I want to write a transformer for converting svg basic types into worldwind shapes like polyline, polygon etc.
Since svg gives coordinates on canvas and I need to convert them to position, I am looking for a method in the api which can do this.
I see there is Vec4 for point but I am not sure how it relates to canvas coordinates.
Will it be a correct representation if for say point x=100,y=100, I do the following
Vec4 vec=new Vec4(x,y,0.0f);
Globe g=view.getGlobe();
Position p=g.computePositionFromPoint(vec);
Will this correspond position will be the position at point(x=100,y=100) on the screen. If i bring my mouse to x=100 and y=100 for the current view the position should be p.
Globe.computePositionFromPoint(vec) takes Cartesian coordinates as input, not screen coordinates. To go from screen coordinates to position you want something like this:
Vec4 screenCoords = new Vec4(x,y);
Vec4 cartesian = view.unProject(screenCoords);
Globe g=view.getGlobe();
Position p=g.computePositionFromPoint(cartesian);
A better way to do this would be:
Point screenPoint = dragContext.getPoint();
View view = dragContext.getView();
double latitude = view.computePositionFromScreenPoint(screenPoint.getX(), screenPoint.getY()).getLatitude().degrees;
double longitude = view.computePositionFromScreenPoint(screenPoint.getX(), screenPoint.getY()).getLongitude().degrees;
Position objectPosition = new Position(LatLon.fromDegrees(latitude, longitude), 0);
Then you can set the position of the object to objectPosition.
If you want to go from canvas to 3D, you're probably looking for View#computeRayFromScreenPoint(double, double). This will give you a ray (in Vec4 format) from the eye through the given pixel on the canvas. You'll have to intersect this ray with something to generate a meaningful 3D point, since each pixel is an infinite line in space.
about the Globe#computePointFromPosition:
Position - A latitude, longitude, altitude position, with the altitude being in MSL (alt above sea level)
Vec4 - Just a directional vector (often used for Cartesian coordinates). For the Cartesian coordinate system, the z-axis comes out of the earth through 0deg/0deg lat/lon, x-axis through 0deg/90deg, and y-axis through the north pole.
As Chris mentioned, The Globe#computePointFromPosition() and Globe#computePointFromPosition() switch from a Position to a Cartesian Vec4 and vice versa, using your globe as the reference frame.
Related
Problem:
Vulkan right handed coordinate system became left handed coordinate system after applying projection matrix. How can I make it consistent with Vulkan coordinate system?
Details:
I know that Vulkan is right handed coordinate system where
X+ points toward right
Y+ points toward down
Z+ points toward inside the screen
I've this line in the vertex shader: https://github.com/AndreaCatania/HelloVulkan/blob/master/shaders/shader.vert#L23
gl_Position = scene.cameraProjection * scene.cameraView * meshUBO.model * vec4(vertexPosition, 1.0);
At this point: https://github.com/AndreaCatania/HelloVulkan/blob/master/main.cpp#L62-L68 I'm defining the position of camera at center of scene and the position of box at (4, 4, -10) World space
The result is this:
As you can see in the picture above I'm getting Z- that point inside the screen but it should be positive.
Is it expected and I need to add something more or I did something wrong?
Useful part of code:
Projection calculation: https://github.com/AndreaCatania/HelloVulkan/blob/master/VisualServer.cpp#L88-L98
void Camera::reloadProjection(){
projection = glm::perspectiveRH_ZO(FOV, aspect, near, far);
isProjectionDirty = false;
}
Camera UBO fill: https://github.com/AndreaCatania/HelloVulkan/blob/master/VisualServer.cpp#L403-L414
SceneUniformBufferObject sceneUBO = {};
sceneUBO.cameraView = camera.transform;
sceneUBO.cameraProjection = camera.getProjection();
I do not use or know Vulcan but perspective projection matrix (at lest in OpenGL) is looking in the Z- direction which inverts one axis of your coordinate system. That inverts the winding rule of the coordinate system.
If you want to preserve original winding than just invert Z axis vector in the matrix for more info see:
Understanding 4x4 homogenous transform matrices
So just scale the Z axis by -1 either by some analogy to glScale(1.0,1.0,-1.0); or by direct matrix cells access.
All the OpenGL left coordinate system vs Vulkan right coordinate system happens during the fragment shader in NDC space, it means your view matrix doesn't care.
