have implemented Smart dustbin which will detect the waste materials and will help in segregation.

I had applied for an early access to BrainyPi as it is right now available for selected audience only.

For more resources and documentation refer https://brainypi.com/docs/13-opencv-examples/

• BrainyPi
• Terminal

We are on the track to making smart cities but are neglecting the most common issue on this path, waste management. Waste generation is increasing daily, and cities are allocating more resources to waste collection. Smart bins eliminate unnecessary spending on waste management. They will have much more use cases than a traditional dustbin. Smart bins will help in segregating the waste into bio-degradable and non-biodegradable and also eliminate the risk of overflowing bins by sending the alarm on phone. We will continuously monitor the bins and the data collected will also help in better waste management. It is perfect for areas with a large volume of human traffic such as parks, academic institutes, malls, hospitals, etc.

This project is implemented on BrainyPi which is an alternative to Raspberry Pi4.

Data set for training the model is collected from various sources such as Kaggle, Youtube videos, Google images.

Data set was preprocessed and finally dataset was annotated.

for training the model we have ssd MobileNet v1.

Loss function during the model training:

First, we will establish the remote connection to BrainyPi

Remote Connection is established by the command

ssh -X pi@auth.iotiot.in -p 65530

After entering Password remote connection will be established.

Clone the git repository for the smart-dustbin model by running the command

git clone https://gitlab.iotiot.in/PIAI23JAN1015/smart-dustbin.git

Changed the directory to smart-dustbin by command:

cd smart-dustbin

After cloning, Its time to run the model. Before running the python file, we need to download the required libraries in BrainPi eg: object-detection-api, numpy, tensorflow etc.

Run the model by command:

python app.py

To view the output go to the the folder saved_detections in images folder. There you will find the detect images with labelled output.

Our Application has been successfully deployed and run on BrainyPi

• https://brainypi.com/docs/13-opencv-examples/
• https://gitlab.iotiot.in/PIAI23JAN1015/smart-dustbin