Profile grinding is widely used for precision machining of curved parts, such as tools and molds. Traditional methods are manual-dependent and struggle with autonomous measurement and compensation of contour errors. This paper introduces a novel method for the direct measurement and compensation of contour errors in large-size profile grinding. A new image stitching algorithm for complex smooth curves is proposed, enabling robust image registration in low-feature regions. Furthermore, an approach for on-machine sub-pixel detection of contour errors is investigated, and an in-situ compensation algorithm for contour errors is designed. Experimental results show that the proposed method achieves high-precision stitching with minimal overlap, significantly improving contour error measurement. The compensation algorithm optimizes the tool path and implement the online compensation of the full contour errors during secondary machining, effectively improving the machining accuracy of curve parts. Additionally, a negative correlation between the number of image stitches and machining accuracy is revealed, and the possible explanations are discussed. The impact of various factors on machining accuracy will be further investigated in future studies.
Read full abstract