Getting RGB pixel data from CGImage
Asked Answered
D

1

18

I am trying to access pixel data from a CGImage. I want to be able to access the RGB values as integers. I think I am nearly there with this code:

UIImage* theImage = [UIImage imageNamed:@"rgb.png"];
CGImageRef cgImageRef = CGImageRetain(theImage.CGImage);
CFDataRef* imageData = CGDataProviderCopyData(CGImageGetDataProvider(cgImageRef));
NSLog(@"the data = %@", imageData);

This then logs:

the data = <010002fe fffdff02 0200fe04 0003fc>

The image is a 5x1 png containing a black, white, red, green and blue pixel in that order.

I don't really understand what I am looking at here. How can I get an array of RGB values, or something similar so I can work with them.

Thanks, Rich

Dead answered 20/5, 2011 at 14:18 Comment(1)
possible duplicate of thisDebonair
L
58

Here's some sample code that will print a bunch of relevant info about an image, as well as a dump of the image's pixel data. It will work on images with alpha channels, as well as images without. The code will even work on images that are not in the RGB color space, though I doubt you'd be likely to get any of those on iOS.

Copy-paste it into your project and try to run it against a few of your image files, and then perhaps you can adapt it to your needs

-(void)imageDump:(NSString*)file
{
    UIImage* image = [UIImage imageNamed:file];
    CGImageRef cgimage = image.CGImage;

    size_t width  = CGImageGetWidth(cgimage);
    size_t height = CGImageGetHeight(cgimage);

    size_t bpr = CGImageGetBytesPerRow(cgimage);
    size_t bpp = CGImageGetBitsPerPixel(cgimage);
    size_t bpc = CGImageGetBitsPerComponent(cgimage);
    size_t bytes_per_pixel = bpp / bpc;

    CGBitmapInfo info = CGImageGetBitmapInfo(cgimage);

    NSLog(
        @"\n"
        "===== %@ =====\n"
        "CGImageGetHeight: %d\n"
        "CGImageGetWidth:  %d\n"
        "CGImageGetColorSpace: %@\n"
        "CGImageGetBitsPerPixel:     %d\n"
        "CGImageGetBitsPerComponent: %d\n"
        "CGImageGetBytesPerRow:      %d\n"
        "CGImageGetBitmapInfo: 0x%.8X\n"
        "  kCGBitmapAlphaInfoMask     = %s\n"
        "  kCGBitmapFloatComponents   = %s\n"
        "  kCGBitmapByteOrderMask     = 0x%.8X\n"
        "  kCGBitmapByteOrderDefault  = %s\n"
        "  kCGBitmapByteOrder16Little = %s\n"
        "  kCGBitmapByteOrder32Little = %s\n"
        "  kCGBitmapByteOrder16Big    = %s\n"
        "  kCGBitmapByteOrder32Big    = %s\n",
        file,
        (int)width,
        (int)height,
        CGImageGetColorSpace(cgimage),
        (int)bpp,
        (int)bpc,
        (int)bpr,
        (unsigned)info,
        (info & kCGBitmapAlphaInfoMask)     ? "YES" : "NO",
        (info & kCGBitmapFloatComponents)   ? "YES" : "NO",
        (info & kCGBitmapByteOrderMask),
        ((info & kCGBitmapByteOrderMask) == kCGBitmapByteOrderDefault)  ? "YES" : "NO",
        ((info & kCGBitmapByteOrderMask) == kCGBitmapByteOrder16Little) ? "YES" : "NO",
        ((info & kCGBitmapByteOrderMask) == kCGBitmapByteOrder32Little) ? "YES" : "NO",
        ((info & kCGBitmapByteOrderMask) == kCGBitmapByteOrder16Big)    ? "YES" : "NO",
        ((info & kCGBitmapByteOrderMask) == kCGBitmapByteOrder32Big)    ? "YES" : "NO"
    );

