Extracting images from a PDF

I am having a little query regarding Extracting specifically images ( only images ) from a supplied PDF document in iPhone Application.

I have gone through the documentation of apple - But I am failure to find it.

I have done following efforts to have the Image from PDF Document.

-(IBAction)btnTappedImages:(id)sender{

    // MyGetPDFDocumentRef is custom c method 
    // & filePath is path to pdf document.
    CGPDFDocumentRef document = MyGetPDFDocumentRef ([filePath UTF8String]);

    int pgcnt = CGPDFDocumentGetNumberOfPages( document );

    for( int i1 = 0; i1 < pgcnt; ++i1 ) {
        // 1. Open Document page 
        CGPDFPageRef pg = CGPDFDocumentGetPage (document, i1+1);
        if( !pg ) {
            NSLog(@"Couldn't open page.");
        }
        // 2. get page dictionary
        CGPDFDictionaryRef dict = CGPDFPageGetDictionary( pg );
        if( !dict ) {
            NSLog(@"Couldn't open page dictionary.");
        }
        // 3开发者_C百科. get page contents stream
        CGPDFStreamRef cont;
        if( !CGPDFDictionaryGetStream( dict, "Contents", &cont ) ) {
            NSLog(@"Couldn't open page stream.");
        }
        // 4. copy page contents steam
        //  CFDataRef contdata = CGPDFStreamCopyData( cont, NULL );

        // 5. get the media array from stream
        CGPDFArrayRef media;
        if( !CGPDFDictionaryGetArray( dict, "MediaBox", &media ) ) {
            NSLog(@"Couldn't open page Media.");
        }

        // 6. open media & get it's size
        CGPDFInteger mediatop, medialeft;
        CGPDFReal mediaright, mediabottom;
        if( !CGPDFArrayGetInteger( media, 0, &mediatop ) || !CGPDFArrayGetInteger( media, 1, &medialeft ) || !CGPDFArrayGetNumber( media, 2, &mediaright ) || !CGPDFArrayGetNumber( media, 3, &mediabottom ) ) {
            NSLog(@"Couldn't open page Media Box.");
        }

        // 7. set media size 
        double mediawidth = mediaright - medialeft, mediaheight = mediabottom - mediatop;

        // 8. get media resources
        CGPDFDictionaryRef res;
        if( !CGPDFDictionaryGetDictionary( dict, "Resources", &res ) ) {
            NSLog(@"Couldn't Open Page Media Reopsources.");
        }

        // 9. get xObject from media resources  
        CGPDFDictionaryRef xobj;
        if( !CGPDFDictionaryGetDictionary( res, "XObject", &xobj ) ) {
            NSLog(@"Couldn't load page Xobjects.");
        }

        char imagestr[16];
        sprintf( imagestr, "Im%d", i1 );

        // 10. get x object stream
        CGPDFStreamRef strm;
        if( !CGPDFDictionaryGetStream( xobj, imagestr, &strm ) ) {
            NSLog(@"Couldn't load stream for xObject");
        }

        // 11. get dictionary from xObject Stream
        CGPDFDictionaryRef strmdict = CGPDFStreamGetDictionary( strm );
        if( !strmdict ) {
            NSLog(@"Failed to load dictionary for xObject");
        }

        // 12. get type of xObject
        const char * type;
        if( !CGPDFDictionaryGetName( strmdict, "Type", &type ) || strcmp(type, "XObject" ) ) {
            NSLog(@"Couldn't load xObject Type");
        }

        // 13. Check weather subtype is image or not
        const char * subtype;
        if( !CGPDFDictionaryGetName( strmdict, "Subtype", &subtype ) || strcmp( subtype, "Image" ) ) {
            NSLog(@"xObject is not image");
        }

        // 14. Bits per component
        CGPDFInteger bitsper;
        if( !CGPDFDictionaryGetInteger( strmdict, "BitsPerComponent",&bitsper ) || bitsper != 1 ) {
            NSLog(@"Bits per component not loaded");
        }

        // 15. Type of filter of image
        const char * filter;
        if( !CGPDFDictionaryGetName( strmdict, "Filter", &filter ) || strcmp( filter, "FlateDecode" ) ) {
            NSLog(@"Filter not loaded");
        }

