nvidia-orin-nx 5.10.104-tegra Cant install d435 SDK camera with ubuntu

[Thanosrom]

Hi all im trying to install my camera with orin nx ,i have been trying this for like 2-3 days !

I tried many ways-steps and still got nothing.

System : Linux nvidia-orin-nx 5.10.104-tegra aarch64 Ubuntu 22.04

1)I tried to install from here : librealsense/distribution_linux.md at master · IntelRealSense/librealsense · GitHub

still i got : E: Unable to locate package librealsense2-utils E: Unable to locate package librealsense2-dkms

Also when i try to change kernels and boot from another one i cant configure grubs because guess what,there is none!

2)Also i tried with ros2 and i cant install humble ros2! Still cant find packages !

3)Also my version of jetpack is 35.0 so again i cant install it from there and i cant install sdk from nvidia official website! System errors …

4)CMake cant do it either i have missing files that in intel github described that there must have been

5)I’ve tried again to boot and pass again ubuntu 18.04 and my usb-stick cant rezognize !

What is happening ?!

[MartyG-RealSense]

Hi @Thanosrom The packages in the distribution_linux.md instructions are not suited for use with Arm-based architecture like that of Nvidia Jetson boards as they are designed for computers with an x86 / x64 architecture. The librealsense SDK has a dedicated Jetson installation instructions page at the link below with package and source code installation guidelines.

https://github.com/IntelRealSense/librealsense/blob/master/doc/installation_jetson.md

Nvidia's JetPack Archive page does not list a 35.0 version, with the '35' versions starting at 35.1, also known as JetPack 5.0.2.

https://developer.nvidia.com/embedded/jetpack-archive

The RealSense SDK does not yet officially support JetPack 5.0.2 but will do in the next librealsense SDK version, likely numbered 2.54.0. In the meantime, the SDK can be installed on Orin by building from source code, and either not patching the kernel or building from source with the RSUSB Backend installation method in order to bypass the kernel. More information about these approaches to a successful Orin installation can be found at #10722 (comment)

---

Once the librealsense SDK has been successfully installed then you can afterwards progress to installing ROS2 Humble and the RealSense ROS wrapper using the instructions here:

https://github.com/IntelRealSense/realsense-ros#--installation

Expand open Step 1, 2 and 3 to reveal the instructions for each step. Under Step 3, use the Option 2: Install from source instructions to build the ROS2 wrapper from source code after ROS2 Humble has been installed.

Step 1 has a link to instructions for installing Humble.

---

Given the number of different problems that you have been experiencing though, it may be first doing a complete wipe of the Orin (including Ubuntu) and freshly installing everything again, starting with Ubuntu 22.04.

[Thanosrom]

Hi @Thanosrom The packages in the distribution_linux.md instructions are not suited for use with Arm-based architecture like that of Nvidia Jetson boards as they are designed for computers with an x86 / x64 architecture. The librealsense SDK has a dedicated Jetson installation instructions page at the link below with package and source code installation guidelines.

https://github.com/IntelRealSense/librealsense/blob/master/doc/installation_jetson.md

Nvidia's JetPack Archive page does not list a 35.0 version, with the '35' versions starting at 35.1, also known as JetPack 5.0.2.

https://developer.nvidia.com/embedded/jetpack-archive

The RealSense SDK does not yet officially support JetPack 5.0.2 but will do in the next librealsense SDK version, likely numbered 2.54.0. In the meantime, the SDK can be installed on Orin by building from source code, and either not patching the kernel or building from source with the RSUSB Backend installation method in order to bypass the kernel. More information about these approaches to a successful Orin installation can be found at #10722 (comment)

Once the librealsense SDK has been successfully installed then you can afterwards progress to installing ROS2 Humble and the RealSense ROS wrapper using the instructions here:

https://github.com/IntelRealSense/realsense-ros#--installation

Expand open Step 1, 2 and 3 to reveal the instructions for each step. Under Step 3, use the Option 2: Install from source instructions to build the ROS2 wrapper from source code after ROS2 Humble has been installed.

Step 1 has a link to instructions for installing Humble.

Given the number of different problems that you have been experiencing though, it may be first doing a complete wipe of the Orin (including Ubuntu) and freshly installing everything again, starting with Ubuntu 22.04.

---

Thanks for the reply,

Sorry but i tried all of that installation guides from your old post too,its kind the same...

Nvidia give me a good reply from the forum and it seems to work , i just post this so everybody can find solution :

Those are the steps :

git clone https://github.com/IntelRealSense/librealsense.git
cd librealsense/
sudo apt install libudev-dev pkg-config libgtk-2-dev
sudo apt install libudev-dev pkg-config libgtk-3-dev (For me this worked)
sudo apt install libusb-1.0-0-dev pkg-config
sudo apt install libglfw3-dev
sudo apt instlal libssl-dev
sudo apt install libssl-dev
sudo apt-get install libglfw3-dev libgl1-mesa-dev libglu1-mesa
sudo apt install cmake
sudo cp config/99-realsense-libusb.rules /etc/udev/rules.d/.
sudo udevadm control --reload-rules && udevadm trigger
mkdir build
cd build/
cmake ../ -DBUILD_EXAMPLES=true
make
sudo make install
cd tools/realsense-viewer
./realsense-viewer

Also i cant find a way to install humble , for some reason i can only install foxy .... and i am not sure about the results.

