Removing Watermark from a PDF using iTextSharp

Question

I added a watermark on pdf using Pdfstamper. Here is the code:

for (int pageIndex = 1; pageIndex <= pageCount; pageIndex++)
{
    iTextSharp.text.Rectangle pageRectangle = reader.GetPageSizeWithRotation(pageIndex);
    PdfContentByte pdfData = stamper.GetUnderContent(pageIndex);
    pdfData.SetFontAndSize(BaseFont.CreateFont(BaseFont.HELVETICA, BaseFont.CP1252, 
        BaseFont.NOT_EMBEDDED), watermarkFontSize);
    PdfGState graphicsState = new PdfGState();
    graphicsState.FillOpacity = watermarkFontOpacity;
    pdfData.SetGState(graphicsState);
    pdfData.SetColorFill(iTextSharp.text.BaseColor.BLACK);
    pdfData.BeginText();
    pdfData.ShowTextAligned(PdfContentByte.ALIGN_CENTER, "LipikaChatterjee", 
        pageRectangle.Width / 2, pageRectangle.Height / 2, watermarkRotation);
    pdfData.EndText();
}

This works fine. Now I want to remove this watermark from my pdf. I looked into iTextSharp but was not able to get any help. I even tried to add watermark as layer and then delete the layer but was not able to delete the content of layer from the pdf. I looked into iText for layer removal and found a class OCGRemover but I was not able to get an equivalent class in iTextsharp.

Answer 1

I'm going to give you the benefit of the doubt based on the statement "I even tried to add watermark as layer" and assume that you are working on content that you are creating and not trying to unwatermark someone else's content.

PDFs use Optional Content Groups (OCG) to store objects as layers. If you add your watermark text to a layer you can fairly easily remove it later.

The code below is a full working C# 2010 WinForms app targeting iTextSharp 5.1.1.0. It uses code based on Bruno's original Java code found here. The code is in three sections. Section 1 creates a sample PDF for us to work with. Section 2 creates a new PDF from the first and applies a watermark to each page on a separate layer. Section 3 creates a final PDF from the second but removes the layer with our watermark text. See the code comments for additional details.

When you create a PdfLayer object you can assign it a name to appear within a PDF reader. Unfortunately I can't find a way to access this name so the code below looks for the actual watermark text within the layer. If you aren't using additional PDF layers I would recommend only looking for /OC within the content stream and not wasting time looking for your actual watermark text. If you find a way to look for /OC groups by name please let me kwow!

using System;
using System.Windows.Forms;
using System.IO;
using iTextSharp.text;
using iTextSharp.text.pdf;

namespace WindowsFormsApplication1 {
    public partial class Form1 : Form {
        public Form1() {
            InitializeComponent();
        }

        private void Form1_Load(object sender, EventArgs e) {
            string workingFolder = Environment.GetFolderPath(Environment.SpecialFolder.Desktop);
            string startFile = Path.Combine(workingFolder, "StartFile.pdf");
            string watermarkedFile = Path.Combine(workingFolder, "Watermarked.pdf");
            string unwatermarkedFile = Path.Combine(workingFolder, "Un-watermarked.pdf");
            string watermarkText = "This is a test";

            //SECTION 1
            //Create a 5 page PDF, nothing special here
            using (FileStream fs = new FileStream(startFile, FileMode.Create, FileAccess.Write, FileShare.None)) {
                using (Document doc = new Document(PageSize.LETTER)) {
                    using (PdfWriter witier = PdfWriter.GetInstance(doc, fs)) {
                        doc.Open();

                        for (int i = 1; i <= 5; i++) {
                            doc.NewPage();
                            doc.Add(new Paragraph(String.Format("This is page {0}", i)));
                        }

                        doc.Close();
                    }
                }
            }

            //SECTION 2
            //Create our watermark on a separate layer. The only different here is that we are adding the watermark to a PdfLayer which is an OCG or Optional Content Group
            PdfReader reader1 = new PdfReader(startFile);
            using (FileStream fs = new FileStream(watermarkedFile, FileMode.Create, FileAccess.Write, FileShare.None)) {
                using (PdfStamper stamper = new PdfStamper(reader1, fs)) {
                    int pageCount1 = reader1.NumberOfPages;
                    //Create a new layer
                    PdfLayer layer = new PdfLayer("WatermarkLayer", stamper.Writer);
                    for (int i = 1; i <= pageCount1; i++) {
                        iTextSharp.text.Rectangle rect = reader1.GetPageSize(i);
                        //Get the ContentByte object
                        PdfContentByte cb = stamper.GetUnderContent(i);
                        //Tell the CB that the next commands should be "bound" to this new layer
                        cb.BeginLayer(layer);
                        cb.SetFontAndSize(BaseFont.CreateFont(BaseFont.HELVETICA, BaseFont.CP1252, BaseFont.NOT_EMBEDDED), 50);
                        PdfGState gState = new PdfGState();
                        gState.FillOpacity = 0.25f;
                        cb.SetGState(gState);
                        cb.SetColorFill(BaseColor.BLACK);
                        cb.BeginText();
                        cb.ShowTextAligned(PdfContentByte.ALIGN_CENTER, watermarkText, rect.Width / 2, rect.Height / 2, 45f);
                        cb.EndText();
                        //"Close" the layer
                        cb.EndLayer();
                    }
                }
            }

