EVA SDK Installation Guide for Windows

Manual Rev.: 1.2

Revision Date: July xx, 2022

Part Number: 50M-00012-1020

Preface

Copyright

Copyright © 2022 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.

• NVIDA® CUDA, NVIDIA® cuDNN, and NIVIDA® TensorRT
• Intel® OpenVINO, Intel® Media SDK
• Pylon Software
• GStreamer
• ADLINK EVA SDK

The following table lists the recommended software versions.

Notes

¹ If the NVIDIA solution is used for inference, NVIDIA® CUDA®, NVIDIA® cuDNN 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.

³ OpenCV is required by the ADLINK EVA SDK. Please install Intel® OpenVINO™ toolkit that includes the corresponding version of OpenCV.

The following table lists the estimated installation space required when installing the software under Windows 10 64-bit.

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

2 Uninstall Previous Software

If the system has a previous version of the software installed, it must first be uninstalled before the new version can be installed. The following content is for reference only (except for the software developed by ADLINK). Actual uninstall procedures vary depending on the vendor software.

2.1 Uninstall via Software Installer Program

Run the program’s installer and follow the screen prompts to uninstall the software. The following table lists the software installers with an uninstall feature, and a sample of the uninstall window.

2.2 Uninstall via Programs and Features

1. Press <Windows> + <Q> and type Control Panel, and then click Control Panel. Click Uninstall a program to open Programs and Features.
   
2. Double-click the program to be uninstalled. Click Yes to begin uninstalling the program.

The following table lists the software that can be uninstalled through Windows Programs and Features.

Note: *** refers to the software version or other information listed in Programs and Features.

2.3 Manually Remove Files

Some software does not support being completely uninstalled with an uninstaller. These must be uninstalled manually be deleting the files from the system.

2.4 Remove Environment Settings

When installing NVIDIA TensorRT, the binary and library are added to the environment PATH. When uninstalling TensorRT, the environment PATH must be deleted manually.

Please use the following steps to delete TensorRT’s path setting.

1. In the Windows Search box, type edit the system environment variables and press <Enter>.
   
2. Click Environment Variables
3. Under System variables, click Path and then Edit.
   
4. Select TensorRT’s bin and lib path and click Delete.

3 Installing Required Dependencies

This chapter describes how to install the following software:

• Microsoft Visual C++ Redistributable for Visual Studio 2015-2019
• Python 3.7 64-bit
• Microsoft Build Tools for Visual Studio 2019/2017
• CMake 3.14 or higher 64-bit

If the system already has Python and CMake installed, make sure they are the correct version.

3.1 Installing Microsoft Visual C++ Redistributable for Visual Studio

Download the install file for Microsoft Visual C++ Redistributable for Visual Studio 2015, 2017 and 2019 x64 from https://support.microsoft.com/en-us/help/2977003/the-latest-supported-visual-c-downloads.

After downloading, run the installer to install the software.

3.2 Installing Python

Download the Python 3.7 64-bit for Windows from https://www.python.org/downloads/windows/. As of January 22, 2021, the latest Python 3.7 release version is 3.7.9.

Uninstall all previous versions of Python from the system. For more details, refer to Uninstall Previous Software.

After downloading, run the installer, select Add Python 3.7 to PATH to add the application to the environment PATH, and then click Install Now.

The default path is C:\Users\<USER_ID>\AppData\Local\Programs\Python\Python37

Click Disable path length limit to allow the Python default path.

After installation, open a command prompt and entert the following command to verify the environment PATH.

If the terminal cannot find the python command, the system might not be able to find the program. For the instructions to add Python to the system environment variables, see Set Windows Environment Variables.

Run the following command to install additional required packages.

3.3 Installing Microsoft Build Tools for Visual Studio

Open a web browser and navigate to https://visualstudio.microsoft.com/downloads/, click Tools for Visual Studio 2019, and click Download next to Build Tools for Visual Studio 2019.

After downloading, run the installer, and choose C++ build tools and Universal Windows Platform build tools for installation.

To use Visual Studio 2017, select MSVC v141- VS 2017 C++ x64/x86 build tools in the C++ build tools.

3.4 Installing CMake

Download CMake 3.14 or higher (64-bit) for Windows from https://cmake.org/download/.

Uninstall all previous versions of CMake from the system. For more details, refer to Uninstall Previous Software.

After downloading, run the installer to install the software.

Note: Choose Add CMake to the system PATH for all users.

