航空发动机叶片异常检测数据集 (AeBAD)
收藏资源简介:
航空发动机叶片异常检测数据集(AeBAD)由西安交通大学机械工程学院创建,包含两个子数据集:单叶片数据集(AeBAD-S)和叶片视频异常检测数据集(AeBAD-V)。AeBAD-S包含不同尺度的单叶片图像,样本未对齐,存在训练集与测试集间的域转移问题,主要由光照和视角变化引起。AeBAD-V包含安装在航空发动机上的叶片视频,用于检测叶片在旋转过程中的异常。该数据集旨在解决实际工业应用中叶片异常检测的问题,强调同一数据类别内的域多样性。
Aero-engine Blade Anomaly Detection Dataset (AeBAD) was created by the School of Mechanical Engineering, Xi'an Jiaotong University. It consists of two sub-datasets: the Single Blade Dataset (AeBAD-S) and the Blade Video Anomaly Detection Dataset (AeBAD-V). AeBAD-S contains single-blade images of varying scales, with unaligned samples and domain shift between the training and test sets, which is mainly caused by variations in illumination and viewing angles. AeBAD-V contains videos of blades installed on aero-engines, which is used for detecting anomalies during blade rotation. This dataset aims to address the problem of blade anomaly detection in real-world industrial applications and emphasizes domain diversity within the same data category.




