Soil colour and spectral analysis employing linear regression models I. Effect of organic matter
 
More details
Hide details
1
Laboratory of Soil Science and Agricultural Chemistry, Agricultural University of Athens, 75 Iera Odos, Botanikos 11855, Athens, Greece
 
Int. Agrophys. 2004, 18(1): 1–10
KEYWORDS
ABSTRACT
This work comprises an investigation into whether soil reflectance spectral analysis which is employed to calculate the colour characteristics (hue, value, chroma) of soil can be carried out using linear regression models, so that comparison of colour characteristics subsequently becomes possible, and also statistically documented. To this end the colour of soil samples was calculated through spectrum reflectance in the visible region of dry smooth-rubbed soil samples smaller than 250 mm. The colour parameters of the CIE system assessed by analysis of the spectrum reflectance were converted into Munsell colour system characte- ristics. Regression in accordance with the piecewise linear model was then applied to the spectrum data. The processing indicated that this model is capable of making satisfactory predictions - above all of the value and secondarily of the chroma of the soil samples. Detection of statistically significant differences in the colour characteristics of horizons of the same profile was effected through the application of the nested model. These differences cannot be detected using the tables of the Munsell colour system. Finally, in each region of the spectrum, qualitative analysis of the effect of the organic matter on the soil colour characteristics was performed, demonstrating its active role in determining the readings for value and chroma.
eISSN:2300-8725
ISSN:0236-8722