Validation of a spatter detection algorithm for remote laser welding applications

The use of laser scanners in industrial laser welding applications is increasing. Welding spatters are challenging if a high quality is required. A control system for spatters is currently being developed, and this study focuses on the validation of a robust spatter detection algorithm. To validate the detection algorithm, artificial spatter images are created. These images provide the ground truth concerning the number of spatters and can therefore be used to improve the recognition quality of the detection algorithm. As a result, this paper describes a method for validating a spatter detection algorithm, which can recognize spatters with a high accuracy.