A research group focusing on the applications and development of image recognition technologies to waste management.
HeadquartersKensington High St, Kensington, London W8 5SF, UK
We are a research group (part of Imperial). Our research focuses on the applications and development of image recognition technologies to waste management. It is divided in two streams:
Stream 1: Exploring the commercial possibilities. By partnering with the British Institution of Wastes Management (CIWM) we are identifying which waste streams and plant configurations offer the highest economic potential. We are investigating the possibility for a ‘reversed’ Ocado style warehouse where all waste can be sorted and some returned to the producer (in line with the EUs producer pays scheme). The sorting purity of such a system would be a momentous competitor to todays energy intensive and inefficient mechanical based sorting processes (wind blowing, vibrating tables, etc). We are also looking into a smart ‘home bin’ version that would have the additional capability of ordering products when they are thrown in the bin.
Stream 2: Developing a machine vision sorting system (convolutional neural net) in partnership with two phd's from the computing departement. Applying machine vision to waste brings a few problems due to the distortion/damage of most objects and the excessive presence of dust and sludge. The aim is to develop a highly modular machine which can be used on a local scale and provide a high sorting purity for a wide range of applications.