Normalize Cross Correlation Algorithm in Pattern Matching Based on 1-D Information Vector

All previous published study in pattern matching based on normalized cross correlation worked in 2-D image. In this study, we propose a pattern matching algorithm using 1-D information vector. The proposed algorithm consists of three main steps: First, the pattern image is scanned in two directions to convert the pattern image from 2-D image into 1-D information vector. Secondly, all blocks (having same size of pattern) in the reference image also are scanned in two directions to obtain 1-D information vector for each block in the reference image. Thirdly, the normalized cross correlation between 1-D information vector of pattern image and all 1-D information vectors in the reference images are established. Finally, we can determine the correct position of pattern in the reference image. Experimentally, we compared the proposed algorithm with three 2-D pattern matching algorithms. The results shown that, the proposed algorithm is more efficient and outperforms the others. Also, we examined the proposed under three different types of noise and we found that it is very robust against noise.