Open AccessIC Microscopic Image Stitching Based on Improved Line Matching
Author(s) -
Zhibin Li,
Yong Yin
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1607/1/012069
Subject(s) - image stitching , artificial intelligence , computer vision , line (geometry) , computer science , position (finance) , process (computing) , matching (statistics) , image registration , image (mathematics) , mathematics , geometry , statistics , finance , economics , operating system
Integrated Circuit(IC) microscopic image stitching is an important process in IC reverse engineering. However, IC is highly integrated and miniaturized, with many similar structures and dense distribution, and the overlapping region of some IC microimages may be the low-texture region. As a result, the traditional image stitching method is prone to mismatch, and the stitching effect is poor. This paper proposes an IC microscopic image stitching method based on improved line matching. The method first extract line segments from two IC microscopic images to be stitched by using Edline algorithm. Secondly, the intersection of line segments is generated by using the position relation of line segments in a certain strategy. Then, combined with the 2D capture method of the IC microscopic image, we limit the matching region. Finally, image registration is completed by combining points and lines, and a transformation matrix is obtained to stitch IC microimages. Experimental results demonstrate that the proposed method reduces the possibility of mismatching in the stitching process while using the stable line features and the position constraints between line segments in IC microscopic images. Moreover, the image stitching guided by line features can also achieve good stitching results in low-texture IC micrograph.