EVA SDK Installation Guide for x86

Manual Rev.: 1.3

Revision Date: September xx, 2023

Part Number: 50M-00040-1030

Preface

Copyright

Copyright © 2023 ADLINK Technology, Inc. This document contains proprietary information protected by copyright. All rights are reserved. No part of this manual may be reproduced by any mechanical, electronic, or other means in any form without prior written permission of the manufacturer.

Disclaimer

The information in this document is subject to change without prior notice in order to improve reliability, design, and function and does not represent a commitment on the part of the manufacturer. In no event will the manufacturer be liable for direct, indirect, special, incidental, or consequential damages arising out of the use or inability to use the product or documentation, even if advised of the possibility of such damages.

Trademarks

Product names mentioned herein are used for identification purposes only and may be trademarks and/or registered trademarks of their respective companies.

Revision History

1 Introduction

This chapter describes the installation of the following software.

• GStreamer, Gstreamer RTSP Plugin and GStreamer Python Plugin
• NIVIDA® DeepStream, NIVIDA® CUDA and NIVIDA® TensorRT
• Intel® OpenVINO
• Intel® Media SDK for GStreamer
• Pylon Software
• Hikrobot Software
• FLIR Software
• ADLINK EVA SDK
• Euresys Solution Software

The following table lists the specified software versions.

Notes:

¹ If the NVIDIA solution is used for inference, NVIDIA® DeepStream, NVIDIA® CUDA®, and NVIDIA® TensorRT™ must be installed.

² If the Intel MSDK encoder and decoder are used for GStreamer plugins, Intel® OpenVINO™ and Intel® Media SDK must be installed.

The following table lists the estimated installation space required when installing the software under Ubuntu 18.04 64-bit. The required installation space includes the install file (.deb, .run, .tgz) package size.

For more details about ADLINK EVA, refer to https://eva-support.adlinktech.com.

2 Installing GStreamer on Ubuntu

The recommended installation configuration requires installing GStreamer 1.16.3 on Ubuntu 20.04.

2.1 Install GStreamer on Ubuntu

Reference: https://gstreamer.freedesktop.org/documentation/installing/on-linux.html?gi-language=c

2.2 Check the GStreamer Installation

After executing the command, a window with an animated video pattern should display on-screen. Use CTRL+C in the terminal to stop the program.

2.3 Install RTSP Plugin for GStreamer (Optional)

If the data will be sent over an RTSP server, the system must have the RTSP plugin installed with the following command for GStreaming.

Use the following command to verify the installation.

3 Installing the GStreamer Python Plugin

GStreamer is built on the GLib and GObject portable libraries which are compatible with Python. Gstreamer can be used in two ways: by writing GStreamer applications, or by writing GStreamer Python elements that can be scanned by the GStreamer Python plugin loader and registered as a plugin in GObject. The following sections describe how to install the GStreamer Python plugin.

3.1 Install the Required Packages

If your platform is using Python 2.7, change the python3-gst-1.0 package to python-gst-1.0.

3.2 Install the GStreamer Python Plugin Loader

Currently, the Debian GStreamer Python package includes GObject Introspection files but does not include the GStreamer Python plugin loader, requiring the use of a GStreamer Python application.

Install python plugin loader package

Install the required Python packages 

 Notes:

If the CPU is an Intel^® Atom™ series processor, installation of opencv-python will take about 1 to 1.5 hours. Do not interrupt the installation process.

Check for installed Python plugins

If the Python plugins are found, the installation has been successful.

4 Installation Process for Intel Solution

The ADLINK EVA SDK also supports the Intel OpenVINO inference solution based on your specific system configuration requirements.

For ADLINK EVA SDK with the Intel solution, the following components must be installed on Ubuntu 18.04:

• OpenVINO 2022.3
• Intel Media SDK plugin for GStreamer (optional)

If you are not using the Intel® OpenVINO™, this chapter can be skipped.

4.1 Uninstall Non-specified Version

If the system has a non-specified version of OpenVINO, uninstall it.

OpenVINO 2021.3 or earlier version removal:

use the OpenVINO toolkit installation package to uninstall it.

Choose Uninstall the product and follow the instructions on the screen to uninstall.

