Shadow of an ellipsoid - Uniform Direction Light

Geometry Level 5

An ellipsoid has its first semi-axis of length 15 units along the unit vector u 1 = 1 9 ( 8 , 4 , 1 ) \mathbf{u_1} = \dfrac{1}{9} (8,-4,1) , and its second semi-axis of length 30, along the unit vector u 2 = 1 9 ( 4 , 7 , 4 ) \mathbf{u_2} = \dfrac{1}{9} (4, 7, -4) , and its third semi-axis of length 10 along the unit vector u 3 = 1 9 ( 1 , 4 , 8 ) \mathbf{u_3} = \dfrac{1}{9} (1, 4, 8 ) . It is positioned such that it touches the x y xy plane at the origin (that is, all of the ellipsoid lies above the x y xy plane, except for one point, the origin). The ellipsoid is subjected to uniform direction light (like sun rays) that has a direction vector ( 1 , 1 , 4 ) (1, -1, -4) , producing a shadow of the ellipsoid on the x y xy plane. Find the area of that shadow.


The answer is 1400.688.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

2 solutions

K T
Jul 22, 2019

By checking u i u j = δ i j \vec{u_i}\cdot\vec{u_j}=\delta_{ij} , it is easily verified that ( u 1 , u 2 , u 3 ) (\vec{u_1},\vec{u_2},\vec{u_3}) form an orthonormal basis. It will be convenient to work in this basis, because it is aligned with the ellipsoid's axes. Let's put our origin at the centre of the ellipsoid. If the light rays have direction r \vec{r} , this vector can be expressed in the u-basis as r u 1 , r u 2 , r u 3 = 1 9 ( 8 , 13 , 35 ) \vec{r}\cdot \vec{u_1}, \vec{r} \cdot \vec{u_2}, \vec{r} \cdot \vec{u_3} = \frac{1}{9}(8, 13, -35) , or, because the length does not matter, just r = ( 8 , 13 , 35 ) \vec{r}=(8,13,-35)

Consider the function f ( u ) = 4 u 1 2 + u 2 2 + 9 u 3 2 f(\vec{u})= 4u_1^2 + u_2^2+ 9u_3^2 then our ellipsoid is the surface where f ( u ) = 900 f(\vec{u})=900 . The gradient f \nabla f is orthogonal to the surface, so f = ( f u 1 , f u 2 , f u 3 ) = ( 8 u 1 , 2 u 2 , 18 u 3 ) \nabla f =(\frac{\partial f}{\partial u_1}, \frac{\partial f}{\partial u_2}, \frac{\partial f}{\partial u_3}) = (8u_1, 2u_2, 18u_3) , so that a normalvector to the surface at a point u \vec{u} on the surface is given by n = ( 4 u 1 , u 2 , 9 u 3 ) \vec{n}=( 4u_1, u_2, 9u_3)

We are interested in the light rays that are tangent to the ellipsoid. These rays have to be normal to n \vec{ n} , so we demand r n = 0 \vec{r} \cdot \vec{n}=0 or, expressing the dot product in coordinates: 32 u 1 + 13 u 2 315 u 3 = 0 32u_1 + 13u_2 - 315u_3=0

This linear equation defines a plane, that I will call P, and allows us to express u 3 2 = ( 1024 u 1 2 + 832 u 1 u 2 y + 169 u 2 2 ) 31 5 2 u_3^2 = \frac{(1024u_1^2+832u_1u_2y +169u_2^2)}{315^2} so that the quadratic equation 4 u 1 2 + u 2 2 + 9 u 3 2 = 900 4u_1^2 + u_2^2 + 9u_3^2 = 900 can be written as 4 u 1 2 + u 2 2 + 9 31 5 2 ( 1024 u 1 2 + 832 u 1 u 2 + 169 u 2 2 ) = 900 4u_1^2 + u_2^2 + \frac{9}{315^2}(1024u_1^2+832u_1u_2 +169u_2^2)  = 900 or 45124 u 1 2 + 832 u 1 u 2 + 11194 u 2 2 = 9922500 45124u_1^2 + 832u_1u_2 + 11194u_2^2 = 9922500

This ellipse can be rotated in the u 1 u_1 - u 2 u_2 -plane to get it into the standard form, with half axes 14.82... 14.82... and 29.77... 29.77... and has area 1387.231501... 1387.231501... .

