Computer Vision for Waste Management

Every year we generate 2 billion tonnes of waste globally, and the World Bank forecasts that this number could rise to 3.4 billion tonnes by 2050. To transform this matter into valuable resources, we need to tackle the problem of efficient recycling. The recycling process comprises multiple steps from waste collection to material transformation, but the most difficult is sorting, or recognizing value in a waste mix. At WasteFlow, we are developing a SaaS for recycling facilities to help them optimize their recycling process by understanding the waste mix at different steps of the process.

A sorting facility is composed of multiple sorting machines that isolate different types of materials, with waste collected in cells at the facility’s entry and exit. When the facility underperforms, it is very complex to understand the source of the problem. Thus, at WasteFlow we are developing a modular camera solution that can be placed at strategic points in the process, such as conveyor belts, sorting machine inputs and outputs, and waste cells, to characterize the waste mix, analyze sorting performance, and detect performance loss or material blocking risks.

Through this project, you will develop a solution for a real-world application that will be used within WasteFlow’s service and will help optimize recycling. Depending on current needs, the project may involve real-time detection and classification of waste on conveyor belts, efficiency measurements, estimating the mass of detected material, or assessing the purity of sorted output and identifying foreign materials. Solutions will need to deliver real-time results and remain robust to vision challenges such as illumination variations, overlapping materials, and the wide diversity of waste types. Deploying models on edge devices to improve scalability is also of interest.

The work will be done in collaboration with a member of WasteFlow team and a research supervisor at CVLAB.

Key Domains:

• Deep Learning
• Object detection
• Recycling

Prerequisite:

• Proficiency in Python
• Basics of Computer Vision and Machine Learning with experience in Pytorch or similar libraries

Contact

• Valentin Ibars @ WasteFlow – [email protected]
• Mathieu Salzmann @ CVLab – [email protected]