© 2019 EDGEMATRIX Inc.

PDF is here.

CONTENTS

1. Information about this document

Document Organization

1. Information about this document

This document describes details of this manual.

2. Overview of the Edge AI Device Console

This section provides an overview and describes principal functions of the Edge AI [y2] Device Console.

3. Basic Operations

This section describes the series of processes starting with the creation of projects to the installation of AI apps and to the status monitoring.

4. Details of Functions

This section describes details of individual functions, such as the project management function, device management function and stream management function.

Related Documents

2. Overview of the Edge AI Device Console

What is the Edge AI Device Console?

The Edge AI Device Console is a managed console for central management of the Edge AI BOX. This console is also equipped with functions for managing status and update of AI apps.

Using the Edge AI Device Console makes following features available:

• Edge AI Box setup

• Remote operation of installed cameras

• Monitoring status of Edge AI Box and cameras

• Verification of images captured by cameras

• AI app setup

• Installation of AI apps

Supported operating environments

The console supports following web browsers:

• Google Chrome/PC

• Firefox/PC (supported video codecs are VP8 and VP9)

• Edge/Windows (Chromium-based Edge)

• Safari/iOS (*)

• Google Chrome/Android (*)

Note*: A portion of the screen may not be displayed correctly. Supported video codecs are VP8 and H.264.

Supported languages

Japanese and English

How to switch languages

The display ordinarily follows the OS and browser language settings. Setup a browser to specify a language. If the specified language is not supported, the display will be in English.

○ Chrome

1. Open the Chrome menu and click “Environment Settings”.

2. Go to the bottom of the screen and click “Detailed Settings”.

3. Click “Language” under the “Language” option.

4. If the desired language is not available, select “Add Language”.

5. Click the “:” (three dots) icon at the right edge and move the desired language to the top.

○ Firefox

Perform settings according to the Firefox Support Page.

○ Edge

Display the Language Settings page according to the Edge Support Page and set the priority language. It is not necessary to execute “Display Microsoft Edge in this language”.

Explanation of key terms

Edge AI Box

The Edge AI Box is a hardware that features the performance and functions for AI processing large data on the edge side.

Using the Edge AI Box makes it possible to process AI apps for images of cameras on site.

Edge AI Device Console

The Edge AI Device Console is a managed console for the central management of the Edge AI Box and AI apps. This console enables a complete control for setting and status monitoring of the Edge AI Box, installation of AI apps, etc.

AI Apps

AI Apps operate on the Edge AI Box, for AI processing of images captured by cameras on site.

Streams

Streams are a source of image data used by the Edge AI Box.

Projects

Projects are created and managed on the Edge AI Device Console. Streams and AI apps can be managed for each project.

Others

ONVIF: An abbreviation for the “Open Network Video Interface Forum” and represents a remote control process for network cameras (compatible with Axis, Bosch and SONY specifications).

PTZ: An abbreviation for the “Pan Tilt Zoom” for controlling camera field of view. The term “Pan” refers to the horizontal camera motion, while “Tilt” refers to the camera rotation on the vertical axis and “Zoom” refers to the enlarging and shrinking the field of view.

ROI: An abbreviation for the “Region of Interest” and refers to the area selected for manipulating, within the image data.

WebRTC: An abbreviation for the “Web Real-Time Communication” and refers to the standard for the real-time handling of video and audio data via a web browser.

The following chapter outlines basic operations of these functions, as well as details of various features.

3. Basic Operations

- From registration settings to adopting AI applications and verifying status -

This chapter explains the basic operations for enabling the use of AI apps, after the Edge AI Box is installed.

General Flow

The flow of steps for enabling the use of AI apps on the Edge AI Box is outlined below:

1. Registration of cameras

2. Stream setup

3. Start of camera operation

4. Verifying operation via monitor view (AI apps not installed)

5. Installation of AI apps

6. AI app manifest setup

7. Start of camera operation

8. Verifying operation via monitor view (AI apps installed)

9. Registration of projects

10. Adding cameras to projects

Enter camera stream registration information.

A stream ID can be created using alphanumeric characters, “_” (underscore) and “-” (hyphen).

