AI and machine-vision litter auditing for objective measurement
#00098
Use camera surveys (handheld, vehicle- or drone-mounted) with machine-vision models to automatically detect, count and classify litter over large areas — producing objective, repeatable density data to target action and rigorously measure whether interventions worked.
Parent issue
#00081 Litter sources and hotspots are poorly measured at an actionable granularity
Sustainable Development Goals
Location
Description
Mechanism
Deploy cameras — handheld, vehicle-mounted, or airborne — with machine-vision models trained to detect, count and classify litter automatically across streets, parks or coastlines. This yields objective, repeatable measurements of litter density and composition at scale, without the labour and inconsistency of manual surveys, and can be re-run before and after an intervention to quantify its effect precisely.
Where it fits
A professional-grade complement to citizen mapping within the measurement facet, well-suited to evaluating specific interventions and monitoring large or repeated areas where rigorous, comparable numbers are needed.
Limits
Requires hardware, trained models and technical capacity; detection accuracy varies with conditions and item type. Less accessible than smartphone crowdsourcing — best used where precise measurement justifies the setup cost.
Sub-issues
0Case studies
2