morphology and HSV model for motion detection

yes @sakryukov, i just convert RGB to HSV and then i tried to extract frames .. and difference between frames .. but i am getting some bugs ..

ohh no Sir, I am trying to fix that .. I think i did not mention that its my first time in image processing work. So i am doing mistakes and getting errors.. i am trying to handle it ..

Actually i just want to know that morphology can help me in motion detection if so then how and why should i use morphology..
i saw one project thats was related to hand mouse motion. there coder applied first time dilate 5x5 mask and then erode . dilate .erode . dilate eroide five times with 7x7 mask and then in last coder applied dilate with 5x5 mask..


I did not understand his logic.. Why did you apply morphology like this ..

As i know that dilate expand the object and remove the noise .and erode shrink the object .

but to be really frank that how and why i should use morphology..Should i do like above coder thinking..

I did not understand his logic..

Can you please guide me thanks