Geometry Nodes: improve Point Distribute node

This greatly simplifies the Point Distribute node. For a poisson disk
distribution, it now uses a simpler dart throwing variant. This results
in a slightly lower quality poisson disk distribution, but it still
fulfills our requirements: have a max density, minimum distance input
and stability while painting the density attribute.

This new implementation has a number of benefits over the old one:
* Much less and more readable code.
* Easier to extend with other distribution algorithms.
* Easier to transfer more attributes to the generated points later on.
* More predictable output when changing the max density and min distance.
* Works in 3d, so no projection on the xy plane is necessary.

This is related to T84640.

Differential Revision: https://developer.blender.org/D10104
This commit is contained in:
Jacques Lucke 2021-01-13 12:34:48 +01:00
parent ed1042ee06
commit d985751324
5 changed files with 181 additions and 471 deletions

View File

@ -436,7 +436,8 @@ static const EnumPropertyItem rna_node_geometry_point_distribute_method_items[]
"POISSON",
0,
"Poisson Disk",
"Project points on the surface evenly with a Poisson disk distribution"},
"Distribute the points randomly on the surface while taking a minimum distance between "
"points into account"},
{0, NULL, 0, NULL, NULL},
};
@ -6291,7 +6292,8 @@ static void rna_def_cmp_output_file_slot_file(BlenderRNA *brna)
prop = RNA_def_property(srna, "save_as_render", PROP_BOOLEAN, PROP_NONE);
RNA_def_property_boolean_sdna(prop, NULL, "save_as_render", 1);
RNA_def_property_ui_text(prop, "Save as Render", "Apply render part of display transform when saving byte image");
RNA_def_property_ui_text(
prop, "Save as Render", "Apply render part of display transform when saving byte image");
RNA_def_property_update(prop, NC_NODE | NA_EDITED, NULL);
prop = RNA_def_property(srna, "format", PROP_POINTER, PROP_NONE);

View File

@ -154,7 +154,6 @@ set(SRC
geometry/nodes/node_geo_join_geometry.cc
geometry/nodes/node_geo_object_info.cc
geometry/nodes/node_geo_point_distribute.cc
geometry/nodes/node_geo_point_distribute_poisson_disk.cc
geometry/nodes/node_geo_point_instance.cc
geometry/nodes/node_geo_point_separate.cc
geometry/nodes/node_geo_rotate_points.cc

View File

@ -46,11 +46,6 @@ void update_attribute_input_socket_availabilities(bNode &node,
CustomDataType attribute_data_type_highest_complexity(Span<CustomDataType>);
void poisson_disk_point_elimination(Vector<float3> const *input_points,
Vector<float3> *output_points,
float maximum_distance,
float3 boundbox);
Array<uint32_t> get_geometry_element_ids_as_uints(const GeometryComponent &component,
const AttributeDomain domain);

