Maths - Representing rotations as 2 reflections

Introduction

reflected image

When an object is reflected one of the components of its position is reversed while the other components are not.

The left or right handedness of the object is reversed.

The maths is explained on this page.

reflected image

If we now get two mirrors and put them at 90° to each other we can get a view that has been reflected in both mirrors. The object will appear to have been rotated by 180° which is twice the angle between the mirrors.

The left or right handedness of the object is preserved.

So we can see that 2 reflections in different planes are equivalent to a rotation.

Geometry showing relationship between reflection and rotation

Rotations can be represented as 2 reflections.

For example if objects are reflected in the two red lines (which are θ degrees apart) then the objects will be rotated by 2 * θ degrees.

For example, if the blue line along the x-axis is deflected in the x-axis it does not move, if this is then reflected in the θ line this will result in a line at an angle of 2 * θ.

Or, a second example, the green line is first reflected in the x-axis (1), then reflected in the θ line (2). Again this results in a rotation by an angle of 2 * θ.

Derivation of rotation matrix.

On this page we have derived the matrix for reflection, so the matrices are:

Reflection in yz

-1 0 0
0 1 0
0 0 1

 

Reflection in θ line

Px2-Pz* Pz- Py* Py 2 * Px * Py 2 * Px * Pz
2 * Py * Px Py2-Px*Px- Pz*Pz 2 * Py * Pz
2 * Pz * Px 2 * Pz * Py Pz2-Py*Py- Px*Px

where Px,Py and Pz are the endpoints of a unit vector along the θ line.

So to get the rotation by 2 * θ we need to apply these translations one after the other, to represent this we multiply the matrices:

-1 0 0
0 1 0
0 0 1
*
Px2-Pz* Pz- Py* Py 2 * Px * Py 2 * Px * Pz
2 * Py * Px Py2-Px*Px- Pz*Pz 2 * Py * Pz
2 * Pz * Px 2 * Pz * Py Pz2-Py*Py- Px*Px

So the following matrix represents the rotation by 2 * θ if Px,Py and Pz are the endpoints of a unit vector along the θ line:

-Px2 + Pz* Pz + Py* Py -2 * Px * Py -2 * Px * Pz
2 * Py * Px Py2-Px*Px- Pz*Pz 2 * Py * Pz
2 * Pz * Px 2 * Pz * Py Pz2-Py*Py- Px*Px

Using Quaternions

We can use quaternions to represent both reflections and rotations, as described here, we therefore have the option to use quaternions to show how to calculate rotation from two reflections.

For Reflection & scaling: P2=q * P1 * q

For Rotation & scaling: P2=q * P1 * q'

So we apply the first reflection 'a' to get:

P2=a * P1 * a

then we apply the second reflection 'b' to get:

P2=b * a * P1 * a * b

Expanding out all the terms gives:

out.w= ( b.w*a.w -b.x*a.x -b.y*a.y -b.z*a.z)*in.w +( -b.w*a.x -b.x*a.w -b.y*a.z +b.z*a.y)*in.x +( -b.w*a.y +b.x*a.z -b.y*a.w -b.z*a.x)*in.y +( -b.w*a.z -b.x*a.y +b.y*a.x -b.z*a.w)*in.z
out.x= ( b.w*a.x +b.x*a.w +b.y*a.z -b.z*a.y)*in.w +( b.w*a.w -b.x*a.x -b.y*a.y -b.z*a.z)*in.x +( -b.w*a.z -b.x*a.y +b.y*a.x -b.z*a.w)*in.y +( b.w*a.y -b.x*a.z +b.y*a.w +b.z*a.x)*in.z
out.y= ( b.w*a.y -b.x*a.z +b.y*a.w +b.z*a.x)*in.w +( b.w*a.z +b.x*a.y -b.y*a.x +b.z*a.w)*in.x +( b.w*a.w -b.x*a.x -b.y*a.y -b.z*a.z)*in.y +( -b.w*a.x -b.x*a.w -b.y*a.z +b.z*a.y)*in.z
out.z= ( b.w*a.z +b.x*a.y -b.y*a.x +b.z*a.w)*in.w +( -b.w*a.y +b.x*a.z -b.y*a.w -b.z*a.x)*in.x +( b.w*a.x +b.x*a.w +b.y*a.z -b.z*a.y)*in.y +( b.w*a.w -b.x*a.x -b.y*a.y -b.z*a.z)*in.z

setting in.w=0 gives:

out.w= ( -b.w*a.x -b.x*a.w -b.y*a.z +b.z*a.y)*in.x +( -b.w*a.y +b.x*a.z -b.y*a.w -b.z*a.x)*in.y +( -b.w*a.z -b.x*a.y +b.y*a.x -b.z*a.w)*in.z
out.x= ( b.w*a.w -b.x*a.x -b.y*a.y -b.z*a.z)*in.x +( -b.w*a.z -b.x*a.y +b.y*a.x -b.z*a.w)*in.y +( b.w*a.y -b.x*a.z +b.y*a.w +b.z*a.x)*in.z
out.y= ( b.w*a.z +b.x*a.y -b.y*a.x +b.z*a.w)*in.x +( b.w*a.w -b.x*a.x -b.y*a.y -b.z*a.z)*in.y +( -b.w*a.x -b.x*a.w -b.y*a.z +b.z*a.y)*in.z
out.z= ( -b.w*a.y +b.x*a.z -b.y*a.w -b.z*a.x)*in.x +( b.w*a.x +b.x*a.w +b.y*a.z -b.z*a.y)*in.y +( b.w*a.w -b.x*a.x -b.y*a.y -b.z*a.z)*in.z

if a.w=0 and b.w=0 then we get:

out.w= ( -b.y*a.z +b.z*a.y)*in.x +( b.x*a.z -b.z*a.x)*in.y +( -b.x*a.y +b.y*a.x)*in.z
out.x= ( -b.x*a.x -b.y*a.y -b.z*a.z)*in.x +( -b.x*a.y +b.y*a.x)*in.y +( -b.x*a.z +b.z*a.x)*in.z
out.y= ( b.x*a.y -b.y*a.x)*in.x +( -b.x*a.x -b.y*a.y -b.z*a.z)*in.y +( -b.y*a.z +b.z*a.y)*in.z
out.z= ( b.x*a.z -b.z*a.x)*in.x +( b.y*a.z -b.z*a.y)*in.y +( -b.x*a.x -b.y*a.y -b.z*a.z)*in.z

now if we compare this with an expansion for a rotation we get:

P2.x
P2.y
P2.z
=
qx*qx+qw*qw-qy*qy- qz*qz 2*qx*qy- 2*qw*qz 2*qx*qz+ 2*qw*qy
2*qw*qz + 2*qx*qy qw*qw - qx*qx+ qy*qy - qz*qz -2*qw*qx+ 2*qy*qz
-2*qw*qy+ 2*qx*qz 2*qw*qx+ 2*qy*qz qw*qw - qx*qx- qy*qy+ qz*qz
P1.x
P1.y
P1.z

which is nothing like it! what did I do wrong here?

Further Reading

This can be used to generate the matrix representation of the axis angle representation of rotation:


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 Engineering Mathematics - This book has been going for a long time and it is now in its 5th edition, so it is tried and tested.

Other Math Books

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

Copyright (c) 1998-2023 Martin John Baker - All rights reserved - privacy policy.