Quality control of effect painting by means of picture development software

In the automotive industry car bodies are painted with effect coatings to get a valuable appearance.

Of course, the single building components must always have exactly the same color without the slightest deviation. To guarantee this, a color matching system should help to notice deviations as soon as possible. This is an important measure because if deviations are noticed at an early stage the operator can run countermeasures to save time and money. This is possible with a special picture development system that contains camera sensors, hardware, software tools and picture development software, and that ensures the right lightning and handling of the building components.

A few months ago divis has made several test series for such color analyses. The aim was to get more information on the delivered paint batches. We wanted to find out if there are any deviations in the formulation concerning basic color, pigments or additives (flakes) and—should that be the case—if it is possible to achieve the color value required by the customer. During the test series various coatings were tested for their color values by taking pictures and analyzing them with the mentioned picture development software. The results showed color deviations between components that actually should have the same color. We could prove changings in the formulation concerning amount, size and color value of the flakes which influenced the basic color. The currently used measurement techniques do not reveal such background because they only measure the basic color valu. The influence of different positions of light and camera on the color determination is subject to further analyses.

If such a control system is integrated into the production process, it is possible to notice errors and the related cause-and-effect-relations already at an early stage due to an online-analysis. So it is possible to initiate countermeasures in time because if the error is noticed at the end of the production line, when the single components are already put together, it can be very time and money consuming to find and eliminate the error source.

Fig.1 Samples of different coatings from a running production

Fig.2 These samples are actually supposed to have the same color values. However, the analysis revealed deviations of the RGB-color values.