open3d使用fpfh做点云配准

2024-02-12 00:10
文章标签 使用 配准 做点 fpfh open3d

本文主要是介绍open3d使用fpfh做点云配准,希望对大家解决编程问题提供一定的参考价值,需要的开发者们随着小编来一起学习吧!

目录

  • 写在前面
  • 准备
    • 编译open3d
    • 配准example
  • 编译demo
  • 配准测试
  • 参考

写在前面

1、环境:win10,cmake3.22.0-rc1,环境:win10,cmake3.22.0-rc1,已编译好的open3d 0.13
2、源码使用的是官方example
https://github.com/isl-org/Open3D/blob/master/examples/cpp/RegistrationRANSAC.cpp
3、建议先看一下open3d的demo:
open3d 0.13的c++版本使用demo https://blog.csdn.net/qq_41102371/article/details/12106527
4、所有资源已给出链接

准备

编译open3d

windows10编译open3d 0.13 https://blog.csdn.net/qq_41102371/article/details/121014372

配准example

将官方demo的CMakeLists.txt与配准源码RegistrationRANSAC.cpp
下载下来,放在文件夹open3d_fpfh_registration下
https://github.com/isl-org/open3d-cmake-find-package/blob/master/CMakeLists.txt
https://github.com/isl-org/Open3D/blob/master/examples/cpp/RegistrationRANSAC.cpp
并将CMakeLists.txt里面的Draw全部换成RegistrationRANSAC
CMakeLists.txt

# On Ubuntu 18.04, get the latest CMake from https://apt.kitware.com/.
cmake_minimum_required(VERSION 3.18)project(Open3DCMakeFindPackage LANGUAGES C CXX)# The options need to be the same as Open3D's default
# If Open3D is configured and built with custom options, you'll also need to
# specify the same custom options.
option(STATIC_WINDOWS_RUNTIME "Use static (MT/MTd) Windows runtime" ON)
if(STATIC_WINDOWS_RUNTIME)set(CMAKE_MSVC_RUNTIME_LIBRARY "MultiThreaded$<$<CONFIG:Debug>:Debug>")
else()set(CMAKE_MSVC_RUNTIME_LIBRARY "MultiThreaded$<$<CONFIG:Debug>:Debug>DLL")
endif()# Find installed Open3D, which exports Open3D::Open3D
find_package(Open3D REQUIRED)add_executable(RegistrationRANSAC)
target_sources(RegistrationRANSAC PRIVATE RegistrationRANSAC.cpp)
target_link_libraries(RegistrationRANSAC PRIVATE Open3D::Open3D)# On Windows if BUILD_SHARED_LIBS is enabled, copy .dll files to the executable directory
if(WIN32)get_target_property(open3d_type Open3D::Open3D TYPE)if(open3d_type STREQUAL "SHARED_LIBRARY")message(STATUS "Copying Open3D.dll to ${CMAKE_CURRENT_BINARY_DIR}/$<CONFIG>")add_custom_command(TARGET RegistrationRANSAC POST_BUILDCOMMAND ${CMAKE_COMMAND} -E copy${CMAKE_INSTALL_PREFIX}/bin/Open3D.dll${CMAKE_CURRENT_BINARY_DIR}/$<CONFIG>)endif()
endif()

