Now the methods of computational image analysis allow us to automate the techniques of routine analysis of digital images in various fields of research. The capillary dynamolysis method reflects the spectrum of organic matter in a biological sample. Last decades it is widely applied in ecology and food industry to study the properties of soil and products.
This paper presents an approach to the analysis and classification of soil samples obtained by capillary dynamolysis. Detectors of brightness differences help to define a special boundary curve on the image, and distances from the central point of the image to the points on the curve are the classifying signs for the image under study.
A set of 128 images is divided into 8 classes, which are used to assess the qualitative properties of soil samples. Computed features make it possible to distinguish samples of the soil of the worst quality and to compare images from different classes.