Grouping and classification of wheat from organic and conventional production systems by combining three image forming methods

Abstract

Using image forming methods, products from organic and conventional origin have been discriminated with high accuracy, but the meaning of these differences in terms of product quality is so far poorly understood. The aim of the presented study is therefore to gain further insights into the suitability of image forming methods for food quality evaluation based on wheat samples from a long-term field trial on the comparison of different organic and conventional production systems (DOC-trial). The images of the encoded samples were (1) grouped into pairs with similar image features, (2) characterized based on reference images (e.g. high resistance to degradation – low resistance to degradation), (3) ranked (according to the quality characterization), and (4) assigned to the different production methods (classified). The encoded samples from the production methods mineral fertilization, conventional production (combination of mineral fertilization and farmyard manure) and the class of organic production methods (biodynamic, bioorganic and unfertilized control) could be grouped and classified in both years. Within the class of organic production methods grouping (assigning of similar samples to image categories) and classification (assigning of categories to production methods) was partially possible. The correct grouping and classification of samples from organic and conventional production shows that different fertilization systems influence image structures in a typical and reproducible manner. The evaluation approach followed in the presented research can provide a considerable contribution to advance our understanding of quality differences between products from different farming systems or plant production measures.