Nvidia Jetson TX2+Intel Realsense D435i跑ORB_SLAM3

柔情只为你懂 2023-10-01 10:32 80阅读 0赞

前言：网上的教程实在是太多，从诸多教程中找到一个适合自己的实属不易。将此记录下来，希望能够帮助到有需要的人。因为时间紧迫，没时间写特别详细的内容，只能引用一些他人的步骤。请见谅！不过经过亲测，过程是完整的。

本机环境：Nvidia Jetson TX2+ROS Melodic+Intel Realsense D435i

**注意：ORB\_SLAM3是可以不用ROS的，但本篇文章是基于ROS使用的**

基本步骤如下：  
  1.TX2刷机  
  2.下载ROS Melodic  
  3.下载Realsense D435i的SDK及ros\_realsense  
  4.配置ORB\_SLAM3环境及编译ORB\_SLAM3  
  5.相机标定  
  6.修改源文件，并运行

### 1.TX2刷机 ###

详细步骤请参考：[https://blog.csdn.net/gls\_nuaa/article/details/122050243][https_blog.csdn.net_gls_nuaa_article_details_122050243]

### 2.下载Melodic ###

详细步骤请参考：[https://blog.csdn.net/gls\_nuaa/article/details/122101482][https_blog.csdn.net_gls_nuaa_article_details_122101482]

### 3.下载Realsense D435i的SDK及ros\_realsense ###

注意：因为TX2是ARM核，而realsense官网上对应的是x86，所以不能按照官网上来，realsense专门针对于nvidia的开发设置了jetsonhacks，弄完之后更方便啦~  
安装SDK步骤如下：

sudo apt-get install libudev-dev pkg-config libgtk-3-dev
    sudo apt-get install libusb-1.0-0-dev pkg-config
    sudo apt-get install libglfw3-dev
    sudo apt-get install libssl-dev
    git clone https://github.com/jetsonhacks/installRealSenseSDK.git
    sudo chmod u+x ./installLibrealsense.sh
    ./installLibrealsense.sh

能运行realsense-viewer就说明成功了

安装realsense\_ros步骤如下：

创建工作空间catkin_ws
    git clone https://github.com/IntelRealSense/realsense-ros.git
    git clone https://github.com/pal-robotics/ddynamic_reconfigure.git
    cd ~/catkin_ws && catkin_make

参考：[https://blog.csdn.net/wm3vdy88/article/details/109184411][https_blog.csdn.net_wm3vdy88_article_details_109184411]

### 4.配置ORB\_SLAM3环境及编译ORB\_SLAM3 ###

配置ORB\_SLAM3的环境要安装一些东西，包括(opencv3.2, Pangolin, Eigen3, boost)

**安装opencv3.2**  
详细步骤参考：[https://my.oschina.net/u/4347613/blog/4627699][https_my.oschina.net_u_4347613_blog_4627699]

编译时opencv3.2时报错：/usr/include/c++/7/cstdlib:75:15: fatal error: stdlib.h: No such file or directory \#include\_next <s  
解决：[https://blog.csdn.net/u010003609/article/details/100086151][https_blog.csdn.net_u010003609_article_details_100086151]

编译opencv时可能会报很多依赖项的错误，在上面的博文中也有描述（好长一段那个）。这个其实取决于你刷机的时候是否刷入了cuda，如果没刷cuda就不会报这个错的。问题的解决可以参考：[http://blog.csdn.net/u014613745/article/details/78310916][http_blog.csdn.net_u014613745_article_details_78310916]

再留一个tx2安装Opencv的教程吧：[https://www.codeleading.com/article/65432877315/][https_www.codeleading.com_article_65432877315]

**Pangolin安装教程**  
详细步骤参考：[https://blog.csdn.net/u012986684/article/details/52860849][https_blog.csdn.net_u012986684_article_details_52860849]

