Image processing for OCR with emgu

See more:

C#

Emgu-CV

image-processing

I am trying to make algorithm to process image for better accuracy of tesseract OCR.

I am using the EmguCV for C#.

First i am applying a binarization of the image
the using the emguCV edge detection

then how to make hole filling using the emgucv
that is a bit unclear to me

Here are some results:

binarization

edge detection


What I have tried:

Here is the code that I am using for the egde detection.
For the binarization I iterate the pixels and make the pixels whiter and darker.

            bmp = ImageLibrary.FilterImage(bmp, 95);


Image<gray, byte=""> input = new Image<gray, byte="">(bmp);

Image<gray, byte=""> output = input.ThresholdAdaptive(new Gray(255), Emgu.CV.CvEnum.AdaptiveThresholdType.MeanC, Emgu.CV.CvEnum.ThresholdType.Binary, 3, new Gray(0.03));

input.ThresholdAdaptive(new Gray(255), AdaptiveThresholdType.GaussianC, ThresholdType.Binary, 11, new Gray(2));