RegistrationRANSAC.cpp,我在官方代码上做了些修改

// ----------------------------------------------------------------------------
// -                        Open3D: www.open3d.org                            -
// ----------------------------------------------------------------------------
// The MIT License (MIT)
//
// Copyright (c) 2018 www.open3d.org
//
// Permission is hereby granted, free of charge, to any person obtaining a copy
// of this software and associated documentation files (the "Software"), to deal
// in the Software without restriction, including without limitation the rights
// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
// copies of the Software, and to permit persons to whom the Software is
// furnished to do so, subject to the following conditions:
//
// The above copyright notice and this permission notice shall be included in
// all copies or substantial portions of the Software.
//
// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
// FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
// IN THE SOFTWARE.
// ----------------------------------------------------------------------------#include <Eigen/Dense>
#include <iostream>
#include <memory>#include "open3d/Open3D.h"using namespace open3d;
using namespace std;
std::tuple<std::shared_ptr<geometry::PointCloud>,std::shared_ptr<pipelines::registration::Feature>>
PreprocessPointCloud(const char *file_name, float voxel_size = 2.0) {//从文件读取点云auto pcd = open3d::io::CreatePointCloudFromFile(file_name);//降采样auto pcd_down = pcd->VoxelDownSample(voxel_size);std::cout << "read " << pcd->points_.size() << " points from " << file_name<< std::endl;std::cout << "voxel_size=" << voxel_size << ", after downsample "<< pcd_down->points_.size() << "points left" << endl;//计算法向量pcd_down->EstimateNormals(open3d::geometry::KDTreeSearchParamHybrid(voxel_size * 2, 30));//计算fpfh特征auto pcd_fpfh = pipelines::registration::ComputeFPFHFeature(*pcd_down,open3d::geometry::KDTreeSearchParamHybrid(voxel_size * 5, 100));return std::make_tuple(pcd_down, pcd_fpfh);
}void PrintHelp() {using namespace open3d;PrintOpen3DVersion();// clang-format offutility::LogInfo("Usage:");utility::LogInfo("    > RegistrationRANSAC source_pcd target_pcd  voxel_size [--method=feature_matching] [--mutual_filter] [--visualize]");// clang-format onutility::LogInfo("");
}int main(int argc, char *argv[]) {using namespace open3d;utility::SetVerbosityLevel(utility::VerbosityLevel::Debug);if (argc < 4 ||utility::ProgramOptionExistsAny(argc, argv, {"-h", "--help"})) {PrintHelp();return 1;}bool visualize = false;if (utility::ProgramOptionExists(argc, argv, "--visualize")) {visualize = true;}//visualize = true;bool mutual_filter = false;//if (utility::ProgramOptionExists(argc, argv, "--mutual_filter")) {//    mutual_filter = true;//}mutual_filter = true;// Prepare inputstd::shared_ptr<geometry::PointCloud> source, target;std::shared_ptr<pipelines::registration::Feature> source_fpfh, target_fpfh;//set voxel_sizefloat voxel_size = std::atof(argv[3]);std::cout << voxel_size << std::endl;float distance_threshold = voxel_size * 1.5;std::tie(source, source_fpfh) = PreprocessPointCloud(argv[1], voxel_size);std::tie(target, target_fpfh) = PreprocessPointCloud(argv[2], voxel_size);pipelines::registration::RegistrationResult registration_result;// Prepare checkersstd::vector<std::reference_wrapper<const pipelines::registration::CorrespondenceChecker>>correspondence_checker;auto correspondence_checker_edge_length =pipelines::registration::CorrespondenceCheckerBasedOnEdgeLength(0.9);auto correspondence_checker_distance =pipelines::registration::CorrespondenceCheckerBasedOnDistance(distance_threshold);correspondence_checker.push_back(correspondence_checker_edge_length);correspondence_checker.push_back(correspondence_checker_distance);registration_result = pipelines::registration::RegistrationRANSACBasedOnFeatureMatching(*source, *target, *source_fpfh, *target_fpfh,mutual_filter, distance_threshold,pipelines::registration::TransformationEstimationPointToPoint(false),3, correspondence_checker,pipelines::registration::RANSACConvergenceCriteria(1000000, 0.999));cout << endl<< "fpfh matrix:" << endl<< registration_result.transformation_ << endl;cout << "inlier(correspondence_set size):"<< registration_result.correspondence_set_.size() << endl;std::vector<std::pair<int, int>> correspondences_ransac;for (int m = 0; m < registration_result.correspondence_set_.size();++m) {std::pair<int, int> pair_(0, 0);pair_.first = registration_result.correspondence_set_[m][0];pair_.second = registration_result.correspondence_set_[m][1];correspondences_ransac.push_back(pair_);}// std::shared_ptr<open3d::geometry::LineSet>auto ransac_lineset =geometry::LineSet::CreateFromPointCloudCorrespondences(*source, *target, correspondences_ransac);std::shared_ptr<geometry::PointCloud> source_transformed_ptr(new geometry::PointCloud);std::shared_ptr<geometry::PointCloud> source_ptr(new geometry::PointCloud);std::shared_ptr<geometry::PointCloud> target_ptr(new geometry::PointCloud);*source_transformed_ptr = *source;*target_ptr = *target;*source_ptr = *source;target->PaintUniformColor({1, 0, 0});                  //红source->PaintUniformColor({0, 1, 0});                  //绿source_transformed_ptr->PaintUniformColor({0, 0, 1});  //蓝source_transformed_ptr->Transform(registration_result.transformation_);if (visualize) {visualization::DrawGeometries({target, source, source_transformed_ptr, ransac_lineset},"Registration result", 960, 900, 960, 100);}return 0;
}

编译demo

cd open3d_fpfh_registration
mkdir build
cd build
cmake -DOpen3D_ROOT="D:/Program Files/open3d" ..
cmake --build . --config Debug

成功生成RegistrationRANSAC.exe
在这里插入图片描述

配准测试

数据就用斯坦福的兔兔,http://graphics.stanford.edu/pub/3Dscanrep/bunny.tar.gz
解压后把bun000.ply和bun045.ply放在open3d_fpfh_registration目录下
回到命令行窗口:

start ./Debug/RegistrationRANSAC.exe ../bun045.ply ../bun000.ply 0.005 --visualize

