使用iText 7，导出Flate编码图像的正确方法是什么？

Question

我正在尝试创建代码，以导出图像内的PDF使用iText版本7.19。我对Flate编码的图像有一些问题。我正在使用的微软免费书籍(参见Moving to Microsoft Visual Studio 2010)中的所有Flate编码图像都是粉色的，这取决于我如何尝试复制字节，它们可能会失真。

如果我试图一次复制所有的图像字节(参见下面代码中的SaveFlateEncodedImage2方法)，它们会像这样失真：

​

​

如果我尝试逐行复制它们(请参阅下面代码中的SaveFlateEncodedImage方法)，它们将是粉红色的，如下所示

​

​

下面是我用来导出它们的代码：

using iText.Kernel;
using iText.Kernel.Pdf;
using iText.Kernel.Pdf.Filters;
using System;
using System.Drawing;
using System.Drawing.Imaging;
using System.IO;
using System.Runtime.InteropServices;

namespace ITextPdfStuff
{
    public class MyPdfImageExtractor
    {
        private readonly string _pdfFileName;

        public MyPdfImageExtractor(string pdfFileName)
        {
            _pdfFileName = pdfFileName;
        }

        public void ExtractToDirectory(string directoryName)
        {
            using (var reader = new PdfReader(_pdfFileName))
            {
                // Avoid iText.Kernel.Crypto.BadPasswordException: https://stackoverflow.com/a/48065052/97803
                reader.SetUnethicalReading(true);

                using (var pdfDoc = new PdfDocument(reader))
                {
                    ExtractImagesOnAllPages(pdfDoc, directoryName);
                }
            }
        }


        private void ExtractImagesOnAllPages(PdfDocument pdfDoc, string directoryName)
        {
            Console.WriteLine($"Number of pdf {pdfDoc.GetNumberOfPdfObjects()} objects");

            // Extract objects https://itextpdf.com/en/resources/examples/itext-7/extracting-objects-pdf
            for (int objNumber = 1; objNumber <= pdfDoc.GetNumberOfPdfObjects(); objNumber++)
            {

                PdfObject currentObject = pdfDoc.GetPdfObject(objNumber);

                if (currentObject != null && currentObject.IsStream())
                {
                    try
                    {                 
                        ExtractImagesOneImage(currentObject as PdfStream, Path.Combine(directoryName, $"image{objNumber}.png"));
                    }
                    catch (Exception ex)
                    {
                        Console.WriteLine($"Object number {objNumber} is NOT an image! -- error: {ex.Message}");
                    }
                }
            }
        }

        private void ExtractImagesOneImage(PdfStream someStream, string fileName)
        {
            var pdfDict = (PdfDictionary)someStream;
            string subType = pdfDict.Get(PdfName.Subtype)?.ToString() ?? string.Empty;

            bool isImage = subType == "/Image";

            if (isImage == false)
                return;

            bool decoded = false;


            string filter = pdfDict.Get(PdfName.Filter).ToString();

            if (filter == "/FlateDecode")
            {
                SaveFlateEncodedImage(fileName, pdfDict, someStream.GetBytes(false));
            }
            else
            {
                byte[] imgData;

                try
                {
                    imgData = someStream.GetBytes(decoded);
                }
                catch (PdfException ex)
                {
                    imgData = someStream.GetBytes(!decoded);
                }

                SaveNormalImage(fileName, imgData);
            }

        }

        private void SaveNormalImage(string fileName, byte[] imgData)
        {
            using (var memStream = new System.IO.MemoryStream(imgData))
            using (var image = System.Drawing.Image.FromStream(memStream))
            {
                image.Save(fileName, ImageFormat.Png);
                Console.WriteLine($"{Path.GetFileName(fileName)}");
            }
        }

        private void SaveFlateEncodedImage(string fileName, PdfDictionary pdfDict, byte[] imgData)
        {
            int width = int.Parse(pdfDict.Get(PdfName.Width).ToString());
            int height = int.Parse(pdfDict.Get(PdfName.Height).ToString());
            int bpp = int.Parse(pdfDict.Get(PdfName.BitsPerComponent).ToString());

            // Example that helped: https://stackoverflow.com/a/8517377/97803
            PixelFormat pixelFormat;
            switch (bpp)
            {
                case 1:
                    pixelFormat = PixelFormat.Format1bppIndexed;
                    break;
                case 8:
                    pixelFormat = PixelFormat.Format8bppIndexed;
                    break;
                case 24:
                    pixelFormat = PixelFormat.Format24bppRgb;
                    break;
                default:
                    throw new Exception("Unknown pixel format " + bpp);
            }

