Standalone Setting ​

In the Standalone Setting the Scene2Model environment the Image Provider and the Scene2Model Modelling Tool are running on the users' device. Under Windows installation, the user can start the image provider directly from the modelling tool. For other variations the user must install and start the image provider separately.

In this setting, the Recognition Component runs on a reachable server. Therefore, the endpoint which is used for the recognition must be configured.

The figure below visualizes Standalone Setting's components and their relations.

The Image Provider consists of two parts, a script that takes picture with a Logitech C920 camera and sends them to Bucket Service Component. From this other applications, like the Recognition Component, can access the picture. The script uses ffmpeg to take the picture and curl to send the picture. There exists scripts for Windows, Linux and Mac OS. Therefore, any device running one of these operating systems can be used to take a picture and send it to the bucket. The scripted are tailored to only use the Logitech C920, but with knowledge in ffmpeg they can be extended to allow the usage of other cameras.

The Bucket Service Component and Recognition Component can run on one or multiple servers. It is important that the Recognition Component can reach the Bucket Service Component over the network over internet. The Scene2Model Modelling Tool must be able to reach the Recognition Component to load the needed information from there.

Therefore, the following connections must be enabled in the set-up:

• The Image Provider must be able to make HTTP calls to the Bucket Service Component.
• The Recognition Component must be able to make HTTP calls to the Bucket Service Component.
• The Scene2Model Modelling Tool must be able to make HTTP calls to the Recognition Component.
  • The Recognition Component also offers a web interface, to help the users of the Scene2Model Modelling Tool. Therefore, a browser on the user device should also be able to open web pages of Recognition Component.

The description of this setting focuses on starting the picture stream of the Image Provider directly from the modelling tool, which is only possible from installation on Windows. If you want to install the picture stream locally on another operating system you can do this via our S2M Bucket Image Provider project. Information on installation and usage are contained in the project

Configuration: End Point of the Recognition Component ​

Before a haptic model can be imported into the Scene2Model Modelling Tool, it must be configured to the prepared components.

Hints:

• Provided URLs should not have a / at the end.
• The concrete endpoints must be adapted to the concrete installation of the setting.
• The lines in the configuration window, which start with # are comments and ignored by the tool:

Configuration Parameters: ​

The following parameters regarding the recognition endpoint must be set:

• OliveEndpoint: URL of the Olive endpoint, over which the information for creating models can be gathered. It is the URL of the controller user interface, plus adding a /rest at the end.
• OliveMicroservice: One Olive instance offers multiple services. In this parameter the identifier of the service created for this instance must be provided.
• BucketAndRecognitionServiceEndpoint: URL of the Recognition Component must be provided
• BucketEndpoint: URL of the Bucket Service Component
• DefaultBucket: name of the bucket that should be used within the Bucket Service Component
• Camera: Name of the camera that should be used (at the moment only a Logitech C920 is supported, which has the name HD Pro Webcam C920)
• Scene2Model_Library_Directory: Path to the currently loaded Domain-specific library directory. This parameter is set automatically if a new library is loaded.

Template for configuration: ​

Scene2Model_Library_Directory=<path-to-domain-specific-library-directory>

OliveEndpoint=<olive-endpoint>
OliveMicroservice=<olive-microservice-identifier>

BucketAndRecognitionServiceEndpoint=<bucket-and-recognition-endpoint>
BucketEndpoint=<bucket-endpoint>
DefaultBucket=<default-bucket-name>

Camera=<camera-name>

Scene2Model_Library_Directory=<path-to-domain-specific-library-directory>

OliveEndpoint=<olive-endpoint>
OliveMicroservice=<olive-microservice-identifier>

BucketAndRecognitionServiceEndpoint=<bucket-and-recognition-endpoint>
BucketEndpoint=<bucket-endpoint>
DefaultBucket=<default-bucket-name>

Camera=<camera-name>

Configuration of the Endpoints ​

1. Open configuration Window

   • Click System tools
   • Click Edit internal configuration file
   • Mark scene2model.properties
   • Click OK

2. Change the needed parameters

3. Confirm by clicking Apply

Configuration of the Camera ​

1. Configure the camera

   • Click System tools
   • Click Edit internal configuration file
   • Click scene2model.properties
   • Click OK
   • Enter the Name of your camera next to the Camera=
     • Camera=HD Pro Webcam C920
   • Click Apply

   

Local Image Provider (Windows) ​

For windows a local image provide can be directly started from the modelling tool. To use this provider option, the webcam must be connected to the device which is running the modelling tool. The camera should be position to get a bird's eye view, so that the attached tags can be recognised. For this setting the top edge of the camera should be pointing in the same direction as the figures will face, once they are positioned. If the top edge of the camera is facing the wrong direction, the recognition still works, but figures in the modelling tool will be shown mirror-inverted.

ATTENTION: This functionality only works with a Logitech C920 camera, out of the box.

Start the Local Windows Image Provider ​

1. Start the picture stream (this description is only for windows)

   • Open the ´Scene2Model Wizard´
   • Choose Run local camera stream (Windows only with Logitech C920)
   • Click Select
   • Check your picture identifier (and update if necessary)
     • IMPORTANT: remove the .jpg from the name
   • A terminal window is opened, showing information about the picture stream. This window should only be closed, once the picture stream is not needed anymore.

   

2. (Optional) Check if the stream works, by clicking Showing recognition results (Yellow button with sun)

Information: To adapt this setting to a new camera, knowledge in FFmpeg are needed. The script starting the local camera stream is located in the installation folder of the modelling tool, under s2m\bucket-image-provider\local-ffmpeg-win. If the runWebStream.bat script is adapted to a new camera, it can also be started with the new camera out of the modelling tool.