A new signal processing algorithm of pulsed infrared thermography

• Published in: Infrared physics & technology Vol. 68; pp. 173 - 178
• Main Authors: Tang, Qingju, Bu, Chiwu, Liu, Yuanlin, Qi, Litao, Yu, Zongyan
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.01.2015
• Subjects: Image sequence; Markov; Principal component analysis; SNR; Image sequence; Markov; SNR; Principal component analysis
• ISSN: 1350-4495; 1879-0275
• DOI: 10.1016/j.infrared.2014.12.002

•Markov–PCA algorithm was proposed to process the pulsed infrared thermography.•The analysis window has a great effect on the processing results.•Markov–PCA algorithm improves the SNR of feature images. A new signal processing algorithm combining Markov and principal component analysis (PCA) algorithm, which was named as Markov–PCA algorithm, was proposed to process the pulsed infrared thermography. First, the image sequence was reconstructed using Markov algorithm, then the original complex data dimensionality was reduced using PCA algorithm, which can remove the noise and redundancy of the infrared image sequences, and thus improve the detectability of defects. Results show that both the starting frame position and size of analysis window has an obvious effect on the processing results of Markov–PCA algorithm. And the proposed Markov–PCA algorithm improves the signal to noise ratio (SNR) of feature images more significantly than the commonly used PPT algorithm.