OtsuThresholdCommand Class

Summary

Performs image segmentation and color reduction using Otsu clustering on grayscale images.

Syntax

Objective-C

C++/CLI

Java

Python

public class OtsuThresholdCommand : RasterCommand

@interface LTOtsuThresholdCommand : LTRasterCommand

public class OtsuThresholdCommand 
    extends RasterCommand

public ref class OtsuThresholdCommand : public RasterCommand

class OtsuThresholdCommand(RasterCommand):

Remarks

This command works only with grayscale images. If the input image is colored it will be converted to a grayscale image.
Otsu clustering is performed by making each cluster as compact as possible so as to minimize overlap. If one adjusts a threshold one way, the spread of one cluster gets larger and the spread for the second cluster gets smaller. Different threshold values are tried until the one is found which produces the minimum combined spread.
Otsu thresholding is typically used as a way to binarize an image. It is best when used on an image that has a bimodal histogram (Clusters = 2).
This command supports 8, 16-bit grayscale images and 24, 32-bit colored images. Support for 16-bit grayscale is available in the Document/Medical Imaging toolkits.
This command supports signed/unsigned images.

Otsu Thresholding Function - Before

Otsu Thresholding Function - After

View additional platform support for this Otsu Thresholding function.

Example

using Leadtools; 
using Leadtools.Codecs; 
using Leadtools.ImageProcessing.Core; 
 
 
public void OtsuThresholdCommandExample() 
{ 
   RasterCodecs codecs = new RasterCodecs(); 
   codecs.ThrowExceptionsOnInvalidImages = true; 
 
   //Load an image 
   RasterImage image = codecs.Load(Path.Combine(LEAD_VARS.ImagesDir, @"ImageProcessingDemo\NaturalFruits.jpg")); 
 
   //Prepare the command 
   OtsuThresholdCommand command = new OtsuThresholdCommand(); 
 
   //Apply  
   command.Clusters = 4; 
   command.Run(image); 
} 
 
static class LEAD_VARS 
{ 
   public const string ImagesDir = @"C:\LEADTOOLS22\Resources\Images"; 
}