            //SECTION 3
            //Remove the layer created above
            //First we bind a reader to the watermarked file, then strip out a bunch of things, and finally use a simple stamper to write out the edited reader
            PdfReader reader2 = new PdfReader(watermarkedFile);

            //NOTE, This will destroy all layers in the document, only use if you don't have additional layers
            //Remove the OCG group completely from the document.
            //reader2.Catalog.Remove(PdfName.OCPROPERTIES);

            //Clean up the reader, optional
            reader2.RemoveUnusedObjects();

            //Placeholder variables
            PRStream stream;
            String content;
            PdfDictionary page;
            PdfArray contentarray;

            //Get the page count
            int pageCount2 = reader2.NumberOfPages;
            //Loop through each page
            for (int i = 1; i <= pageCount2; i++) {
                //Get the page
                page = reader2.GetPageN(i);
                //Get the raw content
                contentarray = page.GetAsArray(PdfName.CONTENTS);
                if (contentarray != null) {
                    //Loop through content
                    for (int j = 0; j < contentarray.Size; j++) {
                        //Get the raw byte stream
                        stream = (PRStream)contentarray.GetAsStream(j);
                        //Convert to a string. NOTE, you might need a different encoding here
                        content = System.Text.Encoding.ASCII.GetString(PdfReader.GetStreamBytes(stream));
                        //Look for the OCG token in the stream as well as our watermarked text
                        if (content.IndexOf("/OC") >= 0 && content.IndexOf(watermarkText) >= 0) {
                            //Remove it by giving it zero length and zero data
                            stream.Put(PdfName.LENGTH, new PdfNumber(0));
                            stream.SetData(new byte[0]);
                        }
                    }
                }
            }

            //Write the content out
            using (FileStream fs = new FileStream(unwatermarkedFile, FileMode.Create, FileAccess.Write, FileShare.None)) {
                using (PdfStamper stamper = new PdfStamper(reader2, fs)) {

                }
            }
            this.Close();
        }
    }
}

Answer 2

As an extension to Chris's answer, a VB.Net class for removing a layer is included at the bottom of this post which should be a bit more precise.

1. It goes through the PDF's list of layers (stored in the OCGs array in the OCProperties dictionary in the file's catalog). This array contains indirect references to objects in the PDF file which contain the name
2. It goes through the properties of the page (also stored in a dictionary) to find the properties which point to the layer objects (via indirect references)
3. It does an actual parse of the content stream to find instances of the pattern /OC /{PagePropertyReference} BDC {Actual Content} EMC so it can remove just these segments as appropriate

The code then cleans up all the references as much as it can. Calling the code might work as shown:

Public Shared Sub RemoveWatermark(path As String, savePath As String)
  Using reader = New PdfReader(path)
    Using fs As New FileStream(savePath, FileMode.Create, FileAccess.Write, FileShare.None)
      Using stamper As New PdfStamper(reader, fs)
        Using remover As New PdfLayerRemover(reader)
          remover.RemoveByName("WatermarkLayer")
        End Using
      End Using
    End Using
  End Using
End Sub

Full class:

Imports iTextSharp.text
Imports iTextSharp.text.io
Imports iTextSharp.text.pdf
Imports iTextSharp.text.pdf.parser

Public Class PdfLayerRemover
  Implements IDisposable

  Private _reader As PdfReader
  Private _layerNames As New List(Of String)

  Public Sub New(reader As PdfReader)
    _reader = reader
  End Sub

  Public Sub RemoveByName(name As String)
    _layerNames.Add(name)
  End Sub

  Private Sub RemoveLayers()
    Dim ocProps = _reader.Catalog.GetAsDict(PdfName.OCPROPERTIES)
    If ocProps Is Nothing Then Return
    Dim ocgs = ocProps.GetAsArray(PdfName.OCGS)
    If ocgs Is Nothing Then Return

    'Get a list of indirect references to the layer information
    Dim layerRefs = (From l In (From i In ocgs
                                Select Obj = DirectCast(PdfReader.GetPdfObject(i), PdfDictionary),
                                       Ref = DirectCast(i, PdfIndirectReference))
                     Where _layerNames.Contains(l.Obj.GetAsString(PdfName.NAME).ToString)
                     Select l.Ref).ToList
    'Get a list of numbers for these layer references
    Dim layerRefNumbers = (From l In layerRefs Select l.Number).ToList

    'Loop through the pages
    Dim page As PdfDictionary
    Dim propsToRemove As IEnumerable(Of PdfName)
    For i As Integer = 1 To _reader.NumberOfPages
      'Get the page
      page = _reader.GetPageN(i)

      'Get the page properties which reference the layers to remove
      Dim props = _reader.GetPageResources(i).GetAsDict(PdfName.PROPERTIES)
      propsToRemove = (From k In props.Keys Where layerRefNumbers.Contains(props.GetAsIndirectObject(k).Number) Select k).ToList

