Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Ok, I completely revised my question. For those interested about my purpose with this question, see the older versions.

So, I would like to calculate the area of a cross-section of a tetrahedron. The only things I know are the $4$ points, with each their own $(x,y,z)$ coordinates.

Under this tetrahedron is a line, the $x$-axis. The cross-sections should be perpendicular to the $x$-axis. A cross-section of a tetrahedron can either be a triangle or a quadrilateral. Between the projection of the two left-most points it will be a triangle, and also between the projections of the third and fourth point. Between the second and third point it will be a quadrilateral. For illustrative purposes:

enter image description here

My question is, how can I find the area of a cross-section perpendicular to the $x$-axis?

share|improve this question
1  
Let $X_i=(x_i,y_i,z_i)$ for $1\le i\le4$ & $x_1\le x_2\le x_3\le <x_4$ not all equal. Take a cross section $x=x_k+t(x_{k+1}-x_k)=(1-t)x_k+tx_{k+1}$ between $x_k$ and $x_{k+1}$ for $t\in(0,1)$. Tetrahedron $T$ in homegeneous coordinates is $$T=\left\{\sum_{i=1}^{4}t_iX_i~\Big\lvert~ t_i\ge0,~\sum t_i\le1\right\}$$ and our cross section has the additional constraint $$\sum t_ix_i=x=x_k+t(x_{k+1}-x_k)=(1-t)x_k+tx_{k+1}.$$ Might be easier if we could map $T$ onto the standard simplex & the $x$ axis to a line. How'd you make the nice img? –  bgins Mar 17 '12 at 22:32
    
Thanks @bgins, I'll have a look at that. I thought of something similar myself, at least for the first part (triangle). Represent the three edges going out from Point $1$ (= $P_1$)as $(1-t)P_1 + tP2$, and the same for the other points. Then take $t \in [0,1]$ for the first edge, and scale for the other two. This way you can find the three points of a triangle, and calculate the area with e.g. the cross product of two edges of this triangle. Problem is, I'd like to get a result like $a+bx+cx^2$ for the first part (and a separate one for the second and third). I made the image in Inkscape :) –  Ailurus Mar 17 '12 at 22:42
    
I once worked out something similar for the easier case of slicing a $d$-cube perpendicular to $(1,1,\cdots,1)$ (which is symmetric and divides into $d$ intervals, with ${d\choose k}$ vertices on one side and ${d\choose k+1}$ on the other) and seem to recall I needed to use $\min()$ functions. Anyway if I get a chance I'll look at it more tomorrow. –  bgins Mar 17 '12 at 23:10

1 Answer 1

up vote 1 down vote accepted

Ok, I more or less solved it -- but I'm not very happy with the solution. I would like to end up with an expression for the area between the first and second points, dependent on $y$ of course. This should have the form $a + by + cy^2$ (in this case I chose to project it on the $y$-axis instead of the $x$-axis). I explain why.

Lets create a tetrahedron (having volume 1) with the points $P_1 = (1,0,2/3)$, $P_2 = (4,1,0)$, $P_3 = (1,2,4)$ and $P_4 = (0,3,1)$. Note that the $y$-position is respectively $0,1,2,3$. The tetrahedron looks like this:

enter image description here

For the cross-sections between $P_1$ en $P_2$, we need points on the lines $P_1 P_2$, $P_1 P_3$ and $P_1 P_4$ (so three points in total). The equations for the relevant parts of these lines are

$$\begin{align} P_1 P_2 & = (1-y)P_1 + y P_2 & y \in [0,1] \\ P_1 P_3 & = (1-\frac{y}{2})P_1 + \frac{y}{2} P_3 & y \in [0,\frac{1}{2}] \\ P_1 P_4 & = (1-\frac{y}{3})P_1 + \frac{y}{3} P_3 & y \in [0,\frac{1}{3}] \\ \end{align}$$

The yellow cross-section above is for $y=0.8$. The area of the triangle can now be calculated as follows. Define two sides of the yellow triangle (using the blue dots):

$$\begin{align} Side1 = P_1 P_2 (0.8) - P_1 P_3 (0.8)\\ Side2 = P_1 P_4 (0.8) - P_1 P_2 (0.8) \end{align}$$

The area now follows from the cross product between these two, i.e. $A = \frac{1}{2} | Side1 \times Side2 |$.

If you apply this procedure for $y = 0$ with small steps until $y = 1$ and plot the area as a function of $y$, you end up with (blue solid line):

enter image description here

Which is actually the same as the first part of the uniform quadratic B-Spline, i.e. $.5 t^2$ (yellow dashed line). Perhaps you know that these B-Splines can be defined as projections of simplices, so they are also called "Simplex Splines".

New Question: I should be able to derive this $.5 t^2$ somehow, using the steps described above. How should I do this?

share|improve this answer
    
Ah, got it! Just writing out the equations (i.e. substituting the coordinates) does the trick, of course. –  Ailurus Mar 18 '12 at 1:32

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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