LabVIEW
How to save TIFF 64 bit
Solved!
It depends what you assuming under "64-bit TIFF format".
Not every image processing library will support all possible image types, and if under "64 bit" you assuming 8 byte Double representation or 64-bit integer, then you have to use third-party library, like libtiff (as far as i remember, it could be possible to save 64 bit floats in double precision, not sure about integers). But both representations are very seldom used, usually 32 bit is fully sufficient for most processing needs.
On the other hand, if under "64-bit TIFF format" you assuming Big TIFF (sometimes known as Large TIFF) which is a TIFF variant file format which uses 64-bit offsets and supports much larger files than 4GB (or >2GB in case of signed offsets), then I can share small library, based on the message from Nico_MPI and modified by me to stay compliant with Big TIFF. It is multipage implementation (usually when you go over 4GB limit you have stack of the images)
Use it like this:
Multipage TIFFs can be opened in Fiji:
This library still pure LabVIEW implementation (don't expect huge performance), with no external libraries needs,except OpenG LargeFile used somewhere.
Feel free to use it as starting point and modify if needed. Downgraded to LV2018.
@ugg88 wrote:
I'm using the GigE vision camera and I want to save the captured image in 64-bit TIFF format. There are the options to save up to u16 bit TIFF for IMAQ. Is there any other way that we can save the images in 64-bit TIFF format?
As others have already said, it is not really clear what you want. In the subject, you are talking about "saving", but your attached VI is about "reading". There is a 64bit RGB version, but I seriously doubt that any camera, eye, or display device can support or distinguish 18446744073709551615 different levels of 64bit greyscale.
Reading a tiff is hard, unless you only want to deal with one specific type. Writing a tiff using plain LabVIEW is relatively simple. Just stick to the common tags and avoid compression. All documentation can be found on the internet.
@Andrey_Dmitriev wrote:
This library still pure LabVIEW implementation (don't expect huge performance), with no external libraries needs,except OpenG LargeFile used somewhere.
You mention the OpenG Large File library. i wonder as original author of that library why you would do that. LabVIEW 8.0 and newer supports full 64-bit file offsets in the file functions. The use of floating point numbers in that library as offsets is because there was no official U64 and I64 support in LabVIEW before 8.0 but it was really suboptimal to do like that.
in LabVIEW 8.0 and newer I would recommend to use the native file functions instead. Also will make the library work on other platforms than Windows.
@ugg88 wrote:
I'm getting different Image values when I save the image directly from NI max and VI.
Usually this happened if you have different camera settings when acquired in MAX or in LabVIEW using IMAQ like pixel depth, etc. In most cases it is sufficient to save camera's settings in MAX, then they will be used in LabVIEW:
The differences between 4095/65535 grays can be easily demonstrated with simple VI:
Two simple gradient images - one will have the only 4095 grays per channel, another one is full range.
Some cameras can stretch histogram to utilize full range between black and white points automatically. Just curious - which camera you have? I've used RGB U64 in the past only for my Nikon camera to analyze 14 bit RGB RAW files.
Another interesting point is that the image acquired in MAX will show "Bits per pixel: 16 (unsigned, grayscale LUT)", but the image saved from LabVIEW will be "Bits per pixel: 16 (unsigned, color LUT)":
but I don't know how to supply TIFF with 16 bit gray LUT. Are you 100% sure that your camera is really RGB with 16 bits per channel and not grayscale?