OpenVINO 2021.4 or later version removal:

Refer to https://docs.openvino.ai/latest/openvino_docs_install_guides_uninstalling_openvino.html

Choose your version and follow the instructions to uninstall the OpenVINO.

4.2 Installing the OpenVINO Toolkit

EVA SDK only needs the OpenVINO runtime, the section will describe how to download and use archive files to install OpenVINO Runtime.

For more details, refer to https://docs.openvino.ai/2022.3/openvino_docs_install_guides_overview.html

1. Download and Install the OpenVINO Core Components

Open a terminal, and enter the following commands.

Create link

2. Configure the Environment

Set environment variables for OpenVino

Optional: The OpenVINO environment variables are removed when you close the shell. As an option, you can permanently set the environment variables as follows:

1. Open the .bashrc file
   
   
2. Add this line to the end of the file

   source /opt/intel/openvino/setupvars.sh

3. To test your change, open a new terminal. You should see [setupvars.sh] OpenVINO environment initialized.

For more details of the installation, refer to https://docs.openvino.ai/2022.3/openvino_docs_install_guides_installing_openvino_from_archive_linux.html

4.3 Installing the Intel Media SDK Plugin for Gstreamer (Optional)

If you would like to use Media SDK plugin in GStreamer, install OpenVino^TM first. It will include Intel Media SDK and the related libraries such as VAAPI, LibVA, and gmmlib.

The Intel Media SDK plugin source code is in gst-plugins-bad module, but as of Aug. 23, 2019, the GStreamer package installed via apt-get does not contain the msdk plugin, so it must be built and installed.

For more details, refer to https://github.com/Intel-Media-SDK/MediaSDK/wiki/Build-GStreamer-MSDK..

1. Prepare the Environment and Dependencies

2. Download and Build the Source Code

Download the source code

Before building, install the following package

Build and install the source code

Note:

After configure log must contains:
    Run-time dependency libmfx found: YES 1.32
    Has header "mfx/mfxdefs.h": YES
   Has header "mfxvp9.h": YES 

Check for the msdk plugin

The following is an example of the installed package.

5 Installation Process for NVIDIA Solution (Optional)

The ADLINK EVA SDK also supports the NVIDIA inference solution based on your specific system configuration requirements.

For ADLINK EVA SDK with the NVIDIA solution, the following components must be installed on Ubuntu 20.04:

• GStreamer 1.16.x
• NVIDIA Driver 515 or higher
• CUDA 11.7
• TensorRT 8.4.1.5
• DeepStream 6.1.1 (optional)

This chapter will describe the installation and uninstallation steps provided on the NVIDIA official website.

If you are not using an NVIDIA inference solution, this chapter can be skipped.

5.1 Uninstall Non-specified Version

This section describes how to unistall a non-specified version of the NVIDIA solution software. After the components have been removed, you must reboot the system.

5.1.1 Remove DeepStream

Check the DeepStream Version

DeepStream 3.0 or earlier version removal:

DeepStream 4.0 or later version removal:

1. Open the uninstall.sh file in /opt/nvidia/deepstream/deepstream/
2. Set PREV_DS_VER as the version number (e.g: 4.0)
3. Run the following command script

   $ sudo ./uninstall.sh

Reference: https://docs.nvidia.com/metropolis/deepstream/5.1/dev-guide/text/DS_Quickstart.html#remove-all-previous-deepstream-installations

5.1.2 Remove TensorRT

Check the TensorRT version

Remove TensorRT

Reference: https://docs.nvidia.com/deeplearning/tensorrt/install-guide/index.html#uninstalling.

5.1.3 Remove CUDA

Check the CUDA version

Remove the CUDA Toolkit

Notes:

1. If using a runfile for installation, refer to https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#handle-uninstallation
2. Reference: https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#removing-cuda-tk-and-driver

5.1.4 Remove NVIDIA Drivers

Check the driver

Remove NVIDIA Drivers

Reference: https://docs.nvidia.com/cuda/archive/11.0/cuda-installation-guide-linux/index.html#removing-cuda-tk-and-driver

5.2 Installing NVIDIA Drivers

Before installing the NVIDIA graphics driver, use the “Software Updater” to install updated software first.

