Date of Award
Master of Science (MS)
Mears , Laine
Hung , Steve
The work focuses on the qualitative enhancement of Thermographic Inspection data using Principal Component Analysis technique. A new processing tool named Economical Singular Value Decomposition (SVD) is proposed for analyzing the infrared image sequences acquired from the test specimen.
Experiments are carried out on test sample using thermal excitation modalities such as flash, halogen and induction to create temperature variations. The infrared (IR) image sequence containing the information about the sample is captured using IR cameras. The acquired IR image sequences are processed using the proposed Economical SVD which projects the matrix of raw pixel values of image sequence into a set of orthogonal components to extract features and reduce data redundancy. The results from processing revealed that Economical Singular Value Decomposition can be used with various thermal excitation modalities while enhancing the contrast between the defect and non defect areas. The proposed processing routine is bench marked against the available standard techniques such as the Thermal Signal Reconstruction for one single application and specific experimental conditions.
Keywords: Thermographic Inspection, Principal Component Analysis, Economical Singular Value Decomposition, Orthogonal Components, Thermal Signal Reconstruction.
Parvataneni, Rohit, "PRINCIPAL COMPONENT THERMOGRAPHY FOR STEADY THERMAL PERTURBATION SCENARIOS" (2009). All Theses. 702.