关键词: 图像重建, 自动识别, 平均灰度矢量, Zernike矩

Abstract: The industry CT technology can accurately reconstruct the grey-level image of work piece internal structure，thus also can provide the possibility for non-destructively examining assembly defect（leak installing，mistake installing and multi-installing etc.）.This paper proposes a new template matching algorithm to realize the assembly defect detection.First，using the average grey-level vector of template to carry on the thick search，determines the candidate match set of points.Next，using the Zernike moment of template to carry on the accurate search，locates the target.The experimental result indicates that，even if the template has the free angle revolving opposite to the examined work piece，the proposed algorithm also can successfully locates the target.

Key words: image reconstruction, automatical recognition, average grey-level vector, Zernike moment