Estimating the length of a curve with the Cauchy-Crofton formula

266 Views Asked by At

This is a comment from Do Carmo's Differential Geometry of Curves and Surfaces:

Equation $\iint n dp d\theta=2l$ (where $n$ is thenumber of intersection points of a curve $C$ with all straight lines in the plane that meet $C$) can also be used to obtain an efficient way of estimating lengths of curves. Indeed, a good approximation for the integral $\iint n dp d\theta$ is given as follows. Consider a family of parallel straight lines such that two consecutive lines are at a distance $r$. Rotate this family by angles of $\pi/4, 2\pi/4, 3\pi/4$ in order to obtain four families of straight lines. Let $n$ be the number of intersection points of a curve $C$ with all these lines. Then $\frac{1}{2}{nr} \frac{\pi}{4}$ is an approximation to the integral $ \iint n dp d\theta=\hbox{length of }C$.

My question: How can I prove it and if it is true that $\lim_{r \to 0}\frac{1}{2}n(r)r\frac{\pi}{4}=\hbox{length of }C$?

I couldn't understand how the author arrived at the formula $\frac{1}{2}{nr} \frac{\pi}{4}$.

1

There are 1 best solutions below

2
On BEST ANSWER

In the ultimate reference on integral geometry, Santaló's Integral Geometry and Geometric Probability, he gives this formula, originally due to Steinhaus (one reference is "Length, shape and area," Colloq. Math. 3 (1954), 1-13), on pp. 31-32. If we turning the grid through multiples of $\pi/m$, we obtain $$\ell\approx \frac{N\pi r}{2m}.$$ If we let $r\to 0$ and $m\to\infty$, then we can suspect with some hypotheses that this is calculating the integral $\dfrac12\iint n\,dp\,d\theta$.

Here's a reference for the error estimates: P.A.P. Moran, "Measuring the length of a curve," Biometrika 53 (1966), 359-364.