In this investigation, the influence of the wooden species, the thickness of the surface layer, the defect size and the amount of glue on the defect detectability have been studied. 0,5-4,0 mm thick surfaces layers made out of oak, merbau, pine and alder have been glued with urea formaldehyde (UF) glue on pine substrates. The glue defects were simulated with 3, 6, 12 and 24 mm. broad bands with 140 g. glue/m2 (correctly glued adhesive joint), 70 g. glue/m2 and no glue at all respectively. The test pieces were tested with heating up thermography and lock-in-thermography. The evaluation was done quantitatively with the aid of MATLAB.

Areas without glue could be well distinguished from correctly glued areas. In heating up thermography, the resolution was 6 mm beneath 0,5 mm thick surface layers and 24 mm beneath 2,0 mm thick surface layers for all wooden species. The contrast for bands without glue were twice as high for merbau and oak compared to pine and alder. The contrast between areas with and without defects respectively increased with the amount of glue deficiency. In this application, heating up thermography seems to be a more suitable test method than lock-in-thermography.