PL-SLAM stvo-pl-ros 使用OAK双目相机实时运行
下载编译PL-StVO
The repository focuses on implementing stereo visual odometry through the integration of point-based and line segment-based feature matching. (https://github.com/rubengooj/stvo-pl?tab=readme-ov-file#pl-stvo "GitHub - rubengooj/stvo-pl: Stereo Visual Odometry by combining point and line segment features")
2.下载编译stvo-pl-ros
完成项目的下载后创建一个新的src文件夹,并将stvopl_node.cpp、package.xml以及CMakeLists.txt放置到新创建的src文件夹中。随后
cd src
catkin_make
3.在编译之前需要打开 stvopl_node.cpp修改几个路径
初始化场景对象
// 接收输入的视頻流并输出到外部的视頻流
img_sub_l_ = std::shared_ptr<Message_filters::Subscriber<sensor_msgs::Image>>(
new message_filters::Subscriber<sensor_msgs::Image>(
nh_, "/stereoPublisher/left/image", 1) );#167行, 建议将此功能归为个人使用
img_sub_r_ = std::shared_ptr<Message_filters::Subscriber<sensor_msgs::Image>>(
new message_filters::Subscriber<sensor_msgs::Image>(
nh_, "/stereoPublisher/right/image", 1) );#171行, 建议将此功能归为个人使用
4.修改完之后编译即可
在实际编写过程中,在所编写的dataset_params.yaml文件中遵循了Kitti数据集的标准格式
cam 0:
cam_cx = 666.816 , 按照相机标定结果确定
cam_cy = 3 , 按照相机标定结果确定
cam_d : [全部设为零]
rx = ry = rz = [设为零]
tx = ty = tz = [设为零]images_subfolder_l : image_ /
images_subfolder_r : image_ /
6. 进入OAK相机工作区(dai_ws)发布双目信息
7.进入build文件夹下
./vo /home/lzd/study_example/PL-VO/stvo-pl/config/dataset_params/dataset_params.yaml /home/lzd/study_example/PL-VO/stvo-pl/config/config/config_kitti.yaml
指令中的路径按照实际填写

