Abstract
This paper presents a novel quality inspection system based on support vector machine (SVM) for micro laser spot welding. An optical sensor with a long pass optical filter is used to derive signals covering a spectrum ranging from visible to infrared bands. A multi-class SVM classifier is applied to classify the quality of welds using a set of features extracted from the sensor signals. A series of experiments was carried out using a bespoke continuous wave (CW) 515nm laser source and a common SMD (surface mounted devices) as a test component. The results obtained show that this approach can be used to monitor the laser welding quality with an overall classification success rate of 97.6%.
Export citation and abstract BibTeX RIS