Fast wrapped phase unwrapping algorithm for indexing quality map priority queue

Abstract

The invention provides a fast wrapped phase unwrapping algorithm for indexing a quality map priority queue. The method comprises steps of first generating a Laplace phase difference quality map; thenpresetting a quality grading number according to the image size, and establishing a quality grading priority index map according to the quality map; then establishing a two-level priority search structure to replace the traditional adjacency list structure in a flood filling algorithm; and finally, using the flood filling algorithm to perform the wrapped phase unwrapping. The method can improve the unwrapping speed of the wrapped phase and has a more obvious advantage when the image size is increased. At the same time, the high reliability Laplace phase difference quality map is designed to meet the accuracy requirement of the actual wrapped phase unwrapping.

Description

technical field [0001] The invention relates to an optical three-dimensional surface measurement technology—phase measurement profilometry, in particular to a key wrapping phase unwrapping algorithm in the phase measurement profilometry. Background technique [0002] The three-dimensional solder paste detector adopts an optical three-dimensional surface measurement technology - phase measurement profilometry, which solves the height value of the object by measuring the phase information of the sinusoidal grating image projected on the surface of the object, thereby recovering the three-dimensional contour of the object. Its real-time performance is largely restricted by the execution efficiency of the wrapping phase unwrapping algorithm, which greatly limits the application of this technology in the surface mount technology industry. Wrapped phase unwrapping technology is also a key part of many 3D measurement technologies, such as: optical 3D surface measurement technology,...

AI Technical Summary

Problems solved by technology

[0002] The three-dimensional solder paste detector adopts an optical three-dimensional surface measurement technology - phase measurement profilometry, which solves the height value of the object by measuring the phase information of the sinusoidal grating image projected on the surface of the object, thereby recovering the three-dimensional contour of the object. Its real-time performance is largely restricted by the execution efficiency of the wrapped phase unwrapping algorithm, which limits the application of this technology in the surface mount technology industry.

In the traditional quality graph-guided method, the flood filling algorithm is very time-consuming to look up the "adjacency list", which makes the execution efficiency of the quality graph-guided method very low

There are also corresponding improvement methods for this problem, such as the unwrapping method based on heap sorting and the quantitative quality graph priority look-up table method. Both methods can effectively improve the efficiency of wrapping phase unwrapping, but both have certain limitations: the former is in the "adjacency When the size of the "list" gradually increases, the time complexity of heap sorting also increases; because the latter quantifies the original quality map, when the wrapping phase is expanded between two points of the same magnitude, it may appear from a low The reverse expansion from quality to high quality, and when the quality value distribution of the quality map is relatively concentrated, the probability of this phenomenon will increase greatly, and the accuracy of wrapping phase unwrapping will decrease at this time

Images