View File

@ -16,9 +16,11 @@
#include "BLI_float3.hh"
#include "BLI_hash.h"
#include "BLI_kdtree.h"
#include "BLI_math_vector.h"
#include "BLI_rand.hh"
#include "BLI_span.hh"
#include "BLI_timeit.hh"
#include "DNA_mesh_types.h"
#include "DNA_meshdata_types.h"
@ -34,7 +36,7 @@
static bNodeSocketTemplate geo_node_point_distribute_in[] = {
{SOCK_GEOMETRY, N_("Geometry")},
{SOCK_FLOAT, N_("Distance Min"), 0.1f, 0.0f, 0.0f, 0.0f, 0.0f, 100000.0f, PROP_NONE},
{SOCK_FLOAT, N_("Distance Min"), 0.0f, 0.0f, 0.0f, 0.0f, 0.0f, 100000.0f, PROP_NONE},
{SOCK_FLOAT, N_("Density Max"), 1.0f, 0.0f, 0.0f, 0.0f, 0.0f, 100000.0f, PROP_NONE},
{SOCK_STRING, N_("Density Attribute")},
{SOCK_INT, N_("Seed"), 0, 0, 0, 0, -10000, 10000},
@ -67,213 +69,196 @@ static float3 normal_to_euler_rotation(const float3 normal)
return rotation;
}
static Vector<float3> random_scatter_points_from_mesh(const Mesh *mesh,
const float density,
const FloatReadAttribute &density_factors,
Vector<float3> &r_normals,
Vector<int> &r_ids,
const int seed)
static Span<MLoopTri> get_mesh_looptris(const Mesh &mesh)
{
/* This only updates a cache and can be considered to be logically const. */
const MLoopTri *looptris = BKE_mesh_runtime_looptri_ensure(const_cast<Mesh *>(mesh));
const int looptris_len = BKE_mesh_runtime_looptri_len(mesh);
const MLoopTri *looptris = BKE_mesh_runtime_looptri_ensure(const_cast<Mesh *>(&mesh));
const int looptris_len = BKE_mesh_runtime_looptri_len(&mesh);
return {looptris, looptris_len};
}
Vector<float3> points;
static void sample_mesh_surface(const Mesh &mesh,
const float base_density,
const FloatReadAttribute *density_factors,
const int seed,
Vector<float3> &r_positions,
Vector<float3> &r_bary_coords,
Vector<int> &r_looptri_indices)
{
Span<MLoopTri> looptris = get_mesh_looptris(mesh);
for (const int looptri_index : IndexRange(looptris_len)) {
for (const int looptri_index : looptris.index_range()) {
const MLoopTri &looptri = looptris[looptri_index];
const int v0_index = mesh->mloop[looptri.tri[0]].v;
const int v1_index = mesh->mloop[looptri.tri[1]].v;
const int v2_index = mesh->mloop[looptri.tri[2]].v;
const float3 v0_pos = mesh->mvert[v0_index].co;
const float3 v1_pos = mesh->mvert[v1_index].co;
const float3 v2_pos = mesh->mvert[v2_index].co;
const float v0_density_factor = std::max(0.0f, density_factors[v0_index]);
const float v1_density_factor = std::max(0.0f, density_factors[v1_index]);
const float v2_density_factor = std::max(0.0f, density_factors[v2_index]);
const float looptri_density_factor = (v0_density_factor + v1_density_factor +
v2_density_factor) /
3.0f;
const int v0_index = mesh.mloop[looptri.tri[0]].v;
const int v1_index = mesh.mloop[looptri.tri[1]].v;
const int v2_index = mesh.mloop[looptri.tri[2]].v;
const float3 v0_pos = mesh.mvert[v0_index].co;
const float3 v1_pos = mesh.mvert[v1_index].co;
const float3 v2_pos = mesh.mvert[v2_index].co;
float looptri_density_factor = 1.0f;
if (density_factors != nullptr) {
const float v0_density_factor = std::max(0.0f, (*density_factors)[v0_index]);
const float v1_density_factor = std::max(0.0f, (*density_factors)[v1_index]);
const float v2_density_factor = std::max(0.0f, (*density_factors)[v2_index]);
looptri_density_factor = (v0_density_factor + v1_density_factor + v2_density_factor) / 3.0f;
}
const float area = area_tri_v3(v0_pos, v1_pos, v2_pos);
const int looptri_seed = BLI_hash_int(looptri_index + seed);
RandomNumberGenerator looptri_rng(looptri_seed);
const float points_amount_fl = area * density * looptri_density_factor;
const float points_amount_fl = area * base_density * looptri_density_factor;
const float add_point_probability = fractf(points_amount_fl);
const bool add_point = add_point_probability > looptri_rng.get_float();
const int point_amount = (int)points_amount_fl + (int)add_point;
for (int i = 0; i < point_amount; i++) {
const float3 bary_coords = looptri_rng.get_barycentric_coordinates();
const float3 bary_coord = looptri_rng.get_barycentric_coordinates();
