Modern forming technology allows the production of highly sophisticated free form sheet material components, affording great flexibility to the design and manufacturing processes across a wide range of industries. This increased design and manufacturing potential places an ever growing demand on the accompanying inspection metrology. As a consequence of their surface shape, these parts underlie a reversible geometrical deformation caused by variations of the material and the manufacturing process, as well as by gravity. This distortion is removed during the assembly process, usually performed in automated robotic processes. For this reason, the part's tolerated parameters have to be inspected in a defined state, simulating the assembly process' boundary conditions. Thus, the inspection process chain consists of six steps: picking the workpiece up, manual fixation of the workpiece, tactile measurement of the surface's coordinates using a defined measurement strategy, manual removal of the fixation and removal of the workpiece from the inspection area. These steps are both laborious and time consuming (for example, the inspection of a car door can take up to a working day to complete). Using optical measuring systems and virtual distortion compensation, this process chain can be dramatically shortened. Optical measuring systems provide as a measurement result a point cloud representing a sample of all nearest surfaces in the measuring range containing the measurand. From this data, a surface model of the measurand can be determined, independent of its position in the measuring range. For thin sheet material parts an approximating finite element model can be deduced from such a surface model. By means of pattern recognition, assembly relevant features of the measurand can be identified and located on this model. Together with the boundary conditions given by the assembly process, the shape of the surface in its assembled state can be calculated using the finite elements method. In application these methods culminate in a shortened inspection process chain (which can now also be automated): picking the workpiece up, placing it in the measuring range, optical measurement, virtual distortion compensation and removal of the workpiece from the inspection area.
This work discusses the methodology of our approach in detail and also provides and analyses experimental results. The underlying research was greatfully funded by the German Research Foundation (DFG).