Maths - Curvature

On this page we will look at the curvature of lines, surfaces and higher order objects. In order to see this curvature we will embed these objects in a space with one extra dimension. For instance:

Curvature is a scalar field, that is there will be a scalar value corresponding to every point on the object.

Curvature of a Line

If the curvature (K) is the same at every point on the line then the line forms a circle. The radius of this circle will be r = 1/K.

So how can we calculate the curvature at every point on any line? One first guess might be the second differential (the rate of change of the slope of the line) but this does not work as we shall see:

try curvature = ∂² x
∂ y²
line second differential

This formula would give a negative curvature on one side of a circle and a positive curvature on the other side of the circle which is obviously wrong. We need to amend the formula as follows to make it work.

curvature = ∂² x
___∂ y²___
[ 1 + ( ∂ x )² ]3/2
∂ y
At this point the second differential is positive, that is: as we travel along the line the slope increases. line positive curvature
At this point the second differential is negative, that is: as we travel along the line the slope decreases. line negative curvature
At this point the curvature is zero, that is: as we travel along the line the slope is at an inflection point between increasing and decreasing. line zero curvature

 

Curvature of a Surface

Despite the fact that the surface is two dimensional, the curvature is a scalar (I should say real) value at every point, not a vector.

In the same way that: for a line, a constant (non zero) curvature will produce a circle so, for a surface, a constant (positive) curvature will produce the surface of a sphere.

However, unlike a line, a surface (at each given point) can have either positive or negative curvature.

To start with the simple case of a flat plane then the curvature at every point is zero. (equicalent to an infinite diameter sphere) surface plane
If we now take part of the surface of a sphere, which meets the plane, with the curvature in both dimensions turning toward us, we will have a positive curvature. surface sphere front
If we now take a similar subsection of a sphere, but on the other side of the plane, with the curvature in both dimensions turning away from us, we will have a positive curvature. surface sphere back
If we now take a surface with the curvature in one dimension turning toward us and the other dimension turning away from us, so we have a saddle, we will have a negative curvature. surface sphere negative

Non Euclidean Geometry

So far we have visulalised these curved spaces by looking at them from a higher dimensional space. However it is possible to investigate them within their own dimensions by using a definition of distance (known as the metric) which gives the required curvature.

When we investigate the geometry of curved space we find that it is very similar to Euclidean geometry except that they don't have the property that there is only one line, parallel to a given line, that goes through a given point.

This leads to many variations on the geometry, for example Pythagoras:

see non Euclidean geometry.

Hyperbolic Geometry

Can we have a surface, with a constant negative curvature everywhere, in the same way that a sphere has a constant positive curvature everywhere?

see hyperbolic geometry.


metadata block
see also:

 

Correspondence about this page

Book Shop - Further reading.

Where I can, I have put links to Amazon for books that are relevant to the subject, click on the appropriate country flag to get more details of the book or to buy it from them.

cover Curves and Surfaces in Geometric Modeling: Theory and Algorithms

cover The NURBs Book (Monographs in Visual Communication) - A NURBS book for software developer, it covers the theory and also the practical side.

Other Books about Curves and Surfaces

This site may have errors. Don't use for critical systems.

Copyright (c) 1998-2017 Martin John Baker - All rights reserved - privacy policy.