I'm trying to learn OCR processing. I have decided to create a .net project which use EmguCv wrapper to use OpenCV library.
I have wrote a little piece of code (most of lines come from here) :
public static Image<Bgr, byte> FindImage(string Imgtemplate, string Imgsource, float coeff)
{
Image<Bgr, byte> source = new Image<Bgr, byte>(Imgsource); // Image B
Image<Bgr, byte> template = new Image<Bgr, byte>(Imgtemplate); // Image A
Image<Bgr, byte> imageToShow = source.Copy();
using (Image<Gray, float> result = source.MatchTemplate(template, Emgu.CV.CvEnum.TemplateMatchingType.CcoeffNormed))
{
double[] minValues, maxValues;
//double minValues = 0, maxValues = 0;
Point[] minLocations, maxLocations;
//Point minLocations = new Point(), maxLocations = new Point();
//CvInvoke.Normalize(result, result, 0, 1, Emgu.CV.CvEnum.NormType.MinMax, Emgu.CV.CvEnum.DepthType.Default, new Mat());
//CvInvoke.MinMaxLoc(result, ref minValues, ref maxValues, ref minLocations, ref maxLocations, new Mat());
result.MinMax(out minValues, out maxValues, out minLocations, out maxLocations);
// You can try different values of the threshold. I guess somewhere between 0.75 and 0.95 would be good.
if (maxValues[0] > coeff)
{
// This is a match. Do something with it, for example draw a rectangle around it.
Rectangle match = new Rectangle(maxLocations[0], template.Size);
//Rectangle match = new Rectangle(maxLocations, template.Size);
imageToShow.Draw(match, new Bgr(Color.Green), 1);
}
}
// Show imageToShow in an ImageBox (here assumed to be called imageBox1)
//imageBox1.Image = imageToShow;
return imageToShow;
}
The source image is :
The template image is :
Both images are jpg. The result (with 100 % match) of the method is :
As you can see the result isn't the good one :/. However, when i run my method with the example given on OpenCV site I have a good result (67 % match) :
Anyone have an idea about my problem ? For the OpenCV example result is there any way to improve the matches percent ? Which option/parameter is the best ?
