Plastic Waste Detection on Cox’s Bazar Sea Beach,Bangladesh

The Cox’s Bazar Beach Plastic Waste Detection Dataset (CoxPWD 2025) is an object-detection dataset of 4,195 RGB images (374 MB) collected along Cox’s Bazar, Bangladesh, the world’s longest natural sea beach. Images were captured during September 2025 at three popular tourist sites which are Kolatoli Beach, Sugandha Beach, and Laboni Beach. Pictures were taken using an iPhone 15 Pro Max, Google Pixel 6, and Canon 650D under varying weather (sunny, cloudy, rainy) and different time-of-day (morning, noon, afternoon) conditions. Each scene includes portrait and landscape views from multiple angles (~15° and 45° from left and right, and ~1 m and 6 m above ground) to introduce realistic variation in scale, occlusion and background.
All images are stored after resizing to 640×640 JPEGs and annotated for bounding-box object detection in COCO format with 15 classes which are packet, polythene, cup, spoon, straw, rope, bag, bottle, bottle_cap, net, sunglass, toy, fishing_item, others, objects. The dataset is split into 80% train (3,367 images), 10% validation (414) and 10% test (414) subsets, each with its own _annotations.coco.json file. This structure makes the dataset plug-and-play for YOLOv8 and other deep-learning detectors targeting coastal plastic-waste monitoring, model benchmarking, and deployment on mobile or edge devices.