UDC 004.932
MULTICRITERIA SYNTHESIS OF THE ALGORITHM FOR DETECTING RAIN PARTICLES ON VIDEO IMAGES
IN DIGITAL TELEVISION SYSTEMS
V. T. Dmitriev, Ph.D. (Tech.), associate professor, Head of Radio Control and Communications Department, RSREU, Ryazan, Russia;
orcid.org/0000-0001-5521-6886, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
A. A. Baukov, post-graduate student, RSREU, Ryazan, Russia;
orcid.org/0000-0002-0003-058X, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
The problem of reducing the visibility of precipitation traces on video sequences is considered. An important
component of solving this problem is the detection of precipitation particles in a frame. The aim of
the work is to create an algorithm for detecting raindrops on video images in digital television systems when
shooting with static cameras. Time distributions of pixel intensity values exposed and unaffected by rain particles
have been studied. An algorithm for detecting rain particles is proposed, which is a three-stage pixels
of a group of n frames classification into areas with moving objects (cars, people, etc.), rain and non-rain
areas, as well as rain areas pixels of current frame into rain (distorted by a drop) and non-rain with the ability
to automatically select local threshold values based on accumulated information. The method of multicriteria
synthesis was used to select the optimal value of number of accumulated frames. It is established
that the application of proposed algorithm with optimal values of 100 and 130 frames when detecting drops
on video images with high and medium intensity rain, respectively, reduces normalized error levels of the
first and second kind to 2,7 % and 3,1 %. The gain of the developed algorithm in comparison with known
ones up to 1 % and by 4 %...9 % is obtained according to these quality criteria.
Key words: video processing, raindrop detection, classification, pixel intensity distribution, local thresholds, quality criteria, multi-criteria synthesis, normalized error levels of the first and second kind.