Quantification of Crystal Conglomerates Using Image Analysis

A regular crystal form, in particular the absence of conglomerates, is an important aspect of refined sugar quality because irregularities increase the quantity of mother liquor which remains on the surface of the crystals or is contained within occlusions in the crystals themselves. This aspect of sugar quality is of greater importance when processing lower purity liquors due to the increased impurity load of the mother liquor retained by the crystals in these circumstances. The current technique for the quantification of conglomerates in sugar samples, namely the crystal regularity index, is based on the subjective classification of photographed crystals in terms of three or four groupings. Each of the groupings is given an arbitrary weighting in terms of crystal regularity and the total score for the sample is expressed as a percentage of the maximum score which could be obtained by a sample of crystals of perfectly regular shape. This technique is both slow and dependent on subjective classification by one or more observers. An attempt has been made to develop an improved technique for conglomerate quantification, automating the procedure as far as possible and replacing the subjective element of the original analytical method by using computerised image analysis procedures. After the extensive investigation of a number of possible classification criteria, some preliminary methods have been obtained which are based on the analysis of crystal shape and the characteristics of the crystal edges.