There are finite number of points on a 2-D plane each on integer coordinates (x,y) such that 0<=x,y<100.
Now what could be done to find out all the boundary points of this set? (Algorithm)
Visualisation of boundary points:
Imagine a rubber band in your hand.
Imagine pins sticked at all the points on the plane.
Now if you release the rubber band such that all the points/pins are inside it, after releasing it will contract.
After contraction, the points that deviate the rubber band(make perfect corners) will be the boundary points.
Image Visualisation: Image (Posting link to image coz i do not have enough reputation points)
Consider the yellow line. The image is not perfect visualisation. There's a point at ~(0.8,1) which should lie inside the boundary.
Related
So, I am using ARKit to display feature points in the session. I am able to get the current frame, then its rawFeaturePoints and place geometries in the world space so the user can see them on screen. That is working great.
In the app I then have a quadrant on screen. My objective is to show in screen coordinates feature points that projected would fall inside the 2D quadrant on screen. To do that, I tried this:
get feature points as an array of vector_float3
for each of those points I then get a SCNVector3 setting the Z component to 0 (near plane)
I then call on the ARSCNView:
public func projectPoint(_ point: SCNVector3) -> SCNVector3
This approach does give me 2D points back, but, depending on where the camera is they seem to be way off.
So then, since in ARKit the camera keeps moving around, do I need to take that into account to achieve what I explained?
EDIT:
About flipping the Y of the CGPoint retrieved from the projectPoint call on the camera:
/**
Project a 3D point in world coordinate system into 2D viewport space.
#param point 3D point in world coordinate system.
#param orientation Viewport orientation.
#param viewportSize Viewport (or image) size.
#return 2D point in viewport coordinate system with origin at top-left.
*/
open func projectPoint(_ point: vector_float3, orientation: UIInterfaceOrientation, viewportSize: CGSize) -> CGPoint
Remy San mentioned flipping the Y. I tried that and it does seem to work. One difference between what he's doing and what I am doing is that I am not using an SKScene, but I am using SCNScene. Looking at the docs it says:
...The projection of the specified point into a 2D pixel coordinate space
whose origin is in the upper left corner...
So, what throws me off is that if I don't flip the Y it seems like it's not really working properly. (I'll try to post images to show what I mean). But then if flipping the Y though makes things look better, it goes against the docs. No?
I get you are using the intrinsics matrix for you projection. ARkit technology may also give you some extra information. These are the cameraPoseARFrame, the projectionMatrix and the transformToWorldMap matrices. Are you taking them into consideration when transforming from world coordinates to pixel coordinates?
If anyone has a methodology for applying these matrices to the point cloud coordinates to convert them into screen coordinates, could you contribute to my answer please? I think they may provide more precision and accuracy to the final result.
Thank you!
Suppose we have an image (pixel buffer) that is in black and white, so each pixel is either black or white (not gray scale).
Now somewhere in the middle of that images, place a green dot. It may have a radius of n for rendering purposed, but it is really a just point. Give the dot a randomly selected direction and speed, and start it moving. If the image is all white pixels, the dot will bounce off the edges of the image, infinitely wandering around the picture. This is quite easy... just reverse either the rise or run of the dot's vector.
Next, suppose the image has some globs of black pixels. As the dot encounters these globs of black pixels, the angle of reflection needs to be calculated. This is also quite easy of the the black pixels have a fixed slope, as in my sketch (blue X represents black pixels). You can find the slope of the blue Xs and easily calculate the new vector.
But how about the case where the black pixels form really unfriendly surfaces? What are some approaches to figuring out this angle?
This is the subject that I am interested in.
There must be some algorithms that exist for this kind of purpose, but I never ran across any in school. I am not asking how to code this, rather approaches to writing the algorithm to do this. I have a few ideas that I'll try, but if there are some standard ways to do this that exist, I'd like to learn about them.
Obviously I'd like to start with Black and White then move into RGBA.
I am looking for any reference material on just this sort of subject. Websites, books, or other references are very very welcome.
Also, if there are different StackOverflow tags that could be good, let me know.
Thanks much!
Edit********** More pics and information
Maybe I wasn't clear what I meant by "unfriendly surfaces". In the previous picture, our blue X's happened to just be a line. Picture a case where it is not a line, rather a wierd shape.