If you are using glm, everything you do in world space or view space is done via a right handed coordinate system.
GLM, a very popular math library that every beginner uses, uses right-handed coordinate system by default.
Your view matrix must be set accordingly, the only way to get a right handed system with x from left to right and y from bottom to top is if to set your z looking direction looking down at the negative values. If you don't provide a right handed system to your glm::lookat call, glm will convert it with one of your axis getting flipped via a series of glm::cross see glm source code
the proper way:
glm::vec3 eye = glm::vec3(0, 0, 10);
glm::vec3 up = glm::vec3(0, 1, 0);
glm::vec3 center = glm::vec3(0, 0, 0);
// looking in the negative z direction
glm::mat4 viewMat = glm::lookAt(eye, up, center);
Personnaly I store all information for coordinate system conversion in the projection matrix because by default glm doest it for you for the z coordinate
from songho: http://www.songho.ca/opengl/gl_projectionmatrix.html
Note that the eye coordinates are defined in the right-handed coordinate system, but NDC uses the left-handed coordinate system. That is, the camera at the origin is looking along -Z axis in eye space, but it is looking along +Z axis in NDC. Since glFrustum() accepts only positive values of near and far distances, we need to NEGATE them during the construction of GL_PROJECTION matrix.
Because we are looking at the negative z direction glm by default negate the sign.
It turns out that the y coordinate is flipped between vulkan and openGL so everything will get turned upside down. One way to resolve the problem is to negate the y values aswell:
glm::mat4 projection = glm::perspective(glm::radians(verticalFov), screenDimension.x / screenDimension.y, near, far);
// Vulkan NDC space points downward by default everything will get flipped
projection[1][1] \*= -1.0f;
If you follow the above step you must end up with something very similar to old openGL applications and with the up vector of your camera with the same sign than most 3D models.
I'm working on a 3D mapping application, and I've got to do some work with things like figuring out the visible region of a sphere (Earth) from a given point in space for things like clipping mapped regions and such.
Several things get easier if I can project the outline of Earth into screen space, clip polygons there, and then project back to the surface of the Earth (lat/lon), but I'm lost as to how to do that.
Is there a reasonable way to compute the outline of a sphere after perspective projection, and then a reasonable way to project things back onto the sphere?
You can clip the polygons in 3D. The silhouette of the sphere - back-projected into 3D - will always be a circle on a plane. Perspective projection does not change that. Thus, you can clip all polygons at the plane.
Calculating the plane is not too hard. If you consider the sphere's center the origin, then the plane could be represented in normal form as:
dot(n, x) = d
n is the normal. This one is easy. It is just the unit direction vector from the sphere center to the observer.
d is the distance from the sphere center. This is a bit harder but not too hard. If l is the distance of the observer to the sphere center and r is the sphere radius, then
d = r^2 / l
This is the plane which you can use to clip your polygons in 3D. If you need the radius of the circle on it, you can use the following formula:
r_c = r / sqrt(1 - r^2/(l-d)^2)
Let us take a point on a sphere in spherical coordinates (cos(u)sin(v),sin(u)sin(v),cos(v)) and an arbitrary projection center (x,y,z).
We express that a projecting line is tangent to the sphere by the perpendicularity condition of the direction of the line and the vector from the origin of the sphere:
(x-cos(u)sin(v))cos(u)sin(v) + (y-sin(u)sinv))sin(u)sin(v) + (z-cos(v)) cos(v) = 0
This simplifies to
x cos(u)sin(v) + y sin(u)sin(v) + z cos(v) = 1
which is a curve in the longitude/latitude coordinates. You can solve u as a function of v or conversely.
I have a question regarding the projection of an image over a set of 3D points. The image is given to me as a JPG, together with position and attitude information of the camera relative to a cartesian coordinate system (Xc,Yc,Zc and yaw, pitch, roll), as well as the horizontal and vertical field of view (in degrees).
Points are given using solely their 3d position in the same coordinate system (Xp,Yp,Zp).
