Research and Application of Artificial Intelligence in the Automatic Annotation of Contact Profilometers

This research aims at the problem of measurement data distortion in contact profilometers caused by factors such as probe contamination and scanning position deviation. An automatic annotation method based on signal processing and geometric correction is proposed. Signal denoising is achieved through Fourier transform. The derivative operation is combined to extract contour characteristic parameters, and the rotation matrix is used to correct the tilt error of the workpiece.Experiments show that, on the premise of retaining the original contour features,the annotation error of the notch width is less than 0.05 mm, and the calculation accuracy of the arc radius reaches 99.2%. This significantly improves the automatic annotation efficiency of the contours of complex workpieces. The research results provide reliable technical support for high-precision measurement in the field of industrial inspection.