Automated Detection of CFRP Defects by Infrared Thermography and Image Analysis

Carbon fiber reinforced polymers (CFRP) have experienced widespread use in the aerospace industry. However, the very different internal structure and properties of woven composites presents a challenge to defect detection in such structures. An automated scheme for determination of the location and size of defects using image processing techniques is reported. The scheme includes local thresholding to binarize the image and reduce non-uniform thermal energy effect, morphology procedures to reduce larger noises and separate the defect from the specimen. Test of the scheme on various specimens having artificial defects embedded in different depth is shown. The contrast is lower as the defect is embedded deeper. Nevertheless, the defects can be detected accurately from the IRT images while comparing to the observed ones.