Abstract

It is important for the European sugar industry to reduce its production costs in order to deal with the reforms of sugar agreements, by being ever mindful of the quality of the raw material. A new image technique has been developed to this end: it allows determination of certain quality criteria, such as: top and leafstalk rates in beet. The device consists of a color camera set in a closed space under constant lighting. The camera is linked to a computer where the images are processed by a software providing on-line predictions. To integrate all the variability proper to the samples taken at the sugar factory, the algorithm presented here is based on a PLS regression on the basis of color image characteristics. The correlations obtained between the visual predictions and the manual references are close to the maximum limits that could be hoped for, taking account of the variability of the manual determination of the beet top. The following factors have been identified as favoring the quality of the forecasts: images from large samples, the turning of those samples and the individual presentation of the beet. The determination of the beet leafstalks is also possible using the same algorithms. This is a less complex task yielding higher correlations

Keywords

sugarbeet; top and leaf; image analysis and processing

Published in

Zuckerindustrie
2006, Volume: 131, number: 1, pages: 21-27
Publisher: VERLAG DR ALBERT BARTENS

SLU Authors

• Geladi, Paul

  • Department of Forest Biomaterials and Technology, Swedish University of Agricultural Sciences
    

UKÄ Subject classification

Food Science
Agricultural Science


Permanent link to this page (URI)

https://res.slu.se/id/publ/11106