13,054,779 members (67,231 online)
Rate this:
See more: , +
I am bit confused how to calculate dx and dy in sobel edge detection method ....
I googled it and i found that.......
```public class SobelEdgeDetector : ImageFilter<SobelEdgeDetectorParms>
{
private Color grayscale(Color cr)
{
return Color.FromArgb(cr.A, (int)(cr.R * .3 + cr.G * .59 + cr.B * 0.11),
(int)(cr.R * .3 + cr.G * .59 + cr.B * 0.11),
(int)(cr.R * .3 + cr.G * .59 + cr.B * 0.11));
}
public override Bitmap FilterProcessImage(SobelEdgeDetectorParms parms, Bitmap image)
{
Bitmap ret = new Bitmap(image.Width, image.Height);
for (int i = 1; i < image.Width - 1; i++)
{
for (int j = 1; j < image.Height - 1; j++)
{
Color cr = image.GetPixel(i + 1, j);
Color cl = image.GetPixel(i - 1, j);
Color cu = image.GetPixel(i, j - 1);
Color cd = image.GetPixel(i, j + 1);
Color cld = image.GetPixel(i - 1, j + 1);
Color clu = image.GetPixel(i - 1, j - 1);
Color crd = image.GetPixel(i + 1, j + 1);
Color cru = image.GetPixel(i + 1, j - 1);
int dx = 0, dy = 0;
switch (parms.Channel)
{
case Channels.R:
dx = cld.R + 2 * cd.R + crd.R - (clu.R + 2 * cu.R + cru.R);
dy = crd.R + 2 * cr.R + cru.R - (cld.R + 2 * cl.R + clu.R);
break;
case Channels.G:
dx = cld.G + 2 * cd.G + crd.G - (clu.G + 2 * cu.G + cru.G);
dy = crd.G + 2 * cr.G + cru.G - (cld.G + 2 * cl.G + clu.G);
break;
case Channels.B:
dx = cld.B + 2 * cd.B + crd.B - (clu.B + 2 * cu.B + cru.B);
dy = crd.B + 2 * cr.B + cru.B - (cld.B + 2 * cl.B + clu.B);
break;
case Channels.RGB:
dx = grayscale(cld).B + 2 * grayscale(cd).B + grayscale(crd).B - (grayscale(clu).B + 2 * grayscale(cu).B + grayscale(cru).B);
dy = grayscale(crd).B + 2 * grayscale(cr).B + grayscale(cru).B - (grayscale(cld).B + 2 * grayscale(cl).B + grayscale(clu).B);
break;
}
double power = Math.Abs(dx) + Math.Abs(dy);
if (power > parms.Threshold)
ret.SetPixel(i, j, parms.EdgeColor);
else
{
if (parms.CopyOriginal)
{
Color c = image.GetPixel(i, j);
if (parms.ConvertToGrayscale)
{
ret.SetPixel(i, j,
Color.FromArgb(255,
(int)(c.R * 0.3 + c.G * 0.59 + c.B * 0.11),
(int)(c.R * 0.3 + c.G * 0.59 + c.B * 0.11),
(int)(c.R * 0.3 + c.G * 0.59 + c.B * 0.11)));
}
else
ret.SetPixel(i, j, c);
}
else
ret.SetPixel(i, j, Color.White);
}
}
}
return ret;
}
public override System.Windows.Forms.Control GetParameterWindow()
{
return new SobelEdgeDetectorParmForm();
}
}```

this is for 3X3 Matrix .........I want to implement 5X5 Matrix..........
Actually I want slop of tangent line to selected point. . . .
tangent= Atan2(dx,dy).
Using 3X3 matrix i am getting correct tangent, but it not more accurate tangent.
I am interested in calculation of dx and dy for 5X5 matrix. I don't want to detect edge, i want to detect tangent that is Atan2(dx,dy)........
Posted 5-Nov-12 19:42pm
Updated 5-Nov-12 21:15pm
v2

Rate this:

## Solution 1

You can use the coefficients from this source to compute dx and dy: http://www.cim.mcgill.ca/~image529/TA529/Image529_99/assignments/edge_detection/references/sobel.htm[^]

Top Experts
Last 24hrsThis month
 Graeme_Grant 193 OriginalGriff 150 Jochen Arndt 110 Dave Kreskowiak 85 RickZeeland 60
 OriginalGriff 5,731 RickZeeland 2,014 ppolymorphe 1,858 F-ES Sitecore 1,646 Dave Kreskowiak 1,494