topic Re: Thoughts on if this is achievable in .NET Forum

Thoughts on if this is achievable

BrentBurgess1980 — Mon, 28 Jun 2010 08:14:10 GMT

I have been asked if it is possible to develop a tool that will determine the maximum slew between 2 alignments and minimum clearance between 2 alignments. I have attached 2 images to help describe what is to be achieved.

From my research, I haven't found such a function, although, apparently microstation has one similar.

Picking the lines is the easy bit. I would assume that I would need to set an interval at which to check along the proposed alignment, and then iterate along that, determining the perpendicular point of the existing alignment, storing the value, and comparing it to the largest/smallest. At a quick glance, the only limitation I can see is the tolerance at which is set.

Has anyone tried somthing similar to this?

Re: Thoughts on if this is achievable

Anonymous — Mon, 28 Jun 2010 12:32:33 GMT

The Curve2d/3d classes in the Geometry namepsace has

GetClosestPointTo() that takes another curve as input,

and is supposed to return the point on the input curve

that's closest to the curve the method is called on.

Whether it actually works or not, i can't say and because

it is known that there are some methods in the 2d geometry

classes that are not implemented, caveat emptor.

Re: Thoughts on if this is achievable

BrentBurgess1980 — Mon, 28 Jun 2010 13:11:58 GMT

Thanks Tony - I'll take a look at that

Re: Thoughts on if this is achievable

Anonymous — Mon, 28 Jun 2010 21:55:09 GMT

I can think of several methods to accomplish this.

What is your required accuracy?

If the curves are large compared to the slew distance you could use the .net equivalent to the vlax-curve-getclosestpointto lisp expression.

I wrote and tested the lisp routine and it works OK.

You are welcome to the lisp code if you like.

Another woul be to use the .net equivalent to the vlax-curve-getfirstderiv of the point on the curve.

Once you have the first derivative (slope) you can create a line or ray that could be created at pi / 2 rotated from the slope, and then use the intersectwith method to calculate the distance.

Re: Thoughts on if this is achievable

GTVic — Mon, 28 Jun 2010 22:47:17 GMT

@Anonymous wrote:
The Curve2d/3d classes in the Geometry namepsace has
GetClosestPointTo() that takes another curve as input,

I might need something similar for my code. Would the Curve2d.GetClosestPointTo(Point2d) return the intersection point on the polyline nearest the 2d point specified, where a line drawn between is always at right angles to the polyline?

I guess if the point is beyond the endpoints of the polyline then the line would not be at 90° so you would need some type of test to check for that condition. Is there a way to find the distance along the polyline to that intersection point?

Re: Thoughts on if this is achievable

Anonymous — Mon, 28 Jun 2010 23:32:21 GMT

No, it returns the point on the curve that is closest to the point you pass as the argument,

so any input coordinate you pass would have to be checked first to see if it is already on

the input curve.

The OP's problem is relatively simple if all curve/polyline segments are straight.

When curved segments are involved,it becomes a bit more complicated.

Re: Thoughts on if this is achievable

GTVic — Mon, 28 Jun 2010 23:39:40 GMT

Yeah, mine is a separate question, bit of a hijack.

Anyway, that is what I want, the input point is not on the curve but the returned point is on the curve. The line between the two will be at right angles to the polyline.

Then I just want to find the distance along the polyline to the returned point. Is there a function for that?

Thanks

Re: Thoughts on if this is achievable

Anonymous — Mon, 28 Jun 2010 23:46:04 GMT

See the methods of Curve2d class in the Geometry namespace, and the

Curve class in the DatabaseServices namespace. A Polyline derives from

the latter, so all of the methods of the Curve class can be used with a

Polyline.

The names of the methods of both classes make what they do fairly

obvious.

Re: Thoughts on if this is achievable

GTVic — Mon, 28 Jun 2010 23:57:17 GMT

Ok, thanks, Polyline.GetDistAtPoint(Point3d) appears to be the one I want.