主分量分析法在红外数字图像序列处理中的应用

Application of Principal Component Analysis in Infrared Image Sequence Processing

为了提高红外热像无损检测中所重构的数字图像的缺陷对比度,从而提高红外热像无损检测的缺陷探测能力,研究了主分量分析法及其在红外数字图像序列处理中的应用。介绍了主分量分析法的原理;用该法对实验中采集的红外数字图像序列进行处理;以信噪比为指标对图像处理效果进行了定量评定。研究表明主分量分析法不仅消除了红外热像无损检测中的加热不均效应,而且使图像序列中图像的信噪比得到不同程度的提高,使图像的质量得到了很大改善;因此是红外热像序列处理的一种有效方法。

The principal component analysis （PCA） and its application in IR digital image sequence processing were studied. In order to improve the thermal contrast for underlying structural flaws of digital images reconstructed in infrared thermographic nondestructive testing（IR NDT）, and to enhance the flaw detectivity of IR NDT. The principle of PCA was introduced. The IR digital image sequence obtained in experiments was processed by PCA and the image processing effects were quantitatively evaluated by the signal-to-noise ratio（SNR）. The study shows that PCA not only removes the uneven heating effect in IR NDT, but also increases the SNR of images to some extent. Moreover, PCA greatly improves the image quality. The results show that PCA is an effective method in IR image sequence processing.