float3 point_pos;
interp_v3_v3v3v3(point_pos, v0_pos, v1_pos, v2_pos, bary_coords);
points.append(point_pos);
/* Build a hash stable even when the mesh is deformed. */
r_ids.append(((int)(bary_coords.hash()) + looptri_index));
float3 tri_normal;
normal_tri_v3(tri_normal, v0_pos, v1_pos, v2_pos);
r_normals.append(tri_normal);
interp_v3_v3v3v3(point_pos, v0_pos, v1_pos, v2_pos, bary_coord);
r_positions.append(point_pos);
r_bary_coords.append(bary_coord);
r_looptri_indices.append(looptri_index);
}
}
return points;
}
struct RayCastAll_Data {
void *bvhdata;
BVHTree_RayCastCallback raycast_callback;
/** The original coordinate the result point was projected from. */
float2 raystart;
const Mesh *mesh;
float base_weight;
FloatReadAttribute *density_factors;
Vector<float3> *projected_points;
Vector<float3> *normals;
Vector<int> *stable_ids;
float cur_point_weight;
};
static void project_2d_bvh_callback(void *userdata,
int index,
const BVHTreeRay *ray,
BVHTreeRayHit *hit)
BLI_NOINLINE static KDTree_3d *build_kdtree(Span<float3> positions)
{
struct RayCastAll_Data *data = (RayCastAll_Data *)userdata;
data->raycast_callback(data->bvhdata, index, ray, hit);
if (hit->index != -1) {
/* This only updates a cache and can be considered to be logically const. */
const MLoopTri *looptris = BKE_mesh_runtime_looptri_ensure(const_cast<Mesh *>(data->mesh));
const MVert *mvert = data->mesh->mvert;
KDTree_3d *kdtree = BLI_kdtree_3d_new(positions.size());
for (const int i : positions.index_range()) {
BLI_kdtree_3d_insert(kdtree, i, positions[i]);
}
BLI_kdtree_3d_balance(kdtree);
return kdtree;
}
const MLoopTri &looptri = looptris[index];
const FloatReadAttribute &density_factors = data->density_factors[0];
BLI_NOINLINE static void update_elimination_mask_for_close_points(
Span<float3> positions, const float minimum_distance, MutableSpan<bool> elimination_mask)
{
if (minimum_distance <= 0.0f) {
return;
}
const int v0_index = data->mesh->mloop[looptri.tri[0]].v;
const int v1_index = data->mesh->mloop[looptri.tri[1]].v;
const int v2_index = data->mesh->mloop[looptri.tri[2]].v;
KDTree_3d *kdtree = build_kdtree(positions);
for (const int i : positions.index_range()) {
if (elimination_mask[i]) {
continue;
}
struct CallbackData {
int index;
MutableSpan<bool> elimination_mask;
} callback_data = {i, elimination_mask};
BLI_kdtree_3d_range_search_cb(
kdtree,
positions[i],
minimum_distance,
[](void *user_data, int index, const float *UNUSED(co), float UNUSED(dist_sq)) {
CallbackData &callback_data = *static_cast<CallbackData *>(user_data);
if (index != callback_data.index) {
callback_data.elimination_mask[index] = true;
}
return true;
},
&callback_data);
}
BLI_kdtree_3d_free(kdtree);
}
BLI_NOINLINE static void update_elimination_mask_based_on_density_factors(
const Mesh &mesh,
const FloatReadAttribute &density_factors,
Span<float3> bary_coords,
Span<int> looptri_indices,
MutableSpan<bool> elimination_mask)
{
Span<MLoopTri> looptris = get_mesh_looptris(mesh);
for (const int i : bary_coords.index_range()) {
if (elimination_mask[i]) {
continue;
}
const MLoopTri &looptri = looptris[looptri_indices[i]];
const float3 bary_coord = bary_coords[i];
const int v0_index = mesh.mloop[looptri.tri[0]].v;
const int v1_index = mesh.mloop[looptri.tri[1]].v;
const int v2_index = mesh.mloop[looptri.tri[2]].v;
const float v0_density_factor = std::max(0.0f, density_factors[v0_index]);
const float v1_density_factor = std::max(0.0f, density_factors[v1_index]);
const float v2_density_factor = std::max(0.0f, density_factors[v2_index]);
/* Calculate barycentric weights for hit point. */
float3 weights;
interp_weights_tri_v3(
weights, mvert[v0_index].co, mvert[v1_index].co, mvert[v2_index].co, hit->co);
const float probablity = v0_density_factor * bary_coord.x + v1_density_factor * bary_coord.y +
v2_density_factor * bary_coord.z;
float point_weight = weights[0] * v0_density_factor + weights[1] * v1_density_factor +
weights[2] * v2_density_factor;
