Assume that we have a video sequence and want to detect moving shadows.
it is clear that image intensity in each(R,G,B) channels decrease in shadow because light source is partially blocked by a moving object. But, my question is how to find the amount of intensity changing in all channels. If, image intensity value in location x before shadow is (50,100,150) and now(there is a shadow) we know that intensity in R channel is 20. what is intensity in G and B channels?
is it (70,120) or (40,60) or anything else?
Thank in advance
Hello,
just a small idea: why not using another color space as HSV ?
In this space shadow doen't affect the color but only the brightness ...
In fact the color of an object is not affected by the shadow ...
I am not an expert of this topic but I thing this is a possible way ...
Kind regards
Pierre
It is logical to assume that image intensity will change approximately equal value for all RGB channels.
Dear Ruslan, It seems to be logical but, it does not happen in real images!!
Hello,
sorry but I do not agree. there is nothing 'logical'. For example if you want to convert a color RGB image to grayscale you should take in consideration the human sensibility and one of the formulas to obtain a gray image from a color one in RGB space is
0.21 R + 0.72 G + 0.07 B
So I propose to pass in HSV color space where one component is dedicated to brigthness and so doen't affect 'color'. In this color space the shadow should affct only this coordinate.
I think some work is already done in this direcction
http://www.atlantis-press.com/php/download_paper.php?id=10471
http://thesai.org/Downloads/Volume4No1/Paper_26-Shadow_Suppression_using_RGB_and_HSV_Color_Space_in_Moving_Object_Detection.pdf
Regards
- Badges
- Science method
- Similar topics
- Biological Science
- Genetics
- Genetic Analysis
- genetic techniques
- Sequencing