A minimum of 8 characters are required. Set any desired name. Other items may be left unentered.

Once the entry is completed, press the Add button.

Once a camera is added, the name will be displayed on the Device Details screen.

Click on the name of the camera to set the camera stream.

2．Stream setup

On the Camera Details screen, set the stream by pressing the Set button to get to the Stream setup.

The URL may be acquired automatically with ONVIF-compatible devices. Wait until the underline motion stops.

If the device is not ONVIF-compatible, look up the URL of the RTSP for the stream on the camera manual, etc.

The entry should look like “rtsp://nnn.nnn.nnn.nnn/mpeg4/media.amp”.

Set the RTSP user name and password.

3．Start of camera operation

Click the Start button on the Camera Details screen to start capturing images with the camera.

4．Verifying operation via monitor view (AI apps not installed)

Select the monitor view on the Camera Details screen to open the monitor view screen, and verify that the camera is operating normally.

5．Installation of AI apps

If an app is not already installed, the “Not installed” test message and the Install button are displayed. Click the Install button to display the Package Installation screen.

Select an app on the Package Installation screen, and start installation.

6．Manifest setup

Typically, there is no need to edit the manifest directly. Refer to details by navigating to Manifest edit from Camera management.

7．Start of camera operation

Click the Start button on the Camera Details screen to start capturing images with the camera.

8．Verifying operation via monitor view (AI apps installed)

Select the monitor view on the Camera Details screen to open the monitor view screen, and verify that the camera is operating normally.

Furthermore, adjust the angle of view in the PTZ settings as needed.

* Procedures of 1 to 8 described above are performed as a part of the Edge AI Box and camera installation, conducted by installation personnel as requested by the device administrator.

9．Registration of projects

10．Adding cameras to projects

Click on the switch at the right edge of the created project name and add the video stream to the project.

This concludes the flow of steps for enabling the use of AI apps on the Edge AI Box.

4. Details of Functions

Project Management

Project Overview Screen

This screen will normally appear after login. The following operations are available on the Project Overview screen:

Addition of projects

New projects can be created with this feature.

A project ID can be created using alphanumeric characters, “_” (underscore) and “-” (hyphen). A minimum of 8 characters are required.

Set any desired name.

Enabling the user support operation allows the user to request for projects and video stream operations by the Device Console operation staff. When this is disabled, only the project members can operate it.

Others

• Navigate to respective Project Details screens

• Navigate to camera maps for each registered project

• Navigate to the Mounted Device Overview screen (*only the device manager is permitted to operate this function)

• Navigate to the mounted device map (*only the device manager is permitted to operate this function)

Project Details Screen

Information of respective projects can be edited and added video streams can be verified on this screen.

Editing projects (circle icon at upper right)

Project name, descriptions and user support operation settings can be changed with this feature.

List of users affiliated with the project

Users can be added, roles set and deleted from the project.

A confirmation request message is sent to the email address of the user and the user must acknowledge this by accessing the URL indicated on this message with a web browser to complete the procedure for adding the user. The user will not be able to access the project until this acknowledgment is completed.

○ Information on user roles

• Project Owner

A project owner is a project member who can operate all available functions on the device console.

• Project User

A project user is a project member with access to following operations limited in comparison with the project owner:

• Changing project name and descriptions

• Adding and deleting project notifications

• Adding and deleting project members

• Project Viewer

A project viewer are only able to browse

Overview of project notification recipients

The project owner can register email addresses to the project, for sending notifications on the status of the Edge AI Box linked with cameras that are added to the project.

A confirmation request message is sent to the email address for registering, and the user of the address must acknowledge this by accessing the URL indicated on this message with a web browser to complete the procedure for the registration.

○ Email notifications

The following combinations can be specified:

• Device events

A notification is sent when a device is determined to be normal or offline, based on the verification results on the status of the Edge AI Box (an offline status alert is deemed to be specified also).

Currently, a notification is sent once every 5 minutes. This time interval may be changed.

• Offline alert

A notification is sent when the Edge AI Box is determined to be offline (this action is same as previously available rule and device alerts).

• Storage alert

A notification is sent when a problem is detected with the storage of the Edge AI Box.

• Camera alert

A notification is sent when a problem is detected with a camera registered to a project.

○ Notification registration acknowledgment request email message

• Do not send

Select “Do not send” to start a simultaneous registration of multiple addresses. Performing registration of email addresses with “Send” option selected when registering the final of all email addresses for registering, the notification registration acknowledgment request email message is sent to all unacknowledged email addresses.

• Send

The notification registration acknowledgment request email message is sent to all unacknowledged email addresses.

Overview of cameras added to a project

This feature facilitates navigation to the Camera Details screens for cameras registered to a project.

Deleting projects

If the logged-in user is the only member of a project in question, and if that user is the project owner, the user may delete the project.

Camera map screen

The logged-in user can verify installation sites and statuses of cameras managed by the user, with icons.

Click on the name of the camera to navigate to the Camera Details screen.

Enter a check mark in the check box located to the left of each camera, to display monitor views. Entering a check mark in multiple check boxes results in a combined display of all selected monitor views.

Device management

Mounted device overview screen

The logged-in user can verify statuses of Edge AI Box and cameras associated with it, managed by the user.

Clicking on a device navigates to the Device Details screen, and clicking on a camera navigates to the Camera Details screen.

Mounted device map screen

Statuses of devices

Statuses of devices are displayed with gray, green, yellow or red icons. These icons represent offline, online, or warning status for a given device.

Editing of devices (circle icon at upper right)

This feature enables setting of names, descriptions, addresses, latitudes and longitudes of devices. Set longitudes and latitudes are used to display locations of devices on the map displayed on the map view.

Status display

This feature provides details on statuses of devices for reference. The boot drive usage, secondary drive usage, OS load average, OS memory usage, GPU temperature and statuses of added streams can be verified with this feature.

Reboot

This feature is used to reboot the device.

HDMI output settings

Refer to the HDMI Output Guide.

License verification

This feature is used to verify applications licensed for use on the AI Box.

Camera overview

This feature is used to navigate to Camera Details screen of the added camera.

Camera status display

This feature displays device and stream statuses. Statuses of devices are displayed with gray, green, yellow or red icons. The following 4 statuses are normally displayed:

- Offline/status unknown

- Verifying status

- Online/stopped

- Online/running

If statuses of devices require attention, they are displayed.

Editing camera (circle icon at upper right)

This feature can be selected by a device manager.

This feature enables setting of names, descriptions, addresses, latitudes and longitudes of devices.

Set longitudes and latitudes are used to display locations of devices on the map displayed on the map view. Longitudes and latitudes set to devices are set as default values.

Editing manifest (circle icon at upper right)

This feature can be selected by a device manager and a project owner.

This feature enables detailed operating settings for streams and AI apps. Typically, there is no need to edit the manifest directly.

Acquire details that should be entered to a manifest from the app vendor.

Direct editing of manifest may be warranted in cases described below:

• Setting up special RTSP URL.

• Manually updating revision number for troubleshooting purposes.

• Copying previous settings for action_rules. Refer also to details by navigating to Camera action rules from Application management.

• Setting ROI or reverting to 0. Refer also to details on the Region of Interest (ROI) Setup screen.

Details display

This feature is displayed for a device manager.

This feature displays descriptions, addresses, latitudes and longitudes, as well as URL, LoginID, Password information, UUID, ONVIF and other information on the cameras.

Manifest display

This feature is displayed for a device manager and a project owner.

This feature enables referencing of the original text on the manifest for a camera.

Start

This feature starts the operation of cameras and AI apps.

Stop

This feature stops the operation of cameras and AI apps.

Monitor view

This feature opens the monitor view. A camera must be operating.

Recording view

This feature searches and plays back recorded video images. A camera must be operating. Refer to the recording view for details.

Record settings

This feature is used to perform settings relating to recording. Refer to the Record Settings screen for details.

License verification

This feature is used to verify applications licensed for use on the AI Box to which the camera is registered.

