Kinect2.0+ORBSLAM2_ with_ pointcloud_ map

---

One 、kinect2.0 drive

1.libfreenect2

git clone https://github.com/OpenKinect/libfreenect2.git
cd libfreenect2
cd depends; ./download_debs_trusty.sh // download upgrade deb  file 
sudo apt-get install build-essential cmake pkg-config// Install the compiler 
sudo dpkg -i debs/libusb*deb // install libsub
sudo apt-get install libturbojpeg libjpeg-turbo8-dev// install TurboJPEG
sudo dpkg -i debs/libglfw3*deb// install OpenGL
sudo apt-get install -f
sudo apt-get install libgl1-mesa-dri-lts-vivid

//Build
cd ..
mkdir build && cd build
cmake .. 
make
make install

Set up udev rules: sudo cp …/platform/linux/udev/90-kinect2.rules /etc/udev/rules.d/, Then plug and unplug again Kinect2

Now you can try to run Demo Program ：./bin/Protonect
You should see that kinect Camera picture

2.iai-kinect2

from Github Download the project source code to the workspace src In the folder

cd ~/catkin_ws/src/
git clone https://github.com/code-iai/iai_kinect2.git
cd iai_kinect2
rosdep install -r --from-paths .
cd ~/catkin_ws
catkin_make -DCMAKE_BUILD_TYPE="Release"

Be careful
For the last line of the above command , It should be noted that , If the previous libfreenect2 Your installation location is not under the standard two paths , Parameter specification is required libfreenect2 Location path :

catkin_make -Dfreenect2_DIR=path_to_freenect2/lib/cmake/freenect2 -DCMAKE_BUILD_TYPE="Release"

End of compilation , Can be in ROS In order to get Kinect2 The data of , If the command cannot be executed , The corresponding environment should be written to ～/.bashrc file .

roslaunch kinect2_bridge kinect2_bridge.launch

roslaunch Initiate related nodes , Another terminal input rostopic list You can view related Topic, You can also enter rosrun kinect2_viewer kinect2_viewer sd cloud , Turn on Viewer View the data

Two 、ORBSLAM2_with_pointcloud_map

1. install Pangolin

Dependencies ：

sudo apt-get install libglew-dev
sudo apt-get install ffmpeg libavcodec-dev libavutil-dev libavformat-dev libswscale-dev libavdevice-dev
sudo apt-get install libdc1394-22-dev libraw1394-dev
sudo apt-get install libjpeg-dev libpng12-dev libtiff5-dev libopenexr-dev

Source code compilation and installation

git clone https://github.com/stevenlovegrove/Pangolin.git
cd Pangolin
mkdir build
cd build
cmake ..
cmake --build .
sudo make install

2.ORBSLAM2_with_pointcloud_map

git clone https://github.com/gaoxiang12/ORBSLAM2_with_pointcloud_map.git

decompression orbslam2_modified.zip, You'll get two folders , Respectively g2o_with_orbslam2 and ORB_SLAM2_modified

compile g2o
Install dependencies

sudo apt-get intall libsuitesparse-dev qtdeclarative5-dev qt5-default libqglviewer-dev

For later revision, please refer to the blog ：

https://blog.csdn.net/qq_25349629/article/details/88350374

function rgbd_tum, View drawing creation based on dataset ：
The instructions I run are shown in the following figure

The point cloud situation is as follows ：

3. Using one's own Kinect Real time mapping

Reference blog ：

https://blog.csdn.net/qq_25349629/article/details/88528765

According to the blogger's compilation, it passed , But in the end, there was a problem

ERROR While loading shared libraries:cannot open shared object file

I do have some dishes …

Because you can't use your own kinect Real time running , So I plan to make my own data set , Use the data set to build the map , Then there are the following pictures ：

Run with your own data set , I don't think the effect is very good , May be related to the recorded dataset , The point cloud is very sparse .

Prove again in real time ： You are still very good …

summary

Problems are always encountered , Of course, we should try our best to solve it , Various SLAM In fact, there are a lot of related bags , Today's temporary rise is a simple summary of something related ORB Of , Of course, the underlying code is the one that needs to be explored , Look at the code tomorrow , I hope to summarize some code related problems later .