Note that this ellipse is just the projection of the actual ellipse, that lives in plane P, onto the u 1 u_1 - u 2 u_2 -plane. The actual ellipse is bigger because of the inclination of the plane P. Vectors in the plane P are ( 315 , 0 , 32 ) (315,0,32) and ( 0315 , 13 ) (0 315, 13) so a normal vector to the plane is v = ( 32 , 13 , 315 ) \vec{v}=(32,13,-315) Hence the true area of the ellipse in plane P itself is bigger by a factor v v 3 = 100418 315 = 1.00599... \frac{|v|}{|v_3|}=\frac{\sqrt{100418}}{315}=1.00599... . Its area is 1395.546... 1395.546... (actually quite close to the maximum cross section of 450 π = 1413.7... 450π=1413.7... ).

Further note that the light rays are not perpendicular to plane P. The cross section of the shadow (perpendicular to the light bundle) is smaller by a factor r v r v = 11450 1458 100418 = 0.9462828... \frac{\vec{r}\cdot\vec{v}}{|r||v|} = \frac{11450}{\sqrt{1458\cdot100418}} = 0.9462828...

Finally, the lightbundle makes an angle with the z-axis, which makes the shadow on the x-y plane bigger by a factor 1 2 + 1 2 + 4 2 4 = 1.06066... \frac{\sqrt{1^2+1^2+4^2}}{4}=1.06066...

The combined effect of these three projections is a factor 1145 1134 = 1.009700176... \frac{1145}{1134}=1.009700176... so that the shadow on the xy-plane has Area = 1387.231501.. × 1.009700176.. = 1400.68789 \text{Area}= 1387.231501..×1.009700176..=\boxed{1400.68789}

Hosam Hajjir
Nov 25, 2018

If ( x , y , z ) (x', y', z') are the coordinates of a point, on the ellipsoid, with respect to u 1 , u 2 , u 3 \mathbf{u_1}, \mathbf{u_2}, \mathbf{u_3} , then we know that,

x 2 1 5 2 + y 2 3 0 2 + z 2 1 0 2 = 1 \dfrac{ {x'}^2 }{15^2} + \dfrac{ {y'}^2 }{30^2} +\dfrac{ {z'}^2 }{10^2} = 1

Thus if we define r = ( x , y , z ) \mathbf{r'} = (x', y', z') , and r = ( x , y , z ) \mathbf{r} = (x, y, z ) ,then

r T D e r = 1 {\mathbf{r'}}^T D_e \mathbf{r'} = 1 ,

where D e D_e is a diagonal matrix whose diagonal entries are, in order, 1 1 5 2 \dfrac{1}{15^2} , 1 3 0 2 \dfrac{1}{30^2} and 1 1 0 2 \dfrac{1}{10^2} . The relationship between r \mathbf{r'} and r \mathbf{r} is simply,

r = r 0 + R e r \mathbf{r} = \mathbf{r_0} + R_e \mathbf{r'} where R e = [ u 1 , u 2 , u 3 ] R_e = [ \mathbf{u_1}, \mathbf{u_2}, \mathbf{u_3} ] , and r 0 \mathbf{r_0} is the position vector of the center of the ellipsoid.