[MartyG-RealSense]

Thanks so much for sharing your solution with the RealSense community!

Regarding your issue with installing Humble on Jetson Orin, a moderator on the Nvidia forum advises at the link below to source the environment first.

https://forums.developer.nvidia.com/t/unable-to-install-ros2-humble-on-agx-orin/239741

[Thanosrom]

Thanks again, i will try it and i will come back. :)

[MartyG-RealSense]

Hi @Thanosrom Do you require further assistance with this case, please? Thanks!

[Thanosrom]

tance with

Hello , for the moment im ok you can close the question .Thanks .

[MartyG-RealSense]

You are very welcome. As you suggest, I will close the issue. Thanks!

[nakai-omer]

Hi @Thanosrom The packages in the distribution_linux.md instructions are not suited for use with Arm-based architecture like that of Nvidia Jetson boards as they are designed for computers with an x86 / x64 architecture. The librealsense SDK has a dedicated Jetson installation instructions page at the link below with package and source code installation guidelines.
https://github.com/IntelRealSense/librealsense/blob/master/doc/installation_jetson.md
Nvidia's JetPack Archive page does not list a 35.0 version, with the '35' versions starting at 35.1, also known as JetPack 5.0.2.
https://developer.nvidia.com/embedded/jetpack-archive
The RealSense SDK does not yet officially support JetPack 5.0.2 but will do in the next librealsense SDK version, likely numbered 2.54.0. In the meantime, the SDK can be installed on Orin by building from source code, and either not patching the kernel or building from source with the RSUSB Backend installation method in order to bypass the kernel. More information about these approaches to a successful Orin installation can be found at #10722 (comment)
Once the librealsense SDK has been successfully installed then you can afterwards progress to installing ROS2 Humble and the RealSense ROS wrapper using the instructions here:
https://github.com/IntelRealSense/realsense-ros#--installation
Expand open Step 1, 2 and 3 to reveal the instructions for each step. Under Step 3, use the Option 2: Install from source instructions to build the ROS2 wrapper from source code after ROS2 Humble has been installed.
Step 1 has a link to instructions for installing Humble.
Given the number of different problems that you have been experiencing though, it may be first doing a complete wipe of the Orin (including Ubuntu) and freshly installing everything again, starting with Ubuntu 22.04.

Thanks for the reply,

Sorry but i tried all of that installation guides from your old post too,its kind the same...

Nvidia give me a good reply from the forum and it seems to work , i just post this so everybody can find solution :

Those are the steps :

git clone https://github.com/IntelRealSense/librealsense.git cd librealsense/ sudo apt install libudev-dev pkg-config libgtk-2-dev sudo apt install libudev-dev pkg-config libgtk-3-dev (For me this worked) sudo apt install libusb-1.0-0-dev pkg-config sudo apt install libglfw3-dev sudo apt instlal libssl-dev sudo apt install libssl-dev sudo apt-get install libglfw3-dev libgl1-mesa-dev libglu1-mesa sudo apt install cmake sudo cp config/99-realsense-libusb.rules /etc/udev/rules.d/. sudo udevadm control --reload-rules && udevadm trigger mkdir build cd build/ cmake ../ -DBUILD_EXAMPLES=true make sudo make install cd tools/realsense-viewer ./realsense-viewer

Also i cant find a way to install humble , for some reason i can only install foxy .... and i am not sure about the results.

Imho, you should also add -DBUILD_WITH_CUDA=true to cmake params to leverage CUDA on the Jetson

Just for reference, these are our flags for building on the Jetson (we don't need the examples on the Jetson so they are also marked as fale):

 cmake ../ -DFORCE_RSUSB_BACKEND=false -DBUILD_PYTHON_BINDINGS=true -DBUILD_WITH_CUDA=true -DCMAKE_BUILD_TYPE=release -DBUILD_EXAMPLES=false -DBUILD_GRAPHICAL_EXAMPLES=false

[ahmet-f-gumustas]

ERROR Solution

I got this error when I ran this code : cmake ../ -DFORCE_RSUSB_BACKEND=false -DBUILD_PYTHON_BINDINGS=true -DBUILD_WITH_CUDA=true -DCMAKE_BUILD_TYPE=release -DBUILD_EXAMPLES=false -DBUILD_GRAPHICAL_EXAMPLES=false

"""
Tell CMake where to find the compiler by setting either the environment
variable "CUDACXX" or the CMake cache entry CMAKE_CUDA_COMPILER to the full
path to the compiler, or to the compiler name if it is in the PATH.
Call Stack (most recent call first):
CMake/global_config.cmake:67 (include)
CMakeLists.txt:45 (global_set_flags)
"""
as a solution this ;
export PATH=/usr/local/cuda-{cuda-version}/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-{cuda-version}/lib64:$LD_LIBRARY_PATH

and check "nvcc -V" , it works if it outputs.

and run "cmake ../" .... will work

[MartyG-RealSense]

Thanks so much @ahmet-f-gumustas for sharing your solution!