Video Surveillance for Business

Workplace Surveillance with Video Analytics Solutions | Use Case

Introduction to Video Surveillance for Workplace

The main objective of video surveillance is to protect office assets and valuable information from compromising. For this, there is a need to timely detect and recognize authorized and unauthorized persons, track their activity inside the building so that organizations can identify the situations that require immediate response, which will then require interpretation and action.

The latest AI and production combination promise a stronger, more efficient, and hassle-free working environment in factories. Click to explore about, Cloud Video Analytics and Surveillance for Intelligent Outcomes

Features of Video Surveillance for Workplace

Below are the features which we are considering under our solution:

Process Management with Video Surveillance

In a defined workplace, there are so many different processes, the objective of this feature is to provide the capability to manage the processes using the power of Artificial Intelligence, Computer Vision, and CCTV (which will do the surveillance)

The benefits of process management in Workplace Surveillance are:

24*7 process management
Process management in an automated way without/minimal human intervention.
In-depth Analysis

Production Analysis with Video Surveillance

Use of AI, Computer Vision, and CCTV to check and analyze production-related issues

The benefits of production analysis with video analytics in workplace surveillance are:

Analysis of the quantity of production happening during a specific period of time.
Analyze the peak hour of production
Analyze the lacking hours of production(which is at its low).

Automated Analytics for Workplace Optimization

Using Analysis and computer vision/ video analysis to optimize the working under the workplace.

The benefits of workplace optimization are:

Automatic Monitoring of the employees in working hours and its Analysis

Video Surveillance for Behavioral Rules

To track whether every person is following behavioural rules or not

Benefits of surveillance behavioural rules is

Manage the behaviour of employees without human supervision.
24*7 alert on any mishappening

Time Tracking

To track the time/productivity time of the person based on the amount spent on his/her chair or in a specific area

The benefits of time tracking for workplace surveillance:

When we talk about tracking, it should be 24*7, and it is the benefit here that it can provide 24*7 tracking.
Not only track time but also ensure the presence of a person in the area(whether he is there or not during the time tracked) where the time is tracked with respect to that person.
According to the Artificial Intelligence Global Surveillance index, at least 75 out of 176 countries globally are actively using AI-based surveillance technologies. Source – How AI Is Making An Impact On The Surveillance World

Activity Monitoring with Video Surveillance

It is very hard to track the activities of employees every minute but using AI and CCTV cameras. It is easily possible to do it and also to generate alerts if any unusual activity is happening.

The benefits of activity monitoring with video analytics are:

Provide the answer to the following types of questions
What are employers doing?
Are they working, or are they wasting time sitting idle?

Crowd Management with Video Surveillance

The objective of this feature is to capability manage the crowd and person under a specific area. It can handle the following questions as to use cases For example:

Benefits of workplace surveillance

Provide the Analysis of the following questions
How many people are allowed at a specific place at a specific time?
How many people are there in a specific place at a particular time?

What are the requirements of workplace surveillance?

The basic requirements of Workplace surveillance is classified in 2 categories.

Technical Requirements

Python 3.5 or above
dlib==19.21.1
face-recognition==1.3.0
face-recognition-models==0.3.0

Other Requirements

CCTV cameras access
Required Employees Data (front, left, and right side of the face)
CPU-based implementation is very slow. We need a GPU to implement the above scenarios on the CCTV feed.
Connectivity to Database

Why is a Video Surveillance Dashboard necessary?

Most workplaces remain open 24*7 to serve their customers anytime and provide them value. Therefore employees and visitors always visit there, some for their daily work and some for interviews and consulting. It makes that environment busy, chaotic, un-enjoyable, and unsafe because it may be possible that an unknown person visits with wrong intentions. They can harm the infrastructure as well as the person. Therefore Video surveillance is necessary. One of the organizations’ major challenges is to identify, track and detect those anomalies and persons. Some other challenges of the workplace are:

Identification:

Organizations cannot identify known or unknown people as a lot of people visit the workplace.

Paperwork:

A lot of paperwork is required to handle their entry record.

No Anomaly Detection:

Organizations are not able to detect any anomaly at workplaces

No Historical Tracking:

Organizations cannot track or map if a person is visiting again and again.

