Ferrogram image intelligent identification method based on auto-encoding network

The invention discloses a ferrogram image intelligent identification method based on an auto-encoding network, which comprises the following steps: extracting a lubricating oil sample for mechanical equipment, making a ferrogram through a ferrogram analyzer, converting the ferrogram into a digital image through microscope imaging to make a fault diagnosis sample, and designing a stack auto-encoding network on the basis of the fault diagnosis sample; and completing ferrogram image intelligent classification by using the network, and further identifying the type of the wear fault. According to the invention, ferrogram image intelligent identification based on the self-encoding network is realized, and the method has the characteristics of intelligence and high efficiency for equipment wear fault diagnosis. 本发明公开了一种基于自编码网络的铁谱图像智能识别方法，先针对机械设备提取润滑油样本，进而通过铁谱分析仪进行谱片制作，然后将谱片经显微镜呈像而转为数字图像制作故障诊断样本，在此基础上设计一种堆栈自编码网络，利用该网络完成铁谱图像智能分类，进而识别出磨损故障的类型。本发明实现了基于自编码网络的铁谱图像智能识别，对设备磨损故障诊断具有智能和高效的特点。