Infrared thermography processing using Markov-PCA algorithm

摘要

The pulsed infrared thermal image sequence characteristics of the coating structure was analyzed, and the temperature change process of any pixel including status and time parameters was considered as discrete Markov process. A combination of Markov and principal component analysis (PCA) algorithm were proposed to process the pulsed infrared image sequence. First, using the Markov method to achieve the image sequence reconstruction, then using PCA method to achieve the original complex data dimensionality reduction to remove the noise and redundancy, and extract the main components reflecting the main features of the data. Results show that the processed images have higher SNR. Results show that the processed images have much higher SNR than that of the original thermal image with the best contrast.