The previous article describes how to use a dataset to run Lsd-slam, this article describes how to run Lsd-slam with a normal USB camera, the default Ubuntu12.04,fuerte has been installed, Workspace has also been set, if not, Please refer to the post of Lsd-slam running in the previous data set.
The camera I use is a Logitech C310 USB camera, and the other USB cameras should be fine.
1. Test the camera for good or bad. Install cheese, perform
$ sudo apt-get install cheese
Run cheese, execute
$ cheese
Confirm that the camera works correctly under Ubuntu
2. Install the corresponding driver under Ros. Perform
$ sudo apt-get install ROS-FUERTE-CAMERA-UMD
Then execute
$ ROSCD Uvc_camera
If you are in the Uvc_camera folder now, the installation is successful.
3. Find the device number that corresponds to the camera. Perform
$ ls/dev/video*
If there's only one camera, you should see/dev/video0.
4. Change the launch file. Enter the launch folder under Uvc_camera, open the Camera_node.launch file, and change to
<launch><arg name= "Device" default= "/dev/video0"/> <node pkg= "Uvc_camera" type= "Camera_node" Name= "Uvc_camera" output= "screen" > <param name= "width" type= "int" value= "640"/> <param name= " Height "type=" int "value=" 480 "/> <param name=" fps "type=" int "value=" "/> <param name= " Frame " Type= "string" value= "Wide_stereo"/> <param name= "Device" type= "string" value= "/dev/video0"/> < /node></launch>
If you are not at ease, you can back up the original file, which/dev/video0 this to be with their previous query to the corresponding
5. Implementation
$ roscore
If there is already a shell executing this command, executing it again will give an error, allowing only one roscore to be executed.
6. Open a new terminal and execute
$rosrun Lsd_slam_viewer Viewer
The Point Cloud Viewer appears
7. Open a new terminal and execute
$ roslaunch Uvc_camera Camera_node.launch
There may be errors in the middle, no tubes. At this point, if the camera has a work light, the indicator light should be light
8. Open a new terminal. Perform
$ rosrun Lsd_slam_core live_slam/image:=<yourstreamtopic> _calib:=<calibration_file>
Which yourstreamtopic I do not understand, Calibration_file is the correct file, I directly use the data set of the correction file (there may be a problem with the principle, but can run through). It's mine
$ rosrun Lsd_slam_core Live_slam/image:=image_raw _calib:=/home/xtl/fuerte_workspace/sandbox/lsd_room_images/lsd_ Room/cameracalibration.cfg
Results
Point cloud is displayed in the Dot Cloud Viewer:
Ubuntu12.04+fuerte running through lsd-slam--using a USB camera