Imаge Segmentation Of Agricultural Products Using Statistical Indicators

Abstract

Machine inspection is a mandatory technological process in industrial processing agriculture products. The camera detects color and shape based irregularities of the object resulting in a large number of parameters for decision-making and sorting product compliance. The goal was to discover a new criterion for decision-making using only the output signal of the RGB camera. Research employed the digital images of raspberries, blackberries, peas and yellow beans during real processing, obtained from a color sorter machine. The visual texture of the surface of the agricultural products was described via defined statistical indicators of color (color average value (Avg), standard deviation (Stdv), entropy (E), and lacunarity (L) was used from the sphere of image fractal analysis as one of the criteria. By applying the non-parametric tests: Wilcoxon signed rank and Friedman test, statistically significant difference was established for the L and Е criteria between compliant and non-compliant industrial products