Estimation of the weight distribution function of a complex object for portion control in an innovative can filling system

This paper pertains to advanced automation of canning processes, with special application to fish canning. The automated system under consideration consists of three stages: portion control, filling, and post-filling inspection and compensation. The work presented deals particularly with portion control which involves cutting a set of objects into pieces by taking into account the weight distribution of each object (fish) and grouping the pieces into can-filling portions according to an optimality criterion. The objective of optimization would be to make each portion as close as possible to the target weight of can, within a tolerance, so that wastage associated with overfills, and also the regulatory violation and consumer dissatisfaction associated with under-fills would be minimized. An intelligent decision making system is used to match online measurements of object (fish) features to a corresponding model that has the necessary information for cutting the object.