基于树莓派的发动机孔探缺陷智能识别方法研究

Research on Intelligent Detection Method of Engine Hole Detection Defects Based on Raspberry PI

设计一种辅助孔探人员识别发动机缺陷的远程智能便携式系统。该系统采用树莓派作为前端,接收孔探仪的视频信号,视频信号在树莓派中经逐帧编码后,通过局域网被发送至服务器端。服务器接收到视频数据,运行部署好的神经网络模型,对视频帧中的孔探缺陷进行识别、标记。服务器在完成缺陷检测后再将带有标记的视频帧重复上述过程传输到树莓派,在树莓派UI界面显示。最后对某型航空发动机进行实时孔探检测,可成功显示实时识别结果,从而验证在工业检测的应用场景下,所提方法的适用性和有效性。

A remote intelligent portable system is proposed to assist borer to identify engine defects.The system uses rasp⁃berry PI as the front end to receive the video signal from the pothole detector.The video signal is encoded frame by frame in rasp⁃berry PI and sent to the server through local area network.After receiving the video data,the server runs the deployed neural net⁃work model to identify and mark the pore-detection defects in the video frame.After the defect detection is completed,the server transmits the marked video frame to raspberry PI and displays it in raspberry PI UI.Through the actual pore-detection of an aeroengine,the pore-detection defects can be identified successfully in real time,so as to verify the applicability and effectiveness of the proposed method in the application scenarios of industrial testing.