Replies: 5 comments 3 replies
-
|
As you pointed out, the union is returning The input are not invalid so there is nothing to "fix". I don't understand why you are using the experimental Or even the one operating on triangle soups: |
Beta Was this translation helpful? Give feedback.
-
|
You have no idea how much I've been waiting for your answer! Thank you so much for it! In addition, I want to ask how it is possible to combine grids that do not intersect, but stand close to each other? I tried to combine the grids, and if the operation failed, then decompose the second of the two grids into convex components, expand each of them a little (something like the Minkowski sum) and try to connect each decomposed and enlarged part with the first grid, but this created more problems than benefits (which is probably not surprising) So, if you have a lot of figures that stand close to each other, what would be an effective way to combine them? |
Beta Was this translation helpful? Give feedback.
-
|
What do you mean with |
Beta Was this translation helpful? Give feedback.
-
|
Oh, I'm sorry, I don't speak English well and I communicate here through a translator, he mistakenly translated Mesh into Grid, I meant Mesh typedef CGAL::Exact_predicates_inexact_constructions_kernel K;
typedef CGAL::Polyhedron_3<K, CGAL::Polyhedron_items_with_id_3> Mesh;In general, at the moment I am trying to transform Mesh into Nef typedef CGAL::Exact_predicates_exact_constructions_kernel K_e;
typedef CGAL::Nef_polyhedron_3<K_e, CGAL::SNC_indexed_items> Nef;and perform a Boolean operation with it, and then get Mesh from Nef, but so far too often Nef does not fulfill the condition after a Boolean operation .is_simple() Nef pre = mass.difference(constr).regularization();
if (pre.is_simple()) {
}I will try to send here tomorrow a minimally feasible example with the data that I use Thanks for the help! |
Beta Was this translation helpful? Give feedback.
-
|
I get
|
Beta Was this translation helpful? Give feedback.
-
|
The problem was solved by using union rather than difference, as well as using bounding_box for some complex Meshes |
Beta Was this translation helpful? Give feedback.
-
|
Although we previously managed to find a workaround for this problem by increasing one of the Mesh in all directions, I think it is not optimal, so I think we should continue the discussion here on what to do when PMP::corefine_and_compute_union()it returns false
#include <CGAL/Polygon_mesh_processing/polygon_soup_to_polygon_mesh.h>
#include <CGAL/IO/read_off_points.h>
#include <CGAL/Polyhedron_3.h>
#include <CGAL/Surface_mesh.h>
#include <CGAL/Polygon_mesh_processing/triangulate_faces.h>
#include <CGAL/Polygon_mesh_processing/self_intersections.h>
#include <CGAL/Polygon_mesh_processing/orientation.h>
#include <CGAL/Polyhedron_items_with_id_3.h>
#include <CGAL/Nef_polyhedron_3.h>
#include <CGAL/Nef_3/SNC_indexed_items.h>
typedef CGAL::Exact_predicates_exact_constructions_kernel K;
typedef K::Point_3 P;
typedef K::Vector_3 V;
typedef std::vector<std::size_t> Face;
typedef CGAL::Polyhedron_3<K> Mesh;
typedef CGAL::Surface_mesh<P> Surface;
typedef Surface::Halfedge_index Halfedge_index;
typedef Surface::Vertex_index Vertex_index;
typedef CGAL::Nef_polyhedron_3<K, CGAL::SNC_indexed_items> Nef;
namespace PMP = CGAL::Polygon_mesh_processing;
struct VectorHash {
template <class T>
std::size_t operator()(const T& v) const {
// Èñïîëüçóåì õýø-ôóíêöèþ äëÿ x, y è z
std::size_t seed = 0;
seed ^= std::hash<double>()(CGAL::to_double(v.x())) 0x9e3779b9 (seed << 6) (seed >> 2);
seed ^= std::hash<double>()(CGAL::to_double(v.y())) 0x9e3779b9 (seed << 6) (seed >> 2);
seed ^= std::hash<double>()(CGAL::to_double(v.z())) 0x9e3779b9 (seed << 6) (seed >> 2);
return seed;
}
};
struct VectorEqual {
bool operator()(const V& lhs, const V& rhs) const {
// Îïðåäåëèòå ñðàâíåíèå äëÿ íîðìàëåé
return lhs == rhs; // èëè èñïîëüçóéòå ñðàâíåíèå ñ îïðåäåëåííîé òî÷íîñòüþ
}
};
struct PointEqual {
bool operator()(const P& lhs, const P& rhs) const {
// Îïðåäåëèòå ñðàâíåíèå äëÿ íîðìàëåé
return lhs == rhs; // èëè èñïîëüçóéòå ñðàâíåíèå ñ îïðåäåëåííîé òî÷íîñòüþ
}
};
inline bool load_from(Mesh& output, const std::string& path) {
output.clear();
std::ifstream input;
input.open(path);
if (!input) {
return false;
}
else{
input >> output;
}
input.close();
if (!CGAL::is_triangle_mesh(output)) {
if (!PMP::triangulate_faces(output)) {
std::cout << "no_triangulate\n";
}
}
if (!output.is_closed()) {
std::cout << "no_closed\n";
}
if (!PMP::does_bound_a_volume(output)) {
std::cout << "no_bound\n";
}
if (PMP::does_self_intersect(output)) {
std::cout << "self_inter\n";
}
return CGAL::is_valid_polygon_mesh(output);
}
inline bool save_to(const Mesh& input, const std::string path) {
if (!CGAL::is_valid_polygon_mesh(input)) {
return false;
}
if (CGAL::IO::write_polygon_mesh(path, input, CGAL::parameters::stream_precision(17))) {
return true;
}
return false;
}Archive with original off-files needed to union off-file of Polyhedron_3 where need remove degenerated faces inline void union_meshes(
const std::string& path_meshes,
const std::string& path_mesh_output) {
std::vector<Mesh> meshes;
if (!load_dir(path_meshes, meshes)) {
return;
}
Mesh first = meshes[0];
for (size_t i = 1; i < meshes.size(); i ) {
Mesh result;
if (PMP::corefine_and_compute_union(first, meshes[i], result)) {
first = result;
}
else {
// problem here
}
std::cout << i 1 << '/' << meshes.size() << std::endl;
}
remesh_planar(first, first);
save_to(first, path_mesh_output);
}inline bool load_dir(const std::string& directory, std::vector<Mesh>& meshes) {
for (const auto& entry : std::filesystem::directory_iterator(directory)) {
if (entry.is_regular_file() && entry.path().extension() == ".off") {
Mesh loaded;
if (load_from(loaded, entry.path().string())) {
meshes.push_back(loaded);
}
}
}
return meshes.size() > 0;
} |
Beta Was this translation helpful? Give feedback.
-
|
I get a lot of similar cases (44 union-operations out of 1718 fail) As you can see from the screenshots below, the problem lies in the edges that lie on top of each other: Based on this, the question arises - is it technically possible for Polyhedron_3 to combine such meshes, or is expansion in all directions generally the only option to combine them? Thanks @sloriot for the help! |
Beta Was this translation helpful? Give feedback.
-
After an unsuccessful attempt to perform a Boolean union operation via C code, I tried to perform it via CGAL_PMP.exe , but apparently I get similar results there
[18:45:06] WARNING: The result of the requested operation is not manifold and has not been computed.
Source Code
I'm trying to use this:
But it for this two meshes - 59-one and 59-two.zip
<59-one.off - mesh1>
<59-two.off - mesh2>
always returns false, because inside
Previously, I have already checked mesh1 and mesh2 in the special FixMesh.h method
Environment
Can I correct the input data using some additional method to get the result of the union?
Thanks!
Updated (07.08)
At the moment, I managed to get rid of the error of performing a Boolean operation on a Mesh only by applying this method with all its methods inside for both meshes as well as for the result. This is very strong (especially the PMP::experimental::autorefine(mesh) method) performance was hit.
Is there a better way (equally reliable, but less costly) to correct geometry before and after a Boolean operation?
Beta Was this translation helpful? Give feedback.
All reactions