Can't execute "ros2 run realsense2_camera realsense2_camera_node" #3158
Comments
|
Hi @drinkBr That rgb_camera.power_line_frequency message can usually be ignored. The power_line_frequency parameter is usually configured in situations where you have fluorescent lights on the ceiling that are interfering with the image (because they flicker at frequencies that are hard to see with the human eye) and you need to change its value to reduce the interference. Do you experience the same problems if you launch with ros2 launch instead of ros2 run, please?
|
|
I executed |
|
xioctl(UVCIOC_CTRL_QUERY) errors are typically related to a conflict between librealsense and the Linux kernel. Such conflicts may be avoided if librealsense is built from source code with CMake with the flag -DFORCE_RSUSB_BACKEND=TRUE included in the CMake build instruction. Which method did you use to install librealsense on your Pi, please? |
I understand how to avoid the error, but I don't know exactly what actions to take. Since I am a beginner with Ubuntu, I would appreciate a simplified explanation.
The specific steps are as follows. I have also included the URL I referenced. -install librealsense2- -LibUVC-backend installation- Reference Thanks. |
|
To build librealsense from source code in RSUSB = true mode, you need additional flags included in your CMake build command. At the moment you are using this:
This command will not build librealsense in RSUSB mode though. To do so, please try the build command below.
This command comes from my RSUSB installation guide at IntelRealSense/librealsense#9931 (comment) I note that after performing the CMake build, you then used the libuvc backend script. But the CMake build process will have already installed librealsense, so doing a libuvc backend build afterwards means that you could have two separate installations of librealsense on the same computer that could potentially conflict with each other. |
I am currently working on this, so I will report back once it's finished. I have a separate question. I want to obtain RGB and depth images from Realsense based on key inputs. However, I read on another page that images are obtained from the topic, which made me wonder. Since it is not possible to perform key input and image subscription at the same time, I think this might be difficult. Is there any solution to this? I would like to hear your opinion. |
|
Some RealSense Python users write a script that subscribes to ROS topics and has the script use those as the camera data. The link below has an example of doing so. https://stackoverflow.com/questions/62938146/getting-realsense-depth-frame-in-ros The RealSense ROS wrapper also provides some example ROS2 'node scripts' in Python language. https://github.com/IntelRealSense/realsense-ros/tree/ros2-master/realsense2_camera/scripts |
I tested two methods for this. After executing this, I immediately ran realsense-viewer and was able to confirm both RGB and Depth images. Then, when I ran When I checked realsense-viewer, I could no longer get the Depth image, and the following error occurred. Second, As before, when I ran realsense-viewer, I was able to confirm both RGB and Depth images. However, when I tried executing ros2 run realsense2_camera realsense2_camera_node, the same error appeared without success. After that, I checked realsense-viewer again, and I could no longer get the Depth image, and the same error as before was displayed. Are the commands I entered correct? I would like to hear your opinion. |
|
The commands that you entered are correct. If your librealsense version is 2.55.1 though then the firmware version installed in your camera should be the current latest 5.16.0.1 version instead of the 5.13.0.55 currently installed. |
|
Hi @drinkBr Do you require further assistance with this case, please? Thanks! |
|
Sorry for the late reply.
I don't have a Raspberry Pi at hand right now, so I can't verify this. I will send another message later this week. |
|
You're very welcome. I look forward to your next update. Good luck! |
I have verified this, and when I checked with Thanks. |
|
Do you have access to any other computers other than your Pi that you could attempt a firmware update on? |
|
Yes, I have a Windows PC. |
|
The firmware driver is stored inside the camera hardware and not on the computer. So if you are able to install the RealSense Viewer tool on your Windows computer and update the firmware in it then the updated firmware will be carried back to the Pi inside the camera. You can download and run the Viewer on its own on Windows using the Intel.RealSense.Viewer.exe file in the Assets file list at the link below. https://github.com/IntelRealSense/librealsense/releases/tag/v2.55.1
|
|
I installed RealSense Viewer on my Windows computer and verified it. As a result, I was able to confirm both RGB and Depth images. It seems that the firmware is installed correctly. There was a warning message that concerns me, so I would like to hear your opinion on this. |
|
