ADLINK Technology EVA SDK Manual de usuario
EVA SDK User's Manual
Manual Rev.: 1.2
Revision Date: July 20, 2021
Part Number: 50M-00010-1020
ii Preface
Preface
Copyright
Copyright © 2021 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
Revision Description Date
1.0 Initial release 2020-09-30
1.1 Release for EVA SDK R3 (Windows) 2021-04-12
1.2 Release for EVA SDK R3.5 2021-07-20
EVA SDK User's Manual
Preface iii
Table of Contents
Preface................................................................................................................................ii
1Introduction................................................................................................................1
2Setting Environment Variables.................................................................................5
3Image and Video Capture Plugins............................................................................7
3.1 Element Descriptions.....................................................................................................7
4Image Processing Plugins......................................................................................33
4.1 Element Descriptions...................................................................................................35
5Image Analytic Plugins ...........................................................................................75
5.1 Element Descriptions...................................................................................................76
6Connector Plugins.................................................................................................103
6.1 Element Descriptions.................................................................................................103
7Convert for OpenVINO Model...............................................................................105
7.1 Model Optimizer Developer Guide.............................................................................105
7.2 Converting a Caffe* Model.........................................................................................105
7.3 Converting a TensorFlow* Model...............................................................................105
7.4 Converting an ONNX* Model.....................................................................................105
7.5 Converting a Tensorflow YoloV3 Model.....................................................................105
8Convert for TensorRT Model................................................................................107
8.1 Build TensorRT Engine Parameters..........................................................................107
8.2 Convert Tensorflow models.......................................................................................108
8.3 Samples.....................................................................................................................108
Safety Instructions.........................................................................................................112
Getting Service...............................................................................................................113
iv Preface
This page intentionally left blank.
EVA SDK User's Manual
Introduction 1
1 Introduction
(EVA Architecture)
The ADLINK Edge Vision Analytics (EVA) SDK provides an integrated development environment (IDE) for
developers wanting to build computer vision and video analytic solutions with deep learning technology. The SDK
provides the required building blocks to help with the implementation of edge inference solutions. Customers can
select either the Intel® Distribution of OpenVINO™ Toolkit, NVIDIA TensorRT™, or both to accelerate deep
learning inference. This SDK can help customers easily migrate or integrate heterogeneous hardware platforms.
ADLINK offers a selection of pre-verified platforms with specific hardware configurations optimized for vision and
video analytic applications requiring little additional engineering effort by developers. The EVA SDK integrates and
verifies different building blocks required for Edge Vision Analytic applications. The building blocks are
implemented as GStreamer plugins. Customers can select a specific plugin based on their requirements, utilizing it
in their development pipeline to quickly build their application. These plugins can be categorized as follows.
•Image/video capture: Supports different image capture solutions with different camera standards.
•Image processing: Provides the required algorithms needed before or after the inference.
•Image analytics: Provides an optimized INTEL/NVIDIA inference engine to accelerate deep learning inference.
•Connector plugin: Provides network connections for uploading images and inference results to a network
service.
All the plugins can be easily assembled to become a pipeline that can form an application. There are three kinds of
plugins: source providers (image/video capture), filters (image processing/image analytics), and data consumers
(connector plugins). Source providers play a leading role in the pipeline, transmitting data to an intermediate filter
that deals with the data for specific processing. The data can then be consumed by the final plugins for data
storage and transmission.
Under the EVA platform, the image capture plugin provides the video stream data to the image preprocessing
plugin, as shown in the figure below. The image processing plugin then delivers the processed data to the AI
inference engine, and the AI inference engine delivers the result. After the intermediate filter process, the data is
published to the data river via a plugin, where other edge devices like OPC UA or robots can subscribe to it.
2 Introduction
The EVA SDK supports application development in C/C++ as well as in Python, with the ability to implement
customized plugins in both programming languages. (Sample code is included in the EVA SDK package.)
The following use scenarios demonstrate how the EVA SDK can be used to easily create vision analytic
applications.
Use Scenario 1: Single pipeline in one application
This scenario demonstrates a simple edge AI application. With the configured pipeline, the image inference can be
made with a single image source on an ADLINK edge AI platform camera (e.g. Neon).
1. Application (1) captures the image from a GigE camera
2. Application (1) does AI inference on Nvidia TensorRT via adrt plugin
3. Application (1) does the inference result interpretation with the adtranslator plugin
Use Scenario 2: Multiple pipelines in one application
This scenario demonstrates an edge AI application. With multiple pipelines, the image inference can be captured
with multiple image sources and the AI inference performed on both Nvidia and Intel hardware with an ADLINK
edge AI platform (e.g. EOS-i).
1. Application (1) captures the image from the IP camera
2. Application (1) decodes the RTSP via NV H.264 HW accelerator
EVA SDK User's Manual
Introduction 3
3. Application (1) does the AI inference for camera 1 on OpenVINO via the advino plugin
4. Application (1) does the inference result interpretation with the adtranslator plugin
5. Application (1) captures the image from the GigE camera
6. Application (1) does the AI inference for the GigE camera on TensorRT via the adrt plugin
7. Application (1) does the inference result interpretation with the adtranslator plugin
Notes:
•The examples using gst-launch in this User’s Manual are for reference only. The “property” values, such as the
file name and path, must be modified according to the actual system environment.
•Support for the elements in the examples are operating system dependent. For example, ‘xvimagesink’ is only
supported on Linux. It is recommended to use the ‘glimagesink’ or ‘d3dvideosink’ elements instead. Note that
only one glimagesink can be used in a pipeline.
4 Introduction
This page intentionally left blank.
EVA SDK User's Manual
Setting Environment Variables 5
2 Setting Environment Variables
Before using EVA SDK, the environment must be installed and set up. For installation details, refer to the EVA
Installation Guide.
On Linux (include x86 and ARM), after installing EVA SDK and required software, open a terminal and run the
following command to set the environment variables.
$ source /opt/adlink/eva/scripts/setup_eva_envs.sh
If you have changed the install path (INSTALL_DIR), be sure to specify the path.
$ source INSTALL_DIR/scripts/setup_eva_envs.sh
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 terminal.
On Windows, after installing EVA SDK and required software, there are two ways to execute the EVA IDE.
Method 1: Manually run the environment variable settings.
Open a Windows command prompt and then run the following commands to set the environment variables; then
enter the GStreamer command or your application to execute.
> C:\ADLINK\gstreamer\setupvars.bat
> C:\ADLINK\eva\scripts\setup_eva_envs.bat
Note: The environment variables are removed when closing the command prompt.
Method 2: Run the desktop shortcut.
On Windows, run EVACMD on the desktop. The shortcut will automatically set the environment and open the
command prompt. You can then enter the GStreamer command.
6 Setting Environment Variables
This page intentionally left blank.
Tabla de contenidos
Otros manuales de Cámara digital de ADLINK Technology
ADLINK Technology
ADLINK Technology NEON-2000-JNX Series Manual de usuario
ADLINK Technology
ADLINK Technology NEON-2000-JNX Series Manual de usuario
ADLINK Technology
ADLINK Technology NEON-1000-MDX Series Manual de usuario
ADLINK Technology
ADLINK Technology NEON-1021-M Manual de usuario
ADLINK Technology
ADLINK Technology NEON-1020 Manual de usuario