We start with our green pixel traveling at a slope of 2. Suppose it's vector is that of 12 pixels per frame. It would have a projected path like this:
But suppose instead of a nice friendly line, we have this:
In my mind I can kinda of see what is likely to happen if this were a ball and some walls.
Look for edge detection algorithms used in image processing. Some edge detectors also approximate the direction of edges.
You can think of the pixel neighborhood of the green dot, maybe somewhere between 3x3 and 7x7, as a small edge direction detection problem. One approach would take two passes at the pixels:
In the first pass, smooth the sharp black/white pixels using a Gaussian filter.
In the second pass, apply an edge detection operator, such as Sobel, Prewitt or Roberts to produce the X and Y derivatives of the pixels' intensity. You can then approximate the direction as:
angle = arctan(dx/dy)
The motivation for the smoothing pass is to give the edge detection operator information from farther-away pixels.
The Wikipedia page on the Canny edge detector has a good discussion on obtaining the direction (the "gradient") of an edge, including an example of a particular Gaussian filter you can use for smoothing.
Am doing something similar with a ball and randomly generated backgrounds.
The filter and edge detection is highly technical but all other processes using a 5*5 or 3*3 grid seem similarly difficult.
However, I think I may have a cheap way around this. Assuming a ball travelling in any direction, scan all leading edges of the ball - a semicircle. The further to the edge of the ball the collision occurs the closer to vertical is the collision. Again, I think, this should allow you to easily infer the background normal and from there the answer is fairly simple
I am using WPF 3D, but I think this question should apply to any 3d texture mapping.
Suppose I have a model of a cow, and I want to draw a circular spot on the cow (and I want to do this dynamically -- supposed I don't know the location of the spot until run-time). I could do this by coloring the vertexes (vertexes are assigned a color based on their distance from the center of the spot), but if the model is fairly low-poly, that will give a pretty jagged-edged circle.
I could do it using a pixel shader, where the shader colors each pixel based on its distance from the center of the spot. But suppose I don't have access to pixel shaders (since I don't in WPF).
So, it seems that what I want to do is dynamically create a texture with the circle pattern on it, and texture the cow with it.
The question is: As I'm drawing that texture, how can I know what 3d coordinate in model space a given xy coordinate on the texture image corresponds to?
That is, suppose I have already textured my model with a plain white texture -- I've set up texture coordinates, done texture mapping, but don't have the texture image yet. So I have this 1000x1000 (or whatever) pixel image that gets draped nicely over the cow according to some nice texture coordinates that have been set up on the model beforehand. I understand that when the 3D hardware goes to draw a given triangle, it uses the texture coordinates of the vertexes of the triangle to find the corresponding triangular region of the image, and then interpolates across the surface of the triangle to fill displayed model pixels with colors from that triangular region of the image.
How do I go the other way? How do I say, for this given xy point on my texture image, and given the texture coordinates that have already been set up on the model, what's the 3d coordinate in model space that this image pixel is going to correspond to once texture mapping happens?
If I had such a function, I could color my texture map image such that all the points (in 3d space) within a certain distance of the circle center point on the cow would get one color, and all points outside that distance would get another color, and I'd end up with a nice, crisp circular spot on the cow, even with a relatively low-poly model. Does that sound right?
I do understand that given the texture coordinates for the vertexes of each triangle, I can step through the triangles in my model, find the corresponding triangle on the texture image, and do my own interpolation, across the texture pixels in that triangle, by interpolating across the 3d plane determined by the vertex points. And that doesn't sound too hard. But I'm just trying to understand if there is some standard 3d concept/function where I can just call a ready-made function to give me the model space coordinates for a given texture xy.
I did end up getting this working. I walk every point on the texture (1024 x 1024 points). Using the model's texture coordinates, I determine which polygon face, if any, the given u,v point is inside of. If it's inside of a face, I get the model coordinates for each point on that face. I then do a barycentric interpolation as described here: http://paulbourke.net/texture_colour/interpolation/
That is, for each u,v point on the texture, I use an inside-polygon check to determine which quad it's in on the 2D texture sheet and then I use an interpolation on that same 2D geometry as described in the link above, but instead of interpolating colors or normals I'm interpolating 3D coordinates.
I can then use the 3D coordinate to color the point on the texture (e.g., to color a circular spot on the cow based on how far in model space the given texture point is from the spot center point). And then I can apply the texture to the model, and it works.
