问题
I'm using Open CV and I'm able to extract multidimensional data from and image with my codes but what would be an easy way to display it? I have a 3D array:
line start [x] [y] [z]
line end [x] [y] [z]
Anything in between the lines will be filled with a solid flat surface.
Is there some kind of function I can use to easily display a 3d image given these points or whatever other points I'm missing?
Such as this pseudo code:
display_3d(x,y,z,angle)
Any help is highly appreciated. Also I'm using C++.
回答1:
I am not sure if i understand correctly but why do you have 3D array ???
array of 3D points representing surface is usually 1D array of 3D points:
float points[n][3]; // not float points[nx][ny][nz] !!!
I am not CV user so I will stick to (OpenGL) rendering, but you can use any other render instead.
1.if the points represents concave flat surface
- on single plane!!!
- and are in the order (not mixed up)
then you can use GL_TRIANGLE_FAN or anything similar in other renders.
float points[n]={ {x,y,z},{x,y,z},...}; glBegin(GL_TRIANGLE_FAN); glVertex3fv(points[0]); glVertex3fv(points[1]); glVertex3fv(points[2]); ... glVertex3fv(points[n-1]); glEnd();
2.if not then you have to triangulate your data
- it is not an easy task but there are some libs out there that do it.
- not sure if OpenCV can do that too, look for triangulation functions (ears drop,etc)
- I prefer my own triangulation routines (better for debugging and also for additional polygon info) but a little hard to code it efficiently and correctly when you not used to used algorithms.
回答2:
First you need to find the convex hull of your points which means finding non-ambiguous and efficient representation of a set of points. Once you do that you will need to triangulate your mesh, usually using delaunay triangulation. QHull provides such functionality.
来源：https://stackoverflow.com/questions/15445927/how-do-i-take-a-3d-dimensional-array-and-convert-it-into-a-model