【无人机自主导航5 SLAM】Intel Realsense T265C双目相机实现ORB-SLAM3

本文主要是介绍【无人机自主导航5 SLAM】Intel Realsense T265C双目相机实现ORB-SLAM3，希望对大家解决编程问题提供一定的参考价值，需要的开发者们随着小编来一起学习吧！

转载自：https://dgzc.ganahe.top/ganahe/2021/wrjzzdhsjirtsmxj.html

【无人机自主导航5 SLAM】Intel Realsense T265C双目相机实现ORB-SLAM3

C/C++ GanAHE 4个月前 (03-21) 32次浏览 已收录 5个评论 扫描二维码

文章目录[隐藏]

• 一、ORB-SLAM3依赖
• 二、ORB-SLAM3配置编译
• 三、编译ORB-SLAM3的ROS接口
• 四、运行ORB-SLAM3
• 补充

深度相机的SDK驱动安装、ROS功能包编译等内容在本站文章：
【无人机自主导航4-1 视觉传感器】通过ROS发布、订阅话题连接视觉相机已有说明，未配置环境的先进行驱动配置。

一、ORB-SLAM3依赖

（一）Pangolin

1. #安装依赖
2. sudo apt install libglew-dev
3. sudo apt install cmake
4. sudo apt install libpython2.7-dev
5. #可选依赖见github，我没装所以就不贴出来啦
6. #下载并编译
7. git clone https://github.com/stevenlovegrove/Pangolin.git
8. cd Pangolin
9. mkdir build
10. cd build
11. cmake ..
12. cmake --build .
13. #最后不要忘了安装
14. sudo make install

（二）其他

根据Github仓库自行安装其他缺少的库。

二、ORB-SLAM3配置编译

1. cd ~/catkin_ws/src
2. git clone https://github.com/UZ-SLAMLab/ORB_SLAM3.git ORB_SLAM3
3. #编译
4. cd ORB_SLAM3
5. chmod +x build.sh
6. ./build.sh
8. cd ..
9. catkin_make
10. source devel/setup.bash

三、编译ORB-SLAM3的ROS接口

1. chmod +x build_ros.sh
2. ./build_ros.sh

3.1 修正相机 launch

根据下方路径自行修改launch文件，包括相机内外参数、启动的镜头、fps等，启动相机时，启动更改后相应的文件即可：

1. ganahe@ganahe-Nitro-AN515-51:~/catkin_ws/src/realsense-ros/realsense2_camera/launch$ pwd
2. /home/ganahe/catkin_ws/src/realsense-ros/realsense2_camera/launch
3. ganahe@ganahe-Nitro-AN515-51:~/catkin_ws/src/realsense-ros/realsense2_camera/launch$ tree
4. .
5. ├── demo_pointcloud.launch
6. ├── demo_t265.launch
7. ├── includes
8. │   └── nodelet.launch.xml
9. ├── opensource_tracking.launch
10. ├── rs_aligned_depth.launch
11. ├── rs_camera.launch
12. ├── rs_d400_and_t265.launch
13. ├── rs_d435_camera_with_model.launch
14. ├── rs_from_file.launch
15. ├── rs_multiple_devices.launch
16. ├── rs_rgbd.launch
17. ├── rs_rtabmap.launch
18. └── rs_t265.launch

本机的参数文件如下：

1. <!--
2. Important Notice: For wheeled robots, odometer input is a requirement for robust
3. and accurate tracking. The relevant APIs will be added to librealsense and
4. ROS/realsense in upcoming releases. Currently, the API is available in the
5. https://github.com/IntelRealSense/librealsense/blob/master/third-party/libtm/libtm/include/TrackingDevice.h#L508-L515.
6. -->
7. <launch>
8. <arg name="serial_no" default=""/>
9. <arg name="usb_port_id" default=""/>
10. <arg name="device_type" default="t265"/>
11. <arg name="json_file_path" default=""/>
12. <arg name="camera" default="camera"/>
13. <arg name="tf_prefix" default="$(arg camera)"/>
15. <arg name="fisheye_width" default="848"/>
16. <arg name="fisheye_height" default="800"/>
17. <arg name="enable_fisheye1" default="true"/>
18. <arg name="enable_fisheye2" default="true"/>
20. <arg name="fisheye_fps" default="30"/>
22. <arg name="gyro_fps" default="200"/>
23. <arg name="accel_fps" default="62"/>
24. <arg name="enable_gyro" default="true"/>
25. <arg name="enable_accel" default="true"/>
26. <arg name="enable_pose" default="true"/>
28. <arg name="enable_sync" default="false"/>
30. <arg name="linear_accel_cov" default="0.01"/>
31. <arg name="initial_reset" default="false"/>
32. <arg name="unite_imu_method" default=""/>
34. <arg name="publish_odom_tf" default="true"/>
36. <group ns="$(arg camera)">
37. <include file="$(find realsense2_camera)/launch/includes/nodelet.launch.xml">
38. <arg name="tf_prefix" value="$(arg tf_prefix)"/>
39. <arg name="serial_no" value="$(arg serial_no)"/>
40. <arg name="usb_port_id" value="$(arg usb_port_id)"/>
41. <arg name="device_type" value="$(arg device_type)"/>
42. <arg name="json_file_path" value="$(arg json_file_path)"/>
44. <arg name="enable_sync" value="$(arg enable_sync)"/>
46. <arg name="fisheye_width" value="$(arg fisheye_width)"/>
47. <arg name="fisheye_height" value="$(arg fisheye_height)"/>
48. <arg name="enable_fisheye1" value="$(arg enable_fisheye1)"/>
49. <arg name="enable_fisheye2" value="$(arg enable_fisheye2)"/>
51. <arg name="fisheye_fps" value="$(arg fisheye_fps)"/>
52. <arg name="gyro_fps" value="$(arg gyro_fps)"/>
53. <arg name="accel_fps" value="$(arg accel_fps)"/>
54. <arg name="enable_gyro" value="$(arg enable_gyro)"/>
55. <arg name="enable_accel" value="$(arg enable_accel)"/>
56. <arg name="enable_pose" value="$(arg enable_pose)"/>
58. <arg name="linear_accel_cov" value="$(arg linear_accel_cov)"/>
59. <arg name="initial_reset" value="$(arg initial_reset)"/>
60. <arg name="unite_imu_method" value="$(arg unite_imu_method)"/>
62. <arg name="publish_odom_tf" value="$(arg publish_odom_tf)"/>
63. </include>
64. </group>
65. </launch>

