Image Dataset for Classification of Water Quality Based on Total Suspended Solids Using Convolutional Neural Networks

This dataset contains 6,518 labeled images of water samples classified into three pollution levels (low, medium, and high) based on total suspended solids (TSS) concentrations. The dataset was generated under controlled experimental conditions using image acquisition from video recordings of water samples prepared with varying concentrations of suspended solids. Images were extracted, filtered, and processed to ensure quality and consistency, forming a structured dataset suitable for machine learning and computer vision applications. The dataset supports the development and validation of artificial intelligence models, particularly convolutional neural networks, for automated water quality classification. This dataset is associated with the research article focused on the classification of water pollution using deep learning techniques and contributes to the advancement of low-cost, non-invasive monitoring methods for environmental systems.本数据集包含6518张带标注的水样图像，基于总悬浮固体（Total Suspended Solids, TSS）浓度将其划分为低、中、高三个污染等级。本数据集通过受控实验流程生成：针对配制不同浓度悬浮固体的水样录制视频，从中提取并采集图像。 研究人员对图像进行提取、筛选与预处理，以保障数据质量与一致性，最终构建出适用于机器学习与计算机视觉任务的结构化数据集。该数据集可用于开发与验证人工智能模型（尤其是卷积神经网络（Convolutional Neural Networks）），以实现自动化水质分类任务。 本数据集与一篇采用深度学习技术开展水污染分类研究的学术论文相关联，可为环境系统低成本、非侵入式监测方法的发展提供支撑。