如何在Cocoa Mac应用程序中合并大型JPEG

Question

In our Mac application built in Cocoa, we would like to combine 5 large JPEG images into one JPEG. We have the following images:

• 1.jpeg 50000px wide X 5000px high
• 2.jpeg 50000px wide X 5000px high
• 3.jpeg 50000px wide X 5000px high
• 4.jpeg 50000px wide X 5000px high
• 5.jpeg 50000px wide X 5000px high

We would like to combine these images one above the other to form an output jpeg:

50000px wide X 25000px high

The problem is that the resulting JPEG is very large and this results in memory issues when we use the following approach to make output JPEG.

  NSRect imageRect = NSMakeRect(0.0, 0.0, 50000, 25000);

  NSBitmapImageRep *savedImageBitmapRep = [[NSBitmapImageRep alloc]
      initWithBitmapDataPlanes:nil
                    pixelsWide:imageRect.size.width
                    pixelsHigh:imageRect.size.height
                 bitsPerSample:8
               samplesPerPixel:4
                      hasAlpha:YES
                      isPlanar:NO
                colorSpaceName:NSCalibratedRGBColorSpace
                  bitmapFormat:0
                   bytesPerRow:(4 * imageRect.size.width)
                  bitsPerPixel:32];

  [NSGraphicsContext saveGraphicsState];
  [NSGraphicsContext
      setCurrentContext:[NSGraphicsContext graphicsContextWithBitmapImageRep:
                                               savedImageBitmapRep]];

  // Read 1.jpeg, 2.jpeg, 3.jpeg, 4.jpeg, 5.jpeg as NSImage 
  // and draw them on the current context in their respective location

  [NSGraphicsContext restoreGraphicsState];

  NSMutableData *imageData = [NSMutableData data];
  CGImageDestinationRef imageDest = CGImageDestinationCreateWithData(
        (__bridge CFMutableDataRef)imageData, kUTTypeJPEG, 1, NULL);

  CGImageDestinationAddImage(imageDest, [savedImageBitmapRep CGImage],
                             (__bridge CFDictionaryRef)properties);
  CGImageDestinationFinalize(imageDest);
  if (imageDest != NULL) {
    CFRelease(imageDest);
  }
  //write imageData to a JPEG file

How can we achieve our objective without facing memory issues?

Answer 1

"Memory issues" is vague but I assume you simply mean that you want to avoid the 4.5+ GBs of memory required to hold 50,000 x 25,000 pixels.

The only way to avoid that is to drop down to a lower level API which lets you encode each MCU row (each 8 or 16 rows of pixels) separately at a time. libjpeg supports this, for example. You would load and render the minimum required data to produce such a row with whatever minimum memory constraints you have in mind, and then pass that row off to the encoder, before moving to the next row.

Answer 2

You can resize them to be smaller for displaying. but only when saving/storing you can directly store all the sizes