Searching for Uncollected Litter with Computer Vision
This work addresses waste management by providing a tool to map litter for measuring strategies and trends, but it is incremental as it builds on existing datasets and methods.
The study tackled the problem of detecting uncollected litter by using computer vision on images from the TACO dataset, achieving good performance with smartphone photos but struggling with vehicle-mounted camera images, with potential for improvement through dataset diversification.
This study combines photo metadata and computer vision to quantify where uncollected litter is present. Images from the Trash Annotations in Context (TACO) dataset were used to teach an algorithm to detect 10 categories of garbage. Although it worked well with smartphone photos, it struggled when trying to process images from vehicle mounted cameras. However, increasing the variety of perspectives and backgrounds in the dataset will help it improve in unfamiliar situations. These data are plotted onto a map which, as accuracy improves, could be used for measuring waste management strategies and quantifying trends.