Description

Business Analytics 2.0 - this is an interface for working with all video analytics modules, which allows you to aggregate information received from various detectors and present it in a single, understandable format.
There are two main functions in this platform:

• Building (and exporting) reports based on statistics collected by detectors
• Notifications that a detectable event has occurred

Requirements

The platform itself is already included in the TRASSIR software and activation requires at least one included module or neurodetector from the list of compatible ones.

At the moment, this platform works with the following modules and neurodetectors (the list is constantly updated, please contact the commercial department for up-to-date information):

• Neuro Left Object Detector
• Queue Detector
• Pose Detector
• Shelf Detector
• Crowd Detector
• Neuro Detector

Opening the settings interface

To work with the interface of the "Business Analytics" module, open the operator interface, click on the "Modules" tab and select the "Analytics" module

Next, the module interface opens to configure it:

By activating a certain module, we see the settings, you can activate all the modules and set various settings for them: 

Settings

• Data collection schedule - позволяет установить интервал, в пределах которого будет происходить сбор данных с детектора
• Data sources - the channels with the zones of the neural detector from which the data will be collected looks like this (first you need to create it by clicking on + at the bottom of the page):
  
  
• Notifications - setting up rules, when executed, a notification will be generated by e-mail or telegram (for example, for a crowd detector "there are on average more than 5 people in the zone within 30 seconds")
• Configuring Reports - parameters for sending and generating reports.
• Configuring Data Collection - specifying the minimum number of people to collect data is only available for the crowd detector. The settings for each detector are similar.
• Objects - the ability to group data sources according to certain criteria (detector type, specific building, etc.). It is also possible to work from a browser on a PC with Business Analytics.
  To group, click on the button:
  

Notifications

In the settings of any of the modules, it is possible to choose the type of notifications. There are several types:

• Telegram
• E-mail
• Event log
• Cloud notifications (push notifications to the phone)
• SMS

In the notification settings, you can select the time interval for sending. At the end of the interval, the last generated event will be sent. Events before it will be ignored and there will be no notifications.

Dashboard

After setting up the detectors, we proceed to the formation of the desktop by clicking on the "Dashboards" button. At the first launch, the table will be empty.
To visually display statistics, click on the "Add" button:

A menu appears on the right where you can select the type of detector, the sources of this detector, add new sources and select them:

To build a diagram, select the source / sources of the detector and click the "Make a Graph" button - a window with the settings of this graph will appear:

At the top right of the graph we see three dots and a screen zoom button. By clicking on them, you can export the data already collected, edit the graph display or delete the graph itself, as well as examine the graph in more detail by expanding it to the full screen.:

Example of a "Desktop" when using a queue detector:

Queue Detector

Allows you to estimate how the queue length changed during the period.

Notifications

Zone has on average "more than № persons for № seconds

by defult its 5 persons for 30 seconds

you can change this value for your issue

Area protection (Neuro Detector)

a new person appeared in the zone (as well as intrusions into protective zones)

The metric allows you to estimate how many intrusions into the zone occurred during the period.

You can also count visitors who have crossed the line (border)

Notifications 

Parking Analytics

Calculating the congestion of parking zones

Neuro Left Object detector

Zone has more than № of objects (more than, more or equal to, less than, less or equal to, exactly)

The metric allows you to estimate how many abandoned items appeared during the period.

Notifications

Shelf detector

there is a gap on the shelf

The metric shows the number of voids on the shelf.

Notifications

A notification is sent if a void larger than the threshold is found on any of the shelves. Voids are detected in accordance with the clause "How the metric is calculated"

Repeated notification is sent no more often than specified by the parameter "Period for repeated notification"

Example of reports

Pose detector

It informs the user about suspicious behavior of people in the protected area

The metric allows you to estimate how many and what suspicious and/or atypical actions occurred during the period

In the settings, you can select the type of pose to create notifications and reports. Using Dashboard, you can create a report both separately for different poses, as a general table where all the designed poses will be listed

• sitting
• bending over
  lying
• down, the left hand is raised
• the right hand is raised
• both hands are raised
  You can select several.

Notifications

Crowd detector

Allows you to estimate how the number of people changed during the period.

If there is more than the number of people in the zone, we create a report and notification.

The notification is sent if the arithmetic mean "Number of people in the zone" for the period specified in the parameter "Period for calculating the number of people" (rounded to the nearest integer) is greater than the "Threshold number of people in the crowd".

In order to give an example faster, I chose 2 people within two seconds

Cloud tokens for Business Analytics 2.0 scripts

• BA_ALARM_ABANDONMENT_DETECTOR 

• BA_ALARM_NEURO_DETECTOR 

• BA_ALARM_SHELF_DETECTOR

• BA_ALARM_QUEUE_DETECTOR 

• BA_ALARM_POSE_DETECTOR 

• BA_ALARM_CROWD_DETECTOR

• BA_ALARM_POS_INCIDENTS

ⓒ Alexander Savkin