        // 16. Image height width   
        CGPDFInteger width, height;
        if( !CGPDFDictionaryGetInteger( strmdict, "Width", &width ) || !CGPDFDictionaryGetInteger( strmdict, "Height", &height ) ) {
            NSLog(@"Image Height - width not loaded.");
        }

        // 17. Load image bytes & verify it
        CGPDFDataFormat fmt = CGPDFDataFormatRaw;
        CFDataRef data = CGPDFStreamCopyData( strm, &fmt );

        int32_t len = CFDataGetLength( data );
        const void * bytes = CFDataGetBytePtr( data );

        // now I have bytes for images in "bytes" pointer the problem is how to covert it into UIImage

        NSLog(@"Image bytes length - %i",len);  
        int32_t rowbytes = (width + 7) / 8;
        if( rowbytes * height != len ) {
            NSLog(@"Invalid Image");
        }

        double xres = width / mediawidth * 72.0, yres = height / mediaheight * 72.0;
        xres = round( xres * 1000 ) / 1000;
        yres = round( yres * 1000 ) / 1000;
    }
}

Yes ! I found it. But It looks very scary - huge code.

NSMutableArray *aRefImgs;
void setRefImgs(NSMutableArray *ref){
    aRefImgs=ref;
}

NSMutableArray* ImgArrRef(){
    return aRefImgs;
}

CGPDFDocumentRef MyGetPDFDocumentRef (const char *filename) {
    CFStringRef path;
    CFURLRef url;
    CGPDFDocumentRef document;
    path = CFStringCreateWithCString (NULL, filename,kCFStringEncodingUTF8);
    url = CFURLCreateWithFileSystemPath (NULL, path, kCFURLPOSIXPathStyle, 0);
    CFRelease (path);
    document = CGPDFDocumentCreateWithURL (url);// 2
    CFRelease(url);
    int count = CGPDFDocumentGetNumberOfPages (document);// 3
    if (count == 0) {
        printf("`%s' needs at least one page!", filename);
        return NULL;
    }
    return document;
}


CGFloat *decodeValuesFromImageDictionary(CGPDFDictionaryRef dict, CGColorSpaceRef cgColorSpace, NSInteger bitsPerComponent) {
    CGFloat *decodeValues = NULL;
    CGPDFArrayRef decodeArray = NULL;

    if (CGPDFDictionaryGetArray(dict, "Decode", &decodeArray)) {
        size_t count = CGPDFArrayGetCount(decodeArray);
        decodeValues = malloc(sizeof(CGFloat) * count);
        CGPDFReal realValue;
        int i;
        for (i = 0; i < count; i++) {
            CGPDFArrayGetNumber(decodeArray, i, &realValue);
            decodeValues[i] = realValue;
        }
    } else {
        size_t n;
        switch (CGColorSpaceGetModel(cgColorSpace)) {
            case kCGColorSpaceModelMonochrome:
                decodeValues = malloc(sizeof(CGFloat) * 2);
                decodeValues[0] = 0.0;
                decodeValues[1] = 1.0;
                break;
            case kCGColorSpaceModelRGB:
                decodeValues = malloc(sizeof(CGFloat) * 6);
                for (int i = 0; i < 6; i++) {
                    decodeValues[i] = i % 2 == 0 ? 0 : 1;
                }
                break;
            case kCGColorSpaceModelCMYK:
                decodeValues = malloc(sizeof(CGFloat) * 8);
                for (int i = 0; i < 8; i++) {
                    decodeValues[i] = i % 2 == 0 ? 0.0 :
                    1.0;
                }
                break;
            case kCGColorSpaceModelLab:
                // ????
                break;
            case kCGColorSpaceModelDeviceN:
                n =
                CGColorSpaceGetNumberOfComponents(cgColorSpace) * 2;
                decodeValues = malloc(sizeof(CGFloat) * (n *
                                                         2));
                for (int i = 0; i < n; i++) {
                    decodeValues[i] = i % 2 == 0 ? 0.0 :
                    1.0;
                }
                break;
            case kCGColorSpaceModelIndexed:
                decodeValues = malloc(sizeof(CGFloat) * 2);
                decodeValues[0] = 0.0;
                decodeValues[1] = pow(2.0,
                                      (double)bitsPerComponent) - 1;
                break;
            default:
                break;
        }
    }