point_weight *= data->base_weight;
if (point_weight >= FLT_EPSILON && data->cur_point_weight <= point_weight) {
data->projected_points->append(hit->co);
/* Build a hash stable even when the mesh is deformed. */
data->stable_ids->append((int)data->raystart.hash());
data->normals->append(hit->no);
const float hash = BLI_hash_int_01(bary_coord.hash());
if (hash > probablity) {
elimination_mask[i] = true;
}
}
}
static Vector<float3> poisson_scatter_points_from_mesh(const Mesh *mesh,
const float density,
const float minimum_distance,
const FloatReadAttribute &density_factors,
Vector<float3> &r_normals,
Vector<int> &r_ids,
const int seed)
BLI_NOINLINE static void eliminate_points_based_on_mask(Span<bool> elimination_mask,
Vector<float3> &positions,
Vector<float3> &bary_coords,
Vector<int> &looptri_indices)
{
Vector<float3> points;
if (minimum_distance <= FLT_EPSILON || density <= FLT_EPSILON) {
return points;
}
/* Scatter points randomly on the mesh with higher density (5-7) times higher than desired for
* good quality possion disk distributions. */
int quality = 5;
const int output_points_target = 1000;
points.resize(output_points_target * quality);
const float required_area = output_points_target *
(2.0f * sqrtf(3.0f) * minimum_distance * minimum_distance);
const float point_scale_multiplier = sqrtf(required_area);
{
const int rnd_seed = BLI_hash_int(seed);
RandomNumberGenerator point_rng(rnd_seed);
for (int i = 0; i < points.size(); i++) {
points[i].x = point_rng.get_float() * point_scale_multiplier;
points[i].y = point_rng.get_float() * point_scale_multiplier;
points[i].z = 0.0f;
for (int i = positions.size() - 1; i >= 0; i--) {
if (elimination_mask[i]) {
positions.remove_and_reorder(i);
bary_coords.remove_and_reorder(i);
looptri_indices.remove_and_reorder(i);
}
}
}
/* Eliminate the scattered points until we get a possion distribution. */
Vector<float3> output_points(output_points_target);
BLI_NOINLINE static void compute_remaining_point_data(const Mesh &mesh,
Span<float3> bary_coords,
Span<int> looptri_indices,
MutableSpan<float3> r_normals,
MutableSpan<int> r_ids,
MutableSpan<float3> r_rotations)
{
Span<MLoopTri> looptris = get_mesh_looptris(mesh);
for (const int i : bary_coords.index_range()) {
const int looptri_index = looptri_indices[i];
const MLoopTri &looptri = looptris[looptri_index];
const float3 &bary_coord = bary_coords[i];
const float3 bounds_max = float3(point_scale_multiplier, point_scale_multiplier, 0);
poisson_disk_point_elimination(&points, &output_points, 2.0f * minimum_distance, bounds_max);
Vector<float3> final_points;
r_ids.reserve(output_points_target);
final_points.reserve(output_points_target);
const int v0_index = mesh.mloop[looptri.tri[0]].v;
const int v1_index = mesh.mloop[looptri.tri[1]].v;
const int v2_index = mesh.mloop[looptri.tri[2]].v;
const float3 v0_pos = mesh.mvert[v0_index].co;
const float3 v1_pos = mesh.mvert[v1_index].co;
const float3 v2_pos = mesh.mvert[v2_index].co;
/* Check if we have any points we should remove from the final possion distribition. */
BVHTreeFromMesh treedata;
BKE_bvhtree_from_mesh_get(&treedata, const_cast<Mesh *>(mesh), BVHTREE_FROM_LOOPTRI, 2);
float3 bb_min, bb_max;
BLI_bvhtree_get_bounding_box(treedata.tree, bb_min, bb_max);
struct RayCastAll_Data data;
data.bvhdata = &treedata;
data.raycast_callback = treedata.raycast_callback;
data.mesh = mesh;
data.projected_points = &final_points;
data.stable_ids = &r_ids;
data.normals = &r_normals;
data.density_factors = const_cast<FloatReadAttribute *>(&density_factors);
data.base_weight = std::min(
1.0f, density / (output_points.size() / (point_scale_multiplier * point_scale_multiplier)));
const float max_dist = bb_max[2] - bb_min[2] + 2.0f;
const float3 dir = float3(0, 0, -1);
float3 raystart;
raystart.z = bb_max[2] + 1.0f;
float tile_start_x_coord = bb_min[0];
int tile_repeat_x = ceilf((bb_max[0] - bb_min[0]) / point_scale_multiplier);
float tile_start_y_coord = bb_min[1];
