I'm making a procedurally generated minecraft-like voxel terrain in Unity. Mesh generation and albedo channel texturing is flawless; however I need to apply different normal map textures for different cube faces regarding whether they're neighboring to another cube or not. Materials accepts only single normal map file and doesn't provide a sprite-sheet-editor kind of functionality for normal maps. So I have no idea about how to use selected slices out of normal map file as if they were albedo textures. I couldn't find any related resources about the problem. Any help will be appreciated. Thanks...
First of all, I'm not an expert in this area, though I am going to try to help you based on my limited and incomplete understanding of parts of Unity.
If there are a finite number of "normal face maps" that can exist, I suggest something like you indicated ("sprite sheet") and create a single texture (also sometimes called a texture atlas) that contains all these normal maps.
The next step, which I'm not sure whether the Standard material shader will be to handle for your situation is to generate UV/texture coordinates for the normal map and pass those along with your vertex xyz positions to the shader. The UV coordinates need to be for each vertex of each face; they are specified as a 2-D (U, V) offset into your atlas of normal maps and are floating point values with a range of [0.0, 1.0], that map to the full X and Y coordinates of the actual normal texture. For instance, if you had an atlas with a grid of textures in 4 rows and 4 columns, a face that should use the top-left texture would have UV coords of [(0,0), (0.25,0), (0.25,0.25), (0, 0.25)].
The difficulty here may depend if you are you using UV coordinates for doing other texture mapping (e.g. in the Albedo or wherever else). If this is the case, I think the Unity Standard Shader permits two sets of texture coordinates, and if you need more, you might have to roll your own shader or find a Shader asset elsewhere that allows for more UV sets. This is where my understanding of gets shaky, as I'm not exactly sure how the shader uses these two UV coordinate sets, and whether there is some existing convention for how these UV coordinate are used, as the standard shader supports secondary/detail maps, which may mean you have to share the UV0 set with all non-detail maps, so albedo, normal, height, occlusion, etc.
Related
I have a triangle mesh along with a function which defines the material properties at each point in 3d space. Using a given resolution in object space, I generate a triangular texture for each triangle; specifically, these each end up being right triangles with size corresponding to the actual triangle in the mesh. I have two problems, however: 1) the output texture atlas is large and obviously contains large amounts of dead space, and 2) each vertex for each triangle needs to have its own texcoord, since each triangle's texture ends up in a different part of the atlas.
What algorithms exist to generate a texture atlas from a mesh with known textures for each triangle? I'm looking to share as many texcoords as possible across the mesh, which means that adjacent triangles should have correspondingly adjacent textures in the atlas. Not everything can be shared -- since 3d objects can't always flatten into a 2d surface with constant texture resolution -- but I'm hoping to maximize this.
I would like to create a GPS drawing program in Argon and A-Frame which draws lines based upon people's movements.
Lines can be drawn in A-Frame with, for example, the meshline component which uses Cartesian points:
<a-entity meshline="lineWidth: 20; path: -2 -1 0, 0 -2 0</a-entity>
If I were to do this with a GPS device, I would take the GPS coordinates and map them directly to something like Google maps. Does Argon have any similar functionality such that I can use the GPS coordinates directly as the path like so:
<a-entity meshline="lineWidth: 20; path: 37.32299 -122.04185 0, 37.32298 -122.03224</a-entity>
Since one can specify an LLA point for a reference frame I suppose one way to do this would be to conceive of the center LLA point as "0, 0, 0" and then use a function to map the LLA domain to a Cartesian range.
It would be preferable, however, to use the geo-coordinates directly. Is this possible in Argon?
To understand the answer, you need to first understand the various frames of reference used by Argon.
First, Argon makes use of cesiumjs.org's geospatial math libraries and Entity's so that all "locations" in Argon must either be expressed geospatially OR be relative to a geospatial entity. These are rooted at the center of the earth, in what Cesium calls FIXED coordinates, but are also know as ECEF or ECF coordinates. In that system, coordinates are in meters, with up/down going through the poles, east/west going through the meridian (I believe). Any point on the surface of the earth is represented with pretty large numbers.
