I have drawn a 2D representation of the problem, but I will eventually have to solve this in 3 dimensions.
A line is drawn from an origin point to infinity, in a direction given by pitch and yaw. There is an axis-aligned box "in front of" the point.
I want to get the coordinates of the point on box that is closest to the line, or, if it intersects, closest to the origin point.
I.e., if the line were 'turned' towards the box, which point of the box would intersect with the line first?
Make parametric represenation of the ray with base point P0, direction vector D and parameter t
P = P0 + t * D
Get t for intersections of the ray with rectangle edges like this (similar in 3d):
Rect.Right = X0 + t * D.X
Find what intersection occurs first (smaller t), check coordinates of intersection. If inside edge - point found. If not, analyze intersection parameters with edgr continuations to determine what corner (perhaps edge in 3d) is the closest
Note that in 2d case you need to check only two possible edges - depending on ray direction. For example - left and bottom for your right picture. When you see that intersections are out of edges - check what of two corners is closer. The same for 3d - but intersection is possible for three faces and closest for more edges or corners.
Let the position of the point be (x0,y0,z0) and the box have corners (x1,y1,z1) and (x2,y2,z2) with x1 < x2, y1 < y2, z1 < z2. In terms of yaw ψ and pitch θ a unit vector along the line will be give by
(u,v,w) = (cos ψ sin θ, sin ψ, cos ψ cos θ)
The line is (x0,y0,z0) + t (u,v,w)
Finding intersection with one of the plane containing a face of the box is trivial. Say to find the intersection with the plane x=x1, just requires solving
x0 + u t = x1 so t = (x1-x0)/u. Once found its easy to check if the intersection is contained in the face.
The tricky situation happens if the line does not intersect the faces. Here we have a pair of skew lines, and wish to find the closest pair of points one on each line.
Consider the closest point to the edge from (x1,y1,z1) to (x2,y1,z1).
We want to find the parameter s,t such that that the points
(x0,y0,z0)+s(u,v,w)
(x1,y1,z1)+t(1,0,0)
are the closest. The segment joining these points must be perpendicular to both lines. A vector along that line is the cross product
N = (u,v,w) X (1,0,0) = (0,w,-v)
Now consider the plane through (x1,y1,z1) spanned by (1,0,0) and N, this has normal
N2 = (1,0,0) X N
= (1,0,0) X (0,w,-v)
= (0,v,w)
and the plane is defined by
P . N2 = (x1,y1,z1) . N2
Take a point on our ray
( (x0,y0,z0)+s(u,v,w) ) . N2 = (x1,y1,z1) . N2
(x0,y0,z0) . N2 + s (u,v,w) . N2 = (x1,y1,z1) . N2
s (u,v,w) . N2 = ((x1,y1,z1)-(x0,y0,z0)) . N2
s (v^2+w^2) = (y1-y0) v + (z1-z0) w
so
s = [ (y1-y0) v + (z1-z0) w ] / (v^2+w^2)
We can repeat the above for each edge on the box, find the closest points and select the smallest.
Related
Lets say I have a point within a circle(not necessarily the origin) moving at a given vector how would I calculate the x and y coordinate of the point where it hits the edge of the circle.
Shift all coordinates by -cx, -cy. Now circle is centered at origin and has equation
x^2+y^2=R^2
Point coordinate (px, py), unit direction vector is (dx,dy). Equation of ray:
x = px + t * dx
y = py + t * dy
Substitute these variables into the circle equation, solve equation, find parameter t>0, then find intersection point (x,y), shift it back by (cx, cy).
I have a point in 3d P(x,y,z) and a plane of view Ax+By+Cz+d=0 . A point in plane is E.Now i want to project that 3d point to that plane and get 2d coordinates of the projected point relative to the point E.
P(x,y,z) = 3d point which i want to project on the plane.
Plane Ax + By + Cz + d = 0 , so normal n = (A,B,C)
E(ex,ey,ez) = A point in plane ( eye pos of camera )
What i am doing right now is to get nearest point in plane from point P.then i subtract that point to E.I suspect that this is right ???
please help me.Thanks.
The closest point is along the normal to the plane. So define a point Q that is offset from P along that normal.
Q = P - n*t
Then solve for t that puts Q in the plane:
dot(Q,n) + d = 0
dot(P-n*t,n) + d = 0
dot(P,n) - t*dot(n,n) = -d
t = (dot(P,n)+d)/dot(n,n)
Where dot((x1,y1,z1),(x2,y2,z2)) = x1*x2 + y1*y2 + z1*z2
You get a point on the plane as p0 = (0, 0, -d/C). I assume the normal has unit length.