If the system software is not updated, after installing the NVIDIA driver and rebooting, the system can hang with the following message: “Stopping User Manager for UID 121”.

Insert the NVIDIA graphics card and follow the steps below to install the drivers.

1. Add the graphics-drivers repository to advanced package tool (apt) 

2. Check the NVIDIA Driver list

In this example, the system must have an NVIDIA device with a recommend driver version of 470 or later.

3. Install the NVIDIA Driver

4. Check the driver

In this example, the system installed driver version is 470 and supports CUDA driver version 11.4

5.3 Installing CUDA Toolkit

Download the CUDA Toolkit 11.7.1 original archive from https://developer.nvidia.com/cuda-11-7-1-download-archive?target_os=Linux&target_arch=x86_64&Distribution=Ubuntu&target_version=20.04&target_type=deb_local

The target platform must conform to the following requirements:

• Operation System: Linux
• Architecture: x86_64
• Distribution: Ubuntu
• Version: 20.04
• Installer Type: deb(local)

Run the following installation commands.

5.4 Installing TensorRT

Download TensorRT 8.4.1.5 for Ubuntu 20.04 and CUDA 11.7 from https://developer.nvidia.com/compute/machine-learning/tensorrt/secure/8.4.1/local_repos/nv-tensorrt-repo-ubuntu2004-cuda11.6-trt8.4.1.5-ga-20220604_1-1_amd64.deb

Select Log in or Join now.

After downloading, install TensorRT from the debian local repository package with the following commands:

Refer to the Debian installation online document for more information at  https://docs.nvidia.com/deeplearning/tensorrt/archives/tensorrt-841/install-guide/index.html#installing-debian

5.5 Installing DeepStream (Optional)

If you are not using a DeepStream SDK, this section can be skipped.

The NVIDIA DeepStream SDK first requires installation of the following components:

• GStreamer
• NVIDIA Driver
• CUDA
• TensorRT

For more details, refer to https://docs.nvidia.com/metropolis/deepstream/6.1/dev-guide/text/DS_Quickstart.html#dgpu-setup-for-ubuntu

1. Install packages

Enter the following command to install the necessary packages before installing the DeepStream SDK:

2. Install librdkafka

librdkafka enables the Kafka protocol adaptor for message brokering.

Clone the librdkafka repository from GitHub

Configure and build the library 

Copy the generated libraries to the DeepStream directory

3. Install the DeepStream SDK

Download the DeepStream 6.1 dGPU tar package from https://developer.nvidia.com/deepstream_sdk_v6.1.0_x86_64.tbz2.

Navigate to the location of the downloaded DeepStream package to extract and install the DeepStream SDK.

Note:

NVIDIA provides three methods for installing the DeepStream SDK. The tar package MUST be used for installation. Attempting to use the Debian package or apt-server to install, will result in compatibility issues.

5.5.1 Verify DeepStream

After installing DeepStream, follow these steps to verify the installation.

1. Check version information

2. Check the plugin in GStreamer

3. Verify the plugin via gst-launch

The following is an example of a properly installed package.

Note:

When running a GStreamer command or EVA IDE, the following warning messaging will display.

The first message is a harmless warning indicating that the DeepStream’s nvinferserver plugin cannot be used since “Triton Inference Server” is not installed on x86(dGPU) platforms.

Refer to NVIDIA troubleshooting for other solutions. https://docs.nvidia.com/metropolis/deepstream/dev-guide/text/DS_troubleshooting.html#errors-occur-when-deepstream-app-fails-to-load-plugin-gst-nvinferserver-on-dgpu-only

The second message is also a harmless warning because the NVIDIA Rivermax SDK is not necessary.

5.6 Installing ONNX Runtime

After installing the CUDA Toolkit, run the following command to install the ONNX Runtime packages.

Reference: https://www.onnxruntime.ai/

6 Installing Pylon Software (Optional)

If you are not using a Basler camera, you can skip this chapter.

6.1 Uninstall Non-specified Version

If the system has a non-specified version of the Pylon software, uninstall it.

If the installer is installed via deb, refer to the following command to remove pylon.

6.2 Install pylon Software

If Basler cameras are used, the Pylon Software must be installed.

