在C或LabView中生成的16位灰度图像的结构

Question

I have a kind of basic (but still unresolved) problem. I'm building a program in LabView in which I gather a large amount of measurements into anxm matrix, and I need to convert that matrix into an image to be processed/analyzed with ImageJ. Due to resolution and the kind of analysis I need, this has to be a 16-bit grayscale image.

Now, the problems I'm facing are the following: LabView does not have (on the basic package) any VI to transform data into 16-bit grayscale, only 8-bit, which kills my dynamic range. I've read around that the Vision and Image toolboxes of LabView might have more options in that sense, but for company-related reasons I cannot access these toolboxes (ie I have to make do with the Full Developer version only).

My reaction at the moment is to try and build on my own a VI to build a 16-bit grayscale image, whether by using standard LabView functions, entirely with C code, or mixtures of both. However, I haven't found any clear-enough information on the web regarding how a 16-bit grayscale is structured, what data is needed, etc.

So, my question is twofold: does anyone have a better way of creating this 16-bit grayscale image (some LabView VI I've skipped, external software I can invoke, etc.)? If not, does anyone have detailed information or code useful to generating such an image from scratch?

Any help will be really useful! Thanks!

Answer 1

Install IMAQ module (different than the vision toolbox), it is free and has basic image management capabilities. I use it for creating 16 bit grayscale images.

Find IMAQ driver here (4.6.4 as of June 2012).

Answer 2

Since you mention 16-bit grayscale, I interpret that to mean that you need 16 bits of precision per pixel.

So, if you're failing to create that directly as grayscale, perhaps you can create it as a true-color image, and simply encode your sixteen bits by splitting them into eg the red and green components?

So for every 16-bit pixel value x , you would set red = x / 256; , green = x % 256; and blue = 0; .

Answer 3

Did you already try posting this question to the NI Forums as well? There may be some other LabVIEW developers who have tried to accomplish the same thing you're tackling here. Hopefully you won't have to start from scratch. https://bit.ly/NIForums

Have you checked to see if you can get your hands on any trial versions of the LabVIEW vision toolkits/addons?

And if your company has a SSP (paid support option although you typically get one year free when you buy LabVIEW), I'd also suggest you give the support engineers a call. They can coach you through some in's and out's of getting around some of the current limitations.