There is no harm at all in clicking the Enable button to enable frame metadata and I would recommend that you do so. What this message means is that when frame metadata is disabled, some types of camera metadata information will not be available, and the metadata that is provided will be 'software metadata' that has timing based on the computer's clock instead of the camera's firmware (which provides 'hardware metadata'). RGB and depth streaming can still work normally without frame metadata support being enabled. But when it is so easy to enable on Windows in the RealSense Viewer, there is no reason not to enable it. |
|
While using realsense-viewer on the Raspberry Pi, after an error occurred with the Stereo Module, I turned on the RGB Camera and the following error was observed. When using realsense-viewer on the Raspberry Pi, the performance is very slow. Could this be one of the causes of the error? When I started realsense-viewer and first activated the Stereo Module, I was able to see the depth information. However, I couldn't see the RGB information. In addition to the slow performance, the fact that there is only 6GB of free storage might also be affecting it. |
|
RealSense cameras commonly have performance issues on Raspberry Pi boards. The same errno=5 Last Error: Input/output error occurred in a Pi case at #2991 xioctl(VIDIOC_S_FMT) errors are typically related to a conflict between librealsense and the kernel. The libuvc backend SDK installation method that you used earlier works best with Pi, as it enables librealsense to bypass the kernel. |
I am using the libuvc backend SDK installation method, but I am encountering an xioctl(VIDIOC_S_FMT) error. What should I do? Should I check for a conflict between librealsense and the kernel? |
|
I have one more question. Currently, I am using a 32GB microSD on the Raspberry Pi, but should I consider using a larger capacity microSD for Ubuntu? |
|
There is not a way to tell if there is a possible kernel conflict with librealsense other than when xioctl(VIDIOC_S_FMT) errors occur. I would think that the access speed of the microSD card will be more important to performance than the capacity if the Ubuntu OS is being run off the card. An SSD drive would give faster drive access. https://thepihut.com/collections/raspberry-pi-ssd-storage Another limiting factor on performance would be the real memory capacity of your Pi. Once its real memory is used up, it will start using the virtual memory of a swapfile, which converts free storage space on the drive into virtual memory that is slower than the real memory on the Pi board. |
|
Sorry for the late reply. I am currently trying to set up a ROS2 environment on a USB-connected SSD that I have. I will send another message later this week. |
|
It's no problem at all. I look forward to your next report. Thanks very much for the update! |
|
Hello, it's been a while. Sorry for the delay in getting back to you. I have installed Ubuntu 22.04 on an SSD and set up the ROS2 environment. Before I proceed with the installation for RealSense, I would like to confirm the installation steps. Are the following steps correct? -install dependencies- -install librealsense2- -LibUVC-backend installation- Thanks. |
|
The above instructions are correct for the most part. However, if you install dependencies and then install librealsense2 with git clone then you do not have to use the libuvc backend installation instructions. The libuvc backend is a different, much simpler method of installing librealsense with a build script. You do not need to follow the prerequisites instructions with that method, as the libuvc method takes care of installing the prerequisites by using the internet connection instead. However, you will need to edit line 46 of the libuvc_installation.sh script to add flags to the CMake build instruction. https://github.com/IntelRealSense/librealsense/blob/master/scripts/libuvc_installation.sh#L46 Change this:
To this:
|
Since you previously mentioned that using the libuvc backend SDK installation method would be good, I would like to proceed with that. Could you provide the necessary commands for using the libuvc backend SDK installation method in the same format as my previous comment? When using that method, is there any difference compared to installing dependencies and then using git clone to install librealsense2? If I install dependencies first and then use git clone to install librealsense2, is it correct to assume that the steps I outlined in my previous comment for "-install dependencies-" and "-install librealsense2-" are sufficient? I apologize for the many questions, and I appreciate your assistance. |
|
When using the libuvc method, make sure that you have an active internet connection and then unplug the camera, open a terminal and input the short list of instructions below. When the message Librealsense script completed is displayed, plug the RealSense camera in and run rs-enumerate-devices to test the librealsense installation. When using the libuvc backend method instead of installing dependencies and using git clone, the key differences with libuvc are:
Your instructions for installing dependencies and then git cloning are fine, though when the flag |
|
Hi @MiTomMi Does performance in the RealSense ROS wrapper on your Pi 4 improve if you launch with 640x480 and 15 FPS for both streams?
