Error when accessing raw pixel data of resized image - ios

Question

I have some problems with access to camera images (or even images from photalbum).
After resizing the UIImage (i tested several different resize methods, they all lead to the same error) I want to access every individual pixel for handing over to a complex algorithm.
The problem is, that there is often a bytesPerRow value that doesn't match to the image size (eg. width*4) when accessing raw pixel data with CGImageGetDataProvider -> resulting in a EXC_BAD_ACCESS error.
Maybe we have an iOS bug here…
Nonetheless, here is the code:
// UIImage capturedImage from Camera
CGImageRef capturedImageRef = capturedImage.CGImage;
// getting bits per component from capturedImage
size_t bitsPerComponentOfCapturedImage = CGImageGetBitsPerComponent(capturedImageRef);
CGImageAlphaInfo alphaInfoOfCapturedImage = CGImageGetAlphaInfo(capturedImageRef);
CGColorSpaceRef colorSpace = CGColorSpaceCreateDeviceRGB();
// calculate new size from interface data.
// with respect to aspect ratio
// ...
// newWidth = XYZ;
// newHeight = XYZ;
CGContextRef context = CGBitmapContextCreate(NULL, newWidth, newHeight, bitsPerComponentOfCapturedImage,0 , colorSpace, alphaInfoOfCapturedImage);
// I also tried to make use getBytesPerRow for CGBitmapContextCreate resulting in the same error
// if image was rotated
if(capturedImage.imageOrientation == UIImageOrientationRight) {
CGContextRotateCTM(context, -M_PI_2);
CGContextTranslateCTM(context, -newHeight, 0.0f);
}
// draw on new context with new size
CGContextDrawImage(context, CGRectMake(0, 0, newWidth, newHeight), capturedImage.CGImage);
CGImageRef scaledImage=CGBitmapContextCreateImage(context);
// release
CGColorSpaceRelease(colorSpace);
CGContextRelease(context);
theImage = [UIImage imageWithCGImage: scaledImage];
CGImageRelease(scaledImage);
After that, I want to access the scaled image by
CGImageRef imageRef = theImage.CGImage;
NSData *data = (NSData *) CGDataProviderCopyData(CGImageGetDataProvider(imageRef));
unsigned char *pixels = (unsigned char *)[data bytes];
// create a new image from the modified pixel data
size_t width = CGImageGetWidth(imageRef);
size_t height = CGImageGetHeight(imageRef);
size_t bitsPerComponent = CGImageGetBitsPerComponent(imageRef);
size_t bitsPerPixel = CGImageGetBitsPerPixel(imageRef);
size_t bytesPerRow = CGImageGetBytesPerRow(imageRef);
CGDataProviderRef provider = CGDataProviderCreateWithData(NULL, pixels, [data length], NULL);
NSLog(#"bytesPerRow: %f ", (float)bytesPerRow);
NSLog(#"Image width: %f ", (float)width);
NSLog(#"Image height: %f ", (float)height);
// manipulate the individual pixels
for(int i = 0; i < [data length]; i += 4) {
// accessing (float) pixels[i];
// accessing (float) pixels[i+1];
// accessing (float) pixels[i+2];
}
So for example when I access an image with 511x768 Pixel and scale that down to 290x436 I get following output:
Image width: 290.000000
Image height: 436.000000
bitsPerComponent: 8.000000
bitsPerPixel: 32.000000
bytesPerRow: 1184.000000
and you can clearly see that the bytesPerRow (althought choosen automatically by cocoa) does not match to the image width.
I would love to see any help
Using iOS SDK 4.3 on Xcode 4

Answer 1

You are ignoring possible line-padding, hence receiving invalid results. Add the following code and replace your loop;
size_t bytesPerPixel = bitsPerPixel / bitsPerComponent;
//calculate the padding just to see what is happening
size_t padding = bytesPerRow - (width * bytesPerPixel);
size_t offset = 0;
// manipulate the individual pixels
while (offset < [data length])
{
for (size_t x=0; x < width; x += bytesPerPixel)
{
// accessing (float) pixels[offset+x];
// accessing (float) pixels[offset+x+1];
// accessing (float) pixels[offset+x+2];
}
offset += bytesPerRow;
};
Addendum: Underlying reasoning for the row-padding is optimizing the memory access for individual rows to 32bit boundaries. It is indeed very common and is done for optimization purposes.