Comparative Analysis of Digital Underwater Video Image Color Balance Correction Algorithms
Kirillov, S. N., Skonnikov, P. N., Baukov, A. A.
Article language: English
Abstract. Digital video image color balance correction algorithms are analyzed. The procedure of selecting the objects presented in a video sequence is proposed to be used in order to make a comparative analysis of color information: hue and saturation in HSL model, and chroma as well. The indices of color balance correction quality are substantiated. To carry out full-reference assessment, the groups of video sequences were chosen where various underwater vehicles filmed inside and outside the water can be found. The selection of underwater vehicles in images was made; the parameters of color information of the vehicles selected were estimated in the following video images: reference, source (underwater) and processed with the help of color balance correction algorithms under consideration. A comparative analysis of color correction methods for video recordings taken in aquatic en-vironment was carried out in order to identify the best algorithm according to the criterion of the clos-est color information of objects presented in an underwater video image to color information of corresponding objects not being distorted by aquatic environment. The results of the comparison have shown that the usage of Gray World Assumption algorithm leads to an improvement in hue up to 6.12 times with saturation loss of 8.8 %, and the usage of White Patch Retinex algorithm – up to 4.43 times with a loss of saturation of 22.4 %.
Keywords: video processing, color balance, aquatic environment, object selection, hue, saturation, chroma
Pages: 51–58Total pages: 8
Year of publication: 2019