Analyze Live Video Image and Quantify Variations in image.

[corrado33]

Hi all,

I would like to know the best way to do something. Basically I have what is essentially a webcam looking down a microscope. The image I'm looking at is essentially a bright blue spot in the middle of all black. That blue spot changes over time. I would like to quantify the changes in that spot size/shape in real time. The video is NOT saved, so it has to be done with a screen capture or having the program looking directly at what's being displayed on the screen. I can make the window of the program with the video go to the same spot every time, that's not a problem. What's the best way to do this? I don't need the EXACT numbers for shape and size of the spot, but rather I'd like to simply have the program export a number if possible. The spot is weird in that when it gets bigger, it gets dimmer (darker blues) and when it gets smaller, it gets brighter (which will show white if it gets bright enough.) Really I want to use autoit to export a number to another program which will, in turn, keep the spot size the same. 

I really just want to know the functions I need, I can read the manual and figure them out from there. 

I did something sorta similar to this many years ago for a game (Mass Effect 2 IIRC), but I was only looking at a few pixels there, and waiting for a change in color. So it's not quite the same. (For those of you who played, it was for the stupid "resource gathering mini game" where you're looking at the planet and have to fire probes down to it to get resources.)

I would also like to have the ability to "save" the initial image, and compare the subsequent video to that image. 

I'm doing this for a few reasons. Professional versions of what I'm trying to create are many thousands of dollars. I don't have many thousands of dollars at my disposal. 

EDIT: I've attached two pictures showing a "small" spot and a "large" spot.laser.bmplaser2.bmp

Edited January 6, 2015 by corrado33

[Jfish]

There was a similar post recently on someone >observing fish in a tank.   That person was tracking movements in real time. It may be a good place to start.