PREDICTION OF SOC IN CALCIC CHERNOZEM IN THE STEPPE ZONE OF UKRAINE USING BRIGHTNESS AND COLOUR INDICATORS

摘要

Soil organic carbon (SOC) is an important component of any soil which determines many of its properties. Nowadays, more and more attention is being paid to the SOC content determination in soils by not using the conventional, time-consuming and expensive technique, but by using colour image processing of soil samples. In this case, even the camera of modern smartphones can be used as an image source, making this technique very convenient and practical. However, it is important to maintain certain standardised conditions (light intensity, light incidence angle, etc.) when capturing the images of soil samples. In our opinion, it is best to use a regular scanner for this purpose, with subsequent image processing by graphic programs (e.g., Adobe Photoshop). To increase the reliability of the colour information obtained in this way, it is desired (if possible) to use a spec- trograph or a monochromator in the subsequent calculation of reflection or brightness ratios. It is these two approaches that we have implemented in our work. As a result of the experiment, the values of brightness ratios (at 480, 650 and 750 nm wavelengths and integral brightness ratio), colour indicators (the hue, saturation and value HSV, red, green and blue RGB, CIE L*a*b* and cyan, magenta, yellow and key CMYK systems) and SOC content in Calcic Chernozem samples of the steppe zone of Ukraine were obtained. Using correlation analysis of the dataset, the existence of direct (r = 0.88–0.90) and inverse close relationships (r = ?0.75–0.90) between SOC, values of brightness ratios and colour indicators of the soil sam- ?0.75–0.90) between SOC, values of brightness ratios and colour indicators of the soil sam- 0.75–0.90) between SOC, values of brightness ratios and colour indicators of the soil sam- ples were established. This allows us to develop predictive models. Statistical analysis showed that the models were significant when they were based on the values of brightness