    return (CGFloat *)CFMakeCollectable(decodeValues);
}

UIImage *getImageRef(CGPDFStreamRef myStream) {
    CGPDFArrayRef colorSpaceArray = NULL;
    CGPDFStreamRef dataStream;
    CGPDFDataFormat format;
    CGPDFDictionaryRef dict;
    CGPDFInteger width, height, bps, spp;
    CGPDFBoolean interpolation = 0;
    //  NSString *colorSpace = nil;
    CGColorSpaceRef cgColorSpace;
    const char *name = NULL, *colorSpaceName = NULL, *renderingIntentName = NULL;
    CFDataRef imageDataPtr = NULL;
    CGImageRef cgImage;
    //maskImage = NULL,
    CGImageRef sourceImage = NULL;
    CGDataProviderRef dataProvider;
    CGColorRenderingIntent renderingIntent;
    CGFloat *decodeValues = NULL;
    UIImage *image;

    if (myStream == NULL)
        return nil;

    dataStream = myStream;
    dict = CGPDFStreamGetDictionary(dataStream);

    // obtain the basic image information
    if (!CGPDFDictionaryGetName(dict, "Subtype", &name))
        return nil;

    if (strcmp(name, "Image") != 0)
        return nil;

    if (!CGPDFDictionaryGetInteger(dict, "Width", &width))
        return nil;

    if (!CGPDFDictionaryGetInteger(dict, "Height", &height))
        return nil;

    if (!CGPDFDictionaryGetInteger(dict, "BitsPerComponent", &bps))
        return nil;

    if (!CGPDFDictionaryGetBoolean(dict, "Interpolate", &interpolation))
        interpolation = NO;

    if (!CGPDFDictionaryGetName(dict, "Intent", &renderingIntentName))
        renderingIntent = kCGRenderingIntentDefault;
    else{
        renderingIntent = kCGRenderingIntentDefault;
        //      renderingIntent = renderingIntentFromName(renderingIntentName);
    }

    imageDataPtr = CGPDFStreamCopyData(dataStream, &format);
    dataProvider = CGDataProviderCreateWithCFData(imageDataPtr);
    CFRelease(imageDataPtr);

    if (CGPDFDictionaryGetArray(dict, "ColorSpace", &colorSpaceArray)) {
        cgColorSpace = CGColorSpaceCreateDeviceRGB();
        //      cgColorSpace = colorSpaceFromPDFArray(colorSpaceArray);
        spp = CGColorSpaceGetNumberOfComponents(cgColorSpace);
    } else if (CGPDFDictionaryGetName(dict, "ColorSpace", &colorSpaceName)) {
        if (strcmp(colorSpaceName, "DeviceRGB") == 0) {
            cgColorSpace = CGColorSpaceCreateDeviceRGB();
            //          CGColorSpaceCreateWithName(kCGColorSpaceGenericRGB);
            spp = 3;
        } else if (strcmp(colorSpaceName, "DeviceCMYK") == 0) {     
            cgColorSpace = CGColorSpaceCreateDeviceCMYK();
            //          CGColorSpaceCreateWithName(kCGColorSpaceGenericCMYK);
            spp = 4;
        } else if (strcmp(colorSpaceName, "DeviceGray") == 0) {
            cgColorSpace = CGColorSpaceCreateDeviceGray();
            //          CGColorSpaceCreateWithName(kCGColorSpaceGenericGray);
            spp = 1;
        } else if (bps == 1) { // if there's no colorspace entry, there's still one we can infer from bps
            cgColorSpace = CGColorSpaceCreateDeviceGray();
            //          colorSpace = NSDeviceBlackColorSpace;
            spp = 1;
        }
    }

    decodeValues = decodeValuesFromImageDictionary(dict, cgColorSpace, bps);

    int rowBits = bps * spp * width;
    int rowBytes = rowBits / 8;
    // pdf image row lengths are padded to byte-alignment
    if (rowBits % 8 != 0)
        ++rowBytes;

