2
$\begingroup$

Extending my earlier question about linear transformations, what's the easiest way to test if there exists a projective linear transformation that takes one polygon to another (in $\mathbb{R}\mathbb{P}^2$)?

$\endgroup$

2 Answers 2

2
$\begingroup$

The answer Tim Gowers and Will Jagy gave you in response to your earlier question can be extended straightforwardly. A homogeneous linear transformation on R^3 has 3x3 - 1 = 4x2 degrees of freedom, so there is a unique projective-linear transformation between any given pair of non-degenerate quadrilaterals. Thus do with quadrilaterals what was done in the previous question with triangles.

$\endgroup$
0
1
$\begingroup$

A hacky way would be, enumerate the points of the two polygons (with adjacent points adjacent in the enumerations), as in $P_1 = [p_{1,1}, p_{1,2}, \ldots, p_{1,n}]$ and $P_2 = [p_{2,1}, \ldots, P_{2,n}]$, build N systems of linear equations, in the form $\forall i, \mathbf{T} p_{1,i} = p_{2,i+j}$ (indexed by $j$). There will be only N such systems, and if any one of them can be satistied (that is, it there is a $\mathbf{T}$ that takes all points $p_{1,i}$ into $p_{2,i+j}$ for some $j$) you have an answer.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.