|
|
Hi @MartyG-RealSense, thank you for your reply. I have tried the command:
However, the same error occurred on the Pi 4: [realsense2_camera_node-1] [ERROR] [1723363115.466008305] [camera.camera]: ./src/rs_node_setup.cpp:487:An exception has been thrown: std::exception The error has been driving me crazy for several days. At first I thought the problem came from the Raspberry Pi 3B which does not have USB 3.0 ports, but when I upgraded to Pi 4 the problem still happens. I am looking forward to your advice. I appreciate your help a lot! |
|
UPDATE: pi@raspberrypi4:~$ ros2 launch realsense2_camera rs_launch.py depth_module.depth_profile:=640x480x15 rgb_camera.color_profile:=640x480x15
However, the same error showed up when I started over the command again. [realsense2_camera_node-1] [ERROR] [1723363115.466008305] [camera.camera]: ./src/rs_node_setup.cpp:487:An exception has been thrown: std::exception Maybe there are some Intel Realsense's firmware conflicts. Here is the firmware of the camera. pi@raspberrypi3:~/robot_intel$ rs-fw-update -l Connected devices:
Thank you! |
Is there a specific method to achieve this? Thanks! |
|
Hi @drinkBr Do you mean how to avoid having pointclouds, alignment and post-processing active so that processing burden can be reduced? If so then all of these are disabled by default so you do not have to do anything to turn them off. |
|
@MiTomMi If you are using firmware 5.16.0.1 then librealsense 2.55.1 and ROS2 wrapper 4.55.1 should be used with it. Are those the versions that you are using, please? |
|
Hi @MartyG-RealSense, I am using ROS2 Wrapper 4.55.1. pi@raspberrypi4:~$ ros2 pkg xml realsense2_camera | grep '' |
|
@MiTomMi Does the problem after camera restart still occur if you add initial_reset:=true to your launch instruction to reset the camera at launch? |
|
Hi, @MartyG-RealSense |
|
@MartyG-RealSense, the problem still occured. Here is the full log: pi@raspberrypi4:~$ ros2 launch realsense2_camera rs_launch.py depth_module.depth_profile:=640x480x15 rgb_camera.color_profile:=640x480x15 initial_reset:=true I'm sorry drinkBr for causing inconveniences with long comments. |
|
@MartyG-RealSense, I don't have a Raspberry Pi at hand right now, so I can't verify this. I will test it tomorrow and send you a message. |
|
Thanks very much, @drinkBr |
|
Hi @MartyG-RealSense, I tried 2 methods: Method 1: Method 2: Both methods result in these 4 major errors as in this issue: [realsense2_camera_node-1] 12/08 14:55:27,553 ERROR [281472736749792] (librealsense-exception.h:52) get_xu(...). xioctl(UVCIOC_CTRL_QUERY) failed Last Error: No such file or directory Could you explain what are those errors and what cause them? I have tried to debug them with many methods from github but I still get the same error. I'm looking forward to your reply. Thanks a lot! |
|
@MiTomMi It looks as though the error #2991 There were a couple of non-Pi cases of the error where updating the camera firmware driver resolved it. |
|
Hi, @MartyG-RealSense When I previously installed realsense-ros, I followed the instructions on the following page. Is that correct? Are there any necessary changes to consider if I installed realsense-related components using the LibUVC-backend installation method? Installation on Ubuntu Step3 Option2 : https://github.com/IntelRealSense/realsense-ros |
|
If you install librealsense from source code using the libuvc backend method then you can then afterwards build the ROS2 wrapper from source code with colcon build. In other words, use Step 3 Option 2 as you said. You are using a Raspberry Pi 4 though, so there might be complications. Please give a colcon build a try and see what happens. You can always remove the wrapper build easily afterwards without affecting your librealsense installation if the wrapper does not work correctly. |
Is it correct that I should verify the |
|
The recommended install order in the instructions is:
Skip step 1 if Humble is already installed. I am not certain what you mean regarding verifying colcon build before installing realsense-ros, as colcon build is what does install realsense-ros. |
|
I misunderstood. So, I just need to execute Step 3 Option 2 in the Terminal, right? |
|
Yes, that's correct. |
|
OK! I will test it tomorrow and send you a message. |
|
Hi @MartyG-RealSense, I am MiTomMi.
I don't know why the frames are fractured like this. Could you figure it out? Thanks a lot!!! |
|
UPDATE:
Do you know what's the problem in realsense-viewer in my case? Thank you!!! |
|
The rs-capture image is a 2D depth image like the one in the RealSense Viewer when you select the '2D' option in the top corner of the Viewer window. The broken images that you provided above are from the '3D' pointcloud mode of the Viewer. Do you get a better depth image in the Viewer if you click on '2D' to switch to 2D depth mode? I note that in your Viewer's 3D mode with the broken images, the 'Measure' option at the top of the window is greyed out and unselectable. This suggests that the GLSL graphics options in your Viewer are disabled. You can use instructions at IntelRealSense/librealsense#8110 (comment) to check the status of the GLSL options and see whether they are currently enabled or disabled. If the boxes beside the Processing and Rendering settings are black (off) then try clicking on them to turn the boxes blue (on) and click the Apply button, then test the depth pointcloud in 3D mode again. |
|
Hi @drinkBr and @haidaumitom Do either of you require further assistance with this case, please? Thanks! |
|
Hi @MartyG-RealSense, I do not require any addition assistance. Thanks for your really supporting help! I really appreciate that! Best regards. |
|
@haidaumitom Thanks very much. You are very welcome. I'm pleased that I could help! @drinkBr Is your problem resolved, please? |
|
Case closed due to no further comments received. |
Issue Description
I encountered the following error when executing
ros2 run realsense2_camera realsense2_camera_node.I believe that it is necessary to set the value of rgb_camera.power_line_frequency, but I would like to hear your opinion.
Also, after this error was displayed, I executed realsense-viewer, and when I turned on the Stereo Module, the following error appeared. I confirmed that the same error message appeared when executing
ros2 run realsense2_camera realsense2_camera_node, so I believe they are related. I would like to hear your opinion.The text was updated successfully, but these errors were encountered: