• How to manipulate pixels on a bitmap by scanline property(Ma Xiaoguang and Ma Xiaoming)


    We have been developing image processing software for above 14 years with old versions of Delphi, such as Delphi 6 and Delphi 7. We often working on pixels of bitmap by its Scanline property.  We want to do such things with FireMonkey framework in Delphi XE6. But at beginning, we even don't know how to do it with the familiar approach, the Scanline property, because TBitmap class itself in FireMonkey has no the Scanline property anymore.

    We have to study how to work with FireMonkey by searching for information from Internet. And we have found a piece of code at Delphi official website: http://docwiki.embarcadero.com/CodeExamples/XE6/en/FMX.AlphaColorToScanline_(Delphi) 

    The code in the above page illustrated that for accessing the pixels on a bitmap, we have to beg a help from class TBitmapData. The demo code cannot be compiled through XE6, because the compiler reports lack of method GetPixelFormatBytes(). The code is for copying specified amount of lines of pixels from a source bitmap to a destination bitmap written with XE3. We have modified the original code a little bit to copy the all the pixels from the source to a destination, the code could be compiled by XE6 and runs well then.

    First of all, creating a FireMonkey Destop project in XE6, drop two TImage components and a TButton component and a TOpenDialog component onto the form. Double click the TButton component and typing in the following code to its code block.

    ------------------------------

    procedure TForm1.Button1Click(Sender: TObject);
    var
      bd1, bd2 : TBitmapData;
      w, h     : Integer;
    begin
      OpenDialog1.Filter := 'JPEG Image|*.jpg';
      if OpenDialog1.Execute then
      begin

        Image1.Bitmap.LoadFromFile(OpenDialog1.FileName);
        w := Image1.Bitmap.Width;
        h := Image1.Bitmap.Height;
        Image2.Bitmap.PixelFormat := Image1.Bitmap.PixelFormat;
        Image2.Bitmap.SetSize(w, h);
        try
          Image1.Bitmap.Map(TMapAccess.Read, bd1);
          Image2.Bitmap.Map(TMapAccess.Write, bd2);
          AlphaColorToScanline(@PAlphaColorArray(bd1.Data)[0], bd2.Data,
            Round(w * h), Image1.Bitmap.PixelFormat);
        finally
          Image1.Bitmap.Unmap(bd1);
          Image2.Bitmap.Unmap(bd2);
        end;
      end;
    end;

    ------------------------------

    As you can see that the TBitmapData class is really great for being as a delegate to accessing data in a TBitmap object. We can mapping a TBitmap object to be as a TBitmapData object, and then mess up bitmap data around with TBitmapData. And we can even control over the accessing permissions to those bitmap data.

    We can copy all the pixels from a bitmap to another, but how can we manipulating each single pixel on a bitmap in our familiar approach, with the scanline property, separately? The answer is included in TBitmapData class. In the following, we did the same things as above with scanline property.

    ------------------------------

    procedure TForm1.Button1Click(Sender: TObject);
    var
      bd1, bd2   : TBitmapData;
      x, y, w, h : Integer;
      p1, p2     : PAlphaColorArray;
    begin
      OpenDialog1.Filter := 'JPEG Image|*.jpg';
      if OpenDialog1.Execute then
      begin

        Image1.Bitmap.LoadFromFile(OpenDialog1.FileName);
        w := Image1.Bitmap.Width;
        h := Image1.Bitmap.Height;
        Image2.Bitmap.PixelFormat := Image1.Bitmap.PixelFormat;
        Image2.Bitmap.SetSize(w, h);
        try
          Image1.Bitmap.Map(TMapAccess.Read, bd1);
          Image2.Bitmap.Map(TMapAccess.Write, bd2);
          for y := 0 to (h - 1) do
          begin

            p1 := PAlphaColorArray(bd1.GetScanline(y));
            p2 := PAlphaColorArray(bd2.GetScanline(y));
            for x := 0 to (w - 1) do
            begin

              p2[x] := p1[x];
            end;
          end;

        finally

          Image1.Bitmap.Unmap(bd1);

          Image2.Bitmap.Unmap(bd2);
        end;
      end;
    end;

    ------------------------------

    Wow, we can manipulating bitmap's pixels as before in FireMonkey framework. That's cool. The benefits of accessing individual pixel on a bitmap is that we can developing our own image filters now. 

    We haven't appreciating all the power of FireMonkey. But we believe that we can do more cool things with it. We will keep on studying it, and sharing more as we can.  

    Ma Xiaoguang and Ma Xiaoming

    http://blog.csdn.net/gmbros/article/details/39047769

    http://blog.sina.cn/dpool/blog/u/1282411531#type=-1

    https://community.embarcadero.com/answers/how-do-i-get-at-individual-pixels-in-a-firemonkey-tbitmap

  • 相关阅读:
    L1范式和L2范式的区别
    随机森林
    LDA-math-神奇的Gamma函数
    (转)共轭先验的理解
    Hits算法
    朴素贝叶斯分类算法(3)
    朴素贝叶斯分类算法(2)
    朴素贝叶斯分类算法(1)
    多项分布(multinominal distribution)
    从对偶问题到KKT条件
  • 原文地址:https://www.cnblogs.com/findumars/p/6891312.html
Copyright © 2020-2023  润新知