关闭

ROS机器人程序设计(原书第2版)补充资料 (陆) 第六章 点云 PCL

标签: ROS机器人程序设计原书第2版ROS
2753人阅读 评论(0) 收藏 举报
分类:

ROS机器人程序设计(原书第2版)补充资料 (陆) 第六章 点云 PCL

书中,大部分出现hydro的地方,直接替换为indigojadekinetic,即可在对应版本中使用。

RGBD深度摄像头传感器最常用的数据存储,处理和显示方式就是点云。

推荐查阅-PCL官网:http://www.pointclouds.org/

http://ros.org/images/wiki/PCL_logo.png1. http://wiki.ros.org/pcl_ros 2. http://wiki.ros.org/pcl

补充阅读:

http://blog.csdn.net/zhangrelay/article/details/50053733

http://blog.csdn.net/zhangrelay/article/details/50240935


第163页:

简介点云。

第163-165页:

理解点云,包括类型,算法和接口等。

第166-190页:

学习在ROS使用PCL,包括创建点云,可视化,滤波,缩减采样,配准,匹配,分区,分割等。

第191页:

本章小结。

  


思考与巩固:

1 使用深度摄像头采集环境信息,并用点云显示,用本章提及的方法进行处理。

2 在ROSwiki上查阅点云相关功能包并完成编译使用。


附:

How to use a PCL tutorial in ROS

Create a ROS package

$ catkin_create_pkg my_pcl_tutorial pcl_conversions pcl_ros roscpp sensor_msgs 

Then, modify the package.xml to add 

  <build_depend>libpcl-all-dev</build_depend>
  <run_depend>libpcl-all</run_depend>

Create the code skeleton

Create an empty file called src/example.cpp and paste the following code in it: 

切换行号显示
   1 #include <ros/ros.h>
   2 // PCL specific includes
   3 #include <sensor_msgs/PointCloud2.h>
   4 #include <pcl_conversions/pcl_conversions.h>
   5 #include <pcl/point_cloud.h>
   6 #include <pcl/point_types.h>
   7 
   8 ros::Publisher pub;
   9 
  10 void 
  11 cloud_cb (const sensor_msgs::PointCloud2ConstPtr& input)
  12 {
  13   // Create a container for the data.
  14   sensor_msgs::PointCloud2 output;
  15 
  16   // Do data processing here...
  17   output = *input;
  18 
  19   // Publish the data.
  20   pub.publish (output);
  21 }
  22 
  23 int
  24 main (int argc, char** argv)
  25 {
  26   // Initialize ROS
  27   ros::init (argc, argv, "my_pcl_tutorial");
  28   ros::NodeHandle nh;
  29 
  30   // Create a ROS subscriber for the input point cloud
  31   ros::Subscriber sub = nh.subscribe ("input", 1, cloud_cb);
  32 
  33   // Create a ROS publisher for the output point cloud
  34   pub = nh.advertise<sensor_msgs::PointCloud2> ("output", 1);
  35 
  36   // Spin
  37   ros::spin ();
  38 }

The code above does nothing but initialize ROS, create a subscriber and a publisher for PointCloud2 data.

Add the source file to CMakeLists.txt

Edit the CMakeLists.txt file in your newly created package and add: 

add_executable(example src/example.cpp)
target_link_libraries(example ${catkin_LIBRARIES})

Download the source code from the PCL tutorial

PCL has about four different ways of representing point cloud data, so it can get a bit confusing, but we'll try to keep it simple for you. The types are: 

  • sensor_msgs::PointCloud — ROS message (deprecated) 

  • sensor_msgs::PointCloud2 — ROS message 

  • pcl::PCLPointCloud2 — PCL data structure mostly for compatibility with ROS (I think) 

  • pcl::PointCloud<T> — standard PCL data structure 

In the following code examples we will focus on the ROS message (sensor_msgs::PointCloud2) and the standard PCL data structure (pcl::PointCloud<T>). However, you should also note that pcl::PCLPointCloud2 is an important and useful type as well: you can directly subscribe to nodes using that type and it will be automatically serialized to/from the sensor_msgs type. See this example to try PCLPointCloud2 yourself. 

sensor_msgs/PointCloud2

If you'd like to save yourself some copying and pasting, you can download the source file for this example here. Just remember to rename the file to example.cpp or edit your CMakeLists.txt to match.

