Research Of Three-dimensional Reconstruction For Micro-components In Precision Manufacturing

With the rapid development of micro-electronics industry in recent years, different types of integrated circuit micro-components become the core of the industry, which plays a decisive role in the quality of the final product. The traditional artificial visual inspection using a magnifying glass and two-dimensional scanning detection for micro-components in precision manufacturing cannot meet the requirements with comprehensive data any more. So the research of three-dimensional reconstruction has an important engineering significance and value to improve accuracy, efficiency and range with more comprehensive, intuitive three-dimensional inspection data.This paper has been supported by National High Technology Research and Development Program(863 Program) "Precision surface mount technology and complete sets of equipment". This article dedicated to the research of key issues of three-dimensional reconstruction for micro-components in precision manufacturing. Firstly, set up a three-dimensional reconstruction system. Then, have camera calibration and corner detection. Capture the image of micro-components by using color raster fringe projection. After preprocessing, we can realize the three-dimensional reconstruction by designing duty ratio method. This paper is mainly concentrated in the following work:i. Build a three-dimensional reconstruction system for micro-components. On the basis of existing technology, this paper set up the system connecting active vision measurement method and grating projection measuring principles, which are described in detail.ii. Proposed a corner detection algorithm based on 180 ° rotation operator. Compared with other algorithm, this can improve the accuracy of camera calibration better by using forms of mean square error and coincidence degree comprehensively with the corner feature of calibration image.iii. Bilateral filtering algorithm for image pre-processing is used, and an optimization algorithm of improved separated hybrid bilateral filter is proposed. The algorithm can retain the edge feature from the color raster projection image of micro-components better by separating dimensions and combining with a variety of filtering methods.iv. Design a duty ratio method based on color raster stripes. This method can obtain the relationship of strips offset and surface height by the encoding and decoding color raster stripe of micro-components' linear characteristic surface shape. The three-dimensional reconstruction results of experiments with different chips in MATLAB show high precision and good effect.