After installing, open a command prompt and enter the following command to verify the environment PATH.

If the terminal cannot find the cmake command, the system might not be able to find the program. For the instructions to add CMake to the system environment variables, see Set Windows Environment Variables.

3.5 Set Windows Environment Variables

This section describes how to manually set the Windows environment.

Use the following steps to update the Windows PATH if an execute command returns an error message stating that an application cannot be found.

1. In the Windows Search box, type edit the system environment variables and press <Enter>.
   
2. Click Environment Variables.
3. Under System variables, click Path and then Edit.
    
4. Click New and enter the program binary directory or click Browse and choose the directory.
   For example,
   • The CMake default path is C:\Program Files\CMake\bin.
   • The Python default path is C:\Users\<USER_ID>\AppData\Local\Programs\Python\Python37.
   • The Python script tool (such as pip3) default path is C:\Users\<USER_ID>\AppData\Local\Programs\Python\Python37\Script.
   • The Pylon runtime library default path is C:\Program Files\Basler\pylon 6\Runtime\x64.
5. Click OK repeatedly to close each window.

4 Installation Process for Intel Solution

If you are not using the Intel® Distribution of OpenVINO™ toolkit, you still have to install OpenVINO for OpenCV. The ADLINK EVA SDK requires OpenCV from OpenVINO.

4.1 Uninstall Previous Version

Uninstall all previous versions of OpenVINO and/or Media SDK from the system. For more details, refer to Uninstall Previous Software .

4.2 Installing the OpenVINO Toolkit

The Intel® Distribution of OpenVINO™ toolkit quickly deploys applications and solutions that emulate human vision. Based on Convolutional Neural Networks (CNN), the toolkit extends computer vision (CV) workloads across Intel® hardware, maximizing performance. The Intel® Distribution of OpenVINO™ toolkit includes the Intel® Deep Learning Deployment Toolkit (Intel® DLDT).

The Intel® Distribution of OpenVINO™ toolkit for Windows 10:

• Enables CNN-based deep learning inference on the edge
• Supports heterogeneous execution across Intel® CPU, Intel® Integrated Graphics, Intel® Movidius™ Neural Compute Stick, Intel® Neural Compute Stick 2, and Intel® Vision Accelerator Design with Intel® Movidius™ VPUs
• Speeds time-to-market via an easy-to-use library of computer vision functions and pre-optimized kernels
• Includes optimized calls for computer vision standards including OpenCV* and OpenCL™

For more details, refer to https://docs.openvinotoolkit.org/2021.1/openvino_docs_install_guides_installing_openvino_windows.html.

4.2.1 System Requirements

The following software must be installed on the system:

• Microsoft Build Tool for Visual Studio with C++ 2019, 2017, or 2015
• CMake 64-bit
• Python 64-bit

See Installing Required Dependencies for more information.

4.2.2 Install the Intel® Distribution of OpenVINO™ Toolkit Core Components

Download the Intel® Distribution of OpenVINO™ toolkit package file from https://www.intel.com/content/www/us/en/developer/tools/openvino-toolkit-download.html.

Select the following download options:

• Operation System: Windows
• Distribution: Online & Offline
• Version Type: 2021 1
• Installer Type: Offline

Enter your information and click Submit & Begin Download.

After downloading, run the installer. Follow the on-screen instructions to complete the installation.

4.2.3 Set the Environment Variables

Before compiling and running OpenVINO applications, the environment variables must be updated.

Open a command prompt and run the setupvars.bat file to temporarily set the environment variables.

Notes:

• OpenVINO toolkit environment variables are removed when the command prompt window is closed.
• If the Intel® Distribution of OpenVINO™ was not installed to the default install directory, replace C:\Program Files (x86)\Intel with the directory where it is installed.

4.2.4 Optional Steps for Intel® Processor Graphics (GPU) 

The steps in this section are required only if you want to use processor graphics (GPU) on your system.

If your applications offload computation to Intel® Integrated Graphics, the Intel Graphics Driver for Windows version must be 15.65 or higher. To see if you have this driver installed:

1. Type device manager in your Windows Search box. The Device Manager opens.
2. Click the drop-down arrow to view the display adapters. See the installed adapter.
   
3. Right-click the adapter name and select Properties.
4. Click the driver tab to see the driver version. Make sure the version number is 15.65 or higher.
   
5. If your device driver version is lower than 15.65, download and install a higher version from https://downloadcenter.intel.com/product/80939/Graphics-Drivers.

