Study Of Recognitizing And Localization Of Mark On Printed Circuit Board

With the development of the machine vision, the placement machine has widely utilized the vision system in order to place accurately the fine pitch electronic devices. The vision system is mainly used for the accurate localization of PCB, the centering and aiming of components and the testing of the devices. These technologies depend on the improvement and proposition of the digital image processing to satisfy the modern industry's increasing demand on the precision and the speed. This document regards the Mark which performs an important function in the PCB localization and the component centering as the object of study. The in-depth study about the shape identification and improving the localizing precision and the computing speed of the PCB will provide the theory for the advancement of the placement machine.The main work is shown in the following:Adopt the simple geometric features to identify the regions which are not the Mark, take the border shape of Mark as the identification target, classify according to the characteristic extracted by the wavelet descriptor and compare it with other recognition algorithm based on the experimental results.Apply the new edge interpolation and Hough transform to detect the circular Mark for the high precision. At the same time some effective strategies are adopted to decrease the computational complexity because the industrial production has itself characteristic and the experimental results indicate this method has an advantage in the accuracy and the speed.Compute the degree of the rectangular Mark' rotation angle accurately through detecting the line edge by the Hough transform. To increase the speed, the rough-and-fine search is a good choice. Then apply the final result to the template matching algorithm in order to deduce the matching complexity. The comparison among many different methods illustrates the superiority of the proposed method.The paper mainly studies about the recognition and detection algorithms as the Mark is the reference point used for the PCB localization. We modify many kinds of Mark detection algorithms and bring forward some novel methods on the basis of analyzing and comparing the existing algorithms. A lot of experiments demonstrate our research has practical value in the industrial production.