This paper details continuing work on the development of a substantive non-destructive method to recover defaced serial numbers stamped or laser engraved into metallic objects based on lock-in infrared thermography. This method relies on the existence of a local zone of plastic strain created from stamping pressures in mechanically stamped pieces and a heat-affected zone in laser engraved samples, both extending to depths below the visible characters. The grain structure within these zones is dislocated due to the external forces applied. These deformed areas are exposed to the surface when the serial numbers are defaced. Infrared thermography utilises the change in thermal conductivity, based on local variation in the thermal gradient from thermal energy applied at the surface. Observation of the thermal gradient at the surface as it propagates through the object allows for identification of these deformed regions and subsequent recovery of the serial number. Principal component analysis is used to enhance the thermographic images. Pseudo Zernike moments and subsequently multiple similarity measures are utilised to identify the numbers based on a reference library of numbers. Several fusion rules are used to obtain a consensus across the similarity measures. This process is tested on several datasets including stamped numbers on a gun barrel, laser engraved numbers on a needle holder and the defaced vehicle identification number of a stolen motorcycle.

Keywords: serial number restoration, lock-in infrared thermography, principal component analysis, pseudo Zernike moments, similarity measure, fusion