Video Surveillance Solutions for Businesses

As per the scenarios we are covering, there are two phases to be covered in this solution:

Phase 1: Identification

Authorized
Unauthorized
Temporary Access granted

Phase 2: Behaviour Analysis

Anomaly Detection(Live Tracking)
Prohibited Area
Access Control
Exit
Map Person

Video Surveillance Solution Phase 1 for Business Case Identification

The solution’s main objective is to develop a model that identifies and classifies the person as authorized and Unauthorised.

Implementation of the solution

The architecture of the solution
Steps of the solution

The workflow of the solution is given below in the steps.

Step 1: We should have the images data of authorized persons in this folder

Data →RealTimeImages →Encoding Images →(authorized person images)

Step 2: Detection of faces and recognition of the person’s identity by matching the faces from authorized persons’ saved database. If the person is authorized, it will show a green colour rectangular bounding box around the face of an authorized person; if the person is not permitted, it will show a red colour rectangular bounding box around an unauthorised person.

Step 3: Suppose the person recognises it as unauthorised, then will collect the data of a person.

Step 4: Allocate time-based to enter the building; when the time exceeds, the alert will be sent to the person and authority and track the user if the extra allocated time is still exceeding.

Output of the solution

From the above approach, we’ll get the following outputs:

Output Parameters	Output Parameters	Output Parameters
Unique_User_ID	Access period for Authorized(8hrs, 10hrs, 12hrs, etc.)	Time Out
Authorized	Unauthorised(2hrs, 3hrs, 4hrs, etc.)	Unauthorised purpose(Interview/Official Visit)
Unauthorised	Date	Area
Name	Time In

Note: The model will generate these columns as output, decided as per the requirement of the AnalyticsOps Team

Usability of Solution

The figure shows how we can use the Identification Feature for Authorisation and
Un-authorization

Phase 2 Solution: Behaviour Analysis and Exit

This phase of the solution’s main objective is to develop a module that can handle Behaviour Analysis and Exit parts.
The Camera will map the exit of a user by providing the exit time of the user.

Implementation of the solution

Step1: Architecture/Steps of the Solution

Step2: Usability of the Solution

Step3: Output of the model

Prohibited Area, Unique_User_ID, Authorized, Unauthorised, Severity, Date, Time, Area, Area wise Duration, Unique_User_ID, Unauthorised(Y/N), Name, Area, Duration, Purpose, date, Entry time, Exit time

Note: The model will generate these columns as output, decided as per the requirement of the AnalyticsOps Team

Overview of the Case Study - Incident

Camera-based intelligent video analytics open the minds to widespread possibilities via video streams. Source: Intelligent Video Analytics

Investigation of Video Surveillance

One of the organizations’ significant challenges is to identify, track and detect those anomalies and persons. Some other challenges of workplace surveillance are :

Identification: Organizations are not able to identify known or unknown people as a lot of people visit the workplace.
Paperwork: Need a lot of paperwork to handle their entry record. Manual processes are not only inefficient but also increase security risk and costs.
No Anomaly Detection: Organizations are not able to detect any anomaly at workplaces.
No Historical Tracking: Organizations are not able to track or map if a person is visiting again and again.

Incident Response

To solve these problems, a solution is necessary that makes workplaces productive and safe using AI and video analytics. The system should detect a person; based on appearance, it recognizes the visitors or regular employees. Then accordingly, monitor access management, behavior analysis, anomaly detection, and alert management to make the system more reliable. It should also analyze the crowd and presence using self-learning analytics. It should also monitor suspicious activities and send an alert to the respective authorities for the actions if encountered. Overall, the system should monitor the whole process using video analytics.

How can XenonStack Help?

Xenonstack provides a solution that will cover the following scenario to solve a problem for workplace surveillance : Real Time Video Analytics

Covering those scenarios, it will allow monitoring the following cases:

Boundary Breach Detection

Virtual Boundary: It allows for tracking the entry and exit of people in a particular area. Thus, users can track the number of visitors, their time of visits, and the duration of their stay there. It allows interlopers to access those areas.
Forbidden Area Violation: The system generates an alert a person doesn’t have access to a certain area but enters there. It detects real-time violations and malicious activities, thus enhancing the safety and security of the place.