4.3 Installing the Intel Media SDK for Gstreamer (Optional)

If you are not using the MSDK encoder and decoder GStreamer plugins, this chapter can be skipped.

To use MSDK encoder and decoder GStreamer plugins in GStreamer, the required software must be installed in the following order:

1. Intel® OpenVINO Toolkit (see Installing the OpenVINO Toolkit)
2. Intel® Media SDK (this section)
3. GStreamer (see Installing GStreamer on Windows)
4. MSDK GStreamer plugin (see Install Third-party Plugins)

This section describes how to install the Intel® Media SDK.

1. Download the Intel® Media SDK for Windows from https://software.intel.com/content/www/us/en/develop/tools/media-sdk/choose-download/client.html.
2. Click Register & Download and complete the registration form.
3. On the Intel® Media SDK for Windows website, choose version 2020 R1 and click MediaSDK2019RD.exe.
   
4. After downloading, run the installer. Follow the on-screen instructions to complete the installation.

5 Installation Process for NVIDIA Solution (Optional)

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

5.1 Uninstall Previous Version

Uninstall all previous versions of CUDA Toolkit from the system and remove cuDNN and/or TensorRT files. For more details, refer to Uninstall Previous Software.

5.2 Installing NVIDIA Driver

According the NVIDIA product type, download the corresponding driver installer from the NVDIA Driver download website https://www.nvidia.com/download/index.aspx?lang=en-us.

For example, for NVIDIA Quadro P2000, the search items are as follows,

After searching for the appropriate product, follow the step on the website to download the driver.

After downloading, run the driver installer and follow the screen prompts to install the driver. It is recommended to reboot after installation to ensure the normal operation of the driver. After rebooting, open Device Manager, click Display adapters, and check that the NVIDIA device is working properly.

5.3 Installing CUDA Toolkit

Download the NVIDIA® CUDA toolkit 11.0 from https://developer.nvidia.com/cuda-11.0-download-archive?target_os=Windows&target_arch=x86_64&target_version=10&target_type=exelocal.

After downloading, run the installer and follow the screen prompts to install the software.

5.4 Installing cuDNN

Download cuDNN v8.0.5 from https://developer.nvidia.com/compute/machine-learning/cudnn/secure/8.0.5/11.0_20201106/cudnn-11.0-windows-x64-v8.0.5.39.zip.

Log in or join the NVIDIA Developer Program as a member.

After downloading, unzip the cuDNN package and copy the following files into the following CUDA Toolkit folders:

• Copy cuda\bin\*.dll to C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.0\bin
• Copy cuda\include\*.h to C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.0\include
• Copy cuda\lib\x64\*.lib to C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.0\lib\x64

Refer to the NVIDA cuDNN documentation for more information: https://docs.nvidia.com/deeplearning/cudnn/install-guide/index.html#install-windows.

5.5 Installing TensorRT

Download and install TensorRT 7.1.3 for Windows from https://developer.nvidia.com/compute/machine-learning/tensorrt/secure/7.1/zips/TensorRT-7.1.3.4.Windows10.x86_64.cuda-11.0.cudnn8.0.zip.

Log in or join the NVIDIA Developer Program as a member.

After downloading, unzip and copy the TensorRT-7.1.3.4 folder to C:\Program Files\NVIDIA GPU Computing Toolkit.

Add the following variables to the PATH environment variable.

• C:\Program Files\NVIDIA GPU Computing Toolkit\TensorRT-7.1.3.4\lib
• C:\Program Files\NVIDIA GPU Computing Toolkit\TensorRT-7.1.3.4\bin

For the instructions to add the system environment, refer to Set Windows Environment Variables.

5.6 Installing ONNX Runtime

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

Refer to https://www.onnxruntime.ai/

6 Installing Pylon Software (Optional)

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

6.1 Uninstall Previous Versions

Uninstall all previous versions of pylon or items where the publisher is Basler from the system. For more details, refer to Uninstall Previous Software.

6.2 Install pylon Software

If Basler cameras are used for the system, 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: 6.1.1
• Operating System: Windows 64-bit

Alternately, use the link: https://www.baslerweb.com/en/sales-support/downloads/software-downloads/#type=pylonsoftware;language=all;version=6.1.1;os=windows64bit

Choose Camera Software Suite Windows.

After downloading, run the camera software suite installer.

Follow the screen prompts to install the software.

On the Profiles page, choose Developer.