In my coordinate system, Z is up. To project the image onto the points, I
compute the vector from camera to each point
Vector3 c2p = (Xp,Yp,Zp)-(Xc,Yc,Zc);
rotate c2p according to my camera's attitude (quaternion):
Vector3 c2pCamFrame = getCamQuaternion().conjugate().rotate(c2p);
compute azimuth and elevation from the camera's "center ray" to the point:
float azimuth = atan2(c2pCamFrame.x(),c2pCamFrame.y()));
float elevation = atan2(c2pCamFrame.z(),sqrt(pow(c2pCamFrame.x(),2)+pow(c2pCamFrame.y(),2)));
if azimuth and elevation are within the field of view, I assign the color of the corresponding pixel to the point.
This works almost perfectly, and the "almost" motivates my question. Let me show you:
I cannot figure out why the elevation of the projection is distorted. In the bottom right of the image, you can see that points outside the frustum (exceeding the elevation) actually become colored - and this distortion is null at an azimuth of 0 degrees and peaks at the left and right edges of the image, creating the pillow distortion.
Why does this distortion appear? I'd love to understand this problem both in geometrical as well as mathematical terms. Thank you!
The field of view angles are only valid on the principal axes. But you can do it the other way around. I.e. calculate the x/y bounds from the angles:
maxX = tan(horizontal_fov / 2)
maxY = tan(vertical_fov / 2)
And check
if(abs(c2pCamFrame.x() / c2pCamFrame.z()) <= maxX
&& abs(c2pCamFrame.y() / c2pCamFrame.z()) <= maxY)
Additionally, you might want to check if the points are in front of the camera:
... && c2pCamFrame.z() > 0
This assumes a left-handed coordinate system.
I have a problem with creating 3D cylinders (without OpenGL). I understand that a mesh is used to create the cylinder surface and triangle fans are used to create the top and bottom caps. I have already implemented the mesh but not the planar triangle fans, so currently my 3D object looks like a cylinder without the bottom and top cap.
I believe this is what I need to do in order to create the bottom and top caps. First, find the center point of the cylinder mesh. Second, find the vertices of the mesh. Third, using the center point and the 2 vertex points, create the triangle. Fourth, repeat the steps until a planar circle is created.
Are the above steps a sufficient way of creating the caps or is there a better way? And how do I find the vertices of the mesh so I can create the triangle fans?
First some notes:
you did not specify your platform
gfx interface
language
not enough info about your cylinder either
is it axis aligned?
what coordinate system (Cartesian/orthogonal/orthonormal)?
need additional dimensions like color or texture coordinates?
So I can provide just generic info then
Axis aligned cylinder
choose the granularity N
number of points along your cap's circle
usually 20-36 is OK but if you need higher precision then sometimes you need even 1000 points or more
all depends on the purpose,zoom, angle and distance of view ...
and performance issues
for now let N=32
you need BR (boundary representation)
you did not specify gfx interface but your text implies BR model (surface polygons)
also no pivot point position so I will choose middle point of cylinder to be (0,0,0)
z axis will be the height of cylinder
and the caps will be coplanar with xy plane
so for cylinder is enough set of 2 rings (caps)
so the points can be defined in C++ like this:
const int N=32; // mesh complexity
double p0[N][3],p1[N][3]; // rings`
double a,da,c,s,r,h2; // some temp variables
int i;
r =50.0; // cylinder radius
h2=100.0*0.5; // half height of cyliner
da=M_PI/double(N-1);
for (a=0.0,i=0;i<N;i++,a+=da)
{
c=r*cos(a);
s=r*sin(a);
p0[i][0]=c;
p0[i][1]=s;
p0[i][2]=+h2;
p1[i][0]=c;
p1[i][1]=s;
p1[i][2]=-h2;
}
the ring points are as closed loop (p0[0]==p0[N-1])
so you do not need additional lines to handle it...
now how to draw
cant write the code for unknown api but
'mesh' is something like QUAD_STRIP I assume
so just add points to it in this order:
QUAD_STRIP = { p0[0],p1[0],p0[1],p1[1],...p0[N-1],p1[N-1] };
if you have inverse normal problem then swap p0/p1
now for the fans
you do not need the middle point (unless you have interpolation aliasing issues)
so similar:
TRIANGLE_FAN0 = { p0[0],p0[1],...p0[N-1] };
TRIANGLE_FAN1 = { p1[0],p1[1],...p1[N-1] };
if you still want the middle point then:
TRIANGLE_FAN0 = { (0.0,0.0,+h2),p0[0],p0[1],...p0[N-1] };
TRIANGLE_FAN1 = { (0.0,0.0,-h2),p1[0],p1[1],...p1[N-1] };
if you have inverse normal problem then reverse the points order (middle point stays where it is)
Not axis aligned cylinder?
just use transform matrix on your p0[],p1[] point lists to translate/rotate to desired position
the rest stays the same
we are programming a 2D game in XNA. Now we have polygons which define our level elements. They are triangulated such that we can easily render them. Now I would like to write a shader which renders the polygons as outlined textures. So in the middle of the polygon one would see the texture and on the border it should somehow glow.
