This module presents how background subtraction is applied and what are the mathematical tools being used.
Background subtraction
What is background subtraction?
Background Subtraction is a process to detect a movement or significant differences inside of the video frame, when compared to a reference, and to remove all the non-significant components (background). Background subtraction is applied in many areas, such as surveillance system (to effectively segment the only moving object).
Steps to implement background subtraction
Learning Background–we captured ten background frames and calculated the mean(µ) and the standard deviation(σ) with the below equations
We assumed the value of the background was iid-normal distribution
calculation of mean and standard deviation. xi is the value at a particular pixel of ith image
Processing data (real time)–Distinguishing background and non-background objects
First, we set the threshold of background asµ±2σ
Each pixel in the background ->N ~ (µ,σ), if that pixel is part of the background, its value will lie withinµ±2σrange 95% of the time.
Any pixels that go beyond the threshold are considered as parts of non-background object.
Normal distribution of background pixels
The below figures are the result of implementing background subtraction
Figure a is the background frame we set. When non-background object is introduced (figure b), we can see the object is segmented out (figure d).
