pLitterStreet - Street Level Plastic Litter Detection Dataset

pLitterStreet dataset comprises of more than 13,000 images. These images were captured using vehicle-mounted cameras that were strategically positioned to focus on the sides of streets. The primary objective of this dataset is to facilitate research related to street litter and its impact on the environment.

Annotations for the images are provided in the widely-used Microsoft COCO JSON format. Image in the dataset is fully annotated, enabling the identification and categorization of various types of litter found along urban and rural streets. These annotations include precise labeling of litter items, making the dataset an invaluable resource for developing and evaluating object detection and image recognition models.