The three columns of R e R_e are orthonormal, hence R e 1 = R e T {R_e}^{-1} = {R_e}^T , and thus, the equation of the ellipsoid is,

( r r 0 ) T Q e ( r r 0 ) = 1 ({\mathbf{r}-\mathbf{r_0}})^T Q_e (\mathbf{r}-\mathbf{r_0}) = 1 , where Q e = R e D e R e T Q_e = R_e D_e {R_e}^T

First thing we need to do is find the coordinates of the center of the ellipsoid, which is the vector r 0 \mathbf{r_0} . To that end, we know that at the tangency point (which is given by r = ( 0 , 0 , 0 ) \mathbf{r} = (0,0,0) ), the normal vector to the surface of the ellipsoid is pointing in the negative k \mathbf{k} direction, where k = ( 0 , 0 , 1 ) \mathbf{k} = (0, 0, 1) . Now, the normal vector is parallel to Q e ( r r 0 ) = Q e r 0 Q_e (\mathbf{r}-\mathbf{r_0} ) = - Q_e \mathbf{r_0} , therefore, we have, Q e r 0 = α k Q_e \mathbf{r_0} = \alpha \mathbf{k} for some positive constant α \alpha . From which, r 0 = α Q e 1 k \mathbf{r_0} = \alpha {Q_e}^{-1} \mathbf{k} . Plugging this, and r = 0 \mathbf{r} = \mathbf{0} , into the ellipsoid equation, we obtain,

α = 1 k T Q e 1 k \alpha = \dfrac{1}{\sqrt{ \mathbf{k}^T {Q_e}^{-1} \mathbf{k} }}

Hence,

r 0 = 1 k T Q e 1 k Q e 1 k \mathbf{r_0} = \dfrac{1}{\sqrt{ \mathbf{k}^T {Q_e}^{-1} \mathbf{k} }} {Q_e}^{-1} \mathbf{k}

So, now the ellipsoid is completely specified.

The shadow of the ellipsoid on the x y xy plane is generated by the intersection of the x y xy plane and the non-uniformly scaled cylinder that is tangent to the ellipsoid, and which has its axis parallel to the fixed direction of light rays d 0 = ( 1 , 1 , 4 ) \mathbf{d_0} = (1, -1, -4) .

Tangency points on the ellipsoid are characterized by the vector equation,

d 0 T Q e ( r r 0 ) = 0 \mathbf{d_0}^T Q_e (\mathbf{r} - \mathbf{r_0} ) = 0

This is an equation of a plane that passes through the center of the ellipsoid. Intersecting a plane with an ellipsoid was addressed here . The resulting ellipse can be parameterized as,

r e ( t ) = r 0 + v 1 cos t + v 2 sin t \mathbf{r_e}(t) = \mathbf{r_0} + \mathbf{v_1} \cos t + \mathbf{v_2} \sin t

where t R t \in \mathbb{R} . Points on the surface of the tangent cylinder to the ellipsoid can be expressed as,

r c ( t ) = r e ( t ) + s d 0 \mathbf{r_c}(t) = \mathbf{r_e}(t) + s \mathbf{d_0}

where t R t \in \mathbb{R} .

The above is a vector equation of the lateral surface of the cylinder. To obtain an algebraic equation, we note that,

r c r 0 = A w \mathbf{r_c} - \mathbf{r_0} = A \mathbf{w}

where,

A = [ v 1 , v 2 , d 0 ] A = [ \mathbf{v_1} , \mathbf{v_2}, \mathbf{d_0} ] , and, w = [ cos t sin t s ] \mathbf{w} = \begin{bmatrix} \cos t \\ \sin t \\ s \end{bmatrix}

Now, let

Q 0 = [ 1 0 0 0 1 0 0 0 0 ] Q_0 = \begin{bmatrix} 1 && 0 && 0 \\ 0 && 1 && 0 \\ 0 && 0 && 0 \end{bmatrix}

Then, w T Q 0 w = 1 \mathbf{w}^T Q_0 \mathbf{w} = 1

But w = A 1 ( r c r 0 ) \mathbf{w} = A^{-1} (\mathbf{r_c} - \mathbf{r_0} )

Hence, the algebraic equation of the shadow cylinder is

( r c r 0 ) T Q c ( r c r 0 ) = 1 ( 1 ) (\mathbf{r_c} - \mathbf{r_0} )^T Q_c (\mathbf{r_c} - \mathbf{r_0} ) = 1 \hspace{30pt} (1)

where Q c = A T Q 0 A 1 Q_c = A^{-T} Q_0 A^{-1}

The final step is to intersect this cylinder with the x y xy plane. This is discussed below.

