The Segmentation Of Infrared Image Based On Advanced Material Defect Detection

Nowadays, advanced material plays more and more important role in aircraft and design. It is significant to find its flaws and damage detection. The essence of image segmentation is to extract the special areas of the target image. It is the key to make the flaw location and damage identification.A variety of flaws and injury prone to be found in the process of manufacturing and assembling. It contains many types and several kinds of expression. Different defect damage show different fault diagnosis mechanism. In the process of infrared data acquisition, a lot of noise is produced. Because of the influence of device and environmental factors, the image has low SNR. Besides, many fault section are uncertain. These factors can increase the difficulty to enhance the detecting precision. Nowadays, traditional infrared image algorithm has many limitation and do not have high precision. It can not meet the practical needs of infrared image segmentation.Combined with the infrared thermography, we analyze the theory of water detect damage in detail. Aiming at the possible problems in image segmentation, we raise a new segmentation algorithm: mathematical morphology image segmentation algorithm.We introduce the infrared thermal image and related defect detecting principle to verify the feasibility of experiment. Combined with the characteristics of infrared thermal image, we analyze the water infrared image. First of all, we need to preprocess the image and improve the structural elements. Then, we conduct the background suppression by using the mathematical morphology method in order to remove background noise and get a new image. Aiming at the new image, we should do histogram equalization. It can enhance the contrast. Combined with improving threshold and morphological edge detection algorithm, we can get the destination image. Finally, compared with several classical edge algorithm, we can conclude that the algorithm we raised have good characteristics. It contain excellent validity and practicability.