Practical challenges of using vision systems for new applications in a sugar factory.

Abstract

Vision systems, or so-called image processing or image analysis systems, are very common in the industry, most often to control product quality. Within the sugar industry the amount of commercially available applications are however still rather limited. Of course, on the sugar end the image processing systems for sugar color and crystal size characterization are well developed and integrated in the existing production process. Further, in the past for example a vision system was developed to determine the top tare of the beet in the tare house to skip manual topping.  Since the introduction of the smartphone and the related apps, image processing is very common and makes part of our normal live. As a result, new ideas for vision system solutions and applications within the sugar factory popped-up by employees in recent years. Three possible applications were further investigated and developed in the Dutch factories of Cosun Beet Company: beet yard volume estimation, foreign object detection between unwashed beets and the size measurement of beet cossettes. This paper describes the problem/solution, the method and results of the performed plant trials and the related practical challenges. It is certain that a joint approach from the sugar industry is required to develop commercially available applications.