int tile_repeat_y = ceilf((bb_max[1] - bb_min[1]) / point_scale_multiplier);
for (int x = 0; x < tile_repeat_x; x++) {
float tile_curr_x_coord = x * point_scale_multiplier + tile_start_x_coord;
for (int y = 0; y < tile_repeat_y; y++) {
float tile_curr_y_coord = y * point_scale_multiplier + tile_start_y_coord;
for (int idx = 0; idx < output_points.size(); idx++) {
raystart.x = output_points[idx].x + tile_curr_x_coord;
raystart.y = output_points[idx].y + tile_curr_y_coord;
data.cur_point_weight = (float)idx / (float)output_points.size();
data.raystart = raystart;
BLI_bvhtree_ray_cast_all(
treedata.tree, raystart, dir, 0.0f, max_dist, project_2d_bvh_callback, &data);
}
}
r_ids[i] = (int)(bary_coord.hash()) + looptri_index;
normal_tri_v3(r_normals[i], v0_pos, v1_pos, v2_pos);
r_rotations[i] = normal_to_euler_rotation(r_normals[i]);
}
}
return final_points;
static void sample_mesh_surface_with_minimum_distance(const Mesh &mesh,
const float max_density,
const float minimum_distance,
const FloatReadAttribute &density_factors,
const int seed,
Vector<float3> &r_positions,
Vector<float3> &r_bary_coords,
Vector<int> &r_looptri_indices)
{
sample_mesh_surface(
mesh, max_density, nullptr, seed, r_positions, r_bary_coords, r_looptri_indices);
Array<bool> elimination_mask(r_positions.size(), false);
update_elimination_mask_for_close_points(r_positions, minimum_distance, elimination_mask);
update_elimination_mask_based_on_density_factors(
mesh, density_factors, r_bary_coords, r_looptri_indices, elimination_mask);
eliminate_points_based_on_mask(elimination_mask, r_positions, r_bary_coords, r_looptri_indices);
}
static void geo_node_point_distribute_exec(GeoNodeExecParams params)
@ -309,25 +294,37 @@ static void geo_node_point_distribute_exec(GeoNodeExecParams params)
density_attribute, ATTR_DOMAIN_POINT, 1.0f);
const int seed = params.get_input<int>("Seed");
Vector<int> stable_ids;
Vector<float3> normals;
Vector<float3> points;
Vector<float3> positions;
Vector<float3> bary_coords;
Vector<int> looptri_indices;
switch (distribute_method) {
case GEO_NODE_POINT_DISTRIBUTE_RANDOM:
points = random_scatter_points_from_mesh(
mesh_in, density, density_factors, normals, stable_ids, seed);
sample_mesh_surface(
*mesh_in, density, &density_factors, seed, positions, bary_coords, looptri_indices);
break;
case GEO_NODE_POINT_DISTRIBUTE_POISSON:
const float min_dist = params.extract_input<float>("Distance Min");
points = poisson_scatter_points_from_mesh(
mesh_in, density, min_dist, density_factors, normals, stable_ids, seed);
const float minimum_distance = params.extract_input<float>("Distance Min");
sample_mesh_surface_with_minimum_distance(*mesh_in,
density,
minimum_distance,
density_factors,
seed,
positions,
bary_coords,
looptri_indices);
break;
}
const int tot_points = positions.size();
Array<float3> normals(tot_points);
Array<int> stable_ids(tot_points);
Array<float3> rotations(tot_points);
compute_remaining_point_data(
*mesh_in, bary_coords, looptri_indices, normals, stable_ids, rotations);
PointCloud *pointcloud = BKE_pointcloud_new_nomain(points.size());
memcpy(pointcloud->co, points.data(), sizeof(float3) * points.size());
for (const int i : points.index_range()) {
*(float3 *)(pointcloud->co + i) = points[i];
PointCloud *pointcloud = BKE_pointcloud_new_nomain(tot_points);
memcpy(pointcloud->co, positions.data(), sizeof(float3) * tot_points);
for (const int i : positions.index_range()) {
*(float3 *)(pointcloud->co + i) = positions[i];
pointcloud->radius[i] = 0.05f;
}
@ -355,9 +352,7 @@ static void geo_node_point_distribute_exec(GeoNodeExecParams params)
Float3WriteAttribute rotations_attribute = point_component.attribute_try_ensure_for_write(
"rotation", ATTR_DOMAIN_POINT, CD_PROP_FLOAT3);
MutableSpan<float3> rotations_span = rotations_attribute.get_span();
for (const int i : rotations_span.index_range()) {
rotations_span[i] = normal_to_euler_rotation(normals[i]);
}
rotations_span.copy_from(rotations);
rotations_attribute.apply_span();
}

