How do I use this formula for images?

[Pandamütze]

Hi there!

I'm new to Labview and I don't get on with the following problem.

What I'd like to do is use the following formula with three color bmps

I=sqrt[(I1-I2)^2+(I1-I3)^2+(I2-I3)^2]            I is the resulting intensity, I1,I2,I3 are the intensities from the three images.

At the moment, I use already saved images, later I'd like to get them from a camera (get 1. image, move Piezo; get 2. image, move Piezo...)

I tried to convert the images to arrays and apply the formula to them, but that didn't work.

So how do I do this calculation?

Is there a possibility to pass the images to Matlab? (I use Matlab for the calculation now and its working fine, but I need to automate the whole process)

I hope someone can help ^^

Pandamütze

PS: I use Labview Version 8.6 including LabviewVision

Message Edited by Pandamütze on 08-20-2009 03:01 AM

[muks]

Did you try math script in labVIEW?

[Pandamütze]

Hm... I tried but there was only the possibility to pass an array to matlab. it would have been nice if i was able to pass the image as it is, because the matlab script works with the images without converting them to arrays.

[muks]

---

Hm... I tried but there was only the possibility to pass an array to matlab. it would have been nice if i was able to pass the image as it is, because the matlab script works with the images without converting them to arrays

---

Yes you need to convert them to array and feed it. (sadly)

[GerdW]

Hi pandamütze,

you have 3 images. Are they always the same size?

If so I would get the image content as array, calc the "intensity" for each (you should define "intensity" clearly) and then do the error calculation with standard math without any loops/matlab/mathscript involved...

[Pandamütze]

the images are all of the size 2080x1542

I just read this "The basic data format for an image is a two-dimensional (2D) matrix, where the value of each element (a pixel) is the intensity at a physical location." which means I have to do the calculation for each Pixel. 

to get the "conventionel" image (the project is called "Structured illumination"), I use the imageadd to caculate (I1+I2+I3)/3.

[GerdW]

Hi pandamütze,

see example for calculation using standard math...

Your description of intensity is still vague. When using 24bit color images you usually get a 32bit value per pixel containing the RGB colors (each 8 bit). This is no intensity! Intensity can be seen as gray level of the pixel, but there are different conversion from RGB to graylevel possible...

[Pandamütze]

Hi!

this is what I came up with ^^ I'm not sure if its working like this, but after all, the result looks a bit better than before

thats what I got  yes, there is something ... ^^

and this is what I expected as the result of the calculation

the images I load in look similar, just with stripes on them (the point is to calculate an image from the focal plane by using three images with a projected grid pattern [pattern moved 1/3 of a period between the images])

[Pandamütze · Accepted solution]

problem almost solved. I get the right image, just a bit darker than the resulting image of the matlab calculation

I attache my latest VI if someone is interested

thanks to you, thinking about your questins helped a lot ^^

Kind regards

Pandamütze

[GerdW]

Hi Pandamütze,

you see the calculation is the smallest problem in LabView here. Most work is loading and saving the images

(Don't have IMAQ installed so can't comment on that. I would have used the standard picture functions...)