Download the pylon camera software from the Basler website at: https://www.baslerweb.com/en/sales-support/downloads/software-downloads/

The recommended software versions are:

• Software Category: pylon Software
• Version: 7.2.1
• Operating System: Linux x86-64 bit

Download the Debian Installer Package and use the following install command.

After the installation is completed, set the pylon library path environment variable with LD_LIBRARY_PATH.

For example, if the pylon library path is /opt/pylon/lib, execute the following command:

Note:

You may need to set the environment for the system permanently.

1. Open the .bashrc file in your home directory, for example:
   $ gedit ~/.bashrc
2. Add export environment variables to the last line of the file, for example:
   • export LD_LIBRARY_PATH=/opt/pylon/lib:$LD_LIBRARY_PATH
3. Save the .bashrc file.
4. Restart your terminal.

Verify the Basler's USB3/GigE Vision cameras with pylon

• After installing pylon, run the pylonviewer tool to test the camera.
• The file should be at /opt/pylon/bin. 
• Refer to the following Basler document on using the pylonviewer: https://docs.baslerweb.com/overview-of-the-pylon-viewer.html

7 Installing Hikrobot Software (Optional)

If you are not using a Hikrobot camera, you can skip this chapter.

7.1 Uninstall Non-specified Version 

If the system has a non-specified version of the Pylon software, uninstall it.

If the installer is installed via deb, refer to the following command to remove pylon.

7.2 Install Hikrobot Software

If Hikrobot cameras are used, the Hikrobot MVS Software must be installed.

Download Machine Vision Software MVS2.1.2（Linux） from: https://www.hikrobotics.com/en/machinevision/service/download?module=0

Run the following commands to install.

 Note:

The above installation steps are for reference only. For more specific installation instructions, refer to the official documentation, or contact Hikrobot at https://en.hikrobotics.com/contactus.

8 Installing FLIR Software (Optional)

This chapter covers the installation of required software for use with FLIR cameras.

If you are not using a FLIR camera, this chapter can be skipped.

8.1 Uninstall Non-specified Version

If the system has a non-specified version of the FLIR Spinnaker software, uninstall it.

Run the uninstall script with the Spinnaker installation package.

8.2 Install FLIR Spinnaker Software

If FLIR cameras are used, the FLIR Spinnaker Software Suite for Windows must be installed.

Download the FLIR Spinnaker installer from: https://www.flir.asia/products/spinnaker-sdk/

The recommended software versions are:

• Version: 3.0.0.118
• Operating System: Linux 20.04, 64-bit

Run the following commands and then follow the on-screen instructions to complete the installation.

Note:

During installation, you MUST enter a “username” to add a new member.

For example, enter ‘adlink’ for the user name.

After installation, connect the camera and run the following command to verify that it works.

Note:

For more information, refer to https://www.flir.com/support-center/iis/machine-vision/application-note/using-spinnaker-on-arm-and-embedded-systems/

If you have any questions, go to https://www.flir.com/.

9 Installing Process for Euresys Solution (Optional)

This chapter covers the installation of the required software for using Euresys Open eVision and GigE Vision Cameras.

If you are not using Euresys solution, this chapter can be skipped.

If you would like to use Euresys, contact your regional distributor to get the Euresys licenses. https://www.euresys.com/en/Products/Accessory/Neo-USB-Dongle-(empty)

9.1 Installing GigE Vision Camera Software

If Gige vision cameras are used, the Memento and eGrabber software must be installed.

The recommended software versions are:

• Version: 23.02.0.68
• Operating System: Linux 20.04, 64-bit

Download Memento and eGrabber from https://www.euresys.com/en/Support/Software,-drivers-and-documentation?series=105d06c5-6ad9-42ff-b7ce-622585ce607f&os=Linux&packages=20d16334-0c64-45e0-8ba0-0d099597937f:23.02;e55e4c47-ed00-45cf-bc7e-ebd6fb2e4c23:23.02

Run the following commands to install.

 For more details, refer to memento-linux-x86_64-22.10.1.45/INSTALL and egrabber-linux-x86_64-22.10.1.45/INSTALL.

 Before using the Euresys eGrabber driver, you must set the following environment variables.