            // .NET docs https://api.itextpdf.com/iText7/dotnet/7.1.9/classi_text_1_1_kernel_1_1_pdf_1_1_filters_1_1_flate_decode_strict_filter.html
            // Java docs have more detail: https://api.itextpdf.com/iText7/java/7.1.7/com/itextpdf/kernel/pdf/filters/FlateDecodeFilter.html
            imgData = FlateDecodeStrictFilter.FlateDecode(imgData, true);
            //  byte[] streamBytes = FlateDecodeStrictFilter.DecodePredictor(imgData, pdfDict);

            // Copy the image one row at a time
            using (var bmp = new Bitmap(width, height, pixelFormat))
            {
                BitmapData bmpData = bmp.LockBits(new Rectangle(0, 0, width, height), ImageLockMode.WriteOnly, pixelFormat);

                int length = (int)Math.Ceiling(width * bpp / 8.0);
                for (int i = 0; i < height; i++)
                {
                    int offset = i * length;
                    int scanOffset = i * bmpData.Stride;
                    Marshal.Copy(imgData, offset, new IntPtr(bmpData.Scan0.ToInt64() + scanOffset), length);
                }

                bmp.UnlockBits(bmpData);
                bmp.Save(fileName, ImageFormat.Png);
            }

            Console.WriteLine($"FlateDecode! {Path.GetFileName(fileName)}");
        }


        /// <summary>This method distorts the image badly</summary>
        private void SaveFlateEncodedImage2(string fileName, PdfDictionary pdfDict, byte[] imgData)
        {
            int width = int.Parse(pdfDict.Get(PdfName.Width).ToString());
            int height = int.Parse(pdfDict.Get(PdfName.Height).ToString());
            int bpp = int.Parse(pdfDict.Get(PdfName.BitsPerComponent).ToString());

            // Example that helped: https://stackoverflow.com/a/8517377/97803
            PixelFormat pixelFormat;
            switch (bpp)
            {
                case 1:
                    pixelFormat = PixelFormat.Format1bppIndexed;
                    break;
                case 8:
                    pixelFormat = PixelFormat.Format8bppIndexed;
                    break;
                case 24:
                    pixelFormat = PixelFormat.Format24bppRgb;
                    break;
                default:
                    throw new Exception("Unknown pixel format " + bpp);
            }

            // .NET docs https://api.itextpdf.com/iText7/dotnet/7.1.9/classi_text_1_1_kernel_1_1_pdf_1_1_filters_1_1_flate_decode_strict_filter.html
            // Java docs have more detail: https://api.itextpdf.com/iText7/java/7.1.7/com/itextpdf/kernel/pdf/filters/FlateDecodeFilter.html
            imgData = FlateDecodeStrictFilter.FlateDecode(imgData, true);
            // byte[] streamBytes = FlateDecodeStrictFilter.DecodePredictor(imgData, pdfDict);

            // Copy the entire image in one go
            using (var bmp = new Bitmap(width, height, pixelFormat))
            {
                BitmapData bmpData = bmp.LockBits(new Rectangle(0, 0, width, height), ImageLockMode.WriteOnly, pixelFormat);
                Marshal.Copy(imgData, 0, bmpData.Scan0, imgData.Length);
                bmp.UnlockBits(bmpData);
                bmp.Save(fileName, ImageFormat.Png);
            }

            Console.WriteLine($"FlateDecode! {Path.GetFileName(fileName)}");
        }
    }
}

可以像这样在.NET核心控制台应用程序中实例化和调用代码：

  string existingFileName = @"c:\temp\ReallyLongBook1.pdf";
  var imageExtractor = new MyPdfImageExtractor(existingFileName);
  imageExtractor.ExtractToDirectory(@"c:\temp\images");

