LAUSR

Abstract With the development of powder bed fusion (PBF) additive manufacturing technique for functional parts production, process monitoring and diagnosis is highly demanded to ensure its process reliability and repeatability. An off-axis vision monitoring method using high-speed camera is proposed in this paper. An optical filter with 350 nm–800 nm cut-off was used to enhance the image contrast between the plume and the melt pool. A new image processing method was designed to extract features from the melt pool, plume and spatters, respectively. Kalman filter tracking was used to pinpoint the exact melt pool position, and image segmentation algorithm was developed to segment the melt pool, plume and spatters from each other; a new tracking method was utilized to remove the spatters generated in the previous frame. After image processing, the features of melt pool intensity, plume area, plume orientation, spatter number, spatter area, spatter orientation and spatter velocity were extracted and their correlations with the scanning quality were investigated. The results indicated that these features were potential indicators for scanning quality assessment. The proposed method could be used to further study the characteristics of plume and spatter and to explore the diagnosis performance based on the fusion of melt pool, plume and spatter information. It provides a promising means for in-situ monitoring and control of PBF process.