Settings - ONVIF settings

This feature is used to perform the IP, user name and password for using a camera that supports ONVIF. These are required to use the PTZ settings.

Settings - Stream editing

This feature is used to set the URL for connecting cameras.

Settings - PTZ settings

This feature is used to open the Camera Angle of View Adjustment screen. Refer to the Pan-Tilt-Zoom (PTZ) Settings screen for details.

Settings - ROI settings

This feature is used to specify an area for AI app processing within the screen of the camera. Refer to the Region of Interest (ROI) Setup screen for details.

Settings - WebRTC settings

The WebRTC settings are settings that are relevant to monitor views.

• The encode option is used to select a video codec for use during a monitor view.

Set the encoding to VP8 or VP9 to use Firefox on a PC.

Set the encoding to VP8 to use iOS or Android.

• The expire option is a parameter for the communication protocol for sending video images for the monitor view and it is used to set the number of seconds to timeout until communication is established. Ordinarily, there is no need to make changes.

• The iceTransportPolicy is set to “all” when using all servers, including the turn server. Select “relay” to communicate only with the turn server. For normal use, leave everything set to “all”.

• The userelay is a setting used to determine whether or not to use the relay (turn) described earlier. Setting this to “false” results in communications being done with servers other than the turn server. For normal use, leave everything set to “true”.

Projects

The project overview, which lists projects associated with the device manager that logged into the system, is displayed when accessed from the mounted device overview. Use the switch at the right edge to register and unregister cameras to relevant projects.

Cameras that are already registered to a project are indicated with the switch on the right edge showing that it is already registered to a project.

A registration to a project is available only when the device manager is the project owner or a project user of the project. The switch is displayed in light gray and disabled for registration in case of project viewers.

Setting AI apps

○ Installation

Select and install an AI application licensed for the AI Box.

Numerical values displayed above the right side button of the application list indicate the number of installed/installable applications.

○ Device app name acquisition

Acquire the name of installed apps.

Camera action rules

This is a feature for setting a desired motion for a camera, according to determinations made by the AI app. The settings are according to the action_rules of the manifest. This feature can be set using the GUI. Navigating to Camera action rules from Application management for details.

Options settings

This is a feature for specifying numerical values that are unique to AI apps. Navigating to Option settings from Application management for details.

SNMP action rules

This is a feature for specifying desired movements, according to the SNMP trap received by the AI Box. Navigating to SNMP action rules from Camera management for details.

Delete this camera

This feature is used to delete the Camera Details screen. All AI apps must be uninstalled and unregistered from all projects before using this feature.

Note: Deleting a camera also deletes all video images recorded by that camera and stored in the AI Box.

Connection ended

The monitoring of the Edge AI Box has stopped if the “x connection ended” message is displayed on the screen. Refer to Connection ended accessed via Device Details screen for details.

Recording view

The recorded video image search settings dialog opens first when the recording view is opened. Set the desired date and time range and set the targeted video recording and initiate a search.

A list of found video recordings, as well as the Search Results screen for video playback are displayed. Individual files can be played back and verified on this screen.

Press the Save/Delete buttons for recorded video images to display check boxes for selecting files.

Save

Enter a check mark and press the Save button to open a dialog box for saving recorded video files.

○ Navigating from mounted device overview

Enter the desired email address. Enter one email address and press the Return key to complete the address. Multiple addresses can be added.

○ Navigating from project setup

Click on the email address entry field to select project owners, project users and notification addresses of the relevant project. Multiple items may be selected. Another email address may also be entered.

Click on the link provided on the received email to open the screen displaying a list of files.

Download files on this screen.

Delete

Enter a check mark and press the Delete button to delete a recorded video file.

Record settings screen

Deleted due to storage capacity limitations

Recordings will be deleted according to creation dates, starting with older dates, if remaining storage capacity on the Edge AI Box is limited, regardless of the settings on this screen.

• AI Box Light and AI Box Standard: Files are deleted when the capacity reaches 69% full.

• AI Box Standard Mk2 and AI Box Advance: Files are deleted when the capacity reaches 95% full.

These operations triggered by capacity limitations take priority over settings made on this screen.

