Dataclusterlabspvtltd/Domestic-Trash-Garbage-Dataset

该数据集包含印度各地城市和农村环境中用手机拍摄的真实家庭垃圾和废物图像。它非常适合用于训练垃圾检测、材料分类和垃圾分类模型，应用于智能废物管理、自动回收线、市政清洁度监测、环保消费者应用和碳足迹估算系统。此外，它还针对生成式AI、视觉问答（VQA）、图像分类和大型多模态模型（LMM）开发进行了优化，为在真实世界废物图像上实现稳健模型性能提供了坚实基础。场景涵盖了多种捕捉情境——家庭垃圾桶、路边垃圾堆、垃圾场、回收中心、公共空间和室内废物容器——在不同的光照条件（白天、夜晚）、距离和材料视角下。数据集捕捉了印度家庭环境中发现的各种垃圾物品的多样性，包括塑料、纸张、金属、有机废物和混合材料。

This dataset contains real-world images of domestic trash and garbage, captured on mobile phones across diverse urban and rural environments throughout India. It is well-suited for training trash detection, material classification, and garbage segregation models used in smart waste management, automated recycling lines, municipal cleanliness monitoring, eco-friendly consumer apps, and carbon-footprint estimation systems. It is also optimized for Generative AI, Visual Question Answering (VQA), Image Classification, and Large Multimodal Model (LMM) development, providing a strong basis for achieving robust model performance on real-world waste imagery. Scenes cover a wide variety of capture scenarios — household trash bins, street-side garbage piles, dump yards, recycling centers, public spaces, and indoor waste containers — under varied lighting conditions (day, night), distances, and material viewpoints. The dataset captures the full diversity of trash items found in Indian domestic settings, including plastics, paper, metals, organic waste, and mixed materials.