2025年全国家电电子玻璃瑕疵图谱数据集

该数据为杭州地区企业杭州聚玻科技有限公司在为玻璃加工厂提供工业视觉瑕疵检测服务过程中产生。数据由布设于产线的高分辨率工业相机实时采集，涵盖了玻璃生产全流程中的表面缺陷、微裂纹、杂质及形变等关键信息。其数据形态主要为视频和图片，数据结构既包含半结构化数据也包括非结构化数据。
上述工业数据可广泛应用于玻璃加工厂的质量管控与追溯、生产与工艺优化、智能预测、知识管理和决策支持等业务场景。通过数据沉淀与建模，可支撑建立高精度的质量检测体系，助力企业实现对每一块玻璃的全生命周期追溯；在生产环节，数据驱动的算法能够优化工艺参数、减少能源消耗、降低次品率。

This dataset was generated by Hangzhou Jubo Technology Co., Ltd., a local enterprise in Hangzhou, during the provision of industrial visual defect detection services for glass processing plants. The data is collected in real-time by high-resolution industrial cameras deployed on production lines, covering key information such as surface defects, micro-cracks, impurities and deformations throughout the entire glass production process. The main data modalities are videos and images, and the data structure includes both semi-structured and unstructured data. This industrial dataset can be widely applied to business scenarios such as quality control and traceability, production and process optimization, intelligent prediction, knowledge management and decision support for glass processing plants. Through data accumulation and modeling, it can support the establishment of a high-precision quality inspection system, enabling enterprises to achieve full-life-cycle traceability of each glass panel. In the production stage, data-driven algorithms can optimize process parameters, reduce energy consumption and lower the defective product rate.

搜集汇总

背景与挑战

背景概述

该数据集由工业相机采集于玻璃生产线，包含表面缺陷、裂纹等多种瑕疵的高清图像及视频数据，主要用于支持玻璃制造业的质量检测优化和工艺参数改进。数据可助力实现产品全生命周期追溯，并通过算法模型降低生产能耗与次品率。