Settings

• Continuous recording

Enter a check mark to select continuous recording.

• Recording time per file

When this recording time is reached, the recording is loaded on a new file.

• Maximum number of files

A value of 0 or more. When this value is exceeded, files are processed starting with older ones, in the manner described below:

• Deleted.

• If set upper limit on recorded video files is checked, the process follows the operation for the set upper limit.

• Bit rate

This represents the number of bits per second. Greater values improve video image quality, but the data sizes also increase.

• Set upper limit on recorded video files

Enter a check mark when setting an upper limit on the total number of recorded video files, including recordings made with action rules, and to upload to the AWS S3 or delete.

• Maximum total number of files

If this value is set to 0, after the recording time is completed, the operation for the set upper limit starts within 1 minute. If this value is set to 1 or greater, the interval is 10 minutes.

• Operation for the set upper limit

• Files are deleted, starting from old ones

• Files are uploaded to S3, starting from old ones

Recording by action rules

The maximum total number of recorded video files for action rules is derived by subtracting the “maximum total number of files” for continuous recording from the “maximum total number of files” with the upper limit set on recorded video files.

If no continuous recording value is set and if there are prior continuous and action rule recorded video files, those that exceed the “maximum total number of files” are deleted, starting from the old ones.

Differences in operations according to settings for maximum number of files and maximum total number of files

• Checked for continuous recording

• Checked for upper limit set on recorded video files

Number of continuous recordings: The maximum number is the upper limit (delete/S3)

Number of action recordings: The value derived by subtracting the maximum number from the maximum total number is the upper limit (delete/S3)

• Not checked for upper limit set on recorded video files

Number of continuous recordings: The maximum number is the upper limit (delete)

Number of action recordings: Processed according to the storage capacity limitations (delete)

• Not checked for continuous recording

• Checked for upper limit set on recorded video files

The total maximum number derived by adding the number of prior continuous and action recordings is the upper limit (delete/S3)

• Not checked for upper limit set on recorded video files

Processed according to the storage capacity limitations when the number of prior continuous and action recordings are combined (delete)

S3 settings

Following settings are required when uploading to S3 is selected.

• Access key

This is an AWS access key. Using the IAM user with authority limited to minimum writing is recommended for this operation.

• Secret key

This is an AWS secret key.

• Region

This is an AWS S3 region.

• Packet name

This is an AWS S3 packet name.

• Prefix

This is an AWS S3 packet internal prefix (folder).

This can be identified in such a manner as <device_id>/<stream_id>/<rec_type>/<year>/<month>/<day>/<hour>. Refer to descriptions displayed on the screen for details.

Pan-Tilt-Zoom (PTZ) settings screen

The PTZ settings can be performed while viewing camera images.

When performing PTZ settings, pre-sets must be saved first, if none are saved. This is for returning the camera to the original position prior to changes. Initiating the PTZ operation without saving preset results in an error and no operation can be performed.

ONVIF settings must be performed before performing PTZ settings.

SNMP action rules

SNMP action rules comprise a function for achieving desired operations, in response to the SNMP trap received by the AI Box.

Using SNMP action rules makes it possible, for instance, to receive notifications according to the sensor input*. (Note*: An interface converter will be required for converting sensor signals to SNMP trap).

SNMP action rules are composed of setting for SNMP trap comparison (SNMP input) and action settings (Action).

• • SNMP input

  • oid: An object identifier of the received SNMP. It is specified with the dot oid format, such as 1.3.6.1.4.1.*****.1. (required)

  • sender IP address: It is an IP address, from which trap message is sent. This is not used for comparison if not specified. (arbitrary)

  • var bind key and var bind value are specified as a pair. This is not used for comparison if not specified. (arbitrary)

  • var bind key is specified in the dot oid format.

  • var bind value is specified either a numerical value or a text string. A check mark must be entered in the box to the right of the input field to specify this value as a text string.

  • Action

  • All actions other than the Record action are available from all action rules for the application management (refer to “Application management” for details on respective action settings).

Multiple SNMP action rules may be set. Actions with same SNMP input, as well as different actions, may be set.

