Welded joints constitute a critical point in materials. Cracks, inclusions or lack of fusion might be a reason for rejecting them. Therefore, the detection of indications is vital to ensure the quality in such unions.
In this context, the use of Induction Active Thermography as a Non-Destructive Testing technique in order to detect surface indications implies the need to develop new techniques for processing the images captured by thermal cameras. In this way, the strengthening of the ability of this thermographic method for the automatic characterization and evaluation of surface indications is intended, enhancing the profiles of cracks and facilitating their detection, all without contact with the workpiece to be analysed.
For this purpose, various image processing algorithms based on mathematical operations as well as statistical methods have been developed, allowing the automatic detection of indications.