High-precision displacement measurement algorithm based on a depth fusion of grating projection pattern.

Abstract

Displacement measurement based on an image recognition algorithm has more advantages in the field of displacement measurement. Due to the uncertainty of image location, the accuracy of the displacement measurement algorithm studied in the previous stage is limited. To improve the accuracy of image-type displacement measurement, a high-precision displacement measurement algorithm based on depth fusion of the grating projection pattern is proposed. First, a line-scan image sensor and parallel light source are placed on both sides of the calibration grating. The marking pattern on the calibration grating is projected onto the image sensor through illumination of the parallel light source to realize the projection imaging. Then, the centroid algorithm is used to calculate the centroid of each line in the collected image, and the mean value of the centroid group is determined. Finally, a linear function is used to fit multiple centroids to calculate the endpoints of the subdivision region, and the subdivision operation is carried out according to a ratio algorithm. To test the performance of the proposed algorithm, a linear displacement measurement device with a 250 mm range and an angular displacement measurement device with a 100 mm diameter grating are both used for experiments. After testing, the linear measurement accuracy of the proposed algorithm is improved from 4.63 to 1.26 µm, and the angular displacement measurement accuracy is enhanced from 8.87 to 2.88''. The research in this paper provides a theoretical basis for high-performance displacement measurement.