Research on Automated Defect Classification Based on Visual Sensing and Convolutional Neural Network-Support Vector Machine for GTA-Assisted Droplet Deposition Manufacturing Process

This paper proposes a novel metal additive manufacturing process, which is composition of gas tungsten arc (GTA) and droplet deposition (DDM). Due to complex physical metallurgical processes involved, such as impact, spreading, surface pre-melting, etc., defects, including lack fusion, overflow discontinuity deposited layers always occur. To assure the quality GTA-assisted DDM-ed parts, online monitoring based on visual sensing has been implemented. The current study also focuses automated defect classification avoid low efficiency bias manual recognition by way convolutional neural network-support vector machine (CNN-SVM). best accuracy 98.9%, with an execution time about 12 milliseconds handle image, proved our model can be enough use in real-time feedback control process.