Leadtools.ImageProcessing.Core Send comments on this topic. | Back to Introduction - All Topics | Help Version 15.12.10
GetTransformationParameters Method
See Also  Example
Leadtools.ImageProcessing.Core Namespace > CoreUtilities Class : GetTransformationParameters Method




image
RasterImage object that references the transformed image.
referencePoints
Pointer to an array of reference points. These points represent the center of mass points for the registration marks in the reference image. You have to provide this information.
transformedPoints
Pointer to an array of transformed points. These points are the center of mass points in the image. You have to provide this information.
Computes the rotation angle, XY scaling, and XY translation of the transformed image with comparison to the reference image. These are the transformations that would have to be performed to the reference image to have it match the current image. This method is available in the Document/Medical Toolkits.

Syntax

Visual Basic (Declaration) 
Public Shared Function GetTransformationParameters( _
   ByVal image As RasterImage, _
   ByVal referencePoints() As Point, _
   ByVal transformedPoints() As Point _
) As TransformationParameters
Visual Basic (Usage)Copy Code
Dim image As RasterImage
Dim referencePoints() As Point
Dim transformedPoints() As Point
Dim value As TransformationParameters
 
value = CoreUtilities.GetTransformationParameters(image, referencePoints, transformedPoints)
C# 
public static TransformationParameters GetTransformationParameters( 
   RasterImage image,
   Point[] referencePoints,
   Point[] transformedPoints
)
Managed Extensions for C++ 
public: static TransformationParameters GetTransformationParameters( 
   RasterImage image,
   Point[] referencePoints,
   Point[] transformedPoints
)
C++/CLI 
public:
static TransformationParameters GetTransformationParameters( 
   RasterImage image,
   array<Point> referencePoints,
   array<Point> transformedPoints
)

Parameters

image
RasterImage object that references the transformed image.
referencePoints
Pointer to an array of reference points. These points represent the center of mass points for the registration marks in the reference image. You have to provide this information.
transformedPoints
Pointer to an array of transformed points. These points are the center of mass points in the image. You have to provide this information.

Return Value

TransformationParameters

Example

Visual BasicCopy Code
ImageProcessing.Core.CoreUtilities.GetTransformationParameters
   Public Sub GetTransformationParametersExample()
      RasterCodecs.Startup()
      Dim codecs As New RasterCodecs()
      codecs.ThrowExceptionsOnInvalidImages = True

      Dim leadImage As RasterImage = codecs.Load("C:\Program Files\LEAD Technologies\LEADTOOLS 15\Images\Master.jpg")

      ' Prepare the command
      Dim rmData(2) As SearchRegistrationMarksCommandData
      ' Mark1
      rmData(0) = New SearchRegistrationMarksCommandData
      rmData(0).Rectangle = New Rectangle(680, 20, 941 - 680, 218 - 20)
      Dim pt0(0) As Point
      rmData(0).MarkDetectedPoints = pt0
      rmData(0).Width = 31
      rmData(0).Height = 29
      rmData(0).Type = RegistrationMarkCommandType.TShape
      rmData(0).MinimumScale = 90
      rmData(0).MaximumScale = 110
      rmData(0).SearchMarkCount = 1

      ' Mark2
      rmData(1) = New SearchRegistrationMarksCommandData
      rmData(1).Rectangle = New Rectangle(665, 790, 899 - 665, 961 - 790)
      Dim pt1(0) As Point
      rmData(1).MarkDetectedPoints = pt1
      rmData(1).Width = 31
      rmData(1).Height = 29
      rmData(1).Type = RegistrationMarkCommandType.TShape
      rmData(1).MinimumScale = 90
      rmData(1).MaximumScale = 110
      rmData(1).SearchMarkCount = 1

      ' Mark3
      rmData(2) = New SearchRegistrationMarksCommandData
      rmData(2).Rectangle = New Rectangle(7, 1073, 298 - 7, 1246 - 1073)
      Dim pt2(0) As Point
      rmData(2).MarkDetectedPoints = pt2
      rmData(2).Width = 31
      rmData(2).Height = 29
      rmData(2).Type = RegistrationMarkCommandType.TShape
      rmData(2).MinimumScale = 90
      rmData(2).MaximumScale = 110
      rmData(2).SearchMarkCount = 1
      Dim command1 As New SearchRegistrationMarksCommand(rmData)
      command1.Run(leadImage)

      If ((rmData(2).MarkDetectedCount <> 1) OrElse (rmData(1).MarkDetectedCount <> 1) OrElse (rmData(0).MarkDetectedCount <> 1)) Then
         Return
      End If

      Dim original() As Point = _
      { _
         New Point(81400, 11300), _
         New Point(78600, 87400), _
         New Point(14300, 115400) _
      }

      Dim detected() As Point = _
      { _
         rmData(0).MarkDetectedPoints(0), _
         rmData(1).MarkDetectedPoints(0), _
         rmData(2).MarkDetectedPoints(0) _
      }

      ' Find center of mass for detected registration marks in the transformed image
      Dim transformed() As Point = CoreUtilities.GetRegistrationMarksCenterMass(leadImage, detected)
      ' Find transformation parameters
      Dim parameters As TransformationParameters = CoreUtilities.GetTransformationParameters(leadImage, original, transformed)
      ' Apply transformatin parameters to correct the image
      Dim applyCommand As New ApplyTransformationParametersCommand(parameters.XTranslation, parameters.YTranslation, parameters.Angle, parameters.XScale, parameters.YScale, ApplyTransformationParametersCommandFlags.Normal)

      applyCommand.Run(leadImage)

