An investigation of data fusion for defects inspection in additive manufacturing

Abstract

Additive Manufacturing (AM) technology is considered as one of the most promising manufacturing technologies in the areas of aerospace and defense industries. However, the AM parts are known to have relatively high residual stresses and a variety of defects, such as porosity, balling, cracking, etc. Therefore, it is critically important to monitor the quality of products during the AM manufacturing process. In this paper, we proposed a novel enhanced fusion algorithm based on Finite Discrete Shearlet Transform (FDST) and multi-scale sequential toggle operator (MSSTO) for visible and infrared images fusion in the AM systems. The original images can be decomposed into low-frequency and high-frequency subband images by FDST. Then, the effective bright and dark image informations are extracted from the low frequency coefficients of source images by MSSTO transform, which are injected into the fusing low frequency coefficient to obtain the final low frequency synthetic coefficient. The high frequency sub-band coefficients are fused by using the local spatial frequency weighting and region energy. The fused image can be obtained by the FDST inverse transformation of the high and low fused coefficient. Experiments show that the proposed algorithm can get more texture information while retaining the significant features of the images, acheiving good detection and indetification results of the defects properties.