This coordinate system is nice because we can represent anything on or near the earth precisely using it. Cesium also supports INERTIAL coordinates, which are used to represent near-earth orbital objects, and can convert between the two frames.
But, it is inconvenient when doing AR for a few reasons:
the numbers used to represent the position of the viewer and objects near them are quite large, even if they are very close, which can lead to mathematical accuracy issues, especially in the 3D graphics system.
The coordinates we "think about" when we think about the world around us have the ground as "flat" and "up" as pointing ... well, up. So, in 3D graphics, an object above another object typically has the same X and Z values, but has a Y that's bigger. In ECEF coordinates, all the numbers change because what we perceive as "up" is really a vector from the center of the earth though us, and is only "up" if we're on the north (or south, depending on your +/-) pole. Most 3D graphics libraries you might want to use (e.g., physics libraries, for example), assume a world in which the ground is one plane (typically the XZ plane) and Y is up (some aeronautics and other engineering applications use Z as up and have XY as the ground, but the issue is the same).
Argon deals with this, as do many geospatial AR systems, by creating a local coordinate system for the graphics and application to use. There are really three options for this:
Pick some arbitrary (but fixed) local place as the origin. Some systems, which are built to work in one place, have this hard-coded. Others let the application set it. We don't do this because it would encourage applications to take the easy path and only work in one place (we've seen this in the past).
Set the local place to the camera. This has the advantage that the math is the most "accurate" because all points are expressed relative to the camera. But, this causes two issues. First, the camera tends to move continuously (even if only due to sensor noise) in AR apps. Second, many libraries (again, like physics libraries) assume that the origin of the system is stable and on the earth, with the camera/user moving through it. These issues can be worked around, but they are tedious for application developers to deal with.
Set the origin of the local coordinates to an arbitrary location near the user, and if the user moves far from it, recenter automatically. The advantage of this is the program doesn't necessarily have to do much to deal with it, and it meshes nicely with 3D graphics libraries. The disadvantage is the local coordinates are arbitrary, and might be different each time a program is run. However, the application developer may have to pay attention to when the origin is recentered.
Argon uses open 3. When the app starts, we create a new local coordinate frame at the user's location, on the plane tangent to the earth. If the user moves far from that location we update the origin and emit an event to the application (currently, we recenter if you are 5km away from the origin). In many simple apps, with only a few frames or reference expressed in geospatial coordinates (and the rest of the application data expressed relative to known geospatial locations), the conversion from geospatial to local can just be done each frame, allowing the app developer to ignore the reentering problem. The programmer is free to use either ENU (east-north-up) or EUS (east-up-south) as their coordinate system; we tend to use EUS because it's similar to what most 3D graphics systems use (Y is up, Z points south, and X is east).
One of the reasons we chose this approach is that we've found in the past that if we had predictable local coordinates, application developers would store data using those coordinates even though that's not a good idea (you data is now tied to some relatively arbitrary application-specific coordinate system, and will now only work in that location).
So, now to your question. Your issue is that you want to use geospatial (cesium's coordinates, that argon uses) coordinates in AFrame. The short answer is you can't use them directly, since AFrame is built assuming a local 3D graphics coordinate system. The argon-aframe package binds aframe to argon by allowing you to specify referenceframe components that position an a-entity at an argon/cesium geospatial location, and take care of all the internal conversions for you.
The assumption when I wrote that code was that authors would then create their content using the local, 3D graphics coordinates, and attach those hunks of graphics to a-entity's that were located in the world with referenceframe's.
In order to have individual coordinates in AFrame correspond to geospatial places, you will need to manage that yourself, perhaps by creating a component to do it for you, or (if the data is known at the start) by converting it up front.
Here's what I'd do.
