Intelligent Traffic Management

Monitor traffic intersections via IP cameras to optimize traffic flow. Detect vehicles and pedestrians, record vehicle types and counts, calculate velocity and acceleration, and more.

Highlights

  • Track vehicles and pedestrians across video frames. Estimate the trajectory, speed, and position of moving vehicles, and identify collisions and near misses.

  • Experiment with accuracy and throughput tradeoffs by swapping out object detector models and tracking and collision detection algorithms.

  • Use a pre-trained Single Shot Object Detection (SSD) model from the Intel® Distribution of OpenVINO™ toolkit to analyze data from a traffic camera.

author-image

By

Overview

Intelligent Traffic Management is designed to detect and track vehicles as well as pedestrians and to estimate a safety metric for an intersection. Object tracking recognizes the same object across successive frames, giving the ability to estimate trajectories and speeds of the objects. The reference implementation also detects collisions and near misses. A real-time dashboard visualizes the intelligence extracted from the traffic intersection along with annotated video stream(s).   

This collected intelligence can be used to adjust traffic lights to optimize the traffic flow of the intersection, or to evaluate and enhance the safety of the intersection by allowing emergency services notifications, such as 911 calls, to be triggered by collision detection, reducing emergency response times. 

Try the reference implementation in the Intel® Developer Cloud for the Edge Container Playground powered by Red Hat* OpenShift*, a Kubernetes* environment for testing containers with Intel® hardware.

Launch on Intel® Developer Cloud for the Edge

Select Configure & Download to install and run Intelligent Traffic Management on your device. To get started with installation, view Documentation

Configure & Download

How It Works

The application uses the inference engine and the Intel® Deep Learning Streamer (Intel® DL Streamer) included in the Intel® Distribution of OpenVINO™ toolkit. The solution is designed to detect and track vehicles and pedestrians and upload cloud data to an AWS* S3 storage.

The Intelligent Traffic Management application requires the services pods, database and a visualizer. Once the installation is successful, the application is ready to be deployed using Helm. After the deployment, the application pod takes in the virtual/real RTSP stream addresses and performs inference and sends metadata for each stream to the InfluxDB database. The visualizer in parallel shows the analysis over the metadata like pedestrians detected, observed collisions and processed video feed.

The application has capability to perform inferences over as much as 20 channels. In addition, the visualizer is capable of showing each feed separately as well as all the feeds at the same time using Grafana*. The user can visualize the output remotely over a browser, provided that they are in the same network.

The architecture is represented by a complex block diagram. The upper half of the diagram shows the data flow for the reference implementation (RI). A camera passes data to the RI itself, which is comprised of blocks labelled Video Inference, Analytics, Cloud Connector, Rule Engine, and Dashboard. Next, the data flows to Amazon Web Services in the cloud, where it is displayed using the Amazon Web Services dashboard. The lower half of the diagram shows software and hardware components used by the RI divided into three categories: third-party developed, open source, and Intel developed hardware. Third party components are Kubernetes, Influx DB, Harbor, Grafana, HiveMQ MQTT, and PostgreSQL. The open source component is Linux. The Intel developed hardware is an Edge Compute Node using Intel Atom® processors, Intel® Core™ processors, or Intel® Xeon® processors.
Figure 1: Architecture Diagram

Try It

Launch on Intel® Developer Cloud for the Edge

Try the reference implementation in the Intel® Developer Cloud for the Edge Container Playground powered by Red Hat* OpenShift*, a Kubernetes* environment for testing containers with Intel® hardware.

Launch on Intel® Developer Cloud for the Edge

Features

The following are features are included with Intel® Developer Cloud for the Edge:

  • Container Images: grafana:itm, influxdb:1.8, itm:1.5 
  • Software Stack: Intel® Distribution of OpenVINO™ Toolkit (Intel® Deep Learning Streamer [Intel® DL Streamer]), InfluxDB*, Grafana containers in a Helm-chart 
  • Task: Object Detection, Tracking and Classification with Intel® DL Streamer (gstreamer pipeline) 
  • Model Info: Person, vehicle, and bike detection pytorch model with MobileNetV2 backbone 
  • Precisions Supported: NA 
  • Outputs: Exposes 2 web-services (URLs). One link showing annotated input streams with fps, tracking IDs, bounding boxes for pedestrians and vehicles with collisions, and one for Grafana dashboard.

Download

Select Configure & Download to install and run Intelligent Traffic Management on your device. To get started with installation, view Documentation.

Configure & Download

Features

Time to complete: 30 - 45 minutes

Programming Language: Python* 3

Available Software: The following are included in the Intelligent Traffic Management zip file:  

 

  • Intel® Distribution of OpenVINO™ toolkit 2021 Release
  • Kubernetes*

Recommended Hardware

The hardware below is recommended for use with this reference implementation. For other suggestions, see Recommended Hardware

 

Learn More

To continue your learning, see the following guides and software resources:

Support Forum 

If you're unable to resolve your issues, contact the  Support Forum.