On the Interfaces page, select the interfaces to which the camera(s) is/are connected.

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

• After installing pylon, run the pylon Viewer tool to test the camera.
• Refer to the following Basler document on using the pylon Viewer:

https://docs.baslerweb.com/overview-of-the-pylon-viewer.html

7 Installing Hikrobot Software (Optional)

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

7.1 Uninstall Previous Versions

Uninstall all previous versions of MVS from the system. For more details, refer to Uninstall Previous Software.

7.2 Install Hikrobot MVS Software

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

Download the Hikrobot installer from the website at: https://en.hikrobotics.com/machinevision/service/download?module=0

The recommended software versions are:

• Version: Machine Vision Software MVS3.2.1 (Windows)
• Operating System: Windows 64-bit

Launch the downloaded installer.

The recommended options are:

• Select Driver: GIGE and USB 3.0
• Others: Enable built-in debug features and Enable Jumbo Frame for All NICs

Verify the Hikrobot's USB3/GigE Vision cameras with MVS.

• After installing MVS, run the MVS tool to test the camera.
• The file should be located at C:\Program Files\MVS\Applications\Win64\MVS.exe"
• Refer to the MVS document at C:\Program Files\MVS\Applications\Win64\doc

8 Installing FLIR Software (Optional)

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

8.1 Uninstall Previous Versions

Uninstall all previous versions of Spinnaker SDK from the system. For more details, refer to Uninstall Previous Software.

8.2 Install FLIR Spinnaker Software

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

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

The recommended software versions are:

• Version: SpinnakerSDK_FULL_2.5.0.80_x64 (Windows)
• Operating System: Windows 64 bit

Launch the download installer. In the Profiles page, choose Application Development.

Select all camera drivers and Visual Studio versions. These include the runtime library based on the Visual Studio code.

Enable the GigE Cameras.

Verify the FLIRs USB3/GigE Vision cameras with SpinView:

• After installing Spinnaker, run the SpinView tool to test the camera.
• The file should be located at “C:\Program Files\FLIR Systems\Spinnaker\bin64\vs2015\ SpinView_WPF.exe”.

9 Installing ROS 2 Foxy Fitzroy (Optional)

If you are not using the ROS 2 Foxy Fitzroy, this chapter can be skipped.

This chapter only describe the installation software required to support the rosgstbridge plugin in ADLINK EVA SDK.

9.1 Install System Requriements

The following software must be installed on the system:

• Chocolatey
• Visual C++ Redistributables 2013
• Windows 64-bit OpenSSL
• RTI Connext DDS
• OpenCV 3.4.6

9.1.1 Install Chocolately

Chocolatey is a package manager for Windows.

Download and install Chocolatey from https://chocolatey.org/

9.1.2 Install Visual C++ Redistributables 2013

Open a command prompt and enter the following command to install the package.

9.1.3 After installing, reboot the computer.

9.1.3 Install OpenSSL

Download and install Win64 OpenSSL v1.1.1m from https://slproweb.com/products/Win32OpenSSL.html.

Add the following variable to the PATH environment variable.

• C:\Program Files\OpenSSL-Win64\bin\

For instructions to add the system environment, refer to Set Windows Environment Variables.

9.1.4 Install RTI Connext DDS

Download and install RTI Connext DDS from https://www.rti.com/free-trial/dds-files-5.3.1.

9.1.5 Install OpenCV

Download a precompiled version of OpenCV 3.4.6 from https://github.com/ros2/ros2/releases/download/opencv-archives/opencv-3.4.6-vc16.VS2019.zip.

Unpack it and copy the opencv folder to C:\.

Add the following variable to the PATH environment variable.

• C:\opencv

For instructions to add the system environment, refer to Set Windows Environment Variables.

9.2 Install ROS 2

Download ROS 2 patch release 4 from

https://github.com/ros2/ros2/releases/download/release-foxy-20201211/ros2-foxy-20201211-windows-release.amd64.zip.

Unpack the zip file and copy all files under ‘ros2-windows’ to ‘c:\dev\ros2_foxy’.

Note: 

1. If you need to use ROS 2 commands and tools, you must install other required packages. For more details, refer to https://docs.ros.org/en/foxy/Installation/Windows-Install-Binary.html#.
2. EVA SDK must use Python 3.7. If you have python 3.8 installed, remember to use the following command to run the EVA python application.