View File

@ -1,281 +0,0 @@
/*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software Foundation,
* Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*/
/*
* Based on Cem Yuksel. 2015. Sample Elimination for Generating Poisson Disk Sample
* ! Sets. Computer Graphics Forum 34, 2 (May 2015), 25-32.
* ! http://www.cemyuksel.com/research/sampleelimination/
* Copyright (c) 2016, Cem Yuksel <cem@cemyuksel.com>
* All rights reserved.
*/
#include "BLI_inplace_priority_queue.hh"
#include "BLI_kdtree.h"
#include "node_geometry_util.hh"
#include <cstring>
#include <iostream>
namespace blender::nodes {
static void tile_point(Vector<float3> *tiled_points,
Vector<size_t> *indices,
const float maximum_distance,
const float3 boundbox,
float3 const &point,
size_t index,
int dimension = 0)
{
for (int dimension_iter = dimension; dimension_iter < 3; dimension_iter++) {
if (boundbox[dimension_iter] - point[dimension_iter] < maximum_distance) {
float3 point_tiled = point;
point_tiled[dimension_iter] -= boundbox[dimension_iter];
tiled_points->append(point_tiled);
indices->append(index);
tile_point(tiled_points,
indices,
maximum_distance,
boundbox,
point_tiled,
index,
dimension_iter + 1);
}
if (point[dimension_iter] < maximum_distance) {
float3 point_tiled = point;
point_tiled[dimension_iter] += boundbox[dimension_iter];
tiled_points->append(point_tiled);
indices->append(index);
tile_point(tiled_points,
indices,
maximum_distance,
boundbox,
point_tiled,
index,
dimension_iter + 1);
}
}
}
/**
* Returns the weight the point gets based on the distance to another point.
*/
static float point_weight_influence_get(const float maximum_distance,
const float minimum_distance,
float distance)
{
const float alpha = 8.0f;
if (distance < minimum_distance) {
distance = minimum_distance;
}
return std::pow(1.0f - distance / maximum_distance, alpha);
}
/**
* Weight each point based on their proximity to its neighbors
*
* For each index in the weight array add a weight based on the proximity the
* corresponding point has with its neighbors.
*/
static void points_distance_weight_calculate(Vector<float> *weights,
const size_t point_id,
const float3 *input_points,
const void *kd_tree,
const float minimum_distance,
const float maximum_distance,
InplacePriorityQueue<float> *heap)
{
KDTreeNearest_3d *nearest_point = nullptr;
int neighbors = BLI_kdtree_3d_range_search(
(KDTree_3d *)kd_tree, input_points[point_id], &nearest_point, maximum_distance);
for (int i = 0; i < neighbors; i++) {
size_t neighbor_point_id = nearest_point[i].index;
if (neighbor_point_id >= weights->size()) {
continue;
}
/* The point should not influence itself. */
if (neighbor_point_id == point_id) {
continue;
}
const float weight_influence = point_weight_influence_get(
maximum_distance, minimum_distance, nearest_point[i].dist);
/* In the first pass we just the weights. */
if (heap == nullptr) {
(*weights)[point_id] += weight_influence;
}
/* When we run again we need to update the weights and the heap. */
else {
(*weights)[neighbor_point_id] -= weight_influence;
heap->priority_decreased(neighbor_point_id);
}
}
if (nearest_point) {
MEM_freeN(nearest_point);
}
}
/**
* Returns the minimum radius fraction used by the default weight function.
*/
static float weight_limit_fraction_get(const size_t input_size, const size_t output_size)
{
const float beta = 0.65f;
const float gamma = 1.5f;
float ratio = float(output_size) / float(input_size);
return (1.0f - std::pow(ratio, gamma)) * beta;
}
/**
* Tile the input points.
*/
static void points_tiling(const float3 *input_points,
const size_t input_size,
void **kd_tree,
const float maximum_distance,
