Re : Re: Re : Re: Re : Re: Store a 12-bit gray scale image

[CyGa]

I'm using a video camera to acquire images. This camera delivers 12-bit gray scale images so 4096 levels of grey. However, when I store these images on my HD on PNG format, the maximum value of a pixel is 256. Which is normal because the 3 colors R, V, B are coded onto 8 bits each. How is it possible to keep the signal dynamic of the image delivered by the camera? Does it exist an image format which would allow me to store the pixel values onto 12 bits ?

[zou]

You can't use Write PNG File.vi.  It's for 8 bit gray scale or 32 bit RGB color.

For 12/16 bit gray scale, you should use IMAQ.

George Zou

http://gtoolbox.yeah.net

[CyGa]

Acutally, I'm using IMAQ. But I need to store these images on my Hard Drive and be able to share them. So I need a format that Windows XP can easily read. How can I do that ?

[zou]

IMAQ support 16 bit gray scale.  What image type are you using?

Both PNG and TIFF support 16 gray scale.

George Zou

http://gtoolbox.yeah.net

[colman77]

There's always the last-resort solution, which is to save as a text file...  Not too good for sharing, but better than buying more software.

[CyGa]

I'm using the IMAQ function "IMAQ Write PNG file" to store the images. But then when you open these stored images with a basic software (such as Paint) each pixel is coded onto 256 levels. Do you think I should turn "Use bit depth?" option to True ?

[Neil.Pate]

If you just need to view the files, there are numerous free viewing utilities. Try http://www.irfanview.com/
It can display virtually any format you can think of!

[CyGa]

Indeed I need to display these images but I also have to view the value of each pixel on a 4096 gray scale and not on a scale of 256.

[chrisger]

hi there

well, if you have IMAQ, why don't you write a little viewer in LabVIEW? The IMAQ display delivers built-in tools for viewing a pixels value.

[colman77]

the problem isn't viewing the images; he can already do that well enough with the programs he's got!

The problem is the color (or in this case, greyscale) depth.  He's finding that the standard "output to PNG" function writes his 12-bit images as 8-bit images, which means data loss.

It seems that the outputted image is based on the "color palette" that you select... unfortunately it also seems that the only greyscale palette available limits you to 8 bpp... I'm not exactly a labview expert though, so I'd wait for someone with a cooler-looking rank thingy to confirm that.