我要吃瓜

Keep Scotland Beautiful’s (KSB’s) vision for this KTP is to develop an automatic litter detection tool and web application that categorises litter by type and brand. Litter counting is currently done manually which is a barrier to scaling surveys. KSB, with academic expertise, will develop an AI-based tool to extract litter information from images captured remotely. Current manual processes limit data collection (quantity and locations) and interpretation, and there is a significant time delay in reporting actionable data which hampers measures to remove litter. The KTP project will enable data capture from existing and new locations with the ability to report more efficiently. The KTP will enable KSB to enhance its current services and exploit commercial opportunities associated with emerging environmental policy and regulation.

Key objectives are: ?Specify and build camera hardware requirements, develop image capture methodology; ?Develop, test and optimise a machine vision algorithm(s) for automated identification of litter; ?Develop a client-facing web-based app for reporting and visualisation of litter data; ?Pilot study to demonstrate proof-of-concept.

Implementation of AI-driven automated remote image analysis for litter reporting is a highly novel concept for KSB. The success of this project will have a significant impact on collecting, interpreting, and reducing litter