I have created a function that does exactly what you need a few months ago as a part of a graph layouting algorithm. I wrote that in python and PyQt. I just pasted the code here at codepad. That should be very easily translateable to c#.
Update:
Translated it one-to-one from my python snippet (Love to do that graphics stuff :) ). As my original code was designed for more than two output lines, I just took that into the c# version as well. For two black lines being 20 pixels away from the red one, just pass width = 40
and num = 2
. The returned jagged array represents an array of lines (outer array), with each line represented by an array of points (inner).
public PointF[][] MultiplyLine(PointF[] line, int width, int num)
{
if (num == 1) return new PointF[][] { line };
if (num < 1) throw new ArgumentOutOfRangeException();
if (line.Length < 2) return Enumerable.Range(0, num)
.Select(x => line).ToArray();
Func<float, float, PointF> normVec = (x, y) => {
float len = (float)Math.Sqrt((double)(x * x + y * y));
return len == 0 ? new PointF(1f, 0f) : new PointF(x / len, y / len);
};
PointF[][] newLines = Enumerable.Range(0, num)
.Select(x => new PointF[line.Length]).ToArray();
float numinv = 1f / (float)(num - 1), cor = 0f;
PointF vec1 = PointF.Empty, vec2 = PointF.Empty, vec3 = PointF.Empty;
int j = -1, i = -1;
foreach (PointF p in line)
{
bool first = j == -1, last = j == line.Length - 2; j++;
if (!last)
vec1 = normVec(line[j + 1].Y - p.Y, -line[j + 1].X + p.X);
if (!first)
vec2 = normVec(-line[j - 1].Y + p.Y, line[j - 1].X - p.X);
if (!first && !last)
{
vec3 = normVec(vec1.X + vec2.X, vec1.Y + vec2.Y);
cor = (float)Math.Sin((Math.PI -
Math.Acos(vec1.X * vec2.X + vec1.Y * vec2.Y)) / 2);
cor = cor == 0 ? 1 : cor;
vec3 = new PointF(vec3.X / cor, vec3.Y / cor);
}
i = -1;
foreach (PointF[] newLine in newLines)
{
i++; cor = (float)width * ((float)i * numinv - 0.5f);
vec1 = first ? vec1 : last ? vec2 : vec3;
newLine[j] = new PointF(vec1.X * cor + p.X, vec1.Y * cor + p.Y);
}
}
return newLines;
}
To try it out I took this small sample (The same sample as in my PyQt code):
PointF[] pts = new PointF[] {
new PointF(100f, 100f), new PointF(300f, 200f),
new PointF(500f, 200f), new PointF(300f, 500f),
new PointF(600f, 450f), new PointF(650f, 180f),
new PointF(800f, 180f), new PointF(800f, 500f),
new PointF(200f, 700f)
};
pictureBox1.Image = new Bitmap(pictureBox1.Width, pictureBox1.Height);
using(Graphics g = Graphics.FromImage(pictureBox1.Image)){
g.DrawLines(new Pen(Color.Red), pts);
foreach (PointF[] line in MultiplyLine(pts, 80, 14))
g.DrawLines(new Pen(Color.Black), line);
}
Which resulted in this graphic: