I'm programming a 3D game where the user controls a first-person camera, and movement is constrained to the inside surface of a sphere. I've managed to constrain the movement, but I'm having trouble figuring out how to manage the camera orientation using quaternions. Ideally the camera up vector should point along the normal of the sphere towards its center, and user should be able to free look around - as if we was always on the bottom of the sphere, no matter where he moves.
Presumably you have two vectors describing the camera's orienation. One will be your V'up describing which way is up relative to the camera orientation and the other will be your V'norm which will be the direction the camera is aimed. You will also have a position p', where your camera is located at some time. You define a canonical orientation and position given by, say:
Vup = <0, 1, 0>
Vnorm = <0, 0, 1>
p = <0, -1, 0>
Given a quaternion rotation q you then apply your rotation to those vectors to get:
V'up = qVupq-1
V'norm = qVnormq-1
p' = qpq-1
In your particular situation, you define q to incrementally accumulate the various rotations that result in the final rotation you apply to the camera. The effect will be that it looks like what you're describing. That is, you move the camera inside a statically oriented and positioned sphere rather than moving the spehere around a statically oriented and positioned camera.
Each increment is computed by a rotation of some angle θ about the vector V = V'up x V'norm.
Quaternions are normally used to avoid gimbal lock in free space motion (flight sims, etc.). In your case, you actually want the gimbal effect, since a camera that is forced to stay upright will inevitably behave strangely when it has to point almost straight up or down.
You should be able to represent the camera's orientation as just a latitude/longitude pair indicating the direction the camera is pointing.
Related
I am looking for a way to find the (x, y) pixel position of a point in an image taken by camera. I know the physical position of the object (distance - width, height and depth), the resolution of the image and probably the focal distance (maybe I could also get some others camera parameteres - bbut I want as less information as possible).
In case I am not clear I want a formula/algorithm/procedure to map from (width, heigh, depth) to (x_pixel_position_in_image, y_pixe_position_in_image) - to connect the physical coordates with the pixel ones.
Thank you very much.
If you check the diagram linked below, the perspective projection of a 3d point with a camera depends on two main sources of information.
Diagram
Camera Parameters (Intrinsics) and the where the camera is in a fixed world coordinate (Extrinsics). Since you want to project points in the camera coordinate system, you can assume the world coordinate is coinciding with the camera. Hence the extrinsic matrix [R|t] can be expressed as,
R = eye(3); and t = [0; 0; 0].
Therefore, all you need to know is the camera parameters (focal length and optical center location). You can read more about this here.
Is there a way to convert that data:
Object position which is a 3D point (X, Y, Z),
Camera position which is a 3D point (X, Y, Z),
Camera yaw, pitch, roll (-180:180, -90:90, 0)
Field of view (-45°:45°)
Screen width & height
into the 2D point on the screen (X, Y)?
I'm looking for proper math calculations according to this exact set of data.
It's difficult, but it's possible to do it for yourself.
There are lots of libraries that do this for you, but it is more satisfying if you do it yourself:
This problem is possible and I have written my own 3D engine to do this for objects in javascript using the HTML5 Canvas. You can see my code here and solve a 3D maze game I wrote here to try and understand what I will talk about below...
The basic idea is to work in steps. To start, you have to forget about camera angle (yaw, pitch and roll) as these come later and just imagine you are looking down the y axis. Then the basic idea is to calculate, using trig, the pitch angle and yaw to your object coordinate. By this I mean imagining that you are looking through a letterbox, the yaw angle would be the angle in degrees left and right to your coordinate (so both positive and negative) from the center/ mid line and the yaw up and down from it. Taking these angles, you can map them to the x and y 2D coordinate system.
The calculations for the angles are:
pitch = atan((coord.x - cam.x) / (coord.y - cam.y))
yaw = atan((coord.z - cam.z) / (coord.y - cam.y))
with coord.x, coord.y and coord.z being the coordinates of the object and the same for the cam (cam.x, cam.y and cam.z). These calculations also assume that you are using a Cartesian coordinate system with the different axis being: z up, y forward and x right.
From here, the next step is to map this angle in the 3D world to a coordinate which you can use in a 2D graphical representation.
To map these angles into your screen, you need to scale them up as distances from the mid line. This means multiplying them by your screen width / fov. Finally, these distances will now be positive or negative (as it is an angle from the mid line) so to actually draw it on a canvas, you need to add it to half of the screen width.
So this would mean your canvas coordinate would be:
x = width / 2 + (pitch * (width / fov)
y = height / 2 + (yaw * (height / fov)
where width and height are the dimensions of you screen, fov is the camera's fov and yaw and pitch are the respective angles of the object from the camera.
You have now achieved the first big step which is mapping a 3D coordinate down to 2D. If you have managed to get this all working, I would suggest trying multiple points and connecting them to form shapes. Also try moving your cameras position to see how the perspective changes as you will soon see how realistic it already looks.
In addition, if this worked fine for you, you can move on to having the camera be able to not only change its position in the 3D world but also change its perspective as in yaw, pitch and roll angles. I will not go into this entirely now, but the basic idea is to use 3D world transformation matrices. You can read up about them here but they do get quite complicated, however I can give you the calculations if you get this far.
It might help to read (old style) OpenGL specs:
https://www.khronos.org/registry/OpenGL/specs/gl/glspec14.pdf
See section 2.10
Also:
https://www.khronos.org/opengl/wiki/Vertex_Transformation
Might help with more concrete examples.
Also, for "proper math" look up 4x4 matrices, projections, and homogeneous coordinates.
https://en.wikipedia.org/wiki/Homogeneous_coordinates
I'd need a formula to calculate 3D position and direction or orientation of a camera in a following situation:
Camera starting position is looking directly into center of the Earth. Green line goes straight up to the sky
Position that camera needs to move to is looking like this
Starting position probably shouldn't matter, but the question is:
How to calculate camera position and direction given 3D coordinates of any point on the globe. In the camera final position, the distance from Earth is always fixed. From desired camera point of view, the chosen point should appear at the rightmost point of a globe.
I think what you want for camera position is a point on the intersection of a plane parallel to the tangent plane at the location, but somewhat further from the Center, and a sphere representing the fixed distance the camera should be from the center. The intersection will be a circle, so there are infinitely many camera positions that work.
Camera direction will be 1/2 determined by the location and 1/2 determined by how much earth you want in the picture.
Suppose (0,0,0) is the center of the earth, Re is the radius of the earth, and (a,b,c) is the location on the earth you want to look at. If it's in terms of latitude and longitude you should convert to Cartesian coordinates which is straightforward. Your camera should be on a plane perpendicular to the vector (a,b,c) and at a height kRe above the earth where k>1 is some number you can adjust. The equation for the plane is then ax+by+cz=d where d = kRe^2. Note that the plane passes through the point (ka,kb,kc) in space, which is what we wanted.
Since you want the camera to be at a certain height above the earth, say h*Re where 1 < k < h, you need to find points on ax+by+cz=d for which x^2+y^2+z^2 = h^2*Re^2. So we need the intersection of the plane and a sphere. It will be easier to manage if we have a coordinate system on the plane, which we get from an orthonormal system which includes (a,b,c). A good candidate for the second vector in the orthonormal system is the projection of the z-axis (polar axis, I assume). Projecting (0,0,1) onto (a,b,c),
proj_(a,b,c)(0,0,1) = (a,b,c).(0,0,1)/|(a,b,c)|^2 (a,b,c)
= c/Re^2 (a,b,c)
Then the "horizontal component" of (0,0,1) is
u = proj_Plane(0,0,1) = (0,0,1) - c/Re^2 (a,b,c)
= (-ac/Re^2,-bc/Re^2,1-c^2/Re^2)
You can normalize the vector to length 1 if you wish but there's no need. However, you do need to calculate and store the square of the length of the vector, namely
|u|^2 = ((ac)^2 + (bc)^2 + (Re^2-c^2))/Re^4
We could complicate this further by taking the cross product of (0,0,1) and the previous vector to get the third vector in the orthonormal system, then obtain a parametric equation for the intersection of the plane and sphere on which the camera lies, but if you just want the simplest answer we won't do that.
Now we need to solve for t such that
|(ka,kb,kc)+t(-ac/Re^2,-bc/Re^2,1-c^2/Re^2)|^2 = h^2 Re^2
|(ka,kb,kc)|^2 + 2t (a,b,c).u + t^2 |u|^2 = h^2 Re^2
Since (a,b,c) and u are perpendicular, the middle term drops out, and you have
t^2 = (h^2 Re^2 - k^2 Re^2)/|u|^2.
Substituting that value of t into
(ka,kb,kc)+t(-ac/Re^2,-bc/Re^2,1-c^2/Re^2)
gives the position of the camera in space.
As for direction, you'll have to experiment with that a bit. Some vector that looks like
(a,b,c) + s(-ac/Re^2,-bc/Re^2,1-c^2/Re^2)
should work. It's hard to say a priori because it depends on the camera magnification, width of the view screen, etc. I'm not sure offhand whether you'll need positive or negative values for s. You may also need to rotate the camera viewport, possibly by 90 degrees, I'm not sure.
If this doesn't work out, it's possible I made an error. Let me know how it works out and I'll check.
My question is similar to 3D Scene Panning in perspective projection (OpenGL) except I don't know how to compute the direction in which to move the mesh.
