Automated counting and characterization of dirt particles in pulp
ICCVG'10 Proceedings of the 2010 international conference on Computer vision and graphics: Part II
A sequential machine vision procedure for assessing paper impurities
Computers in Industry
| Hi-index | 0.00 |
This paper presents a Machine Vision System developed for on-line visual pulp inspection under the critical lighting requirements of the UNE-ISO 5350-2 standard. These requirements typically result in a very low dynamic range in the electronic images due to the following: high-speed displacement of the product (up to 2.7m/s), very high resolution (0.004mm^2/pixel), and the backlight requirement. This problem has been solved using TDI technology in the acquisition system, made up of several parallel CCD lines that shift their electrical charges in synchronization with the product movement, therefore producing a single line-image with several times more dynamic range. The obtained image has a cross light non-uniformity that is resolved by an image processing algorithm that divides the image into several sub-images that are processed in parallel. The system has been tested in on-line production plants, and has been proven to provide a great improvement over the objectivity and sampling rate limitations of human inspection, thereby realizing a more reliable and trustworthy quality control measurement.