A point r \mathbf{r} on a plane is given by,

r = r 1 + W u ( 2 ) \mathbf{r} = \mathbf{r_1} + W \mathbf{u} \hspace{30pt} (2)

where r 1 \mathbf{r_1} is any arbitrary point on the plane, and W = [ w 1 , w 2 ] W = [ \mathbf{w_1} , \mathbf{w_2} ] is a 3 × 2 3 \times 2 matrix whose columns are arbitrary unit vectors that together with the normal to the plane form an orthonormal basis for R 3 \mathbb{R}^3 .

In the case of the x y xy -plane, it is natural to choose r 1 = ( 0 , 0 , 0 ) \mathbf{r_1} = (0, 0, 0) , w 1 = ( 1 , 0 , 0 ) \mathbf{w_1} = (1, 0, 0) and w 2 = ( 0 , 1 , 0 ) \mathbf{w_2} = (0, 1, 0) .

Substituting this vector expression (2) into the equation of the cylinder (1), one obtains,

( r 1 r 0 + W u ) T Q c ( r 1 r 0 + W u ) = 1 ( \mathbf{r_1} - \mathbf{r_0} + W \mathbf{u} )^T Q_c (\mathbf{r_1} - \mathbf{r_0} + W \mathbf{u}) = 1

Expanding this quadratic form gives,

u T ( W T Q c W ) u + 2 u T W T Q c ( r 1 r 0 ) + ( r 1 r 0 ) T Q c ( r 1 r 0 ) = 1 \mathbf{u}^T (W^T Q_c W) \mathbf{u} + 2 \mathbf{u}^T W^T Q_c (\mathbf{r_1} - \mathbf{r_0}) + (\mathbf{r_1} - \mathbf{r_0})^T Q_c ( \mathbf{r_1} - \mathbf{r_0}) = 1

Take u 0 = ( W T Q c W ) 1 W T Q c ( r 1 r 0 ) \mathbf{u_0} = - (W^T Q_c W)^{-1} W^T Q_c (\mathbf{r_1} - \mathbf{r_0} ) , then the above equation becomes,

( u u 0 ) T ( W T Q c W ) ( u u 0 ) = C (\mathbf{u} - \mathbf{u_0})^T (W^T Q_c W) (\mathbf{u} - \mathbf{u_0}) = C

where the constant C C is given by C = 1 ( r 1 r 0 ) T Q c ( r 1 r 0 ) + u 0 T ( W T Q c W ) u 0 C = 1 - (\mathbf{r_1} - \mathbf{r_0})^T Q_c (\mathbf{r_1} - \mathbf{r_0}) + {\mathbf{u_0}}^T (W^T Q_c W) \mathbf{u_0}

Dividing by C C , we obtain,

( u u 0 ) T Q 1 ( u u 0 ) = 1 (\mathbf{u} - \mathbf{u_0})^T Q_1 (\mathbf{u} - \mathbf{u_0}) = 1

where Q 1 = 1 C ( W T Q c W ) Q_1 = \dfrac{1}{C} (W^T Q_c W) .

The final step is to diagonalize Q 1 Q_1 into Q 1 = R D R T Q_1 = R D R^T . The diagonal elements of the diagonal matrix D D are the eigenvalues of Q 1 Q_1 which are the reciprocal of the square of the semi-major and semi-minor axes lengths a a and b b of the resulting ellipse of intersection.

Hence, at last the shadow area is given by π a b \pi a b .

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...