**comp.graphics.algorithms**

## Subject: **Estimation of Bezier curve flatness criteria**

Hi all,

Some time ago I have already asked this question, but didn't get

answer, may be because of incomplete problem formulation. I'll try to

explaint problem correctly.

The original problem is building stoke for Bezier curves. In my

situation stoke have to consists of Bezier curves with the same degree.

This realizable only in certain situation when source curve enough flat

otherwise I subdivide source curve with Casteljau method.

Let's look on the picture there is quad curve example

http://len97-1.spb.ru/images/untitled.JPG

(P0,P1,P2) - source curve

(L0,L1,L2) and (P0,P1,P2) - left and right curves

(L0,R0) is perpendicular to (P0,P1) and (L2,R2) is perpendicular to

(P2,P1)

(P0,P1) || (L0,L1) || (R0,R1) and (P1,P2) || (L1,L2) || (R1,R2)

dist(P0,L0) = dist(P0,R0) = dist(P2,L2) = dist(P2,R2) = W

As you see distance between source and left/right curves is about W, so

I can use them as stroke for source one. Ther are some bad examples

when source one isn't flat

http://len97-1.spb.ru/images/untitled2.JPG

http://len97-1.spb.ru/images/untitled3.JPG

I think, I have found effective flatness criteria for source curve, I

suppose that distance between source and left/right curves in "good"

cases is about distance between parametric centers dist(P(0.5),L(0.5)).

Let's write it mathematically (main claim)

If we have a small A >=0 and |dist(P(0.5),L(0.5)) - W| <= A,

so exists a small B >= 0 so W-B <= minDist(P,L) <= maxDist(P,L) <=

W+ B

Practicaly it works very well. Moreover my calculations show that

dist(P(0.5),L(0.5)) depends on angle (P0,P1,P2) only!! It meens we can

easy calculate angle's tangent as (P0-P1)*(P2-P1)/((P0-P1)x(P2-P1)) and

calculate some function C(A) to check that tangent ~ C(A). Really C(A)

is complex but we have to calculate it only once.

Unfortunately I cann't prove the man claim =(

May be exists similar works or literature?

Or may be somebody did similar research?

I would appreciate for any your help.

Thanks, Denis.

Reply

View All Messages in

**comp.graphics.algorithms**

path:

Replies:

Re: Estimation of Bezier curve flatness criteria

Re: Estimation of Bezier curve flatness criteria

Re: Estimation of Bezier curve flatness criteria

Copyright © 2006 WatermarkFactory.com. All Rights Reserved.