FMF-Benchmark
收藏资源简介:
FMF-Benchmark是由东北大学流程工业综合自动化国家重点实验室创建的一个用于融合镁熔炼过程异常检测的跨模态数据集。该数据集包含超过2.2百万个同步采集的视频和电流数据样本,覆盖了1000多个小时的工业生产视频。数据集的创建过程涉及从不同生产批次中选择和整理数据,以确保数据的质量和多样性。该数据集主要用于开发和测试跨模态学习算法,特别是在极端干扰如电流波动和视觉遮挡情况下的异常检测,旨在提高融合镁熔炼过程的安全性和效率。
FMF-Benchmark is a cross-modal dataset for fusion-based anomaly detection in magnesium smelting processes, developed by the State Key Laboratory of Integrated Automation of Process Industry, Northeastern University. It contains over 2.2 million synchronously collected video and current data samples, covering more than 1,000 hours of industrial production video recordings. The dataset construction process involves selecting and curating data from different production batches to ensure its quality and diversity. This dataset is primarily used to develop and test cross-modal learning algorithms, especially for anomaly detection under extreme disturbances such as current fluctuations and visual occlusions, with the goal of improving the safety and efficiency of fusion-based magnesium smelting processes.

- 1Cross-Modal Learning for Anomaly Detection in Fused Magnesium Smelting Process: Methodology and Benchmark东北大学流程工业综合自动化国家重点实验室 · 2024年



