Determination of Besatz in Cereals Using Physical Analysis Instrument Based on Imaging and Artificial Neural Network Technology

Sibel Maraş¹, Rukiye Şahin², Maksut Barış Eminoğlu^3* and Ufuk Türker³

¹Turkish Grain Board, Ankara, Türkiye. ²ABP Measurement and Control Systems Import Marketing Ltd. Co., Ankara, Turkey. ³Department of Agricultural Machinery and Technologies Engineering, Faculty of Agriculture, Ankara University, Ankara, Türkiye. Corresponding Author E-mail: eminoglu@agri.ankar.edu.tr

Abstract:

In Türkiye, cereal products are classified according to their quality and then bought by public and private sector organizations. Physical analyses conducted according to standards focus on the definition of besatz in cereals, relying on the analyst's visual skill and expertise. Laboratory and field studies were carried out over a four-year period using the "Cgrain Value ™ Instrument" developed to detect both besatz and sound grains of common wheat, durum wheat, and barley products, utilizing imaging and artificial neural network technology. In these studies, the results of classical method analyses performed by analysts according to national standards and the results obtained from the instrument were evaluated. While an expert analyst can analyze first-class wheat in 15-20 minutes, it takes 25-30 minutes to analyze low-quality wheat. The Cgrain Value ™ Instrument, regardless of the sample's qualities, completes the analysis in 3.5-5 minutes with over 90% accuracy. When the results were examined, it was observed that the instrument achieved its lowest success rate of 91.3% in the nonvitreous grains within the durum wheat sample; the highest success rate was 99.9% in the pest-damaged grains within the red wheat sample. The repeatability rates between the classical analysis result and the instrument analysis result were determined to vary between 82.7% and 99.8%. It is thought that the analysis success demonstrated in this research will increase with the improvement of the instrument's calibration file.

Keywords:

Artificial Neural Network Technology; Barley; Besatz Analysis in Cereals; Common Wheat; Durum Wheat; Imaging

[ HTML Full Text]