In Mean Background subtraction or mean filter background subtraction the background is the mean of the previous n frames
( Σ i=0 to n-1 (It- i(x,y)) )/ n
In Median background subtraction the background is the median of the previous n frames
median{ I t-i (x,y) }
i ∈ { 0,1,2,3, ... n-1 }
The mathematics that i used in approximate median background subtraction was:
if It (x,y) > B t(x,y) ⇒ B t(x,y) + 1
if It (x,y) < B t(x,y) ⇒ B t(x,y) - 1
This method will eventually converge where half of the pixels are greater than the background and another half less than the background.
Results can be found here:
Result
No comments:
Post a Comment