Accurate measurement method for parameters of quartz pendulous reed based on the vision image

In order to solve the problem that quartz pendulous reed (QPR) is easily scratched and improve the speed and accuracy of the detection, a novel non-contact measurement method based on the visual image to acquire the geometrical plane parameters of QPR is presented in this paper. The images of QPR are obtained by the system of vision, and Sobel operator is used to calculate the gray-level gradient image and locate one pixel accuracy; then the sub-pixel positioning algorithm of Zernike moments operator is used to calculate the accurate locations of object edges; finally, by least-squares fitting to the edge, the geometrical plane parameters of QPR can be obtained. The experimental results show that the locating accuracy of this method is superior to 0.1 pixel.