A synthetically photogrammetric system for precise contour parameters with bi-telecentric imaging

The traditional method is inefficient and expensive, and the traditional machine vision measurement method has issues with inadequate measurement accuracy and low stability. However, the contour distance measurement of the traditional polygonal workpiece plays an important role in ensuring its processing quality. In this research, a high-precision measuring method for polygonal workpiece contour distance based on photographic bi-telecentric lens imaging is created in order to achieve greater manufacturing quality. First, a bi-telecentric imaging method-based analytical modeling and overall structural design of the measuring system is completed; next, a more accurate approach for exact edge extraction of contour images is proposed. By performing quadratic function interpolation on the gradient modulus values of nearby three points in the gradient direction of candidate edge points, this method raises the connection degree of the edge point chain and yields more precise sub-pixel coordinates of edge points. Finally, the RANSAC algorithm improves the contour line. obtaining the contour distance using the multi-sample interactive point-line distance computation method. The average relative error of the measured contour distance obtained by this system is less than 0.007mm, and the repeated measurement accuracy is less than 0.2m, according to an experiment measuring the contour distance of a polygonal workpiece at various points in relation to the nominal value highly reliable measurement of the profile distance.