      'Get the raw content
      Dim contentarray = page.GetAsArray(PdfName.CONTENTS)
      If contentarray IsNot Nothing Then
        For j As Integer = 0 To contentarray.Size - 1
          'Parse the stream data looking for references to a property pointing to the layer.
          Dim stream = DirectCast(contentarray.GetAsStream(j), PRStream)
          Dim streamData = PdfReader.GetStreamBytes(stream)
          Dim newData = GetNewStream(streamData, (From p In propsToRemove Select p.ToString.Substring(1)))

          'Store data without the stream references in the stream
          If newData.Length <> streamData.Length Then
            stream.SetData(newData)
            stream.Put(PdfName.LENGTH, New PdfNumber(newData.Length))
          End If
        Next
      End If

      'Remove the properties from the page data
      For Each prop In propsToRemove
        props.Remove(prop)
      Next
    Next

    'Remove references to the layer in the master catalog
    RemoveIndirectReferences(ocProps, layerRefNumbers)

    'Clean up unused objects
    _reader.RemoveUnusedObjects()
  End Sub

  Private Shared Function GetNewStream(data As Byte(), propsToRemove As IEnumerable(Of String)) As Byte()
    Dim item As PdfLayer = Nothing
    Dim positions As New List(Of Integer)
    positions.Add(0)

    Dim pos As Integer
    Dim inGroup As Boolean = False
    Dim tokenizer As New PRTokeniser(New RandomAccessFileOrArray(New RandomAccessSourceFactory().CreateSource(data)))
    While tokenizer.NextToken
      If tokenizer.TokenType = PRTokeniser.TokType.NAME AndAlso tokenizer.StringValue = "OC" Then
        pos = CInt(tokenizer.FilePointer - 3)
        If tokenizer.NextToken() AndAlso tokenizer.TokenType = PRTokeniser.TokType.NAME Then
          If Not inGroup AndAlso propsToRemove.Contains(tokenizer.StringValue) Then
            inGroup = True
            positions.Add(pos)
          End If
        End If
      ElseIf tokenizer.TokenType = PRTokeniser.TokType.OTHER AndAlso tokenizer.StringValue = "EMC" AndAlso inGroup Then
        positions.Add(CInt(tokenizer.FilePointer))
        inGroup = False
      End If
    End While
    positions.Add(data.Length)

    If positions.Count > 2 Then
      Dim length As Integer = 0
      For i As Integer = 0 To positions.Count - 1 Step 2
        length += positions(i + 1) - positions(i)
      Next

      Dim newData(length) As Byte
      length = 0
      For i As Integer = 0 To positions.Count - 1 Step 2
        Array.Copy(data, positions(i), newData, length, positions(i + 1) - positions(i))
        length += positions(i + 1) - positions(i)
      Next

      Dim origStr = System.Text.Encoding.UTF8.GetString(data)
      Dim newStr = System.Text.Encoding.UTF8.GetString(newData)

      Return newData
    Else
      Return data
    End If
  End Function

  Private Shared Sub RemoveIndirectReferences(dict As PdfDictionary, refNumbers As IEnumerable(Of Integer))
    Dim newDict As PdfDictionary
    Dim arrayData As PdfArray
    Dim indirect As PdfIndirectReference
    Dim i As Integer

    For Each key In dict.Keys
      newDict = dict.GetAsDict(key)
      arrayData = dict.GetAsArray(key)
      If newDict IsNot Nothing Then
        RemoveIndirectReferences(newDict, refNumbers)
      ElseIf arrayData IsNot Nothing Then
        i = 0
        While i < arrayData.Size
          indirect = arrayData.GetAsIndirectObject(i)
          If refNumbers.Contains(indirect.Number) Then
            arrayData.Remove(i)
          Else
            i += 1
          End If
        End While
      End If
    Next
  End Sub

#Region "IDisposable Support"
  Private disposedValue As Boolean ' To detect redundant calls

  ' IDisposable
  Protected Overridable Sub Dispose(disposing As Boolean)
    If Not Me.disposedValue Then
      If disposing Then
        RemoveLayers()
      End If

      ' TODO: free unmanaged resources (unmanaged objects) and override Finalize() below.
      ' TODO: set large fields to null.
    End If
    Me.disposedValue = True
  End Sub

  ' TODO: override Finalize() only if Dispose(ByVal disposing As Boolean) above has code to free unmanaged resources.
  'Protected Overrides Sub Finalize()
  '    ' Do not change this code.  Put cleanup code in Dispose(ByVal disposing As Boolean) above.
  '    Dispose(False)
  '    MyBase.Finalize()
  'End Sub

  ' This code added by Visual Basic to correctly implement the disposable pattern.
  Public Sub Dispose() Implements IDisposable.Dispose
    ' Do not change this code.  Put cleanup code in Dispose(ByVal disposing As Boolean) above.
    Dispose(True)
    GC.SuppressFinalize(Me)
  End Sub
#End Region

End Class