Inspiration

As people who grew up in Haiti and Togo in West Africa, We've always been keenly aware of the impact that waste and pollution can have on our environment and health. We've seen firsthand how trash can pile up on streets and beaches, contaminating our water and soil, and how it can be difficult for people to know how to dispose of it properly. We wanted to create a project that could help people sort their trash more effectively and reduce waste in their communities.

What We Learned

During the course of this project, We learned a lot about machine learning and computer vision. We learned how to preprocess and augment image data to improve model performance, how to train a convolutional neural network (CNN) to classify images of trash, and how to use OpenCV to detect the color of trash cans. We also learned how to build a Flask web app to enable users to interact with the models in real-time.

How We Built the Project

To build the project, We started by gathering a dataset of images of various types of trash and preprocessed them to a fixed size and normalized the pixel values. We then trained a CNN model using TensorFlow to classify the images into six categories: cardboard, glass, metal, paper, plastic, and trash. Next, We gathered a dataset of images of trash cans with different colors and used OpenCV to train a computer vision model to detect the color of the cans. Finally, We built a Flask web app that used the trained models to classify images of trash and detect the color of trash cans.

Challenges Faced

One of the biggest challenges We faced was finding and preprocessing the image datasets. It took a lot of time and effort to find publicly available datasets of trash and trash cans that were of sufficient quality and size. We also had to spend a lot of time cleaning and augmenting the datasets to improve model performance. Another challenge was training the models on the main laptop we were using, which was not powerful enough to train the models quickly or efficiently. We had to use cloud computing resources to speed up the training process.

Statistics

As creators from Haiti and Togo in West Africa, We were motivated to create a project that could have a positive impact on my communities. According to the World Bank, in 2021, Haiti generated 5,000 tons of waste per day, and only 15% of it was collected and properly disposed of. In Togo, the situation is even more dire, with only 5-7% of waste being collected and treated. By helping people sort their trash more effectively, this project has the potential to reduce waste and pollution in these communities and improve public health.

Built With

Share this project:

Updates

Submission history