Prev: relation b/w bits per pixel(bpp) and compression ratio
Next: using QccPack
From: Leon on 17 Mar 2010 03:38
I am running into two issues trying to read and write bmp images.
My images are grayscale 24bit BMPs exported from photoshop.

first I read the file

[I,map] = imread(filename,'bmp');

it runs without error but map is an empty matrix. This creates an error with imwrite if I use

imwrite(I, map, newfilename,'bmp');

I can circumvent the error by
map = gray(256);

But than I cannot find how to make the new file be a 24bit and not 8bit BMP. As far as I could see in the help files, there is no 'bitdepth' paramater for bmp files.

Thanks,

Leon
From: Walter Roberson on 17 Mar 2010 03:59
Leon wrote:
> I am running into two issues trying to read and write bmp images.
> My images are grayscale 24bit BMPs exported from photoshop.
>
> first I read the file
>
> [I,map] = imread(filename,'bmp');
>
> it runs without error but map is an empty matrix.

That tells you that as far as imread() is concerned, the file is a
truecolor image rather than a pseudo-color image; 'I' will be M x N x 3;
if it is truly grayscale then the R, G, and B components will all be equal.

>This creates an error
> with imwrite if I use
> imwrite(I, map, newfilename,'bmp');
>
> I can circumvent the error by map = gray(256);
>
> But than I cannot find how to make the new file be a 24bit and not 8bit
> BMP. As far as I could see in the help files, there is no 'bitdepth'
> paramater for bmp files.

If I remember correctly (and it is nearly 3 AM here, so I might be
wrong), the bit depth that will be used will depend upon the data class
of the image and upon whether it is a truecolor image or a pseudo-color
image. For 24 bit grayscale, you may need to use data of class uint32 .
 | 
Pages: 1
Prev: relation b/w bits per pixel(bpp) and compression ratio
Next: using QccPack