SNMP actions are enabled for execution only when a camera is operating. SNMP actions become disabled when the camera operation stops.

SNMP trap specifications:

• • Trap supports SNMP v2 format.

  • Trap signal port is set to #162 and the community is fixed as “public.”

  • The event argument passed to respective actions is a json text string consisting of all var-bind_key/bind_key pairs included in the trap. The object identifiers for var_bind_key and var_bind_value are specified in the dot oid format, but the leading character is replaced with route node text string described below:

  • 0: "ccitt"

  • 1: "iso"

  • 2: "joint-iso-ccitt"

Application Management

Camera action rules

(Camera action rules are described on the Camera Details screen.)

Camera action rules are features that facilitate desired movements according to decisions made by the AI app. Following types are available:

• Record (record)

• Upload (upload)

• Email (email)

• Line (line)

• SNMP (snmp)

• HTTPS (https)

• Play (play)

• Image Freeze (imagefreeze)

• UDP (udp)

• MQTT (mqtt), Relay (relay), and VACAN (vacan) (*these functions are currently under development; they are currently unavailable.)

The types of features available may vary depending on the AI app being used.

Respective types described above can be set to AND and OR conditions for determining occurrence of events. Conditions are composed as a set of expressions that include a key, an operator, (=, !=, <, <=, >, >=) and a value, such as:

people = 10

What they can be used for and what they mean vary depending on the AI app.

Recording

action_name : record

When an event occurs, a video image is recorded before and after the event.

Parameters:

• duration_in_seconds

Sets the number of seconds for recording before and after an event.

Upload

action_name : upload

When an event occurs, data is uploaded via the AWS Kinesis Firehose.

Data content may vary depending on the AI app.

Refer to the AWS site for more on the Kinesis Firehose.

Parameters:

• deliveryStreamName

Sets the name of delivery stream.

• accessKey

This is the AWS access key. Using the IAM user with authority limited to minimum writing is recommended for this operation.

• secretKey

This is the AWS secret key.

Email notification

action_name : email

Notifications on events are sent by email.

Notification contents may vary depending on the AI app.

Parameters:

• host

This is the SMTP host address

• port

SMTP port number

• sender

This is the sender address for SMTP login

• password

This is the SMTP login password

• recipients

These are email addresses for sending messages. Multiple recipients may be set.

• subject

This is the title of the email message.

• text

This is the body text of the email message.

• interval

This is the minimum number of seconds in a send interval. If set to 0, an email is sent for all events, but care is needed in so doing, as this can lead to large amounts of messages being sent out.

Line Notification

action_name : line

Notifications on events are sent out via LINE Notify.

Notification contents may vary depending on the AI app.

Refer to the corporate site of LINE Corporation for details on LINE Notify.

Parameters:

• token_id

This is the Token ID.

• message

This is the main body.

• stickerId

This is the Sticker ID.

• stickerPackageId

This is the sticker package ID.

• interval

This is the minimum number of seconds in a send interval. If set to 0, an email is sent for all events, but care is needed in so doing, as this can lead to large amounts of messages being sent out.

SNMP

action_name : snmp

The SNMP trap is sent when an event occurs.

Parameters:

• oid

This is the oid of the SNMP trap.

• ipaddress

This is the IP address of the recipient.

• port

This is the port number of the recipient.

• var_bind_key

varBind

• var_bind_value

This is the value of varBind.

• community

This is the community of trap. The default setting is “public.”

• interval

This is the minimum number of seconds in a send interval. If set to 0, an email is sent for all events, but care is needed in so doing, as this can lead to large amounts of messages being sent out.

HTTPS

action_name : https

When an event occurs, the HTTP Post is executed to the specified URL, by performing the Basic authentication.

The request body is data of the JSON format, generated by the application. Content may vary depending on the AI app.

Parameters:

• url

This is the HTTPS server URL.

• user

This is the user name for the Basic authentication.

• password

This is the password for the Basic authentication.

PLAY

action_name : play

The specified RTSP stream is played back by the HDMI output when an event occurs.

Parameters:

• uri

This is the RTSP URL.

