123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141 |
- /***********************************************************************
- * Software License Agreement (BSD License)
- *
- * Copyright 2011-2024 Jose Luis Blanco (joseluisblancoc@gmail.com).
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- *
- * 1. Redistributions of source code must retain the above copyright
- * notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- * notice, this list of conditions and the following disclaimer in the
- * documentation and/or other materials provided with the distribution.
- *
- * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
- * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
- * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
- * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
- * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
- * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- *************************************************************************/
- #include <cstdlib>
- #include <ctime>
- #include <iostream>
- #include <nanoflann.hpp>
- #include "utils.h"
- template <typename num_t>
- void kdtree_demo(const size_t N)
- {
- PointCloud<num_t> cloud;
- // construct a kd-tree index:
- using my_kd_tree_t = nanoflann::KDTreeSingleIndexDynamicAdaptor<
- nanoflann::L2_Simple_Adaptor<num_t, PointCloud<num_t>>,
- PointCloud<num_t>, 3 /* dim */
- >;
- dump_mem_usage();
- my_kd_tree_t index(3 /*dim*/, cloud, {10 /* max leaf */});
- // Generate points:
- generateRandomPointCloud(cloud, N);
- num_t query_pt[3] = {0.5, 0.5, 0.5};
- // add points in chunks at a time
- size_t chunk_size = 100;
- for (size_t i = 0; i < N; i = i + chunk_size)
- {
- size_t end = std::min<size_t>(i + chunk_size, N - 1);
- // Inserts all points from [i, end]
- index.addPoints(i, end);
- }
- // remove a point
- size_t removePointIndex = N - 1;
- index.removePoint(removePointIndex);
- dump_mem_usage();
- {
- std::cout << "Searching for 1 element..." << std::endl;
- // do a knn search
- const size_t num_results = 1;
- size_t ret_index;
- num_t out_dist_sqr;
- nanoflann::KNNResultSet<num_t> resultSet(num_results);
- resultSet.init(&ret_index, &out_dist_sqr);
- index.findNeighbors(resultSet, query_pt, {10});
- std::cout << "knnSearch(nn=" << num_results << "): \n";
- std::cout << "ret_index=" << ret_index
- << " out_dist_sqr=" << out_dist_sqr << std::endl;
- std::cout << "point: ("
- << "point: (" << cloud.pts[ret_index].x << ", "
- << cloud.pts[ret_index].y << ", " << cloud.pts[ret_index].z
- << ")" << std::endl;
- std::cout << std::endl;
- }
- {
- // do a knn search searching for more than one result
- const size_t num_results = 5;
- std::cout << "Searching for " << num_results << " elements"
- << std::endl;
- size_t ret_index[num_results];
- num_t out_dist_sqr[num_results];
- nanoflann::KNNResultSet<num_t> resultSet(num_results);
- resultSet.init(ret_index, out_dist_sqr);
- index.findNeighbors(resultSet, query_pt);
- std::cout << "knnSearch(nn=" << num_results << "): \n";
- std::cout << "Results: " << std::endl;
- for (size_t i = 0; i < resultSet.size(); ++i)
- {
- std::cout << "#" << i << ",\t"
- << "index: " << ret_index[i] << ",\t"
- << "dist: " << out_dist_sqr[i] << ",\t"
- << "point: (" << cloud.pts[ret_index[i]].x << ", "
- << cloud.pts[ret_index[i]].y << ", "
- << cloud.pts[ret_index[i]].z << ")" << std::endl;
- }
- std::cout << std::endl;
- }
- {
- // Unsorted radius search:
- std::cout << "Unsorted radius search" << std::endl;
- const num_t radiusSqr = 1;
- std::vector<nanoflann::ResultItem<size_t, num_t>> indices_dists;
- nanoflann::RadiusResultSet<num_t, size_t> resultSet(
- radiusSqr, indices_dists);
- index.findNeighbors(resultSet, query_pt);
- nanoflann::ResultItem<size_t, num_t> worst_pair =
- resultSet.worst_item();
- std::cout << "Worst pair: idx=" << worst_pair.first
- << " dist=" << worst_pair.second << std::endl;
- std::cout << "point: (" << cloud.pts[worst_pair.first].x << ", "
- << cloud.pts[worst_pair.first].y << ", "
- << cloud.pts[worst_pair.first].z << ")" << std::endl;
- std::cout << std::endl;
- }
- }
- int main()
- {
- // Randomize Seed
- srand(static_cast<unsigned int>(time(nullptr)));
- kdtree_demo<float>(1000000);
- kdtree_demo<double>(1000000);
- return 0;
- }
|