Method for evaluating the surface of a body component, and method for training an artificial neural network

A method for evaluating a surface of a body component of a motor vehicle, the method comprising:

• creating a plurality of virtual polygon networks of the surface;
• determining at least one variable characterizing a curvature of each of the plurality of virtual polygon networks at at least one node of each of the plurality of virtual polygon networks, the at least one variable including a plurality of sub-variables;
• and
• determining, by an artificial neural network, in dependence on the at least one variable characterizing the curvature, at least one output variable characterizing a surface flaw of the surface to evaluate the surface;
• wherein a plurality of two-dimensional geometric maps of the virtual polygon network are formed for each of the plurality of virtual polygon networks, each of the plurality of two-dimensional geometric maps including a plurality of pixels, each of which is assigned at the at least one node;
• and
• wherein each of the plurality of two-dimensional geometric maps is assigned only one of the plurality of sub-variables at at least one pixel of the plurality of pixels, and the output variable characterizing the surface flaw is determined in dependence on the plurality of two-dimensional geometric maps by the artificial neural network.