Development of nondestructive inspection systems using infrared images based on image processing techniques

摘要

The last few years, infrared thermography technique has been applied successfully in petrochemical industry. The main information from this technology acquired by using infrared thermal camera is in the form of thermal image. Any abnormal condition of an inspected object will be reflected as an abnormal spot (often hotspot) at certain location in a thermal image with highest pixel intensity values. In image processing terminology, this spot is called as the defect. Current practice, this hotspot is interpreted manually by human inspector. This way of interpretation is not effective and efficient. This report proposes an automatic and efficient way for interpreting a thermal image, in term of defect detection.