ORB-SLAM2报错集合(数据集测试系列1）

字梦笙

已于 2023-08-24 21:24:12 修改

阅读量742

点赞数

分类专栏： slam学习分享文章标签： slam ubuntu 计算机视觉

于 2023-08-23 16:04:03 首次发布

本文链接：https://blog.csdn.net/mengshenglo/article/details/132453332

版权

slam学习分享专栏收录该内容

7 篇文章 1 订阅

订阅专栏

错误1

报错1：FATAL_ERROR "OpenCV > 2.4.3 not found."的错误
或者 No rule to make target ‘…/Thirdparty/DBoW2/lib/libDBoW2.so’, needed by ‘…/lib/libORB_SLAM2.so’. Stop.

解决：将/ORB_SLAM2_detailed_comments/CMakeLists.txt
和/ORB_SLAM2_detailed_comments/Thirdparty/DBoW2/CMakeLists.txt中的opencv版本改成自己对应的版本

即

# find_package(OpenCV 3.0 QUIET)
# if(NOT OpenCV_FOUND)
#    find_package(OpenCV 2.4.3 QUIET)
#    if(NOT OpenCV_FOUND)
#       message(FATAL_ERROR "OpenCV > 2.4.3 not found.")
#    endif()
# endif()

find_package(OpenCV REQUIRED)
include_directories(${OpenCV_INCLUDE_DIRS})

错误2

报错2：error: ‘CV_LOAD_IMAGE_UNCHANGED’ was not declared in this scope

解决：把Examples文件夹里.cc文件所有的这个名改成cv::IMREAD_UNCHANGED
（因为opencv新版本变量名换了）

错误3

EIGEN_DEPRECATED const unsigned int AlignedBit = 0x80（可以不用管，只是警告而已）
或者Pangolin could not be found because dependency Eigen3 could not be found.
（这两个就是pangolin和Eigen的版本问题）

解决：
关于pangolin v0.5和Eigen 3.2.1以上的版本的卸载与安装（如果装了3.2.1，后续catkin_make会报错，建议3.4.0）
卸载pangolin库

locate pangolin

sudo rm -rf  XXX的路径

sudo updatedb

知道locate pangolin没有输出路径

安装pangolin v0.5库
https://github.com/stevenlovegrove/Pangolin/releases/tag/v0.5

到Pagolin/src文件夹下，在CMakeList.txt中注释掉下面这段代码（屏蔽掉ffmpeg的影响）：

# find_package(FFMPEG QUIET)
# if(BUILD_PANGOLIN_VIDEO AND FFMPEG_FOUND)
#   set(HAVE_FFMPEG 1)
#   list(APPEND INTERNAL_INC  ${FFMPEG_INCLUDE_DIRS} )
#   list(APPEND LINK_LIBS ${FFMPEG_LIBRARIES} )
#   list(APPEND HEADERS ${INCDIR}/video/drivers/ffmpeg.h)
#   list(APPEND SOURCES video/drivers/ffmpeg.cpp)
#   message(STATUS "ffmpeg Found and Enabled")
# endif()


cd Pangolin
mkdir build
cd build

cmake ..
make -j12

sudo make install

由于ORB-SLAM2需要低版本的Eigen(最好3.2.1以上)，不然会报错EIGEN_DEPRECATED const unsigned int AlignedBit = 0x80;

查看自己Eigen版本
pkg-config --modversion eigen3

卸载
locate eigen3

sudo rm -rf /usr/include/eigen3
sudo rm -rf /usr/lib/cmake/eigen3
sudo rm -rf /usr/local/include/eigen3
sudo rm -rf /usr/share/doc/libeigen3-dev 
sudo rm -rf /usr/local/share/pkgconfig/eigen3.pc /usr/share/pkgconfig/eigen3.pc /var/lib/dpkg/info/libeigen3-dev.list /var/lib/dpkg/info/libeigen3-dev.md5sums
sudo rm -rf /usr/local/lib/pkgconfig/eigen3.pc
sudo rm -rf /usr/local/share/eigen3

查看是否删除彻底

sudo updatedb
locate eigen3
pkg-config --modversion eigen3

Eigen3.4.0的安装
https://gitlab.com/libeigen/eigen/-/releases

找到3.4.0下载tar.gz

mkdir build
cd build
cmake..
sudo make install
sudo cp -r /usr/local/include/eigen3 /usr/include

# 检测当前版本
pkg-config --modversion eigen3

错误4

“opencv/cv.h not found”
解决
需要将ORBextractor.h改成：

#include <opencv2/opencv.hpp>

错误5

“orb_slam2/src/PnPsolver.cc:860:37: error: ‘CvMat’ was not declared in this scope”

也是因为opencv新版本的没有继承老版本的变量导致

解决：在System.h, PnPSolver.h, Sim3Solver.h 中都加入

#include <opencv2/imgproc/types_c.h>
#include <opencv2/opencv.hpp>
using namespace cv;