My first idea was to walk along the polygon and draw a quad on each line segment with a specific texture. This works but looks strange for small corners where the textures are forced to overlap.
My second approach was to mark all border vertices with some kind of normal pointing out of the polygon. Passing this to the shader would interpolate the normals across edges of the triangulation and I could use the interpolated "normal" as a value for shading. I could not test it yet but would that work? A special property of the triangulation is that all vertices are on the border so there are no vertices inside the polygon.
Do you guys have a better idea for what I want to achieve?
Here A picture of what it looks right now with the quad solution:
You could render your object twice. A bigger stretched version behind the first one. Not that ideal since a complex object cannot be streched uniformly to create a border.
If you have access to your screen buffer you can render your glow components into a rendertarget and align a full-screen quad to your viewport and add a fullscreen 2D silhouette filter to it.
This way you gain perfect control over the edge by defining its radius, colour, blur. With additional output values such as the RGB values from the object render pass you can even have different advanced glows.
I think rendermonkey had some examples in their shader editor. Its definetly a good starting point to work with and try out things.
Propaply you want calclulate new border vertex list (easy fill example with triangle strip with originals). If you use constant border width and convex polygon its just:
B_new = B - (BtoA.normalised() + BtoC.normalised()).normalised() * width;
If not then it can go more complicated, there is my old but pretty universal solution:
//Helper function. To working right, need that v1 is before v2 in vetex list and vertexes are going to (anti???) cloclwise!
float vectorAngle(Vector2 v1, Vector2 v2){
float alfa;
if (!v1.isNormalised())
v1.normalise();
if (!v2.isNormalised())
v2.normalise();
alfa = v1.dotProduct(v2);
float help = v1.x;
v1.x = v1.y;
v1.y = -help;
float angle = Math::ACos(alfa);
if (v1.dotProduct(v2) < 0){
angle = -angle;
}
return angle;
}
//Normally dont use directly this!
Vector2 calculateBorderPoint(Vector2 vec1, Vector2 vec2, float width1, float width2){
vec1.normalise();
vec2.normalise();
float cos = vec1.dotProduct(vec2); //Calculates actually cosini of two (normalised) vectors (remember math lessons)
float csc = 1.0f / Math::sqrt(1.0f-cos*cos); //Calculates cosecant of angle, This return NaN if angle is 180!!!
//And rest of the magic
Vector2 difrence = (vec1 * csc * width2) + (vec2 * csc * width1);
//If you use just convex polygons (all angles < 180, = 180 not allowed in this case) just return value, and if not you need some more magic.
//Both of next things need ordered vertex lists!
//Output vector is always to in side of angle, so if this angle is.
if (Math::vectorAngle(vec1, vec2) > 180.0f) //Note that this kind of function can know is your function can know that angle is over 180 ONLY if you use ordered vertexes (all vertexes goes always (anti???) cloclwise!)
difrence = -difrence;
//Ok and if angle was 180...
//Note that this can fix your situation ONLY if you use ordered vertexes (all vertexes goes always (anti???) cloclwise!)
if (difrence.isNaN()){
float width = (width1 + width2) / 2.0; //If angle is 180 and border widths are difrent, you cannot get perfect answer ;)
difrence = vec1 * width;
//Just turn vector -90 degrees
float swapHelp = difrence.y
difrence.y = -difrence.x;
difrence.x = swapHelp;
}
//If you don't want output to be inside of old polygon but outside, just: "return -difrence;"
return difrence;
}
//Use this =)
Vector2 calculateBorderPoint(Vector2 A, Vector2 B, Vector2 C, float widthA, float widthB){
return B + calculateBorderPoint(A-B, C-B, widthA, widthB);
}
Your second approach can be possible...
mark the outer vertex (in border) with 1 and the inner vertex (inside) with 0.
in the pixel shader you can choose to highlight, those that its value is greater than 0.9f or 0.8f.
it should work.