**Eigen3安装教程**  
详细步骤参考：[https://www.cnblogs.com/newneul/p/8256803.html][https_www.cnblogs.com_newneul_p_8256803.html]

**boost安装，命令如下**：

sudo apt-get install libboost-all-dev

**编译ORB\_SLAM3**：

git clone https://github.com/UZ-SLAMLab/ORB_SLAM3.git
    cd ORB_SLAM3
    chmod +x build.sh
    ./build.sh
    #安装ros版
    chmod +x build_ros.sh
    ./build_ros.sh

执行build.sh报错：recipe for target ‘CMakeFiles/ORB\_SLAM3.dir/src/LocalMapping.cc.o‘ failed  
解决办法：[https://blog.csdn.net/weixin\_48464886/article/details/118970304][https_blog.csdn.net_weixin_48464886_article_details_118970304]

执行build\_ros.sh会报错(关于sophus和类型转换)，[解决方法在这里][Link 1]

### 5.相机标定 ###

这里提供一种简单的标定方法，具体见该文章第三步：[https://blog.csdn.net/gls\_nuaa/article/details/122104557][https_blog.csdn.net_gls_nuaa_article_details_122104557]

### 6.修改源文件并运行 ###

修改源文件详细步骤参考，修改的内容完全一致：[https://blog.csdn.net/gls\_nuaa/article/details/122104557][https_blog.csdn.net_gls_nuaa_article_details_122104557]

执行命令如下(2,3顺序无所谓)：

roscore
    rosrun ORB_SLAM3 Mono /home/jetson1/test_ORB_SLAM3/src/ORB_SLAM3/Vocabulary/ORBvoc.txt /home/jetson1/test_ORB_SLAM3/src/ORB_SLAM3/Examples/ROS/ORB_SLAM3/Asus.yaml
    roslaunch realsense2_camera rs_camera.launch

过程中遇到两个报错，记录一下：  
报错：CMake Error at /opt/ros/melodic/share/cv\_bridge/cmake/cv\_bridgeConfig.cmake:113 (message)  
解决办法：[https://blog.csdn.net/woodgril/article/details/109155073][https_blog.csdn.net_woodgril_article_details_109155073]  
要根据实际情况做文件上的修改

报错：UBUNTU18.04环境下ROS编译ORB\_SLAM2时遇到问题：\[ROSBUILD\] ROSPACK FOUND PACKAGE “ORB\_SLAM2” AT “”…  
解决：[https://www.freesion.com/article/7597639275/][https_www.freesion.com_article_7597639275]

[https_blog.csdn.net_gls_nuaa_article_details_122050243]: https://blog.csdn.net/gls_nuaa/article/details/122050243
[https_blog.csdn.net_gls_nuaa_article_details_122101482]: https://blog.csdn.net/gls_nuaa/article/details/122101482
[https_blog.csdn.net_wm3vdy88_article_details_109184411]: https://blog.csdn.net/wm3vdy88/article/details/109184411
[https_my.oschina.net_u_4347613_blog_4627699]: https://my.oschina.net/u/4347613/blog/4627699
[https_blog.csdn.net_u010003609_article_details_100086151]: https://blog.csdn.net/u010003609/article/details/100086151
[http_blog.csdn.net_u014613745_article_details_78310916]: http://blog.csdn.net/u014613745/article/details/78310916
[https_www.codeleading.com_article_65432877315]: https://www.codeleading.com/article/65432877315/
[https_blog.csdn.net_u012986684_article_details_52860849]: https://blog.csdn.net/u012986684/article/details/52860849
[https_www.cnblogs.com_newneul_p_8256803.html]: https://www.cnblogs.com/newneul/p/8256803.html
[https_blog.csdn.net_weixin_48464886_article_details_118970304]: https://blog.csdn.net/weixin_48464886/article/details/118970304
[Link 1]: https://github.com/UZ-SLAMLab/ORB_SLAM3/issues/442
[https_blog.csdn.net_gls_nuaa_article_details_122104557]: https://blog.csdn.net/gls_nuaa/article/details/122104557
[https_blog.csdn.net_woodgril_article_details_109155073]: https://blog.csdn.net/woodgril/article/details/109155073
[https_www.freesion.com_article_7597639275]: https://www.freesion.com/article/7597639275/