错误6

报错： static assertion failed: std::map must have the same value_type as its allocator
解决：修改/ORB_SLAM2_detailed_comments/include/LoopClosing.h::50，

typedef map<KeyFrame*,g2o::Sim3,std::less<KeyFrame*>,
Eigen::aligned_allocator<std::pair<const KeyFrame*, g2o::Sim3> > > KeyFrameAndPose;
修改为

typedef map<KeyFrame*,g2o::Sim3,std::less<KeyFrame*>,
    Eigen::aligned_allocator<std::pair<KeyFrame* const, g2o::Sim3> > > KeyFrameAndPose;

错误7

orb rror: ‘usleep’ was not declared in this scope

解决：
在System.h里添加

#include <unistd.h>

错误8

LoopClosing.cc
mnFullBAIdx++;

解决：

        // mnFullBAIdx++;
        if (mnFullBAIdx==0)
        {
            mnFullBAIdx=1;
        }

最后在/ORB_SLAM2_detailed_comments目录下

 ./build.sh

一堆warning没事，这是eigen低版本输出的信息而已，最后可以100%的。

数据集的话，三个都可以测试，其他测试类型可以参考github:GitHub - electech6/ORB_SLAM2_detailed_comments: Detailed comments for ORB-SLAM2 with trouble-shooting, key formula derivation, and diagrammatic drawing

下面进行测试

下面是在/ORB_SLAM2_detailed_comments终端下执行，凡是有mengs的，就是自己数据集的路径

TUM-RGBD测试

./Examples/Monocular/mono_tum Vocabulary/ORBvoc.txt Examples/Monocular/TUM1.yaml /media/mengs/SSD/ubuntu-dateset/slam常用数据集/TUM-RGBD/rgbd_dataset_freiburg1_xyz

KITTI测试

./Examples/Monocular/mono_kitti Vocabulary/ORBvoc.txt Examples/Monocular/KITTI00-02.yaml /media/mengs/SSD/ubuntu-dateset/slam常用数据集/KITTI/data_odometry_gray/00

EuRoC测试

./Examples/Stereo/stereo_euroc Vocabulary/ORBvoc.txt Examples/Stereo/EuRoC.yaml /media/mengs/SSD/ubuntu-dateset/slam常用数据集/EuRoC/MH_01_easy/mav0/cam0/data /media/mengs/SSD/ubuntu-dateset/slam常用数据集/EuRoC/MH_01_easy/mav0/cam1/data Examples/Stereo/EuRoC_TimeStamps/MH01.txt

其他错误

/usr/bin/ld: warning: libopencv_imgcodecs.so.4.2, needed by /opt/ros/noetic/lib/libcv_bridge.so, may conflict with libopencv_imgcodecs.so.3.4

或者

rospack found package “ORB_SLAM2” at “”, but the current directory is
“/home/zhangjian/catkin_ws/src/ORB_SLAM2/Examples/ROS/ORB_SLAM2”. You should
double-check your ROS_PACKAGE_PATH to ensure that packages are found in the correct
precedence order.

参考链接

使用ROS编译ORB-SLAM2时遇到的问题（已解决）_名字真的难想的博客-CSDN博客

报错

_ZNK2cv8FileNodecviEv' /usr/bin/ld: /lib/x86_64-linux-gnu/libopencv_core.so.4.2:

参考链接

编译工程需要Opencv3 与 ROS自带Opencv4冲突解决办法_秃头队长的博客-CSDN博客

报错

No rule to make target '/usr/lib/x86_64-linux-gnu/libIlmImf.so', needed by '../lib/libORB_SLAM2.so'. Stop.
CMakeFiles/Makefile2:178: recipe for target 'CMakeFiles/ORB_SLAM2.dir/all' failed

参考

sudo apt-get install libopenexr-dev

opencv卸载

[详细教程] opencv的完全卸载和重新安装--3.x版本与4.x版本_opencv卸载重装_HeyMountain的博客-CSDN博客

opencv3.4.11安装

Ubuntu 20.04 下安装 OpenCV 3.4.11 并测试使用_ubuntu opencv测试_罐装三斤的博客-CSDN博客

ROS环境下

ros由于个人电脑或者opencv版本的原因，即使./build_ros.sh成功，但运行rosrun ORB_SLAM2 Mono ~/catkin_ws/src/ORB_SLAM2/Vocabulary/ORBvoc.txt ~/catkin_ws/src/ORB_SLAM2/Examples/ROS/ORB_SLAM2/Asus.yaml会立即闪退，换了opencv3,比较容易的版本是3.4.11，但却始终没法通过./build_ros.sh。所以自己又装回了4.7.0，后续ORB-SLAM3可能得一阵子才更新。

正常电脑参考下面这个链接即可

最完整---在ROS用自己电脑的摄像头在运行ORB_SLAM2_orb-slam2 ros_一大块肉松的博客-CSDN博客

2023.08.24