10 Installing GStreamer and ADLINK EVA SDK on Windows

10.1 Uninstall Previous Version

Uninstall all previous versions of EVA from the system. For more details, refer to Uninstall Previous Software.

If gstreamer has been installed, use the following commands to remove msdk, nvdec and nvenc (or nvcodec).

• remove msdk plugin

• remove nvdec and nvenc  (or nvcodec) plugins

10.2 Install GStreamer and EVA SDK

Download the EVA install package and copy it to the Windows 64-bit system. Follow the installation screen prompts.

The installer includes GStreamer and the ADLINK EVA SDK files. After installation, the GStreamer files will be located at C:\ADLINK\gstreamer and the EVA SDK will be located at C:\ADLINK\eva.

The ADLINK EVA SDK must use a python plugin and third-party plugins for GStreamer. ADLINK has created a GStreamer installer to simplify the installation process. If there is any problem with the plugin provided by GStreamer, ask for help from the plugins’ developers. ADLINK cannot guarantee the performance and stability of any third-party plugins.

Note: Due to the limitation of the mason build tool, the ADLINK EVA SDK currently only supports drive C:\ as the installation and build location.

10.3 Install EVA SDK IDE and Samples (Optional)

If you are not using the EVA SDK IDE and Samples, this chapter can be skipped.

Download the EVA_IDE install package and copy it to the Windows 64-bit system. Follow the on-screen installation instructions.

The installer includes ADLINK EVA SDK IDE files and the samples. After installation, the files will be located in C:\ADLINK\eva.

10.4 Install Third-party Plugins (Optional)

The ADLINK GStreamer installer provides third-party plugins including msdk, and nvcodec.

The msdk plugin provides the GStreamer H.264 and HEVC format encoder and decoder elements using the Intel Media SDK. The nvcodec plugins provide the GStreamer H.264 and HEVC elements using NVIDIA Codec SDK and NVIDIA CUDA. The source code for these plugins is in the GStreamer Bad Plug-ins package and must be built by users as NVIDIA and Intel do not provide binary files. However, ADLINK has configured these plugins in advance.

Follow the steps below to install the plugins.

• Install required dependencies.
  • msdk plugin: refer to Installation Process for Intel Solution to install OpenVINO and Media SDK.
  • nvdec and nvenc plugin: refer to Installation Process for NVIDIA Solution to install NVIDIA Driver and CUDA toolkit
• Copy the required binary to C:\ADLINK\gstreamer\lib\gstreamer-1.0
  • for the msdk plugin: gstmsdk.dll and gstmsdk.lib

    > cd C:\ADLINK\gstreamer > xcopy third-party\gstmsdk* lib\gstreamer-1.0 /y

  • for the nvcodec plugin: gstnvcodec.dll and gstnvcodec.lib

    > cd C:\ADLINK\gstreamer > xcopy third-party\gstnv* lib\gstreamer-1.0 /y

Note: When using other third-party plugins, copy the plugin binary (*.lib and *.dll) to C:\ADLINK\gstreamer\lib\gstreamer-1.0 and the libraries that the plugin needs (If any) to C:\ADLINK\gstreamer\bin.

10.5 Set Environment Variables

Open a command prompt and run setupvars.bat to temporarily set the environment variables.

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

If the third-party plugins are installed, use the following commands to check them.

• Check msdk

  > gst-inspect-1.0 msdk

• Check nvcodecand nvenc

  > gst-inspect-1.0 nvcodec

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.

Getting Service

Ask an Expert: http://askanexpert.adlinktech.com

ADLINK Technology, Inc.
Address:No. 66, Huaya 1st Rd., Guishan District,
Taoyuan City 333411, Taiwan
Tel: +886-3-216-5088
Fax: +886-3-328-5706
Email: [email protected]

Ampro ADLINK Technology, Inc.
Address:6450 Via Del Oro
San Jose, CA 95119-1208, USA
Tel: +1-408-360-0200
Toll Free:+1-800-966-5200 (USA only)
Fax: +1-408-600-1189
Email: [email protected]

ADLINK Technology (China) Co., Ltd.
Address:300 Fang Chun Rd., Zhangjiang Hi-Tech Park, Pudong New Area
Shanghai, 201203 China
Tel: +86-21-5132-8988
Fax: +86-21-5132-3588
Email: [email protected]

ADLINK Technology GmbH
Address:Hans-Thoma-Straße 11
D-68163 Mannheim, Germany
Tel: +49-621-43214-0
Fax: +49-621 43214-30
Email: [email protected]

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