A statistical method for the relative comparison of automotive paint colours using digital imaging analysis

Relative colour comparisons were performed using digital imaging techniques and analysis. A statistical method was used to quantify how well a test colour matched a colour standard. Colour histogram comparisons were performed by incorporating a combination of control limits (based on prediction intervals) and threshold limits that were calculated for each curve set. Test colours were imaged and compared to colour standards by calculating the per cent match for each of the RGB curves. Colours that did not show a per cent match of 60 per cent or greater in all three colour curves were considered failures. Some colour families, e. g. reds, required larger control limits to account for colour variability.