//  maskImage = SMaskImageFromImageDictionary(dict);

    if (format == CGPDFDataFormatRaw)
    {
        sourceImage = CGImageCreate(width, height, bps, bps * spp, rowBytes, cgColorSpace, 0, dataProvider, decodeValues, interpolation, renderingIntent);
        CGDataProviderRelease(dataProvider);
        cgImage = sourceImage;
//      if (maskImage != NULL) {
//          cgImage = CGImageCreateWithMask(sourceImage, maskImage);
//          CGImageRelease(sourceImage);
//          CGImageRelease(maskImage);
//      } else {
//          cgImage = sourceImage;
//      }
    } else {
        if (format == CGPDFDataFormatJPEGEncoded){ // JPEG data requires a CGImage; AppKit can't decode it {
            sourceImage =
            CGImageCreateWithJPEGDataProvider(dataProvider,decodeValues,interpolation,renderingIntent);
            CGDataProviderRelease(dataProvider);
            cgImage = sourceImage;
//          if (maskImage != NULL) {
//              cgImage = CGImageCreateWithMask(sourceImage,maskImage);
//              CGImageRelease(sourceImage);
//              CGImageRelease(maskImage);
//          } else {
//              cgImage = sourceImage;
//          }
        }
        // note that we could have handled JPEG with ImageIO as well
        else if (format == CGPDFDataFormatJPEG2000) { // JPEG2000 requires ImageIO {
            CFDictionaryRef dictionary = CFDictionaryCreate(NULL, NULL, NULL, 0, NULL, NULL);
            sourceImage=
            CGImageCreateWithJPEGDataProvider(dataProvider, decodeValues, interpolation, renderingIntent);


            //          CGImageSourceRef cgImageSource = CGImageSourceCreateWithDataProvider(dataProvider, dictionary);
            CGDataProviderRelease(dataProvider);

            cgImage=sourceImage;

            //          cgImage = CGImageSourceCreateImageAtIndex(cgImageSource, 0, dictionary);
            CFRelease(dictionary);
        } else // some format we don't know about or an error in the PDF
            return nil;
    }
    image=[UIImage imageWithCGImage:cgImage];   
    return image;
}

@implementation DashBoard

// Implement viewDidLoad to do additional setup after loading the view, typically from a nib.
- (void)viewDidLoad {
    [super viewDidLoad];
    filePath=[[NSString alloc] initWithString:[[NSBundle mainBundle] pathForResource:@"per" ofType:@"pdf"]];
}



-(IBAction)btnTappedText:(id)sender{
    if(arrImgs!=nil && [arrImgs retainCount]>0 ) { [arrImgs release]; arrImgs=nil; }
    arrImgs=[[NSMutableArray alloc] init];

    setRefImgs(arrImgs);
//  if(nxtTxtDtlVCtr!=nil && [nxtTxtDtlVCtr retainCount]>0) { [nxtTxtDtlVCtr release]; nxtTxtDtlVCtr=nil; }
//  nxtTxtDtlVCtr=[[TxtDtlVCtr alloc] initWithNibName:@"TxtDtlVCtr" bundle:nil];
//  nxtTxtDtlVCtr.str=StringRef();
//  [self.navigationController pushViewController:nxtTxtDtlVCtr animated:YES];

    // 1. Open Document page
    CGPDFDocumentRef document = MyGetPDFDocumentRef ([filePath UTF8String]);

    int pgcnt = CGPDFDocumentGetNumberOfPages( document );

    for( int i1 = 0; i1 < pgcnt; ++i1 ) {

        CGPDFPageRef pg = CGPDFDocumentGetPage (document, i1+1);
        if( !pg ) {
            NSLog(@"Couldn't open page.");
        } else {

            // 2. get page dictionary
            CGPDFDictionaryRef dict = CGPDFPageGetDictionary( pg );
            if( !dict ) {
                NSLog(@"Couldn't open page dictionary.");
            } else {
                // 3. get page contents stream
                CGPDFStreamRef cont;
                if( !CGPDFDictionaryGetStream( dict, "Contents", &cont ) ) {
                    NSLog(@"Couldn't open page stream.");
                } else {
                    // 4. copy page contents steam
                    //  CFDataRef contdata = CGPDFStreamCopyData( cont, NULL );

