Transformation of Colour-Images for precision edge probing

Colour and spectral images gain more interest in all types of applications of digital image processing. Even in geometric measurement, where the colour information is usually of minor importance, the number of tasks in which the colour information becomes useful is rising. Sometimes it is just a question of usability for the human operator of the system, in other cases the measurement could be done more robust or with higher precision than in a greyscale image only. There are a lot of algorithms for digital colour image processing concerning recognition/ detection or colour measurement. But there are very little approaches for subpixel precision edge probing methods as needed for geometric measurements. Subpixeling methods in grey value images enable probing of edges with a resolution larger than the pixel centre distance of the sensors. Depending on the quality of the image data an increase of resolution of 1/10th to 1/100th pixels is attainable. Typically methods [1] for subpixel-accurate edge position determination presume that only one function value, i.e. the intensity respectively the grey value, varies within the image. In colour and spectral images there is more than one value per pixel. In [2] an approach based on vector difference has been shown with the restriction to a single pixel line. The aim is to create a new edge filter for digital colour images with properties that are useful for subsequent edge probing methods. The requirements for the new filter are: