Applying Colorimetric and Image Analysis Methods to Identify the Coal Areas Contributing to the Hydro-Sedimentary Dynamics in Jiu River Basin.
2019
Gabriela Adina MOROŞANU | Liliana ZAHARIA | Philippe BELLEUDY
Jiu River, one of the most important Danube tributaries in South-Western Romania, is about 340 km long, with a catchment area of 10,080 km2. The Jiu River basin is distinguished by the presence of coal mining areas (bituminous coal in the upper sector – the Petroșani depression, and lignite in the middle sector - the Motru-Rovinari basin), which impacts the composition of fine sediments through the incorporation of coal particles. The general goal of the paper is to determine the presence of coal and estimate the proportion of coaly particles in the fine sediments transferred within the Jiu watershed, by using visual-based methods, namely colorimetric and image analysis applied on 88 samples of fine river sediment collected from Jiu River and its main tributaries and on 5 samples of bulk coal taken from mining areas. The colour-based approach conducted to a better characterisation of the river sediments and coal samples in terms of the colour shades that could highlight the presence of coal. The best explanatory index proved to be a*, closely followed by b*, y, C* and v', which were further used in the automatic interaction detection models. There was no evident link between coal content seen from the coal grains surface (%) point of view through image analysis and the distribution of colorimetric indices for Jiu river and its tributaries. Still, the automatic interaction detection analysis succeeded to correctly classify the lignite and for a part of the bituminous coal samples (the fractions <20 µm). The remaining coarser fractions of bituminous coal samples were confused with lignite particles and could not be used as predictors for sediment samples with coal content
Show more [+] Less [-]AGROVOC Keywords
Bibliographic information
This bibliographic record has been provided by Directory of Open Access Journals