Openvslam 学习报告

什么是Openvslam

概念

OpenVSLAM（Visual SLAM）是一个开源的视觉定位与地图构建（Visual Simultaneous Localization and Mapping，Visual SLAM）库。它旨在通过使用单目相机、双目相机、RGB-D相机或者其他传感器，实现实时的相机定位和地图构建。Visual SLAM 是一种技术，可以使机器（例如机器人、车辆或者无人机）在未知环境中通过感知视觉信息，实时地进行定位和建立地图。使开发者能够利用相机捕获的图像数据来定位相机自身并创建环境地图。它结合了特征提取、特征匹配、姿态估计、地图构建等核心技术，以实现准确和实时的相机定位和地图构建。

特点

多传感器的支持
实时性能
开源
易用

安装和运行OpenVSLAM

克隆源代码

git clone https://github.com/lp-research/openvslam

安装依赖库

安装Ubuntu系统依赖

sudo apt update -y
sudo apt upgrade -y --no-install-recommends
# basic dependencies
sudo apt install -y build-essential pkg-config cmake git wget curl unzip
# g2o dependencies
sudo apt install -y libatlas-base-dev libsuitesparse-dev
# OpenCV dependencies
sudo apt install -y libgtk-3-dev
sudo apt install -y ffmpeg
sudo apt install -y libavcodec-dev libavformat-dev libavutil-dev libswscale-dev libavresample-dev
# eigen dependencies
sudo apt install -y gfortran
# other dependencies
sudo apt install -y libyaml-cpp-dev libgoogle-glog-dev libgflags-dev

# (if you plan on using PangolinViewer)
# Pangolin dependencies
sudo apt install -y libglew-dev

安装Eigen矩阵库

cd /Desktop/study/dir
wget -q http://bitbucket.org/eigen/eigen/get/3.3.4.tar.bz2
tar xf 3.3.4.tar.bz2
rm -rf 3.3.4.tar.bz2
cd eigen-eigen-5a0156e40feb
mkdir -p build && cd build
cmake \
    -DCMAKE_BUILD_TYPE=Release \
    -DCMAKE_INSTALL_PREFIX=/usr/local \
    ..
make -j4
sudo make install

安装OpenCV库

cd /Desktop/study/dir
wget -q https://github.com/opencv/opencv/archive/3.4.0.zip
unzip -q 3.4.0.zip
rm -rf 3.4.0.zip
cd opencv-3.4.0
mkdir -p build && cd build
cmake \
    -DCMAKE_BUILD_TYPE=Release \
    -DCMAKE_INSTALL_PREFIX=/usr/local \
    -DENABLE_CXX11=ON \
    -DBUILD_DOCS=OFF \
    -DBUILD_EXAMPLES=OFF \
    -DBUILD_JASPER=OFF \
    -DBUILD_OPENEXR=OFF \
    -DBUILD_PERF_TESTS=OFF \
    -DBUILD_TESTS=OFF \
    -DWITH_EIGEN=ON \
    -DWITH_FFMPEG=ON \
    -DWITH_OPENMP=ON \
    ..
make -j4
sudo make install

安装DBoW2

cd /Desktop/study/dir
git clone https://github.com/shinsumicco/DBoW2.git
cd DBoW2
mkdir build && cd build
cmake \
    -DCMAKE_BUILD_TYPE=Release \
    -DCMAKE_INSTALL_PREFIX=/usr/local \
    ..
make -j4
sudo make install

安装g2o

cd /Desktop/study/dir
git clone https://github.com/RainerKuemmerle/g2o.git
cd g2o
git checkout 9b41a4ea5ade8e1250b9c1b279f3a9c098811b5a
mkdir build && cd build
cmake \
    -DCMAKE_BUILD_TYPE=Release \
    -DCMAKE_INSTALL_PREFIX=/usr/local \
    -DCMAKE_CXX_FLAGS=-std=c++11 \
    -DBUILD_SHARED_LIBS=ON \
    -DBUILD_UNITTESTS=OFF \
    -DBUILD_WITH_MARCH_NATIVE=ON \
    -DG2O_USE_CHOLMOD=OFF \
    -DG2O_USE_CSPARSE=ON \
    -DG2O_USE_OPENGL=OFF \
    -DG2O_USE_OPENMP=ON \
    ..
make -j4
sudo make install

安装Pangolin库

cd /Desktop/study/dir
git clone https://github.com/stevenlovegrove/Pangolin.git
cd Pangolin
git checkout ad8b5f83222291c51b4800d5a5873b0e90a0cf81
mkdir build && cd build
cmake \
    -DCMAKE_BUILD_TYPE=Release \
    -DCMAKE_INSTALL_PREFIX=/usr/local \
    ..
make -j4
sudo make install

编译源码

cd /Desktop/study/openvslam
mkdir build && cd build
cmake \
    -DBUILD_WITH_MARCH_NATIVE=ON \
    -DUSE_PANGOLIN_VIEWER=ON \
    -DUSE_SOCKET_PUBLISHER=OFF \
    -DUSE_STACK_TRACE_LOGGER=ON \
    -DBOW_FRAMEWORK=DBoW2 \
    -DBUILD_TESTS=ON \
    ..
make -j4

测试（环境已经安装成功）

在这里插入图片描述

运行

在这里插入图片描述

运行失败的总结

爆红是我改过代码之后的输出。可能是爆内存了。听取老师的建议后我租了服务器。并且通过查阅资料，发现可以使用Docker一键安装。

运行成功

$ docker pull celinachild/openvslam:latest

$ docker run --gpus all -it --ipc=host --net=host --privileged -e DISPLAY=unix$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix:rw -e NVIDIA_DRIVER_CAPABILITIES=all celinachild/openvslam

# for mapping
$ ./run_video_slam -v ./orb_vocab/orb_vocab.dbow2 -m ./aist_living_lab_1/video.mp4 -c ./aist_living_lab_1/config.yaml --no-sleep --map-db map.msg

# for localization
$ ./run_video_localization -v ./orb_vocab/orb_vocab.dbow2 -m ./aist_living_lab_2/video.mp4 -c ./aist_living_lab_2/config.yaml --no-sleep --map-db map.msg