I have a program in which various meshes can be selected. Once a mesh is selected I want it to translate when click-dragging the cursor. When the cursor moves up, I want the mesh to move up, and so on for the appropriate direction. In other words, I want the mesh to translate in directions along the plane that is perpendicular to the viewing direction.
I have the Vector2 for the Delta (x,y) in cursor postion, and I have the Vector3 viewDirection of the camera and the center of the mesh. How can I figure out which way to translate the mesh in 3d space with the Delta and viewDirection? Will I need other information in order to to this calculation (such as the up, or eye)?
It doesn't matter if if the scale of the translation is off, I'm just trying to figure out the direction right now.
EDIT: for some reason I had a confusion about getting the up direction. Clearly it can be calculated by applying the camera rotation to the specified perspective up vector.
You'll need an additional vector, upDirection, which is the unit vector pointing "up" from your camera. You can now cross-product viewDirection and upDirection to get rightDirection, the vector pointing "right" from your camera.
You want to map y deltas to motion along upDirection (or -upDirection) and x deltas to motion in rightDirection. These vectors are in world-space.
You may want to scale the translation speed to match the mouse speed. If you are using perspective projection you'll want to scale the translation speed with your model's depth with respect to your camera (The further the object is from your camera, the faster you will need to move it if you want it to match the mouse.)
How do I make a infinite/repeating world that handles rotation, just like in this game:
http://bloodfromastone.co.uk/retaliation.html
I have coded my rotating moving world by having a hierarchy like this:
Scene
- mainLayer (CCLayer)
- rotationLayer(CCNode)
- positionLayer(CCNode)
The rotationLayer and positionLayer have the same size (4000x4000 px right now).
I rotate the whole world by rotating the rotationLayer, and I move the whole world by moving the positionLayer, so that the player always stays centered on the device screen and it is the world that moves and rotates.
Now I would like to make it so that if the player reaches the bounds of the world (the world is moved so that the worlds bounds gets in to contact with the device screen bounds), then the world is "wrapped" to the opposite bounds so that the world is infinite. If the world did not rotate that would be easy, but now that it does I have no idea how to do this. I am a fool at math and in thinking mathematically, so I need some help here.
Now I do not think I need any cocos2d-iphone related help here. What I need is some way to calculate if my player is outside the bounds of the world, and then some way to calculate what new position I must give the world to wrap the world.
I think I have to calculate a radius for a circle that will be my foundry inside the square world, that no matter what angle the square world is in, will ensure that the visible rectangle (the screen) will always be inside the bounds of the world square. And then I need a way to calculate if the visible rectangle bounds are outside the bounds circle, and if so I need a way to calculate the new opposite position in the bounds circle to move the world to. So to illustrate I have added 5 images.
Visible rectangle well inside bounds circle inside a rotated square world:
Top of visible rectangle hitting bounds circle inside a rotated square world:
Rotated square world moved to opposite vertical position so that bottom of visible rectangle now hitting bounds circle inside rotated world:
Another example of top of visible rectangle hitting bounds circle inside a rotated square world to illustrate a different scenario:
And again rotated square world moved to opposite vertical position so that bottom of visible rectangle now hitting bounds circle inside rotated world:
Moving the positionLayer in a non-rotated situation is the math that I did figure out, as I said I can figure this one out as long as the world does not get rotate, but it does. The world/CCNode (positionLayer) that gets moved/positioned is inside a world/CCNode (rotationLayer) that gets rotated. The anchor point for the rotationLayer that rotates is on the center of screen always, but as the positionLayer that gets moved is inside the rotating rotationLayer it gets rotated around the rotationLayer's anchor point. And then I am lost... When I e.g. move the positionLayer down enough so that its top border hits the top of the screen I need to wrap that positionLayer as JohnPS describes but not so simple, I need it to wrap in a vector based on the rotation of the rotationLayer CCNode. This I do not know how to do.
Thank you
Søren
Like John said, the easiest thing to do is to build a torus world. Imagine that your ship is a point on the surface of the donut and it can only move on the surface. Say you are located at the point where the two circles (red and purple in the picture) intersect:
.
If you follow those circles you'll end up where you started. Also, notice that, no matter how you move on the surface, there is no way you're going to reach an "edge". The surface of the torus has no such thing, which is why it's useful to use as an infinite 2D world. The other reason it's useful is because the equations are quite simple. You specify where on the torus you are by two angles: the angle you travel from the "origin" on the purple circle to find the red circle and the angle you travel on the red circle to find the point you are interested in. Both those angles wrap at 360 degrees. Let's call the two angles theta and phi. They are your ship's coordinates in the world, and what you change when you change velocities, etc. You basically use them as your x and y, except you have to make sure to always use the modulus when you change them (your world will only be 360 degrees in each direction, it will then wrap around).
