/// <summary>
/// Exports a JPEG image.
/// </summary>
static void ExportJpegImage(PdfDictionary image, bool flateDecode, ref int count)
{
// Fortunately JPEG has native support in PDF and exporting an image is just writing the stream to a file.
var stream = flateDecode ? Filtering.Decode(image.Stream.Value, "/FlateDecode") : image.Stream.Value;
var fs = new FileStream(String.Format("Image{0}.jpeg", count++), FileMode.Create, FileAccess.Write);
var bw = new BinaryWriter(fs);
bw.Write(stream);
bw.Close();
}
private async Task GetImagesFromPage(PdfPage page, ImportParams importParams, ScannedImageSource.Concrete source)
{
if (page.CustomValues.Elements.ContainsKey("/NAPS2ImportedPage"))
{
source.Put(await ExportRawPdfPage(page, importParams));
return;
}
// Get resources dictionary
var resources = page.Elements.GetDictionary("/Resources");
// Get external objects dictionary
var xObjects = resources?.Elements.GetDictionary("/XObject");
if (xObjects == null)
{
return;
}
// Iterate references to external objects
foreach (var item in xObjects.Elements.Values)
{
// Is external object an image?
if (!((item as PdfReference)?.Value is PdfDictionary xObject) ||
xObject.Elements.GetString("/Subtype") != "/Image")
{
continue;
}
// Support multiple filter schemes
var element = xObject.Elements.Single(x => x.Key == "/Filter");
var elementAsArray = element.Value as PdfArray;
var elementAsName = element.Value as PdfName;
if (elementAsArray != null)
{
var arrayElements = elementAsArray.Elements.Select(x => x.ToString()).ToArray();
if (arrayElements.Length == 2)
{
source.Put(DecodeImage(arrayElements[1], page, xObject, Filtering.Decode(xObject.Stream.Value, arrayElements[0]), importParams));
}
}
else if (elementAsName != null)
{
source.Put(DecodeImage(elementAsName.Value, page, xObject, xObject.Stream.Value, importParams));
}
else
{
throw new NotImplementedException("Unsupported filter");
}
}
}
/// <summary>
/// Unfilters the stream.
/// </summary>
void Decode()
{
if (Stream != null && Stream.Value != null)
{
PdfItem item = Elements["/Filter"];
if (item != null)
{
byte[] bytes = Filtering.Decode(Stream.Value, item);
if (bytes != null)
{
Stream.Value = bytes;
Elements.Remove("/Filter");
Elements.SetInteger("/Length", Stream.Length);
}
}
}
}
/// <summary>
/// Unfilters the stream.
/// </summary>
void Decode()
{
if (Stream != null && Stream.Value != null)
{
PdfItem item = Elements[PdfStream.Keys.Filter];
if (item != null)
{
var decodeParms = Elements[PdfStream.Keys.DecodeParms];
byte[] bytes = Filtering.Decode(Stream.Value, item, decodeParms);
if (bytes != null)
{
Stream.Value = bytes;
Elements.Remove(PdfStream.Keys.Filter);
Elements.Remove(PdfStream.Keys.DecodeParms);
Elements.SetInteger(PdfStream.Keys.Length, Stream.Length);
}
}
}
}
private IEnumerable <ScannedImage> GetImagesFromPage(PdfPage page)
{
if (page.CustomValues.Elements.ContainsKey("/NAPS2ImportedPage"))
{
yield return(ExportRawPdfPage(page));
yield break;
}
// Get resources dictionary
PdfDictionary resources = page.Elements.GetDictionary("/Resources");
// Get external objects dictionary
PdfDictionary xObjects = resources?.Elements.GetDictionary("/XObject");
if (xObjects == null)
{
yield break;
}
// Iterate references to external objects
foreach (PdfItem item in xObjects.Elements.Values)
{
var reference = item as PdfReference;
var xObject = reference?.Value as PdfDictionary;
// Is external object an image?
if (xObject != null && xObject.Elements.GetString("/Subtype") == "/Image")
{
// Support multiple filter schemes
// For JPEG: "/DCTDecode" OR ["/DCTDecode", "/FlateDecode"]
// For PNG: "/FlateDecode"
var element = xObject.Elements.Single(x => x.Key == "/Filter");
var elementAsArray = element.Value as PdfArray;
var elementAsName = element.Value as PdfName;
if (elementAsArray != null)
{
// JPEG ["/DCTDecode", "/FlateDecode"]
yield return(ExportJpegImage(page, Filtering.Decode(xObject.Stream.Value, "/FlateDecode")));
}
else if (elementAsName != null)
{
switch (elementAsName.Value)
{
case "/DCTDecode":
yield return(ExportJpegImage(page, xObject.Stream.Value));
break;
case "/FlateDecode":
yield return(ExportAsPngImage(page, xObject));
break;
case "/CCITTFaxDecode":
yield return(ExportG4(page, xObject));
break;
default:
throw new NotImplementedException("Unsupported image encoding");
}
}
else
{
throw new NotImplementedException("Unsupported filter");
}
}
}
}
private IEnumerable <ScannedImage> GetImagesFromPage(PdfPage page, ImportParams importParams)
{
if (page.CustomValues.Elements.ContainsKey("/NAPS2ImportedPage"))
{
yield return(ExportRawPdfPage(page, importParams));
yield break;
}
// Get resources dictionary
PdfDictionary resources = page.Elements.GetDictionary("/Resources");
// Get external objects dictionary
PdfDictionary xObjects = resources?.Elements.GetDictionary("/XObject");
if (xObjects == null)
{
yield break;
}
// Iterate references to external objects
foreach (PdfItem item in xObjects.Elements.Values)
{
var reference = item as PdfReference;
var xObject = reference?.Value as PdfDictionary;
// Is external object an image?
if (xObject != null && xObject.Elements.GetString("/Subtype") == "/Image")
{
// Support multiple filter schemes
var element = xObject.Elements.Single(x => x.Key == "/Filter");
var elementAsArray = element.Value as PdfArray;
var elementAsName = element.Value as PdfName;
if (elementAsArray != null)
{
string[] arrayElements = elementAsArray.Elements.Select(x => x.ToString()).ToArray();
if (arrayElements.Length == 2)
{
yield return(DecodeImage(arrayElements[1], page, xObject, Filtering.Decode(xObject.Stream.Value, arrayElements[0]), importParams));
}
}
else if (elementAsName != null)
{
yield return(DecodeImage(elementAsName.Value, page, xObject, xObject.Stream.Value, importParams));
}
else
{
throw new NotImplementedException("Unsupported filter");
}
}
}
}