Still image display

action_name : imagefreeze

The specified still image is displayed by the HDMI output when an event occurs (the HMDI setting must be set to ON).

Parameters:

• url

This is the URL of the image file.

The file inside the AI Box can be identified in such a manner as: file:///home/kiosk/Downloads/sample.jpeg.

UDP

action_name : udp

The UDP data is sent when an event occurs.

Parameters:

• address

This is the IP address of the recipient.

• port

This is the port number of the recipient.

• userdata

This is the UDP data. Message is encoded with UTF-8 encoding before being sent.

Options settings

(Optional settings are described on the Camera Details screen.)

These optional settings are a feature for specifying numerical values that are unique to AI apps. The value is comprised of a key and a value as a set.

Values may be a numerical value, a text string or a list.

If the name of the key starts with a polygon or line, a dedicated GUI can be opened as the setting is performed using a mouse.

Region of Interest (ROI) settings screen

This screen is opened by navigating to ROI Settings from Settings in the Camera Details screen.

This feature is used to specify the video image area targeted for AI app processing.

The coordinates are specified to the top, left, bottom and right, with the left upper corner as the origin. Values to for top, bottom, right and left can be varied by dragging the cursor.

Settings made on this screen are reflected on the manifest of the stream.

The ROI applies to video images on this screen. In order to perform ROI settings again, set all of top, left, bottom and right of to 0 with the manifest, then release the ROI, then use this screen for settings.

HDMI Output Guide

Connect a display to the HDMI port of the Edge AI Box to view the same video image shown on the monitor view with a much larger screen. Streaming videos specified with the Play action can also be viewed via the HDMI output.

Preparing Edge AI Box Main Unit

Connect a display to the HDMI port of the Edge AI Box, then connect keyboard and mouse to USB ports.

Use the connected display and keyboard to login to the kiosk account. We inform password at delivery. Other than logging in, no further action is required.

Note:

The availability of the HDMI output with an AI app operating may depend on the individual AI app. Consult the app vendor for details.

When logging in to the kiosk account to execute Linux applications, such as a web browser, ensure to close them all beforehand. Leaving a Linux application operating can lead to an unexpected malfunction.

Logout from the kiosk account when not using the HDMI output. If an AI app is not compatible, remaining logged in could result in an unexpected malfunction. Unplugging the HDMI cable is not sufficient to logout of the kiosk account.

HDMI output of monitor view

Note: When changing HDMI settings, all camera apps associated with that Edge AI Box will turn off temporarily. Restart each unit after changing settings.

○ Device settings

• Open the Edge AI Device Console on a web browser, then open the Device Details screen.

• After logging into the kiosk account, standby for 20 seconds.

• Press the HDMI Output Settings button.

• Enter a check mark for “Enable HDMI Output” option, then press the Save button (see figure below). (Remove the check mark to turn HDMI OFF).

• Standby for a few seconds, then confirm when the message indicating “HDMI Output Enabled” appears at the lower left of the screen. Apps of all cameras stop at this time.

The HDMI output settings remain intact, even if the Edge AI Box is rebooted.

Connect the display to the HDMI port. Note that the HDMI output settings are not enabled unless logged into the kiosk account (see figure below).

○ Camera settings

• Open the Camera Details screen for the camera of the video image for the HDMI output.

• If you wish to output monitor view footage to HDMI, press the Start button and then enter a check mark for the “HDMI output”, and then press the Start button for the HDMI output of video images that are on the monitor view (see figure below).

The same video image as the one shown on the monitor view will be shown on the connected display in a few moments.

HDMI output of Play action

There are no required settings for this feature, other than preparing the Edge AI Box main unit. Refer to details by navigating to “Play” from “Camera action rules” accessed from “Application management”.

Account management

Click the user ID on the upper right of the screen, select the account, and navigate to the Account Management screen.

Changing Passwords

This feature can be used to change passwords.

Email

This feature can be used to change email settings.

Authentication method

This feature can be used to change the login authentication method. The use of Password and Email, which is more secure, is recommended.

• Password only

Login status expires after 1 hour.

• Password and Email

Login status expires after 24 hours.