The part of p in the same direction as n is dot(p-n0, n) * n + p0, so the projection is p - dot(p-p0,n)*n.
If you want some coordinates on the plane, you have to provide a basis/coordinate system. Eg two linear independent vectors which span the plane. The coordinates depend on these basis vectors.
I tried using a raycasting-style function to do it but can't get any maintainable results. I'm trying to calculate the intersection between two tangents on one circle. This picture should help explain:
I've googled + searched stackoverflow about this problem but can't find anything similar to this problem. Any help?
Well, if your variables are:
C = (cx, cy) - Circle center
A = (x1, y1) - Tangent point 1
B = (x2, y2) - Tangent point 2
The lines from the circle center to the two points A and B are CA = A - C and CB = B - C respectively.
You know that a tangent is perpendicular to the line from the center. In 2D, to get a line perpendicular to a vector (x, y) you just take (y, -x) (or (-y, x))
So your two (parametric) tangent lines are:
L1(u) = A + u * (CA.y, -CA.x)
= (A.x + u * CA.y, A.y - u * CA.x)
L2(v) = B + v * (CB.y, -CB.x)
= (B.x + v * CB.y, B.x - v * CB.x)
Then to calculate the intersection of two lines you just need to use standard intersection tests.
The answer by Peter Alexander assumes that you know the center of the circle, which is not obvious from your figure http://oi54.tinypic.com/e6y62f.jpg.
Here is a solution without knowing the center:
The point C (in your figure) is the intersection of the tangent at A(x, y) with the line L perpendicular to AB, cutting AB into halves. A parametric equation for the line L can be derived as follows:
The middle point of AB is M = ((x+x2)/2, (y+y2)/2), where B(x2, y2). The vector perpendicular to AB is N = (y2-y, x-x2). The vector equation of the line L is hence
L(t) = M + t N, where t is a real number.
Goal
I want to determine if a test point is within a defined quadrilateral. I'm probably going to implement the solution in Matlab so I only need pseudo-code.
Inputs
Corners of quadrilateral : (x1,y1) (x2,y2) (x3,y3) (x4,y4)
Test point : (xt, yt)
Output
1 - If within quadrilateral
0 - Otherwise
Update
It was pointed out that identifying the vertices of the quadrilateral is not enough to uniquely identify it. You can assume that the order of the points determines the sides of the quadrilateral (point 1 connects 2, 2 connects to 3, 3 connects to 4, 4 connects to 1)
You can test the Point with this condition. Also you can treat quadrilateral as 2 triangles to calculate its area.
Use inpolygon. Usage would be inpolygon(xt,yt,[x1 x2 x3 x4],[y1 y2 y3 y4])
Since it's a simple quadrilateral you can test for a point in triangle for each end and a point in rectangle for the middle.
EDIT Here is some pseudo code for point in triangle:
function SameSide(p1,p2, a,b)
cp1 = CrossProduct(b-a, p1-a)
cp2 = CrossProduct(b-a, p2-a)
if DotProduct(cp1, cp2) >= 0 then return true
else return false
function PointInTriangle(p, a,b,c)
if SameSide(p,a, b,c) and SameSide(p,b, a,c)
and SameSide(p,c, a,b) then return true
else return false
Or using Barycentric technique:
A, B, and C are the triangle end points, P is the point under test
// Compute vectors
v0 = C - A
v1 = B - A
v2 = P - A
// Compute dot products
dot00 = dot(v0, v0)
dot01 = dot(v0, v1)
dot02 = dot(v0, v2)
dot11 = dot(v1, v1)
dot12 = dot(v1, v2)
// Compute barycentric coordinates
invDenom = 1 / (dot00 * dot11 - dot01 * dot01)
u = (dot11 * dot02 - dot01 * dot12) * invDenom
v = (dot00 * dot12 - dot01 * dot02) * invDenom
// Check if point is in triangle
return (u > 0) && (v > 0) && (u + v < 1)
If the aim is to code your own test, then pick any classic point in polygon test to implement. Otherwise do what Jacob suggests.
assuming you the given coordinates are arranged s.t.
(x1,y1) = rightmost coordinate
(x2,y2) = uppermost coordinate
(x3,y3) = leftmost coordinate
(x4,y4) = botoom-most coordinate
You can do the following:
1. calculate the 4 lines of the quadrilateral (we'll call these quad lines)
2. calculate 4 lines, from the (xt, yt) to every other coordinate (we'll call these new lines)
3. if any new line intersects any of the quad lines, then the coordinate is outside of the quadrilateral, otherwise it is inside.
