Aluminum wheels / rims for cars are manufactured today according to individual specifications. Therefore, it is getting more and more difficult to distinguish them due to small characteristic changes. There exist identical wheels (rims) in matted black and glossy chrome. Modern designs also combine color and brilliance elements with a black base color.
During the single manufacturing steps car wheels / rims are separated several times and mixed again. Before they reach single inspection stations the respective wheel / rim type must therefore be recognized free of doubt by wheel geometry and design comparisons, to convey them to the next process step appropriately resp. to control the next step. Syscon, a German system company for machine vision, reached optimized output images using the software solution VisualApplets and the High Dynamic Range (HDR) application on a programmed Baumer camera within the own solution DesignSweeper, which has been used for more than 20 years in many rim plants.
By HDR five grey scale images are acquired with ascending exposure time and merged to a well-balanced exposed whole image. Image parts that are too bright or dark get discarded (linear weighting), whereas the surface of the rim is represented well-balanced illuminated. Thus, even small logos and symbols can be recognized in the image, irrespective of a matted, glossy, bright or dark rim surface.
VisualApplets with the HDR application has been implemented on the camera FPGA using VisualApplets Embedder. Due to image pre-processing and image processing on the FPGA no CPU resources are necessary for image processing.
Identification of wheel / rim types of cars
HDR application with VisualApplets implemented on the camera FPGA using VisualApplets Embedder
|VisualApplets example on the HDR algorithm, to extend effective pixel bit depth and reduce noise|