Assuming you have a list of geospatial coordinates (expressed as LLA), I'd convert each to a local coordinates (by first converting from LLA to Cesium's FIXED ECEF coordinates and creating a Cesium Entity, and then calling Argon's context.getEntityPose() on that entity (which will return it's local coordinates). I would pick one geospatial location in the set (perhaps the first one?) and then subtract it's local coordinates from each of them, so that they are all expressed in local coordinates relative to that known geospatial location.
Then, I'd create an AFrame entity attached to the referenceframe of that unique geospatial entity, and create your graphics content inside of it, using the local coordinates that are expressed relative to it. For example, let's say the geospatial location is LongLat = "-84.398881 33.778463" and you stored those points (local coordinates, relative to LongLat) in userPath, you could do something like this:
<ar-scene>
<ar-geopose id="GT" lla=" -84.398881 33.778463" userotation="false">
<a-entity meshline="lineWidth: 20; path: userPath; color: #E20049"></a-entity>
</ar-geopose>
</ar-scene>
I am working with disparity maps (1024 x 768) obtained via stereo and I am able to get point clouds with XYZRGB pcl::Points. However not all pixels from the disparity map are valid depth hence there will never be 1024x768 = 786432 XYZRGB points. Fortunately I am able to save the point clouds unorganized (i.e. height=1). Unfortunately, some normal estimation methods etc, are tailored for organized pointclouds. How can I create organised pointclouds from this ?
I believe that this is not possible.
First of all unorganized point cloud (PC) is just list of points in random order written in file
On the other hand organized PC carries information of in which order orginal points were obtained by depth camera and some other information. This information is stored in lets call it grid.
Once you destroy this grid omiting some points theres no algorithm that can put it back together as it originally was
You can use other methods which provides PCL that doesnt take OPC as an argument. Result will be same as if you would use organized point cloud only little bit slower (depends on size of your input cloud)
I assume that you do have the calibration parameters that are necessary to transform the image points and their depth into 3D points, right?
In this case, you simply create a 2D point cloud and do the following for each pixel of the disparity map:
If the point is valid:
set the corresponding point in the point cloud to the 3D point
else:
set the corresponding point in the cloud to NaN (i.e. a 3D point with NaN as coordinates)
I'm writing a software renderer which is currently working well, but I'm trying to get perspective correction of texture coordinates and that doesn't seem to be correct. I am using all the same matrix math as opengl for my renderer. To rasterise a triangle I do the following:
transform the vertices using the modelview and projection matrixes, and transform into clip coordinates.
for each pixel in each triangle, calculate barycentric coordinates to interpolate properties (color, texture coordinates, normals etc.)
to correct for perspective I use perspective correct interpolation:
(w is depth coordinate of vertex, c is texture coordinate of vertex, b is the barycentric weight of a vertex)
1/w = b0*(1/w0) + b1*(1/w1) + b2*(1/w2)
c/w = b0*(c0/w0) + b1*(c1/w1) + b2*(c2/w2)
c = (c/w)/(1/w)
This should correct for perspective, and it helps a little, but there is still an obvious perspective problem. Am I missing something here, perhaps some rounding issues (I'm using floats for all math)?
See in this image the error in the texture coordinates evident along the diagonal, this is the result having done the division by depth coordinates.
Also, this is usually done for texture coordinates... is it necessary for other properties (e.g. normals etc.) as well?
I cracked the code on this issue recently. You can use a homography if you plan on modifying the texture in memory prior to assigning it to the surface. That's computationally expensive and adds an additional dependency to your program. There's a nice hack that'll fix the problem for you.
OpenGL automatically applies perspective correction to the texture you are rendering. All you need to do is multiply your texture coordinates (UV - 0.0f-1.0f) by the Z component (world space depth of an XYZ position vector) of each corner of the plane and it'll "throw off" OpenGL's perspective correction.
I asked and solved this problem recently. Give this link a shot:
texture mapping a trapezoid with a square texture in OpenGL
The paper I read that fixed this issue is called, "Navigating Static Environments Using Image-Space Simplification and Morphing" - page 9 appendix A.