我通过这段代码运行了下面这本免费的微软书籍：Moving to Microsoft Visual Studio 2010

有问题的图片在第10页，它是黑白的(不是粉色的)。

我不是PDF专家，几天来我一直在研究这段代码，现在我挑选了一些示例，试图将它们拼凑在一起。任何能帮我摆脱粉色图片的帮助，都将不胜感激。

-更新2020年2月4日

这是MKL建议的修改后的修订版。他的改变提取了比我更多的图像，并产生了出现在我上面提到的书中的适当外观的图像：

using iText.Kernel.Pdf;
using iText.Kernel.Pdf.Canvas.Parser;
using iText.Kernel.Pdf.Canvas.Parser.Data;
using iText.Kernel.Pdf.Canvas.Parser.Listener;
using iText.Kernel.Pdf.Xobject;
using System;
using System.Collections.Generic;
using System.IO;

namespace ITextPdfStuff
{
    public class MyPdfImageExtractor
    {
        private readonly string _pdfFileName;

        public MyPdfImageExtractor(string pdfFileName)
        {
            _pdfFileName = pdfFileName;
        }

        public void ExtractToDirectory(string directoryName)
        {
            using (var reader = new PdfReader(_pdfFileName))
            {
                // Avoid iText.Kernel.Crypto.BadPasswordException: https://stackoverflow.com/a/48065052/97803
                reader.SetUnethicalReading(true);

                using (var pdfDoc = new PdfDocument(reader))
                {
                    ExtractImagesOnAllPages(pdfDoc, directoryName);
                }
            }
        }

        private void ExtractImagesOnAllPages(PdfDocument pdfDoc, string directoryName)
        {
            Console.WriteLine($"Number of pdf {pdfDoc.GetNumberOfPdfObjects()} objects");

            IEventListener strategy = new ImageRenderListener(Path.Combine(directoryName, @"image{0}.{1}"));
            PdfCanvasProcessor parser = new PdfCanvasProcessor(strategy);
            for (var i = 1; i <= pdfDoc.GetNumberOfPages(); i++)
            {
                parser.ProcessPageContent(pdfDoc.GetPage(i));
            }
        }
    }


    public class ImageRenderListener : IEventListener
    {
        public ImageRenderListener(string format)
        {
            this.format = format;
        }

        public void EventOccurred(IEventData data, EventType type)
        {
            if (data is ImageRenderInfo imageData)
            {
                try
                {
                    PdfImageXObject imageObject = imageData.GetImage();
                    if (imageObject == null)
                    {
                        Console.WriteLine("Image could not be read.");
                    }
                    else
                    {
                        File.WriteAllBytes(string.Format(format, index++, imageObject.IdentifyImageFileExtension()), imageObject.GetImageBytes());
                    }
                }
                catch (Exception ex)
                {
                    Console.WriteLine("Image could not be read: {0}.", ex.Message);
                }
            }
        }

        public ICollection<EventType> GetSupportedEvents()
        {
            return null;
        }

        string format;
        int index = 0;
    }
}

Answer 1

PDF在内部支持非常灵活的位图图像格式，特别是对于不同的颜色空间而言。

iText在其解析API中支持导出其子集，本质上是可以轻松导出为常规JPEG或PNG的图像子集。

因此，首先尝试使用iText解析API进行导出是有意义的。您可以按如下方式执行此操作：

Directory.CreateDirectory(@"extract\");
using (PdfReader reader = new PdfReader(@"Moving to Microsoft Visual Studio 2010 ebook.pdf"))
using (PdfDocument pdfDocument = new PdfDocument(reader))
{
    IEventListener strategy = new ImageRenderListener(@"extract\Moving to Microsoft Visual Studio 2010 ebook-i7-{0}.{1}");
    PdfCanvasProcessor parser = new PdfCanvasProcessor(strategy);
    for (var i = 1; i <= pdfDocument.GetNumberOfPages(); i++)
    {
        parser.ProcessPageContent(pdfDocument.GetPage(i));
    }
}

使用帮助器类ImageRenderListener

public class ImageRenderListener : IEventListener
{
    public ImageRenderListener(string format)
    {
        this.format = format;
    }

    public void EventOccurred(IEventData data, EventType type)
    {
        if (data is ImageRenderInfo imageData)
        {
            try
            {
                PdfImageXObject imageObject = imageData.GetImage();
                if (imageObject == null)
                {
                    Console.WriteLine("Image could not be read.");
                }
                else
                {
                    File.WriteAllBytes(string.Format(format, index++, imageObject.IdentifyImageFileExtension()), imageObject.GetImageBytes());
                }
            }
            catch (Exception ex)
            {
                Console.WriteLine("Image could not be read: {0}.", ex.Message);
            }
        }
    }

    public ICollection<EventType> GetSupportedEvents()
    {
        return null;
    }

    string format;
    int index = 0;
}

对于您的示例文档，它成功导出了近400张图像，其中包括上面的示例图像：

​

​

但它也有不到30个图像无法导出，在标准输出上，你会发现“图像无法读取:色彩空间/DeviceN不受支持。”