Surface modeling systems and methods
A method of generating a surface model of a physical environment includes obtaining an image of the physical environment. The method also includes generating a planar polygon mesh from at least the image. The method further includes extracting a boundary polygon of the planar polygon mesh. Moreover, the method includes generating a convex hull for the boundary polygon of the surface mesh. In addition, the method includes generating a minimal area oriented boundary polygon from the convex hull. The method may also include generating a maximal area oriented internal polygon inside of the boundary polygon of the planar polygon mesh.
1. A computer-implemented method of generating a surface model of a physical environment, the method comprising:
a computer processor generating a planar polygon mesh from at least an image of the physical environment;
the computer processor extracting a boundary polygon from the planar polygon mesh;
the computer processor generating a convex hull for the boundary polygon; and
the computer processor generating a minimal area oriented boundary polygon (MAOBP) from the convex hull, the minimal area oriented boundary polygon (MAOBP) comprising a polygon that contains the convex hull and the boundary polygon.
2. The method of claim 1 , further comprising:
the computer processor obtaining the image of the physical environment by:
the computer processor obtaining a 3-D point cloud corresponding to the physical environment using an imaging device, and
the computer processor obtaining pose information for the imaging device; and
the computer processor computing a truncated signed distance function for the 3-D point cloud using the pose information, wherein generating the planar polygon mesh comprises tessellating the truncated signed distance function.
3. The method of claim 2 , further comprising the computer processor combining two smaller planar polygon meshes into one larger planar polygon mesh.
4. The method of claim 1 , further comprising the computer processor obtaining a gravity vector, wherein the generated planar polygon mesh is one of substantially parallel and orthogonal to the gravity vector.
5. The method of claim 1 , wherein generating the convex hull comprises the computer processor using a Graham-Scan algorithm.
6. The method of claim 1 , wherein generating the minimal area oriented boundary polygon (MAOBP) comprises the computer processor using a rotating calipers algorithm.
7. The method of claim 1 , further comprising the computer processor generating a maximal area oriented internal polygon (MAOIP) that is entirely contained inside of the boundary polygon.
8. The method of claim 7 , wherein generating the maximal area oriented internal polygon (MAOIP) comprises the computer processor performing a search in a search area defined by the boundary polygon, and
wherein generating the maximal area oriented internal polygon (MAOIP) comprises forming a grid in the search area,
the method further comprising adjusting a resolution of the grid based on a size of the search area.
9. The method of claim 7 , further comprising:
the computer processor receiving a selection of a point inside of the boundary polygon of the planar polygon mesh wherein generating the maximal area oriented internal polygon (MAOIP) comprises performing a search in a search area defined using the selected point and forming a grid in the search are; and
the computer processor adjusting a resolution of the grid based on a size of the search area,
wherein generating the maximal area oriented internal polygon (MAOIP) comprises the computer processor forming a grid in the search area.
10. The method of claim 7 , wherein the minimal area oriented boundary polygon (MAOBP) and the maximal area oriented internal polygon (MAOIP) have a same shape.
11. The method of claim 1 , wherein the planar polygon mesh is generated by the computer processor based on a marching cubes algorithm.
12. The method of claim 1 , wherein the minimal area oriented boundary polygon (MAOBP) is at least one of a rectangle, a triangle, and a circle.
13. The method of claim 1 , further comprising the computer processor determining a fit between the minimal area oriented boundary polygon (MAOBP) and the boundary polygon,
wherein determining the fit comprises the computer processor calculating a difference between a first area of the minimal area oriented boundary polygon (MAOBP) and a second area of the boundary polygon.
14. The method of claim 1 , further comprising the computer processor storing data representing the minimal area oriented boundary polygon (MAOBP),
wherein the minimal area oriented boundary polygon (MAOBP) is a rectangle,
wherein the data comprises four sets of coordinates corresponding to the rectangle, and
wherein each of the four sets of coordinates is a pair of coordinates.
15. The method of claim 14 , wherein the data further comprises a length of the rectangle, a width of the rectangle, and a center of the rectangle.
16. The method of claim 1 , wherein generating the planar polygon mesh comprises the computer processor capturing static portions of a series of images of the physical environment.
17. The method of claim 1 , wherein the minimal area oriented boundary polygon (MAOBP) comprises a smallest polygon that contains the convex hull and the boundary polygon, and
wherein the minimal area oriented boundary polygon (MAOBP) comprises a smallest rectangle that contains the convex hull and the boundary polygon.
18. The method of claim 17 , further comprising the computer processor generating a maximal area oriented internal polygon (MAOIP) that is contained inside of the boundary polygon.
19. The method of claim 18 , wherein the minimal area oriented boundary polygon (MAOBP) overestimates an extent of a surface being modeled, and the maximal area oriented internal polygon (MAOIP) underestimates the extent of the surface being modeled.