                    // 5. get the media array from stream
                    CGPDFArrayRef media;
                    if( !CGPDFDictionaryGetArray( dict, "MediaBox", &media ) ) {
                        NSLog(@"Couldn't open page Media.");
                    } else {
                        // 6. open media & get it's size
                        CGPDFInteger mediatop, medialeft;
                        CGPDFReal mediaright, mediabottom;
                        if( !CGPDFArrayGetInteger( media, 0, &mediatop ) || !CGPDFArrayGetInteger( media, 1, &medialeft ) || !CGPDFArrayGetNumber( media, 2, &mediaright ) || !CGPDFArrayGetNumber( media, 3, &mediabottom ) ) {
                            NSLog(@"Couldn't open page Media Box.");
                        } else {
                            // 7. set media size 
                            //      double mediawidth = mediaright - medialeft, mediaheight = mediabottom - mediatop;
                            // 8. get media resources
                            CGPDFDictionaryRef res;
                            if( !CGPDFDictionaryGetDictionary( dict, "Resources", &res ) ) {
                                NSLog(@"Couldn't Open Page Media Reopsources.");
                            } else {
                                // 9. get xObject from media resources  
                                CGPDFDictionaryRef xobj;
                                if( !CGPDFDictionaryGetDictionary( res, "XObject", &xobj ) ) {
                                    NSLog(@"Couldn't load page Xobjects.");
                                } else {
                                    CGPDFDictionaryApplyFunction(xobj, pdfDictionaryFunction, NULL);
                                }
                            }
                        }
                    }
                }
            }
        }
    }

    NSLog(@"Total images are - %i",[arrImgs count]);

    if(nxtImgVCtr!=nil && [nxtImgVCtr retainCount]>0 ) { [nxtImgVCtr release]; nxtImgVCtr=nil; }
    nxtImgVCtr=[[ImgVCtr alloc] initWithNibName:@"ImgVCtr" bundle:nil];
    nxtImgVCtr.arrImg=arrImgs;
    [self.navigationController pushViewController:nxtImgVCtr animated:YES];
}

Converting a CGPDFStreamRef to a UIImage doesn't really make conceptual sense; a PDF represents a document with potentially multiple pages and no obvious mapping to a bitmap image.

Even if you know your PDF contains only one page, you'll still need to do some parsing and finagling. There's a very short discussion of that here: http://lists.apple.com/archives/Cocoa-dev/2006/Sep/msg01407.html

//it completely fit in view in uiview,pdf pages are get images in view
  CGPDFDocumentRef PDFfile;    
     CFURLRef pdfURL = CFBundleCopyResourceURL(CFBundleGetMainBundle(), CFSTR("iPhoneAppProgrammingGuide.pdf"), NULL, NULL);
                PDFfile = CGPDFDocumentCreateWithURL((CFURLRef)pdfURL);
                CFRelease(pdfURL);

            CGPDFPageRef page = CGPDFDocumentGetPage(PDFfile,currentpage);

            context = UIGraphicsGetCurrentContext();
            CGContextSaveGState(context);
            CGContextSetRGBFillColor(context, 1.0, 1.0, 1.0, 1.0);
            CGContextFillRect(context,self.bounds);   
            CGContextTranslateCTM(context, -1.0, [self bounds].size.height);
            CGContextScaleCTM(context, 1.0, -1.0);
            CGContextConcatCTM(context, CGPDFPageGetDrawingTransform(page,  kCGPDFArtBox, [self bounds], 0, true));
            CGContextDrawPDFPage(context, page);    
            CGContextRestoreGState(context);
            CGAffineTransform transform = aspectFit(CGPDFPageGetBoxRect(page, kCGPDFMediaBox),
                                                        CGContextGetClipBoundingBox(context));

               CGContextConcatCTM(context, transform);
                 UIGraphicsBeginImageContext(CGSizeMake(self.bounds.size.width, self.bounds.size.height));