启动指令如下：

1. roscore

打开新的终端，启动T265C：

1. # 注意在此之前必须在 rs_t265.launch中设定相机的参数，比如是否开启相机
2. roslaunch realsense2_camera rs_t265.launch

3.2 修改Topic

打开相机节点实现通讯：

1. # 注意在此之前必须在 rs_t265.launch中设定相机的参数，比如是否开启相机
2. roslaunch realsense2_camera rs_t265.launch

查看相机镜头对应的话题信息：

1. rostopic list

根据显示的信息修改ros_stereo.cc文件，比如本文中的数据修正为：

1. message_filters::Subscriber<sensor_msgs::Image> left_sub(nh, "/camera/fisheye1/image_raw", 1);
2. message_filters::Subscriber<sensor_msgs::Image> right_sub(nh, "/camera/fisheye2/image_raw", 1);

再次编译./build_ros.sh

3.3 再次编译功能包

随后执行：

1. cd ~/catkin_ws
2. catkin_make
3. source devel/setup.bash

四、运行ORB-SLAM3

三个终端

• 首先执行：
  1. roscore
• 运行相机：

1. roslaunch realsense2_camera rs_t265.launch

• 运行SLAM:

1. rosrun ORB_SLAM3 Stereo ~/catkin_ws/src/ORB_SLAM3/Vocabulary/ORBvoc.txt
3. ~/catkin_ws/src/ORB_SLAM3/Examples/ROS/ORB_SLAM3/T265.yaml false

补充

根据大家的反馈，下面补充本文实验所用的T265.yaml文件，基本采用官方的改动较小：

1. %YAML:1.0
3. #--------------------------------------------------------------------------------------------
4. # Camera Parameters. Adjust them!
5. #--------------------------------------------------------------------------------------------
6. Camera.type: "PinHole"
8. # Camera calibration and distortion parameters (OpenCV)
9. Camera.fx: 286.419189453125
10. Camera.fy: 286.384307861328
11. Camera.cx: 101.355010986328
12. Camera.cy: 102.183197021484
14. Camera.k1: 0.0
15. Camera.k2: 0.0
16. Camera.p1: 0.0
17. Camera.p2: 0.0
19. Camera.width: 848
20. Camera.height: 800
22. # Camera frames per second
23. Camera.fps: 30.0
25. # IR projector baseline times fx (aprox.)
26. Camera.bf: 40.0
28. # Color order of the images (0: BGR, 1: RGB. It is ignored if images are grayscale)
29. Camera.RGB: 1
31. # Close/Far threshold. Baseline times.
32. ThDepth: 40.0
34. # Deptmap values factor
35. DepthMapFactor: 1.0
37. #--------------------------------------------------------------------------------------------
38. # ORB Parameters
39. #--------------------------------------------------------------------------------------------
41. # ORB Extractor: Number of features per image
42. ORBextractor.nFeatures: 1000
44. # ORB Extractor: Scale factor between levels in the scale pyramid
45. ORBextractor.scaleFactor: 1.2
47. # ORB Extractor: Number of levels in the scale pyramid
48. ORBextractor.nLevels: 8
50. # ORB Extractor: Fast threshold
51. # Image is divided in a grid. At each cell FAST are extracted imposing a minimum response.
52. # Firstly we impose iniThFAST. If no corners are detected we impose a lower value minThFAST
53. # You can lower these values if your images have low contrast
54. ORBextractor.iniThFAST: 20
55. ORBextractor.minThFAST: 7
57. #--------------------------------------------------------------------------------------------
58. # Viewer Parameters
59. #--------------------------------------------------------------------------------------------
60. Viewer.KeyFrameSize: 0.05
61. Viewer.KeyFrameLineWidth: 1
62. Viewer.GraphLineWidth: 0.9
63. Viewer.PointSize:2
64. Viewer.CameraSize: 0.08
65. Viewer.CameraLineWidth: 3
66. Viewer.ViewpointX: 0
67. Viewer.ViewpointY: -0.7
68. Viewer.ViewpointZ: -1.8
69. Viewer.ViewpointF: 500

---

GanAHE 的小创元世纪, 版权所有丨如未注明 , 均为原创丨本网站采用@BY-NC-SA协议进行授权
转载请注明原文链接：【无人机自主导航5 SLAM】Intel Realsense T265C双目相机实现ORB-SLAM3

这篇关于【无人机自主导航5 SLAM】Intel Realsense T265C双目相机实现ORB-SLAM3的文章就介绍到这儿，希望我们推荐的文章对编程师们有所帮助！

---