An Infrared Thermal Imaging Approach with Cooling Pre-treatment for Monitoring Cracks inside Metal Plate

This article introduces the basis theory of infrared thermal non-destructive detection, and builds an infrared thermal wave image system for metal plates. Aiming at the difficulty to observe small defects in infrared thermal wave images, the paper innovatively proposes a method of applying refrigeration pre-processing, so that small defects can be observed on infrared images. Though a number of experiments, we select the optimal refrigeration temperature difference and observation time. In view of the serious interference of background noises in infrared thermal wave images, we study the homomorphic filtering enhancement methods for infrared thermal images. Then we use OTSU, KSW entropy and PSO algorithm to perform threshold segmentation on the enhanced images to effectively extract target defects. Finally, we quantitatively analyze the error of the extracted target defect areas.