      RasterCodecs.Shutdown()
   End Sub
C#Copy Code
ImageProcessing.Core.CoreUtilities.GetTransformationParameters 
      public void GetTransformationParametersExample() 
      { 
         // Load an image 
         RasterCodecs.Startup(); 
         RasterCodecs codecs = new RasterCodecs(); 
         codecs.ThrowExceptionsOnInvalidImages = true; 
 
         RasterImage image = codecs.Load(@"C:\Program Files\LEAD Technologies\LEADTOOLS 15\Images\Master.jpg"); 
 
         // Prepare the command 
         SearchRegistrationMarksCommandData[] rmData = new SearchRegistrationMarksCommandData[3]; 
 
         //Mark1 
         rmData[0] = new SearchRegistrationMarksCommandData(); 
         rmData[0].Rectangle = new Rectangle(680, 20, 941 - 680, 218 - 20); 
         rmData[0].MarkDetectedPoints = new Point[1]; 
         rmData[0].Width = 31; 
         rmData[0].Height = 29; 
         rmData[0].Type = RegistrationMarkCommandType.TShape; 
         rmData[0].MinimumScale = 90; 
         rmData[0].MaximumScale = 110; 
         rmData[0].SearchMarkCount = 1; 
 
         //Mark2 
         rmData[1] = new SearchRegistrationMarksCommandData(); 
         rmData[1].Rectangle = new Rectangle(665, 790, 899 - 665, 961 - 790); 
         rmData[1].MarkDetectedPoints = new Point[1]; 
         rmData[1].Width = 31; 
         rmData[1].Height = 29; 
         rmData[1].Type = RegistrationMarkCommandType.TShape; 
         rmData[1].MinimumScale = 90; 
         rmData[1].MaximumScale = 110; 
         rmData[1].SearchMarkCount = 1; 
 
         //Mark3 
         rmData[2] = new SearchRegistrationMarksCommandData(); 
         rmData[2].Rectangle = new Rectangle(7, 1073, 298 - 7, 1246 - 1073); 
         rmData[2].MarkDetectedPoints = new Point[1]; 
         rmData[2].Width = 31; 
         rmData[2].Height = 29; 
         rmData[2].Type = RegistrationMarkCommandType.TShape; 
         rmData[2].MinimumScale = 90; 
         rmData[2].MaximumScale = 110; 
         rmData[2].SearchMarkCount = 1; 
         SearchRegistrationMarksCommand command1 = new SearchRegistrationMarksCommand(rmData); 
         command1.Run(image); 
         if ((rmData[2].MarkDetectedCount != 1) || (rmData[1].MarkDetectedCount != 1) || (rmData[0].MarkDetectedCount != 1)) 
            return; 
 
         Point[] original = 
         { 
            new Point(81400, 11300), 
            new Point(78600, 87400), 
            new Point(14300, 115400) 
         }; 
 
         Point[] detected = 
         { 
            rmData[0].MarkDetectedPoints[0], 
            rmData[1].MarkDetectedPoints[0], 
            rmData[2].MarkDetectedPoints[0] 
         }; 
 
         //Find center of mass for detected registration marks in the transformed image 
         Point[] transformed = CoreUtilities.GetRegistrationMarksCenterMass(image, detected); 
         //Find transformation parameters 
         TransformationParameters parameters = CoreUtilities.GetTransformationParameters(image, original, transformed); 
         //Apply transformation parameters to correct the image 
         ApplyTransformationParametersCommand applyCommand = new ApplyTransformationParametersCommand(parameters.XTranslation, parameters.YTranslation, parameters.Angle, parameters.XScale, parameters.YScale, ApplyTransformationParametersCommandFlags.Normal); 
         applyCommand.Run(image); 
 
         RasterCodecs.Shutdown(); 
      }

Remarks

  • This method detects transformation parameters of the transformed image by comparing it with the reference image. The number of reference and detected points must be three.
  • Use GetRegistrationMarksCenterMass to fill the referencePoints array with the points representing the center of masses for each of the reference marks from the reference image.
  • Fill the transformedPoints array by performing the following steps:

  • The properties of TransformationParameters which are: XTranslation, YTranslation, Angle, XScale and YScale are internally divided by 100. For example, an Angle of 500 would mean to rotate the image 5 degrees clockwise.
  • The results of this method must be sent without any modification to the ApplyTransformationParameters method in order to correct the image.
  • If you want to correct the image yourself, you have to perform the inverse operations in this order:

  • If you simply want to automatically straighten the image, use the DeskewCommand.
  • This method supports 12 and 16-bit grayscale and 48 and 64-bit color images. Support for 12 and 16-bit grayscale and 48 and 64-bit color images is available only in the Document/Medical toolkits.
  • This command does not support signed images.
For more information, refer to Detecting Registration Marks.

Requirements

Target Platforms: Microsoft .NET Framework 2.0, Windows 98, Windows NT 4.0, Windows Millennium Edition, Windows 2000, Windows XP Home Edition, Windows XP Professional, Windows Server 2003 family

See Also