Font Size: a A A

Research And Application Of Printed Character Recognition Algorithm For IC Chips

Posted on:2016-02-14Degree:MasterType:Thesis
Country:ChinaCandidate:Y HuFull Text:PDF
GTID:2348330479952654Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
IC(Integrated Circuit) chips play a vital role in electronic industries. IC marking is very important as it provides information about the integrated circuit chips, such as product function and classification. IC marking recognition is one of the essential processes in semiconductor fabrication. Traditional inspecting methods such as manual work, template matching are not up to the requirement of Intelligent chip identification. The paper studies the technology of IC chip printed character recognition. An IC chip printed character recognition system is designed by using these technologies. The main work is as follows:(1)The features of IC chip images were analyzed. The IC image is proved to have Gauss noise and speckle noise. Uneven illumination even exists in some images. We approve the method of the median filter and Gauss filter for image filtering and adopts a multiresolution Otsu method to separate IC chips from the background image. It reduces the image noise, enhances the edge features of IC chip image, and provides a high signal-to-noise ratio IC image for the following recognition.(2)A connected-component based method according to the feature of the chip is presented in this paper. The method is divided into three parts. First, connected-component of the image is generated. Then, the chip region is selected by the features of the chips. Then the edge of the chip area is detected and the chip edge is fitting by least square algorithm. The IC chip image is rotated to the correct position, according to the angle of image rotation correction. At last, the marking is located because the IC chip characters are printed on a chip with a relatively fixed position. The robustness and real-time of the proposed algorithm were verified by experiments.(3)A new projection segmentation algorithm based on features of character is presented in this paper after analyzing the existing character segmentation algorithm. This algorithm can segment touching characters and broken characters correctly, due to uneven illumination, motion blur, as well as problems in the printing process. It was successfully applied to IC chip printing character segmentation, which can improve touching and broken character segmentation accuracy.At the end of this paper, an IC chip printed character recognition system is designed with the chip ST2378 E based on the research of IC chip printed character recognition algorithm. Experimental results have shown that marking defects inspection system is able to inspect an IC marking with 20 characters in just 150 ms. The system achieves a maximum recognition rate of 98.5%. As a result, it is an ideal solution for IC chip printed character recognition system.
Keywords/Search Tags:IC chip printed character inspection, character recognition, character localization, character segmentation, neural network
PDF Full Text Request
Related items