Theft Analysis

Theft Alert: Systems always track the inventory of a workplace. If it detects any object's movement at an abnormal time, it alerts the manager.
Loitering: The system keeps track of passers-by and records the time they spend in monitored zones. The system generates an alert when more than usual time is spent by a passerby at the same spot.

Anomaly Detection

Suspicious Movement Analysis: If anyone spends extra time in a specific location, the system will raise the alarm. Thus detect any suspicious activity before it occurs on-premises.
Abandoned Object: The system can identify unattended stationary objects, thus detecting threats of cases where objects are left behind with wrong.
Fire Alert: CCTVs or smoke detectors only detect smoke or fire. But they respond only to smoke or flames after the fire has already begun.

Thermal imaging cameras can see through smoke and have the exact temperature and location of hotspots.

Density/Flow

Bottleneck Analysis: The system may detect a crowded place immediately and alert the organizer. For taking immediate action to make the environment peaceful by reducing the wait time.
Entry/Exit Counts: Using a feature-rich camera placed at different locations, users can keep track of the persons and an object entering or exiting the premises.
Traffic Queueing: The system detects the traffic of people and vehicles at entry gates. And monitor their behavior and early detect the situation before it becomes more cumbersome.

Trip Tracking

Path Analysis within an Area: The system keeps track of the visitor’s journey to monitor their activity always.
Presence Distribution Map: Analysing the location and people density in various locations a different times. Users can make decisions regarding scheduling the break and the work timing of people.

Questions to Discuss for Video Surveillance

Still, some questions come in customers' minds: “ How the system will perform”, “How it track a person?”, “When will it generate an alert?”, To answer all these questions, we will discuss a case study for workplace surveillance that will answer all user questions and explain how the system works for an unauthorized person’s journey from entry to exit.

How does the system prevent the entry of unknown people? How does it get to know that a person is unknown?

The system detects a face and identifies whether a person is authorized or unauthorized. By matching it with already recorded faces from a database when a person visits a workplace. If a person is unknown, it generates an alert at the entry and thus stops entry.

Okay, I got to know how the system will detect and prevent the entry of unknown people. But what if a visitor or interviewee comes for the first time? How does the system allow them to come? How does that person get an entry? Whether he/she will be able to come inside the workplace or not?

If the person is unknown, then it starts with Access Management. Meaning it will give temporary access to him/her.

As answered in the previous question, that system will give access to them. What is necessary to give access?

At the entry time, the system gives temporary permission with some information like Purpose, Name, Floor, Department, Concern Person, Time Slot, Phone Number.

How does the system get to know that the access they were given to is the right person? Because it may be possible that the person is not coming for an interview, he can be the wrong person?

An alert will be sent to the organization about the meeting of a temporary person. So the organization will get to know whether that person has an invitation or not.

Okay, As discussed in the previous question, a person will be allowed entry for some defined time. How will the system get to know whether a person leaves at the correct time or not?

The system will continue monitoring a person; If a person exceeds its time limit in case of an unauthorized person, an alert will be generated to a respected person and the organization.

It is also possible that the person that entered the workplace went to the wrong location by mistake or intentionally. Or it may also be possible that person is harming another person or ruining infrastructure. How may the system get to know about that?

After giving entry, record his path and behavior. So if a person is going to a restricted area or doing any suspicious activity, an alert will be sent to the respective organization.

Are these all cases possible for authorized because he/she may also do the same?

Yes, the system also monitors the path of authorized people. And in all of the possible scenarios discussed above, an alert will be generated to the respective organization.

A person may be visiting again and again, but he/she is not a regular person, so he/she has to register themselves? How did the system get to know?

When an unknown person visits a workplace, his/her data will be saved in the database. So that it can map temporarily accessible people on his/her next entry.

When does the system stop tracking?

The system will successfully exit and close the case of a particular person when he/she leaves the building. Workflow Productivity Measurement with Video Analytics

Workflow Productivity Measurement with Video Analytics

Conclusion

The workplace surveillance solution will track, manage, check and improve employee productivity to keep the workplace safe from theft, sabotage, and malicious activities.