Defect Detection and Calibration System Based on Embedded AI
In order to solve the problem of large error and high missing rate of surface defect identification by manual and traditional methods,this paper designs a defect detection and calibration system based on embedded AI,which uses machine vision technology to replace traditional detection,and can accurately identify and mark the defect part of the board.The system is based on YOLOv5 target detection algorithm and introduces SE attention mechanism,uses 20,275 wood surface defect images of the pipeline as the data set,and deploy-es the model to the embedded AI platform.When the defect is detected,the defect information is sent to the mechanical arm,and the pixel coordinates are converted to the actual coordinates to realize the calibration of the defect position.
万方数据
维普