const float3 boundbox)
{
Vector<float3> tiled_points(input_points, input_points + input_size);
Vector<size_t> indices(input_size);
for (size_t i = 0; i < input_size; i++) {
indices[i] = i;
}
/* Tile the tree based on the boundbox. */
for (size_t i = 0; i < input_size; i++) {
tile_point(&tiled_points, &indices, maximum_distance, boundbox, input_points[i], i);
}
/* Build a new tree with the new indices and tiled points. */
*kd_tree = BLI_kdtree_3d_new(tiled_points.size());
for (size_t i = 0; i < tiled_points.size(); i++) {
BLI_kdtree_3d_insert(*(KDTree_3d **)kd_tree, indices[i], tiled_points[i]);
}
BLI_kdtree_3d_balance(*(KDTree_3d **)kd_tree);
}
static void weighted_sample_elimination(const float3 *input_points,
const size_t input_size,
float3 *output_points,
const size_t output_size,
const float maximum_distance,
const float3 boundbox,
const bool do_copy_eliminated)
{
const float minimum_distance = maximum_distance *
weight_limit_fraction_get(input_size, output_size);
void *kd_tree = nullptr;
points_tiling(input_points, input_size, &kd_tree, maximum_distance, boundbox);
/* Assign weights to each sample. */
Vector<float> weights(input_size, 0.0f);
for (size_t point_id = 0; point_id < weights.size(); point_id++) {
points_distance_weight_calculate(
&weights, point_id, input_points, kd_tree, minimum_distance, maximum_distance, nullptr);
}
/* Remove the points based on their weight. */
InplacePriorityQueue<float> heap(weights);
size_t sample_size = input_size;
while (sample_size > output_size) {
/* For each sample around it, remove its weight contribution and update the heap. */
size_t point_id = heap.pop_index();
points_distance_weight_calculate(
&weights, point_id, input_points, kd_tree, minimum_distance, maximum_distance, &heap);
sample_size--;
}
/* Copy the samples to the output array. */
size_t target_size = do_copy_eliminated ? input_size : output_size;
for (size_t i = 0; i < target_size; i++) {
size_t index = heap.all_indices()[i];
output_points[i] = input_points[index];
}
/* Cleanup. */
BLI_kdtree_3d_free((KDTree_3d *)kd_tree);
}
static void progressive_sampling_reorder(Vector<float3> *output_points,
float maximum_density,
float3 boundbox)
{
/* Re-order the points for progressive sampling. */
Vector<float3> temporary_points(output_points->size());
float3 *source_points = output_points->data();
float3 *dest_points = temporary_points.data();
size_t source_size = output_points->size();
size_t dest_size = 0;
while (source_size >= 3) {
dest_size = source_size * 0.5f;
/* Changes the weight function radius using half of the number of samples.
* It is used for progressive sampling. */
maximum_density *= std::sqrt(2.0f);
weighted_sample_elimination(
source_points, source_size, dest_points, dest_size, maximum_density, boundbox, true);
if (dest_points != output_points->data()) {
memcpy((*output_points)[dest_size],
dest_points[dest_size],
(source_size - dest_size) * sizeof(float3));
}
/* Swap the arrays around. */
float3 *points_iter = source_points;
source_points = dest_points;
dest_points = points_iter;
source_size = dest_size;
}
if (source_points != output_points->data()) {
memcpy(output_points->data(), source_points, dest_size * sizeof(float3));
}
}
void poisson_disk_point_elimination(Vector<float3> const *input_points,
Vector<float3> *output_points,
float maximum_distance,
float3 boundbox)
{
weighted_sample_elimination(input_points->data(),
input_points->size(),
output_points->data(),
output_points->size(),
maximum_distance,
boundbox,
false);
progressive_sampling_reorder(output_points, maximum_distance, boundbox);
}
} // namespace blender::nodes