Hope this helps!
ct
The only correct transformation from UV coordinates to a 3D plane is an homographic transformation.
http://en.wikipedia.org/wiki/Homography
You must have it at some point in your computations.
To find it yourself, you can write the projection of any pixel of the texture (the same as for the vertex) and invert them to get texture coordinates from screen coordinates.
It will come in the form of an homographic transform.
Yeah, that looks like your traditional broken-perspective dent. Your algorithm looks right though, so I'm really not sure what could be wrong. I would check that you're actually using the newly calculated value later on when you render it? This really looks like you went to the trouble of calculating the perspective-correct value, and then used the basic non-corrected value for rendering.
You need to inform OpenGL that you need perspective correction on pixels with
glHint(GL_PERSPECTIVE_CORRECTION_HINT,GL_NICEST)
What you are observing is the typical distortion of linear texture mapping. On hardware that is not capable of per-pixel perspective correction (like for example the PS1) the standard solution is just subdividing in smaller polygons to make the defect less noticeable.
I am creating a 2D sprite game in Unity, which is a 3D game development environment.
I have constrained all translation of objects to the XY-plane and rotation to the Z-axis.
My problem is that the meshes that are used to detect collisions between objects must still be in 3D. I have the need to detect collisions between the player object (a capsule collider) and a sprite (that has its collision volume defined by a polygonal prism).
I am currently writing the level editor and I have the need to let the user define the collision area for any given tile. In the image below the user clicks the points P1, P2, P3, P4 in that order.
Obviously the points join up to form a quadrilateral. This is the collision area I want, however I must then convert that to a 3D mesh. Basically I need to generate an extrusion of the polygon, then assign the vertex winding and triangles etc. The vertex positions is not a problem to figure out as it is merely a translation of the polygon down the z-axis.
I am having trouble creating an algorithm for assigning the winding order of the vertices, especially since the mesh must consist only of triangles.
Obviously the structure I have illustrated is not important, the polygon may be any 2d shape and will always need to form a prism.
Does anyone know any methods for this?
Thank you all very much for your time.
A simple algorithm that comes to mind is something like this:
extrudedNormal = faceNormal.multiplyScale(sizeOfExtrusion);//multiply the face normal by the extrusion amt. = move along normal
for each(vertex in face){
vPrime = vertex.clone();//copy the position of each vertex to a new object to be modified later
vPrime.addSelf(extrudedNormal);//add translation in the direction of the normal, with the amt. used in the
}
So the idea is basic:
clone the face normal and move it in
the same direction by the amt. you
want to extrude by
clone the face vertices and move them
using the moved(extruded) normal
position
For a more complete, feature rich example, refer to the Procedural Modeling Unity samples. They include a nice Mesh extrusion sample too (see ExtrudedMeshTrail.js which uses MeshExtrusion.cs).
Goodluck!
To create the extruded walls:
For each vertex a (with coordinates ax, ay) in your polygon:
- call the next vertex 'b' (with coordinates bx, by)
- create the extruded rectangle corresponding to the line from 'a' to 'b':
- The rectangle has vertices (ax,ay,z0), (ax,ay,z1), (bx,by,z0), (bx,by,z1)
- This rectangle can be created from two triangles:
- (ax,ay,z0), (ax,ay,z1), (bx,by,z0) and (ax,ay,z1), (bx,by,z0), (bx,by,z1)
If you want to create a triangle strip instead, it's even simpler. For each vertex a, just add (ax,ay,z0) and (ax,ay,z1). Whichever vertex you processed first will also need to be processed again after looping over all other vertices.
To create the end-caps:
This step is probably unnecessary for collision purposes. But, one simple technique is here: http://www.siggraph.org/education/materials/HyperGraph/scanline/outprims/polygon1.htm
Each resulting triangle should be added at depth z0 and z1.