Suppose now that your ship is at coordinates (theta_ship,phi_ship) and has orientation gamma_ship. You want to draw a square window with the ship at its center and length/width equal to some percentage n of the whole world (say you only want to see a quarter of the world at a time, then you'd set n = sqrt(1/4) = 1/2 and have the length and width of the window set to n*2*pi = pi). To do this you need a function that takes a point represented in the screen coordinates (x and y) and spits out a point in the world coordinates (theta and phi). For example, if you asked it what part of the world corresponds to (0,0) it should return back the coordinates of the ship (theta_ship,phi_ship). If the orientation of the ship is zero (x and y will be aligned with theta and phi) then some coordinate (x_0,y_0) will correspond to (theta_ship+k*x_0, phi_ship+k*y_0), where k is some scaling factor related to how much of the world one can see in a screen and the boundaries on x and y. The rotation by gamma_ship introduces a little bit of trig, detailed in the function below. See the picture for exact definitions of the quantities.
!Blue is the screen coordinate system, red is the world coordinate system and the configuration variables (the things that describe where in the world the ship is). The object
represented in world coordinates is green.
The coordinate transformation function might look something like this:
# takes a screen coordinate and returns a world coordinate
function screen2world(x,y)
# this is the angle between the (x,y) vector and the center of the screen
alpha = atan2(x,y);
radius = sqrt(x^2 + y^2); # and the distance to the center of the screen
# this takes into account the rotation of the ship with respect to the torus coords
beta = alpha - pi/2 + gamma_ship;
# find the coordinates
theta = theta_ship + n*radius*cos(beta)/(2*pi);
phi = phi_ship + n*radius*sin(beta)/(2*pi));
# return the answer, making sure it is between 0 and 2pi
return (theta%(2*pi),phi%(2*pi))
and that's pretty much it, I think. The math is just some relatively easy trig, you should make a little drawing to convince yourself that it's right. Alternatively you can get the same answer in a somewhat more automated fashion by using rotations matrices and their bigger brother, rigid body transformations (the special Euclidian group SE(2)). For the latter, I suggest reading the first few chapters of Murray, Li, Sastry, which is free online.
If you want to do the opposite (go from world coordinates to screen coordinates) you'd have to do more or less the same thing, but in reverse:
beta = atan2(phi-phi_ship, theta-theta_ship);
radius = 2*pi*(theta-theta_ship)/(n*cos(beta));
alpha = beta + pi/2 - gamma_ship;
x = radius*cos(alpha);
y = radius*sin(alpha);
You need to define what you want "opposite bounds" to mean. For 2-dimensional examples see Fundamental polygon. There are 4 ways that you can map the sides of a square to the other sides, and you get a sphere, real projective plane, Klein bottle, or torus. The classic arcade game Asteroids actually has a torus playing surface.
The idea is you need glue each of your boundary points to some other boundary point that will make sense and be consistent.
If your world is truly 3-dimensional (not just 3-D on a 2-D surface map), then I think your task becomes considerably more difficult to determine how you want to glue your edges together--your edges are now surfaces embedded in the 3-D world.
Edit:
Say you have a 2-D map and want to wrap around like in Asteroids.
If the map is 1000x1000 units, x=0 is the left border of the map, x=999 the right border, and you are looking to the right and see 20 units ahead. Then at x=995 you want to see up to 1015, but this is off the right side of the map, so 1015 should become 15.
If you are at x=5 and look to the left 20 units, then you see x=-15 which you really want to be 985.
To get these numbers (always between 0 and 999) when you are looking past the border of your map you need to use the modulo operator.
new_x = x % 1000; // in many programming languages
When x is negative each programming language handles the result of x % 1000 differently. It can even be implementation defined. i.e. it will not always be positive (between 0 and 999), so using this would be safer:
new_x = (x + 1000) % 1000; // result 0 to 999, when x >= -1000
So every time you move or change view you need to recompute the coordinates of your position and coordinates of anything in your view. You apply this operation to get back a coordinate on the map for both x and y coordinates.
I'm new to Cocos2d, but I think I can give it a try on helping you with the geometry calculation issue, since, as you said, it's not a framework question.
I'd start off by setting the anchor point of every layer you're using in the visual center of them all.
Then let's agree on the assumption that the first part to touch the edge will always be a corner.
In case you just want to check IF it's inside the circle, just check if all the four edges are inside the circle.
In case you want to know which edge is touching the circumference of the circle, just check for the one that is the furthest from point x=0 y=0, since the anchor will be at the center.
If you have a reason for not putting the anchor in the middle, you can use the same logic, just as long as you include half of the width of each object on everything.