Optional: The environment variables are removed when you close the terminal. As an option, you can permanently set the environment variables as follows:

1. Open the .bashrcfile

   $ gedit ~/.bashrc

2. Add this line to the end of the file

   source /opt/euresys/egrabber/shell/setup_gentl_paths.sh source /opt/euresys/egrabber/shell/select-gigelink-producer.sh

After installation, connect the camera and run the following command to verify that it works.

9.2 Installing Euresys Open eVision

If Euresys Open eVision are used, the Open eVision, the Neo License Manager, Codemeter, and the Open eVision Cuda resdist must be installed.

The recommended software versions are:

• Version: 22.12.0.14750
• Operating System: Linux 20.04, 64 bit

Download open_evision-linux-x86_64-22.12.0.14750.deb.tar.gz and    open_evision-cuda-redist-linux-x86_64-22.12.0.14750.deb.tar.gz from https://www.euresys.com/en/Support/Download-area?Series=f97da39d-3c25-404c-aee7-73de1d1867fc

Run the following commands and then follow the on-screen instructions to complete the installation.

10 Installing the ADLINK EVA SDK

10.1 Remove Previous ADLINK EVA SDK Versions

If you have a previously installed version of the ADLINK EVA SDK, remove it with the following command:

If the install path has been changed, specified the correct path with:

INSTALL_DIR enters the specified path.

Note:

When uninstalling the ADLINK EVA SDK, the folder specified with INSTALL_DIR will be deleted.

10.2 Install the Required Package for EVA IDE

10.3 Download and Install the ADLINK EVA SDK

Download the ADLINK EVA SDK installation package and copy it to your Linux Ubuntu 18.04 64-bit system.

Change mode and run install package

xxxx is the version and the install path is /opt/adlink/eva.

Select the EVA SDK plugins to be installed.

For example, to install the TensorRT inference plugin and the Pylon plugin, enter 3,4. To install all plugins, including the OpenVINO inference plugin, TensorRT inference plugin, Pylon plugin, Hik plugin, and Flir plugin, enter 1.

Notes:

1. OpenVINO is only supported on Intel platforms.
2. TensorRT is only supported on NVIDIA platforms.
3. Before installing the EVA SDK, you must ensure the corresponding software packages have been installed. If the plugins are installed but there are no corresponding libraries, when the GStreamer command runs for the first time, it will display messages to warn that the related libraries are missing, and the installed plugins will not work.

Other commands can be used to install the ADLINK EVA SDK.

• Use the -s command option to install all the plugins in silent mode. The install path is /opt/adlink/eva.

  $ sudo ./EVA_SERP_xxxx.run -- -s$ sudo ./EVA_SERP_xxxx.run -- -s

10.4 Set Environment Variables

Set environment variables for the ADLINK EVA SDK

The script will set up the environment variables of the following installed software.

• OpenVINO
• MediaSDK
• Pylon
• ADLINK EVA SDK

If the software has no corresponding libraries, the script will not set up the corresponding environment variables.

Note:

The environment variables are removed when closing the command prompt or terminal.

Safety Instructions

Read and follow all instructions marked on the product and in the documentation before you operate your system. Retain all safety and operating instructions for future use.

• Please read these safety instructions carefully.
• Please keep this User‘s Manual for later reference.
• Read the specifications section of this manual for detailed information on the operating environment of this equipment.
• When installing/mounting or uninstalling/removing equipment, turn off the power and unplug any power cords/cables.
• To avoid electrical shock and/or damage to equipment:
  • Keep equipment away from water or liquid sources.
  • Keep equipment away from high heat or high humidity.
  • Keep equipment properly ventilated (do not block or cover ventilation openings).
  • Make sure to use recommended voltage and power source settings.
  • Always install and operate equipment near an easily accessible electrical socket-outlet.
  • Secure the power cord (do not place any object on/over the power cord).
  • Only install/attach and operate equipment on stable surfaces and/or recommended mountings.
  • If the equipment will not be used for long periods of time, turn off and unplug the equipment from its power source.
• Never attempt to fix the equipment. Equipment should only be serviced by qualified personnel.

Please visit the Contact page at www.adlinktech.com for information on how to contact the ADLINK regional office nearest you.