ROS learning (XX) robot slam function package -- installation and testing of rgbdslam

---

Preface

   In the previous study ,slam The constructed maps are all two-dimensional maps , At present, many algorithms can realize the map construction of three-dimensional information , Robots not only know where obstacles are in the map , And also know what the obstacle is .
  ROS Provides a variety of 3D SLAM Function pack ,rgbdslam That's one of them , This chapter mainly studies rgbdslam Installation and testing of .

One 、 New workspace

The order is as follows ：

mkdir -p ~/catkin_rgbdslam_ws/src
cd ~/catkin_rgbdslam_ws/src
catkin_init_workspace
cd ~/catkin_rgbdslam_ws
catkin_make

Two 、 download rgbdslam_v2 Corresponding to the author g2o

The order is as follows ：

cd ~/catkin_rgbdslam_ws/src
sudo apt-get install libsuitesparse-dev  
git clone https://github.com/felixendres/g2o.git
cd g2o
mkdir build
cd build
cmake ..
make..
sudo make install

among , The second command is responsible for installing dependencies .

3、 ... and 、 install pcl1.8

download , The order is as follows ：

cd ~
wget https://github.com/PointCloudLibrary/pcl/archive/pcl-1.8.0.tar.gz
tar -xvzf pcl-pcl-1.8.0.tar.gz
cd ~/pcl-pcl-1.8.0
gedit CMakeLists.txt

Add the following to CMakeLists.txt Of the 146 That's ok （ stay endif（） after ）, Add C++11 Support ：

SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11")

compile 、 install , The order is as follows ：

cd ~/pcl-pcl-1.8.0
mkdir build
cd build
cmake ../
make VERBOSE=1
sudo make install

Four 、 To configure rgbdslam_v2

The order is as follows ：

cd ~/catkin_rgbdslam_ws/src
git clone https://github.com/felixendres/rgbdslam_v2
cd rgbdslam_v2
gedit CMakeLists.txt

modify CMakeLists The contents of the document , take

find_package(PCL 1.7 REQUIRED COMPONENTS common io)

Change it to ：

find_package(PCL 1.8 REQUIRED COMPONENTS common io)

And add the following at the lowest end ：

SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11")

modify /opt/ros/kinetic/share/pcl_ros/cmake/pcl_rosConfig.cmake file , Will all /usr/lib/x86_64-linux-gnu/libpcl_ Change the beginning to /usr/local/lib/libpcl_（ A total of 34 individual ）

5、 ... and 、 structure siftgpu library

The order is as follows ：

cd ~/catkin_rgbdslam_ws/src/rgbdslam_v2/external/SiftGPU
sudo apt-get install libglew-dev
sudo apt-get install libdevil1c2 libdevil-dev
make

6、 ... and 、 compile rgbdslam_v2

The order is as follows ：

cd ~/catkin_rgbdslam_ws
catkin_make
cd build/rgbdslam_v2
make VERBOSE=1
make install

If compiled opencv error , According to this blog openCV Step on the pit diary To operate , Then recompile .

7、 ... and 、 test

1、 A simple test

Open a new terminal , function roscore

roscore

Switch to /catkin_rgbdslam_ws/devel/lib/rgbdslam/ Under the table of contents , Open another terminal , Run the following command ：

./rgbdslam

The interface is as follows ：

If this interface displays normally , Then prove rgbdslam Installation successful .

2、 Dataset testing

   Download datasets , Download address ：http://vision.in.tum.de/data/datasets/rgbd-dataset/download#freiburg1_room, Will download okay .bag The documents are stored in catkin_rgbdslam_ws/rgdb_data Under the table of contents .
   modify rgdbslam.launch file , Make sure the topic you subscribe to is consistent with the data set , The modification is as follows ：

<param name="config/topic_image_mono"              value="/camera/rgb/image_color"/> 
<param name="config/topic_image_depth"             value="/camera/depth/image"/>

Open the terminal , Enter the command as follows ：

roscore

Switch to rgdbslam_v2/launch Under the table of contents , Open a new terminal , The order is as follows ：

roslaunch rgbdslam rgbdslam.launch

The interface is as follows ：

Switch to catkin_rgbdslam_ws/rgdb_data Under the table of contents , Open a new terminal , The order is as follows ：

rosbag play rgbd_dataset_freiburg1_room.bag

When the data package starts publishing data , You can see the image data and slam The process of , as follows ：

When the construction is finished , Directly select Save as point cloud data in the menu bar , I keep it in rgdb_data Under the table of contents , The file named quicksave.pcd.

   have access to pcl_ros Function pack view the saved point cloud map ：

rosrun pcl_ros pcd_to_pointcloud quicksave.pcd