Getting Started Guide - Metro Vision AI SDK#

Overview#

The Metro Vision AI SDK provides a comprehensive development environment for computer vision applications using Intel’s optimized tools and frameworks. This guide demonstrates the installation process and provides a practical object detection implementation using DLStreamer and OpenVINO.

Learning Objectives#

Upon completion of this guide, you will be able to:

  • Install and configure the Metro Vision AI SDK

  • Execute object detection inference on video content

  • Understand the basic pipeline architecture for computer vision workflows

System Requirements#

Verify that your development environment meets the following specifications:

  • Operating System: Ubuntu 24.04 LTS or Ubuntu 22.04 LTS

  • Memory: Minimum 8GB RAM

  • Storage: 20GB available disk space

  • Network: Active internet connection for package downloads

Installation Process#

Execute the automated installation script to configure the complete development environment:

curl https://raw.githubusercontent.com/open-edge-platform/edge-ai-suites/refs/heads/main/metro-ai-suite/metro-sdk-manager/scripts/metro-vision-ai-sdk.sh | bash

Metro Vision AI SDK Installation

The installation process configures the following components:

  • Docker containerization platform

  • Intel DLStreamer video analytics framework

  • OpenVINO inference optimization toolkit

  • Pre-trained model repositories and sample implementations

Object Detection Implementation#

This section demonstrates a complete object detection workflow using the installed components.

Step 1: Create Working Directory#

Create a dedicated working directory for the tutorial:

mkdir -p ~/metro/metro-vision-get-started-tutorial
cd ~/metro/metro-vision-get-started-tutorial

Step 2: Download Sample Video and model#

Download the required video sample and pre-trained model components:

wget -O sample.mp4 https://github.com/intel-iot-devkit/sample-videos/raw/master/person-bicycle-car-detection.mp4
mkdir -p models/intel/pedestrian-and-vehicle-detector-adas-0001/FP32/
wget -O "models/intel/pedestrian-and-vehicle-detector-adas-0001/FP32/pedestrian-and-vehicle-detector-adas-0001.xml" "https://storage.openvinotoolkit.org/repositories/open_model_zoo/2023.0/models_bin/1/pedestrian-and-vehicle-detector-adas-0001/FP32/pedestrian-and-vehicle-detector-adas-0001.xml?raw=true"
wget -O "models/intel/pedestrian-and-vehicle-detector-adas-0001/FP32/pedestrian-and-vehicle-detector-adas-0001.bin" "https://storage.openvinotoolkit.org/repositories/open_model_zoo/2023.0/models_bin/1/pedestrian-and-vehicle-detector-adas-0001/FP32/pedestrian-and-vehicle-detector-adas-0001.bin?raw=true"
wget -O "models/intel/pedestrian-and-vehicle-detector-adas-0001/pedestrian-and-vehicle-detector-adas-0001.json" "https://raw.githubusercontent.com/dlstreamer/dlstreamer/refs/heads/master/samples/gstreamer/model_proc/intel/pedestrian-and-vehicle-detector-adas-0001.json"

Step 3: Pipeline Execution#

Execute the object detection pipeline using the configured assets.

# Allow X11 connections from local clients
xhost +

# Start the container
docker run --rm -it --name dlstreamer \
  -v $PWD:/data \
  -e DISPLAY=$DISPLAY \
  -v /tmp/.X11-unix:/tmp/.X11-unix \
  intel/dlstreamer

# Run the GStreamer pipeline inside the running container
gst-launch-1.0 filesrc location=/data/sample.mp4 ! \
  decodebin ! videoconvert ! \
  gvadetect model=/data/models/intel/pedestrian-and-vehicle-detector-adas-0001/FP32/pedestrian-and-vehicle-detector-adas-0001.xml \
    model-proc=/data/models/intel/pedestrian-and-vehicle-detector-adas-0001/pedestrian-and-vehicle-detector-adas-0001.json \
    device=CPU ! \
  gvawatermark ! videoconvert ! autovideosink

# To exit the container
exit

Object Detection

Pipeline Architecture Analysis#

The executed command implements a processing pipeline with the following stages:

  1. File Input: Reads the specified video file from the filesystem

  2. Video Decoding: Converts encoded video stream to raw frame data

  3. Object Detection: Applies the pre-trained model to identify pedestrians and vehicles

  4. Visualization: Renders bounding boxes and classification labels on detected objects

  5. Display Output: Presents the annotated video stream through the system display

The resulting output displays the original video content with overlaid detection annotations indicating identified objects and their classification confidence levels.

Technology Framework Overview#

DLStreamer Framework#

DLStreamer provides a comprehensive video analytics framework built on GStreamer technology. Key capabilities include:

  • Multi-format video input support (files, network streams, camera devices)

  • Real-time inference execution on video frame sequences

  • Flexible output rendering and storage options

  • Modular pipeline architecture for custom workflow development

OpenVINO Optimization Toolkit#

OpenVINO delivers cross-platform inference optimization for Intel hardware architectures. Core features include:

  • Model format standardization through Intermediate Representation (IR)

  • Hardware-specific performance optimization

  • Extensive pre-trained model repository

  • Development tools for model conversion and validation

Next Steps#

Continue your learning journey with these hands-on tutorials:

Tutorial 1: OpenVINO Model Benchmark#

Learn to benchmark AI model performance across different Intel hardware (CPU, GPU, NPU) and understand optimization techniques for production deployments.

Tutorial 2: Multi-Stream Video Processing#

Build scalable video analytics solutions by processing multiple video streams simultaneously with hardware acceleration.

Tutorial 3: Real-Time Object Detection#

Develop a complete object detection application with interactive controls, performance monitoring, and production-ready features.

Tutorial 4: Advanced Video Analytics Pipelines#

Create sophisticated video analytics using Intel® DL Streamer framework, including human pose estimation and multi-model integration.

Tutorial 5: Profiling#

Profiling and monitoring performance of Metro Vision AI workloads using command-line tools.

Additional Resources#

Technical Documentation#

  • DLStreamer - Comprehensive documentation for Intel’s GStreamer-based video analytics framework

  • DLStreamer Pipeline Server - RESTful microservice architecture documentation for scalable video analytics deployment

  • OpenVINO - Complete reference for Intel’s cross-platform inference optimization toolkit

  • OpenVINO Model Server - Model serving infrastructure documentation for production deployments

  • Edge AI Libraries - Comprehensive development toolkit documentation and API references

  • Edge AI Suites - Complete application suite documentation with implementation examples

Support Channels#