- Time to Complete: 30-45 minutes
-
Verify you meet the System Requirements.
-
Install Prerequisites such as Docker and other required software.
Note that these operations must be executed when logged in as a standard (non-root) user. Do NOT use root or sudo.
-
Download the Intel® SceneScape software archive from https://github.com/open-edge-platform/scenescape/releases.
-
Extract the Intel® SceneScape archive on the target Ubuntu system. Change directories to the extracted Intel® SceneScape folder.
cd scenescape-<version>
Clone the repository and change directories to the cloned repository:
git clone https://github.com/open-edge-platform/scenescape.git &&
cd scenescape/Note: The default branch is main. To work with a stable release version, list the available tags and checkout a specific version tag:
git tag
git checkout <tag-version>Build container images:
makeThe build may take around 15 minutes depending on target machine. This step generates common base docker image and docker images for all microservices.
By default, a parallel build is being run with the number of jobs equal to the number of processors in the system.
Optionally, the number of jobs can be adjusted by setting the JOBS variable, e.g. to achieve sequential building:
make JOBS=1make list-dependenciesThis step generates dependency lists. Two separate files are created for system packages and Python packages per each microservice image.
Before deploying the demo of Intel® SceneScape for the first time, please set the environment variable SUPASS with the super user password for logging into Intel® SceneScape. Important: This should be different than the password for your system user.
export SUPASS=<password>make demoIf you are running remotely, connect using https://<ip_address> or https://<hostname>, using the correct IP address or hostname of the remote Intel® SceneScape system. If accessing on a local system use https://localhost. If you see a certificate warning, click the prompts to continue to the site. For example, in Chrome click "Advanced" and then "Proceed to <ip_address> (unsafe)".
Note: These certificate warnings are expected due to the use of a self-signed certificate for initial deployment purposes. This certificate is generated at deploy time and is unique to the instance.
Enter "admin" for the user name and the value you typed earlier for SUPASS.
Intel® SceneScape uses Docker Compose profiles to organize services into logical groups. When starting or stopping services, you must specify the same profile(s) used during deployment.
The following profiles are available:
| Profile | Description |
|---|---|
controller |
Scene Controller in default mode (analytics + tracking). Used by make demo. |
analytics |
Scene Controller in analytics-only mode (without tracking). |
experimental |
Enables mapping and cluster-analytics services. |
mapping |
Enables mapping service only. |
cluster-analytics |
Enables cluster-analytics service only. |
vdms |
Enables the VDMS visual database service (used for re-identification). |
tracker |
Enables the tracker service. |
Profiles can be specified on the command line with --profile:
docker compose --profile controller up -dMultiple profiles can be combined:
docker compose --profile controller --profile experimental up -dAlternatively, profiles can be set via the COMPOSE_PROFILES environment variable:
export COMPOSE_PROFILES=controller
docker compose up -dFor multiple profiles, use a comma-separated list:
export COMPOSE_PROFILES=controller,experimental
docker compose up -dFor more details, see the Docker Compose profiles documentation and the COMPOSE_PROFILES environment variable reference.
Note: The
--profileflags used withdocker compose downmust match those used when starting the services. Otherwise, containers started under a specific profile will remain running.
To stop the containers, use the following command in the project directory (see Docker Compose Profiles for details on choosing profiles):
docker compose --profile controller down --remove-orphansTo start after the first time, use the following command in the project directory:
docker compose --profile controller up -dIntel® SceneScape was downloaded, built and deployed onto a fresh Ubuntu system. Using the web user interface, Intel® SceneScape provides two scenes by default that can be explored running from stored video data.
Note: The “Documentation” menu option allows you to view Intel® SceneScape HTML version of the documentation in the browser.
- Check the How-to Guides for step-by-step instructions on how to perform specific tasks in Intel® SceneScape.
-
How to Define Object Properties: Step-by-step guide for configuring the properties of an object class.
-
How to enable reidentification: Step-by-step guide to enable reidentification.
-
Geti AI model integration: Step-by-step guide for integrating a Geti trained AI model with Intel® SceneScape.
-
Running License Plate Recognition with 3D Object Detection: Step-by-step guide for running license plate recognition with 3D object detection.
-
How to Configure DL Streamer Video Pipeline: Step-by-step guide for configuring DL Streamer video pipeline.
-
Model configuration file format: Model configuration file overview.
-
How to Manage Files in Volumes: Step-by-step guide for managing files in Docker and Kubernetes volumes.
-
How to upgrade Intel® SceneScape: Step-by-step guide for upgrading from an older version of Intel® SceneScape.
-
How Intel® SceneScape converts Pixel-Based Bounding Boxes to Normalized Image Space
-
Hardening Guide for Custom TLS: Optimizing security posture for a Intel® SceneScape installation.
