Automatic segmentation of microporous defects in composite film materials based on the improved attention U-Net module

Abstract

ABSTRACT Composite film materials offer several advantages, such as high strength-to-weight ratios, flexibility, durability, or cost-effectiveness, which make them ideal for a wide range of applications across various industries. However, microporous defects in the material compromise its durability and air-tightness. Thus far, few works have been performed on inquiring non-destructive testing methods intended for the detection of microporous defects in composite film materials. In this study, the automatic detection of microporous defects is achieved by using a thermal imaging inspection system developed on the basis of a laser infrared system and the U-Net. Innovatively, an improved attention module is added to U-Net to boost its detection performance. While using a laser as a heat source, a fibre laser beam shaper is chosen as the laser beam expansion to ensure a uniform intensity distribution of the laser beam within a small heating area. Experimental results show that the proposed method can reliably detect microporous defects in the composite film materials with sizes in the order of millimeters/micrometers.