    CGDataProviderRef provider = CGImageGetDataProvider(cgimage);
    NSData* data = (id)CGDataProviderCopyData(provider);
    [data autorelease];
    const uint8_t* bytes = [data bytes];

    printf("Pixel Data:\n");
    for(size_t row = 0; row < height; row++)
    {
        for(size_t col = 0; col < width; col++)
        {
            const uint8_t* pixel =
                &bytes[row * bpr + col * bytes_per_pixel];

            printf("(");
            for(size_t x = 0; x < bytes_per_pixel; x++)
            {
                printf("%.2X", pixel[x]);
                if( x < bytes_per_pixel - 1 )
                    printf(",");
            }

            printf(")");
            if( col < width - 1 )
                printf(", ");
        }

        printf("\n");
    }
}

Here's some sample output on two images I tried. They're both 5x3 rgb. The "a.png" image has an alpha channel, while the "b.rgb" does not.

===== a.png =====
CGImageGetHeight: 5
CGImageGetWidth:  3
CGImageGetColorSpace: <CGColorSpace 0x4d08ff0> (kCGColorSpaceDeviceRGB)
CGImageGetBitsPerPixel:     32
CGImageGetBitsPerComponent: 8
CGImageGetBytesPerRow:      20
CGImageGetBitmapInfo: 0x00000003
  kCGBitmapAlphaInfoMask     = YES
  kCGBitmapFloatComponents   = NO
  kCGBitmapByteOrderMask     = NO
  kCGBitmapByteOrderDefault  = NO
  kCGBitmapByteOrder16Little = NO
  kCGBitmapByteOrder32Little = NO
  kCGBitmapByteOrder16Big    = NO
  kCGBitmapByteOrder32Big    = NO
Pixel Data:
(00,00,00,FF), (FF,FF,FF,FF), (FF,00,00,FF), (00,FF,00,FF), (00,00,FF,FF)
(00,00,00,FF), (FF,FF,FF,FF), (FF,00,00,FF), (00,FF,00,FF), (00,00,FF,FF)
(FF,FF,FF,00), (FF,FF,FF,00), (FF,FF,FF,00), (FF,FF,FF,00), (FF,FF,FF,00)


===== b.png =====
CGImageGetHeight: 5
CGImageGetWidth:  3
CGImageGetColorSpace: <CGColorSpace 0x4d08ff0> (kCGColorSpaceDeviceRGB)
CGImageGetBitsPerPixel:     24
CGImageGetBitsPerComponent: 8
CGImageGetBytesPerRow:      15
CGImageGetBitmapInfo: 0x00000000
  kCGBitmapAlphaInfoMask     = NO
  kCGBitmapFloatComponents   = NO
  kCGBitmapByteOrderMask     = NO
  kCGBitmapByteOrderDefault  = NO
  kCGBitmapByteOrder16Little = NO
  kCGBitmapByteOrder32Little = NO
  kCGBitmapByteOrder16Big    = NO
  kCGBitmapByteOrder32Big    = NO
Pixel Data:
(00,00,00), (FF,FF,FF), (FF,00,00), (00,FF,00), (00,00,FF)
(00,00,00), (FF,FF,FF), (FF,00,00), (00,FF,00), (00,00,FF)
(00,00,00), (FF,FF,FF), (FF,00,00), (00,FF,00), (00,00,FF)
Library answered 20/5, 2011 at 21:29 Comment(3)
bytes[row * bpr + col * bytes_per_pixel] Last time I wrote code like that (admittedly a long time ago on old-by-today's-standards hardware), doing that for every pixel was slow for large images. Incrementing a pointer and using addition rather than multiplication will be faster. You might think that's no big deal but a large image will make the difference, those pixels add up fast.Milden
Thanks for this, useful. I noticed you're printing the width as the height and vice versa.Nate
Don't forget to investigate the AlphaInfoMask in detail, certain graphics functions are fussy about which to accept.Trochee

© 2022 - 2024 — McMap. All rights reserved.