Assume A,B,C,D are the vertices of the quadrilateral and P is the point.
If P is inside the quadrilateral then all dot products dot(BP,BA), dot(BP,BC), dot(AP,AB), dot(AP,AD), dot(DP,DC), dot(DP,DA), dot(CP,CB) and dot(CP,CD) will be positive.
If P is outside the quadrilateral at least one of these products will be negative.
The solution I used to solve this problem was to get the angle of P (in the diagrams the OP posted) for each of the 4 triangles it makes with each side of the quadrilateral. Add the angles together. If they equal (or nearly equal, depending on the error tolerance of the code) 360, the point is inside the quadrilateral. If the sum is less than 360, the point is outside. However, this might only work with convex quadrilaterals.
Let's say I have two points in 3D space (a and b) and a fixed axis/unit vector called n.
I want to create a rotation matrix that minimizes the euclidan distance between point a (unrotated) and the rotated point b.
E.g:
Q := matrix_from_axis_and_angle (n, alpha);
find the unknown alpha that minimizes sqrt(|a - b*Q|)
Btw - If a solution/algorithm can be easier expressed with unit-quaternions go ahead and use them. I just used matrices to formulate my question because they're more widely used.
Oh - I know there are some degenerated cases ( a or b lying exactly in line with n ect.) These can be ignored. I'm just looking for the case where a single solution can be calculated.
sounds fairly easy. Assume unit vector n implies rotation around a line parallel to n through point x0. If x0 != the origin, translate the coordinate system by -x0 to get points a' and b' relative to new coordinate system origin 0, and use those 2 points instead of a and b.
1) calculate vector ry = n x a
2) calculate unit vector uy = unit vector in direction ry
3) calculate unit vector ux = uy x n
You now have a triplet of mutually perpendicular unit vectors ux, uy, and n, which form a right-handed coordinate system. It can be shown that:
a = dot(a,n) * n + dot(a,ux) * ux
This is because unit vector uy is parallel to ry which is perpendicular to both a and n. (from step 1)
4) Calculate components of b along unit vectors ux, uy. a's components are (ax,0) where ax = dot(a,ux). b's components are (bx,by) where bx = dot(b,ux), by = dot(b,uy). Because of the right-handed coordinate system, ax is always positive so you don't actually need to calculate it.
5) Calculate theta = atan2(by, bx).
Your rotation matrix is the one which rotates by angle -theta relative to coordinate system (ux,uy,n) around the n-axis.
This yields degenerate answers if a is parallel to n (steps 1 and 2) or if b is parallel to n (steps 4, 5).
I think you can rephrase the question to:
what is the distance from a point to a 2d circle in 3d space.
the answer can be found here
so the steps needed are as following:
rotating the point b around a vector n gives you a 2d circle in 3d space
using the above, find the distance to that circle (and the point on the circle)
the point on the circle is the rotated point b you are looking for.
deduce the rotated angle
...or something ;^)
The distance will be minimized when the vector from a to the line along n lines up with the vector from b to the line along n.
Project a and b into the plane perpendicular to n and solve the problem in 2 dimensions. The rotation you get there is the rotation you need to minimize the distance.
Let P be the plane that is perpendicular to n.
We can find the projection of a into the P-plane, (and similarly for b):
a' = a - (dot(a,n)) n
b' = b - (dot(b,n)) n
where dot(a,n) is the dot-product of a and n
a' and b' lie in the P-plane.
We've now reduced the problem to 2 dimensions. Yay!
The angle (of rotation) between a' and b' equals the angle (of rotation) needed to swing b around the n-axis so as to be closest to a. (Think about the shadows b would cast on the P-plane).
The angle between a' and b' is easy to find:
dot(a',b') = |a'| * |b'| * cos(theta)
Solve for theta.
Now you can find the rotation matrix given theta and n here:
http://en.wikipedia.org/wiki/Rotation_matrix
Jason S rightly points out that once you know theta, you must still decide to rotate b clockwise or counterclockwise about the n-axis.
The quantity, dot((a x b),n), will be a positive quantity if (a x b) lies in the same direction as n, and negative if (a x b) lies in the opposite direction. (It is never zero as long as neither a nor b is collinear with n.)
If (a x b) lies in the same direction as n, then b has to be rotated clockwise by the angle theta about the n-axis.
If (a x b) lies in the opposite direction, then b has to be rotated clockwise by the angle -theta about the n-axis.