若是其他数据,改一下…/bun045.ply …/bun000.ply 0.005 --visualize这几个参数就好,分别是source点云路径,target点云路径,降采样网格大小(根据点云大小和数据量来自行调节),–visualize加可视化

配准结果,红色是bun000,绿色是bun045,蓝色是bun045配准至bun000的结果,连线是连接的匹配上的fpfh特征
在这里插入图片描述

参考

文中已列出

--------------------------------------------------------------------------------------------诺有缸的高飞鸟202110

这篇关于open3d使用fpfh做点云配准的文章就介绍到这儿,希望我们推荐的文章对编程师们有所帮助!



http://www.chinasem.cn/article/701183

相关文章

一文详解如何使用Java获取PDF页面信息

《一文详解如何使用Java获取PDF页面信息》了解PDF页面属性是我们在处理文档、内容提取、打印设置或页面重组等任务时不可或缺的一环,下面我们就来看看如何使用Java语言获取这些信息吧... 目录引言一、安装和引入PDF处理库引入依赖二、获取 PDF 页数三、获取页面尺寸(宽高)四、获取页面旋转角度五、判断

C++中assign函数的使用

《C++中assign函数的使用》在C++标准模板库中,std::list等容器都提供了assign成员函数,它比操作符更灵活,支持多种初始化方式,下面就来介绍一下assign的用法,具有一定的参考价... 目录​1.assign的基本功能​​语法​2. 具体用法示例​​​(1) 填充n个相同值​​(2)

Spring StateMachine实现状态机使用示例详解

《SpringStateMachine实现状态机使用示例详解》本文介绍SpringStateMachine实现状态机的步骤,包括依赖导入、枚举定义、状态转移规则配置、上下文管理及服务调用示例,重点解... 目录什么是状态机使用示例什么是状态机状态机是计算机科学中的​​核心建模工具​​,用于描述对象在其生命

使用Python删除Excel中的行列和单元格示例详解

《使用Python删除Excel中的行列和单元格示例详解》在处理Excel数据时,删除不需要的行、列或单元格是一项常见且必要的操作,本文将使用Python脚本实现对Excel表格的高效自动化处理,感兴... 目录开发环境准备使用 python 删除 Excphpel 表格中的行删除特定行删除空白行删除含指定

深入理解Go语言中二维切片的使用

《深入理解Go语言中二维切片的使用》本文深入讲解了Go语言中二维切片的概念与应用,用于表示矩阵、表格等二维数据结构,文中通过示例代码介绍的非常详细,需要的朋友们下面随着小编来一起学习学习吧... 目录引言二维切片的基本概念定义创建二维切片二维切片的操作访问元素修改元素遍历二维切片二维切片的动态调整追加行动态

prometheus如何使用pushgateway监控网路丢包

《prometheus如何使用pushgateway监控网路丢包》:本文主要介绍prometheus如何使用pushgateway监控网路丢包问题,具有很好的参考价值,希望对大家有所帮助,如有错误... 目录监控网路丢包脚本数据图表总结监控网路丢包脚本[root@gtcq-gt-monitor-prome

Python通用唯一标识符模块uuid使用案例详解

《Python通用唯一标识符模块uuid使用案例详解》Pythonuuid模块用于生成128位全局唯一标识符,支持UUID1-5版本,适用于分布式系统、数据库主键等场景,需注意隐私、碰撞概率及存储优... 目录简介核心功能1. UUID版本2. UUID属性3. 命名空间使用场景1. 生成唯一标识符2. 数

SpringBoot中如何使用Assert进行断言校验

《SpringBoot中如何使用Assert进行断言校验》Java提供了内置的assert机制,而Spring框架也提供了更强大的Assert工具类来帮助开发者进行参数校验和状态检查,下... 目录前言一、Java 原生assert简介1.1 使用方式1.2 示例代码1.3 优缺点分析二、Spring Fr

Android kotlin中 Channel 和 Flow 的区别和选择使用场景分析

《Androidkotlin中Channel和Flow的区别和选择使用场景分析》Kotlin协程中,Flow是冷数据流,按需触发,适合响应式数据处理;Channel是热数据流,持续发送,支持... 目录一、基本概念界定FlowChannel二、核心特性对比数据生产触发条件生产与消费的关系背压处理机制生命周期

java使用protobuf-maven-plugin的插件编译proto文件详解

《java使用protobuf-maven-plugin的插件编译proto文件详解》:本文主要介绍java使用protobuf-maven-plugin的插件编译proto文件,具有很好的参考价... 目录protobuf文件作为数据传输和存储的协议主要介绍在Java使用maven编译proto文件的插件