The sensor_msgs/PointCloud2 format was designed as a ROS message, and is the preferred choice for ROS applications. In the following example, we downsample a PointCloud2 structure using a 3D grid, thus reducing the number of points in the input dataset considerably. 

To add this capability to the code skeleton above, perform the following steps: 

  • visit http://www.pointclouds.org/documentation/, click on Tutorials, then navigate to the Downsampling a PointCloud using a VoxelGrid filter tutorial (http://www.pointclouds.org/documentation/tutorials/voxel_grid.php

  • read the code and the explanations provided there. You will notice that the code breaks down essentially in 3 parts: 
    • load the cloud (lines 9-19) 
    • process the cloud (lines 20-24) 
    • save the output (lines 25-32) 
  • since we use ROS subscribers and publishers in our code snippet above, we can ignore the loading and saving of point cloud data using the PCD format. Thus, the only relevant part in the tutorial remains lines 20-24 that create the PCL object, pass the input data, and perform the actual computation: 

切换行号显示
   1   // Create the filtering object
   2   pcl::VoxelGrid<pcl::PCLPointCloud2> sor;
   3   sor.setInputCloud (cloud);
   4   sor.setLeafSize (0.01, 0.01, 0.01);
   5   sor.filter (*cloud_filtered);

  • In these lines, the input dataset is named cloud, and the output dataset is called cloud_filtered. We can copy this work, but remember from earlier that we said we wanted to work with the sensor_msgs class, not the pcl class. In order to do this, we're going to have to do a little bit of extra work to convert the ROS message to the PCL type. Modify the callback function as follows: 

切换行号显示
   1 #include <pcl/filters/voxel_grid.h>
   2 
   3 ...
   4 
   5 void 
   6 cloud_cb (const sensor_msgs::PointCloud2ConstPtr& cloud_msg)
   7 {
   8   // Container for original & filtered data
   9   pcl::PCLPointCloud2* cloud = new pcl::PCLPointCloud2; 
  10   pcl::PCLPointCloud2ConstPtr cloudPtr(cloud);
  11   pcl::PCLPointCloud2 cloud_filtered;
  12 
  13   // Convert to PCL data type
  14   pcl_conversions::toPCL(*cloud_msg, *cloud);
  15 
  16   // Perform the actual filtering
  17   pcl::VoxelGrid<pcl::PCLPointCloud2> sor;
  18   sor.setInputCloud (cloudPtr);
  19   sor.setLeafSize (0.1, 0.1, 0.1);
  20   sor.filter (cloud_filtered);
  21 
  22   // Convert to ROS data type
  23   sensor_msgs::PointCloud2 output;
  24   pcl_conversions::fromPCL(cloud_filtered, output);
  25 
  26   // Publish the data
  27   pub.publish (output);
  28 }

Note

Since different tutorials will often use different variable names for their inputs and outputs, remember that you may need to modify the code slightly when integrating the tutorial code into your own ROS node. In this case, notice that we had to change the variable name input to cloud, and output to cloud_filtered in order to match up with the code from the tutorial we copied.

Note that there is a slight inefficiency here. The fromPCL can be replaced with moveFromPCL to prevent copying the entire (filtered) point cloud. However, the toPCL call cannot be optimized in this way since the original input is const.

Save the output file then build: 

$ cd %TOP_DIR_YOUR_CATKIN_HOME%
$ catkin_make

Then run: 

$ rosrun my_pcl_tutorial example input:=/narrow_stereo_textured/points2

or, if you're running an OpenNI-compatible depth sensor, try: 

$ roslaunch openni_launch openni.launch
$ rosrun my_pcl_tutorial example input:=/camera/depth/points

You can visualize the result by running RViz: 

$ rosrun rviz rviz

and adding a "PointCloud2" display. Select camera_depth_frame for the Fixed Frame (or whatever frame is appropriate for your sensor) and select output for the PointCloud2 topic. You should see a highly downsampled point cloud. For comparison, you can view the /camera/depth/points topic and see how much it has been downsampled. 

pcl/PointCloud<T>

As with the previous example, if you want to skip a few steps, you can download the source file for this example here.

The pcl/PointCloud<T> format represents the internal PCL point cloud format. For modularity and efficiency reasons, the format is templated on the point type, and PCL provides a list of templated common types which are SSE aligned. In the following example, we estimate the planar coefficients of the largest plane found in a scene. 

