Method and device for quality control of the seam with a laser butt-welded sheets or strips

The invention relates to a method for quality control of the seam with a laser butt-welded sheets or strips, wherein a plurality of sensor data from at least two arranged around the welding location sensors is measured, and wherein the sensor data includes at least a summary and correlating measurement data processing to assess the quality of the weld are supplied as input. For timely quality control of the weld, which enables an accurate analysis of the welding result is provided that the stored data are supplied with an essentially hierarchical network structure, as input variables at least one trainable artificial neural network that the at least one trainable artificial neural network comprising is formed essentially hierarchical network structure of at least two substantially independent trainable artificial neural subnets that said first artificial neural sub-network of at least two independent, artificial neural subnets is formed, that the first artificial neural subnets respectively, the results of the Datenvorverarbeitungen fed as input variables be that the second artificial neural network part of the results of the first artificial neural subnets are supplied as input variables and that the Erge BNIS the at least one artificial neural network is used for quality control.