Again, it seems like this must be a standard procedure with a name...
One issue is that the result is very sensitive to the quality of the the texturing as set up by the modeler. For instance, if a relatively large quad on the cow corresponds to a small quad on the texture image, there just aren't enough pixels to work with to get a smooth curve within that model quad once the texture is applied. You can of course use a higher-res texture, such as 2048x2048, but then your loop time is 4x.
It's actually a rasterization process if I didn't misunderstand your question. In lightmapping, one may also need to find the corresponding positions and normals in world space for each texel in the lightmap space and then baking irradiance. (which seems similar to your goal)
You can use standard Graphics API to do this task instead of writing your own implementation. Let:
Size of texture -> Size of G-buffers
UVs of each mesh triangle -> Vertex positions vec3(u, v, 0) of the input stage
Indices of each mesh triangle -> Indices of the input stage
Positions (and normals, etc.) of each mesh triangle -> Attributes of the input stage
After the rasterizer stage of the graphics pipeline, all fragments that lie within the UV triangle are generated, and the attributes that have been supplied are interpolated automatically. You can do whatever you want now in pixel shader!
I have read lots of ray tracer algorithm on the web. But, I have no clear understanding of the shading and shadow. Is below pseudocode correct written according to my understanding ?
for each primitive
check for intersection
if there is one
do color be half of the background color
Ishadow = true
break
for each ambient light in environment
calculate light contribution to the color
if ( Ishadow == false )
for each point light
calculate diffuse shading
calculate reflection direction
calculate specular light
trace for reflection ray // (i)
add color returned from i after multiplied by some coefficient
trace for refraction ray // (ii)
add color returned from ii after multiplied by some coefficient
return color value calculated until this point
You should integrate your shadows with the normal ray-tracing path:
For every screen-pixel you send a ray through the scene and you eventually determine the closest object-intersection: at the point of the closest object-intersection you would at first read out the pixel color (texture of the object at that point), aside from calculating reflection-vector etc (using the normal-vector) you would now additionally cast a ray from that intersection-point to each of the light-sources in your scene: if these rays intersect other objects before hitting the light-sources then the intersection-point is in shadow and you can adapt the final color of that point accordingly.
The trouble with pseudocode is that it is easy to get "pseudo" enough that it becomes the same well of ambiguity that we are trying to avoid by getting away from natural languages. "Color be half of the background color?" The fact that this line appears before you iterate through your light sources is confusing. How can you be setting Ishadow before you iterate over light sources?
Maybe a better description would be:
given a ray in space
find nearest object with which ray intersects
for each point light
if normal at surface of intersected object points toward light (use dot product for this)
cast a ray into space from the surface toward the light
if ray intersection is closer than light* light is shadowed at this point
*If you're seeing strange artifacts in your shadows, there is a mistake that is made by every single programmer when they write their first ray tracer. Floating point (or double-precision) math is imprecise and you will frequently (about half the time) re-intersect yourself when doing a shadow trace. The explanation is a bit hard to describe without diagrams, but let me see what I can do.
If you have an intersection point on the surface of a sphere, under most circumstances, that point's representation in a floating point register is not mathematically exact. It is either slightly inside or slightly outside the sphere. If it is inside the sphere and you try to run an intersection test to a light source, the nearest intersection will be the sphere itself. The intersection distance will be very small, so you can simply reject any shadow ray intersection that is closer than, say .000001 units. If your geometry is all convex and incapable of legitimately shadowing itself, then you can simply skip testing the sphere when doing shadow tests.
I would like to calculate distance to certain objects in the scene, I know that I can only calculate relative distance when using a single camera but I know the coordinates of some objects in the scene so in theory it should be possible to calculate actual distance. According to the opencv mailing list archives,
http://tech.groups.yahoo.com/group/OpenCV/message/73541
cvFindExtrinsicCameraParams2 is the function to use, but I can't find information on how to use it?
PS. Assuming camera is properly calibrated.
My guess would be, you know the width of an object, such as a ball is 6 inches across and 6 inches tall, you can also see that it is 20 pixels tall and 25 pixels wide. You also know the ball is 10 feet away. This would be your start.
Extrinsic parameters wouldn't help you, I don't think, because that is the camera's location and rotation in space relative to another camera or an origin. For a one camera system, the camera is the origin.
Intrinsic parameters might help. I'm not sure, I've only done it using two cameras.