To add this capability to the code skeleton above, perform the following steps: 

  • visit http://www.pointclouds.org/documentation/, click on Tutorials, then navigate to the Planar model segmentation tutorial (http://www.pointclouds.org/documentation/tutorials/planar_segmentation.php

  • read the code and the explanations provided there. You will notice that the code breaks down essentially in 3 parts: 
    • create a cloud and populate it with values (lines 12-30) 
    • process the cloud (38-56) 
    • write the coefficients (58-68) 
  • since we use ROS subscribers in our code snippet above, we can ignore the first step, and just process the cloud received on the callback directly. Thus, the only relevant part in the tutorial remains lines 38-56 that create the PCL object, pass the input data, and perform the actual computation: 

切换行号显示
   1   pcl::ModelCoefficients coefficients;
   2   pcl::PointIndices inliers;
   3   // Create the segmentation object
   4   pcl::SACSegmentation<pcl::PointXYZ> seg;
   5   // Optional
   6   seg.setOptimizeCoefficients (true);
   7   // Mandatory
   8   seg.setModelType (pcl::SACMODEL_PLANE);
   9   seg.setMethodType (pcl::SAC_RANSAC);
  10   seg.setDistanceThreshold (0.01);
  11 
  12   seg.setInputCloud (cloud.makeShared ());
  13   seg.segment (inliers, coefficients);

  • In these lines, the input dataset is named cloud and is of type pcl::PointCloud<pcl::PointXYZ>, and the output is represented by a set of point indices that contain the plane together with the plane coefficients. cloud.makeShared()creates a boost shared_ptr object for the object cloud (see the pcl::PointCloud API documentation).

Copy these lines, in the code snippet above, by modifying the callback function as follows: 

切换行号显示
   1 #include <pcl/sample_consensus/model_types.h>
   2 #include <pcl/sample_consensus/method_types.h>
   3 #include <pcl/segmentation/sac_segmentation.h>
   4 
   5 ...
   6 
   7 void 
   8 cloud_cb (const sensor_msgs::PointCloud2ConstPtr& input)
   9 {
  10   // Convert the sensor_msgs/PointCloud2 data to pcl/PointCloud
  11   pcl::PointCloud<pcl::PointXYZ> cloud;
  12   pcl::fromROSMsg (*input, cloud);
  13 
  14   pcl::ModelCoefficients coefficients;
  15   pcl::PointIndices inliers;
  16   // Create the segmentation object
  17   pcl::SACSegmentation<pcl::PointXYZ> seg;
  18   // Optional
  19   seg.setOptimizeCoefficients (true);
  20   // Mandatory
  21   seg.setModelType (pcl::SACMODEL_PLANE);
  22   seg.setMethodType (pcl::SAC_RANSAC);
  23   seg.setDistanceThreshold (0.01);
  24 
  25   seg.setInputCloud (cloud.makeShared ());
  26   seg.segment (inliers, coefficients);
  27 
  28   // Publish the model coefficients
  29   pcl_msgs::ModelCoefficients ros_coefficients;
  30   pcl_conversions::fromPCL(coefficients, ros_coefficients);
  31   pub.publish (ros_coefficients);
  32 }

Notice that we added two conversion steps: from sensor_msgs/PointCloud2 to pcl/PointCloud<T>and from pcl::ModelCoefficients to pcl_msgs::ModelCoefficients. We also changed the variable that we publish from output to coefficients.

In addition, since we're now publishing the planar model coefficients found rather than point cloud data, we have to change our publisher type from: 

  // Create a ROS publisher for the output point cloud
  pub = nh.advertise<sensor_msgs::PointCloud2> ("output", 1);

to: 

  // Create a ROS publisher for the output model coefficients
  pub = nh.advertise<pcl_msgs::ModelCoefficients> ("output", 1);

Save the output file, then compile and run the code above: 

$ rosrun my_pcl_tutorial example input:=/narrow_stereo_textured/points2

or, if you're running an OpenNI-compatible depth sensor, try: 

$ rosrun my_pcl_tutorial example input:=/camera/depth/points

See the output with 

$ rostopic echo output


0
0
查看评论

ROS 学习 (三)

