Περιγραφή:Crystallization patterns emerge when an aqueous dihydrate Copper chloride (CuCl₂⋅2H₂O) solution in the presence of organic additives (juices/extracts) is crystallized on a glass plate. The emerging patterns are additive-specific and reflect physiological processes like maturation and ageing, the effect of processing, feeding regime and production system in a broad range of agricultural products. The patterns and their underlying structures are evaluated visually by means of defined morphological criteria and by means of computerized image analysis, respectively. The currently applied texture analysis algorithm reflects the spatial linear relationships between grey-scale values of the scanned crystallization structures, rendering the zero point arbitrary and constraining data analysis to the ordinal scale. Furthermore the algorithm is non-consistent with the physically defined geometric properties of the crystallization structures. In this article the development of a structure analysis algorithm is described and discussed which allows a quantification of the crystallization structures by computing 15 width-, and length-parameters, introducing a non-arbitrary zero-point and an equidistant scale which permits all statistical measures. The algorithm is applied to crystallization structures produced from carrot samples which shows it reflects the monotonic relation between physically defined geometric properties of the crystallization structures and laboratory procedure parameters influencing the overall morphological features of the crystallization structures. For instance the nucleation time, which is the time elapsed prior to initial nucleation of the crystallization structure, and the circular region of interest (ROI) around the geometric center of the glass plate used in image analysis evaluation. It is concluded that this structure analysis algorithm is a valuable addition to the image evaluation tools applicable for crystallization investigations of agricultural products, augmenting the image analysis with a non-arbitrary zero point and an equidistant scale which permits all statistical measures.