这两天,为了把之前在win7系统下写好的点云分割程序在ROS系统里打包,供其他package调用,依照着官网开始学习pcl_ros.这里,简单记录下自己这两天学习的一点心得.(附:我安装的是ROS Diamondback版本,其他版本请参照之)     1.使用下面语...
  • chen_nan
  • chen_nan
  • 2012-02-13 20:35
  • 3548

ROS_PCL_第一个PCL程序

1.创建ROS程序包 程序包是基于pcl_conversions pcl_ros pcl_msgs sensor_msgs cd catkin_ws catkin_create_pkg pcl_tutorials pcl_conversions pcl_ros pcl_msgs sensor_...
  • hanshuning
  • hanshuning
  • 2016-08-25 14:45
  • 1905

pcl的初步使用(ROS)

ROS里面已经预装好了pcl,和一些与pcl之间的转换,我们接下来看看如何在ros里面使用pcl
  • qq_25491201
  • qq_25491201
  • 2016-04-04 17:29
  • 15327

ROS下使用PCL/kinfu

最近想做机器人平台上的kinect fusion
  • l_h2010
  • l_h2010
  • 2014-08-02 17:28
  • 3907

ROS_PCL+Rviz创建点云并三维显示

1.程序包的配置等参照 第一个PCL程序 2.代码.cpp #include #include #include #include main (int argc, char **argv) { ros::init (argc, argv, "pcl_create")...
  • hanshuning
  • hanshuning
  • 2016-08-28 16:08
  • 3255

ROS机器人程序设计(原书第2版)补充资料 (零) 源代码、资料和印刷错误修订等 2017年02月22日更新

ROS机器人程序设计(原书第2版)补充资料 (零) 源代码等 书中,大部分出现hydro的地方,直接替换为indigo或jade或kinetic,即可在对应版本中使用。 稍后会补充,各章节针对indigo版本和kinetic版本的修正说明,方便读者阅读。 如果长期使用Ubuntu和ROS,推荐LTS...
  • ZhangRelay
  • ZhangRelay
  • 2016-11-25 19:38
  • 9836

ROS机器人程序设计(原书第2版)学习镜像分享及使用说明

ROS机器人程序设计(原书第2版)学习镜像分享及使用说明ROS机器人操作系统在线练习课程正在逐步完善中,目前以ROS官网中文资料制作,CSDN下载ROS基础资源汇总:http://download.csdn.net/album/detail/4431通过网页快速了解Linux(Ubuntu)和ROS...
  • ZhangRelay
  • ZhangRelay
  • 2016-11-24 20:54
  • 13144

ROS机器人程序设计(原书第二版)笔记

前几张基础在网上找博客学的要比看书强,可以看我上一个博客推荐 前8章 indigo zhangrelay的博客:http://blog.csdn.net/ZhangRelay/article/details/49893069 第8章 倔强的不倒翁:http://blog.csdn.net/x...
  • m0_37903426
  • m0_37903426
  • 2017-12-15 12:58
  • 47

ROS机器人程序设计(原书第2版)补充资料 (贰) 第二章 ROS系统架构及概念

ROS机器人程序设计(原书第2版)补充资料 (贰) 第二章 ROS系统架构及概念 由于工作事物繁忙,更新有些慢,抱歉。 已经完成的各章节补充说明,会依据反馈意见持续更新,希望大家多提宝贵意见,非常感谢。 在完成了第一章的学习实现之后,基本已经掌握了ROS系统的安装,那么如何使用ROS,理解系统架构和...
  • ZhangRelay
  • ZhangRelay
  • 2016-11-17 00:08
  • 3836

ROS机器人程序设计(原书第2版)补充资料 (捌) 第八章 导航功能包集入门 navigation

ROS机器人程序设计(原书第2版)补充资料 (捌) 第八章 导航功能包集入门navigation书中,大部分出现hydro的地方,直接替换为indigo或jade或kinetic,即可在对应版本中使用。本章三个非常重要概念:TF,SLAM,AMCL。务必掌握。补充内容:http://blog.csd...
  • ZhangRelay
  • ZhangRelay
  • 2016-11-20 17:13
  • 2311
    个人资料
    • 访问:884070次
    • 积分:11474
    • 等级:
    • 排名:第1612名
    • 原创:237篇
    • 转载:23篇
    • 译文:49篇
    • 评论:660条
    博客专栏
    最新评论