Pages

Sunday, July 29, 2018

Massive Growth of AI in Video Surveillance

Massive Growth of AI in Video Surveillance

In June 29, 2001 release A.I. Artificial Intelligence, co-written by American science fiction drama film, co-produced, and directed by Steven Spielberg and based on the 1969 short story "Supertoys Last All Summer Long" by Brian Aldiss. In the late 22nd century, rising sea levels from global warming have wiped out coastal cities such as AmsterdamVenice, and New York, and drastically reduced the world's population. A new type of robots called Mecha, advanced humanoids capable of thoughts and emotions, have been created.
No noo don’t stop to reading of my article, I am saying this movie come to true in the late 22nd century.The goal of AI is to provide software that can reason on input and explain on output. AI will provide human-like interactions with software and offer decision support for specific tasks, but it’s not a replacement for humans – and won’t be anytime soon. It augments our abilities and makes us better at what we do. Because AI algorithms learn differently than humans, they look at things differently. They can see relationships and patterns that escape us.

Neural Networks give thinking, Machine learning is popular but deep learning breakthroughs drive AI boom” says Arindam Bhadra, eSecurity expert cum blogger.

Now we look at why AI will become an integral part of your video surveillance solution in the coming years.

AI would play a major part in is going to be the analysis of the massive amount of data that is collected by the security system. If you consider the amount of data that is being collected by the device sensors, it is difficult to make heads or tails of data at present and this is where AI will come into the picture.

The rise of AI is beginning to revolutionize technology in various industries, and security is prominent among them.AI has become more popular today thanks to increased data volumes, advanced algorithms, and improvements in computing power and storage. Video surveillance solutions, for instance, are now seeing an AI-based applications.

The AI CCTV is basically a processing of the image with the help of deep learning algorithm and after processing you will get information. Artificial intelligence CCTV camera is just like CCTV operator that will give you information which you mean.

I will give infographics comparison of 100 cases in which CCTV recording as required, you can judge how to use artificial intelligence camera is.
To understand in the better way I will put an example so that this topic will clear in the more better way.

If two camera is installed at the main gate of an industry then what is the use of that camera, you need information and activity which is happening at gate, like how many people entered in the industry, number of vehicle passing from gate, may be you need vehicle number of all vehicle which are entered in the premises, you may wish for how many visitor visiting in premises, you may wish that how many time you regular staff passing from gate. So these are the activities that you need to watch, for this you put CCTV operator who can process all information and send you detail log but with help of AI CCTV camera with AI-Software, you can find the automatic log of these activities from software. So from the above example, it is more clear that what actually artificial intelligence camera is.

Surveillance systems that include video analytics analyze video footage in real-time and detect abnormal activities that could pose a threat to an organization’s security. AI helps security software ‘learn’ what is normal so it can identify unusual, and potentially harmful, behavior that a human alone may miss.

To understand technically AI-CCTV camera is a simple IP camera but it is embedded with AI-Software, so according to need they develop AI-software’s. There are many kind algorithms and software likeFacial reorganization- it will allow you finds people in a crowds place like railway stations, Airports, stadiums.

Intelligent Object, Event and Facial Recognition
In line with the concept of proactive and real-time security, AI-based technologies can offer superior facial and object detection capabilities. In fact, there is massive potential for facial recognition in the law and order sector where specific individuals need to be identified from crowds in large spaces.

And security is not the only place that is betting on this. Marketing departments in large corporations are also considering (if not already employing) facial recognition solutions to identify potential high-value customers who can be offered special services.

The power of AI will also enable “faceless recognition”, wherein a person can be identified not just from visuals of his face, but also from factors such as their height, posture, build, etc. Moreover, solutions could become more aware of the environment that they are monitoring, detecting patterns of activities and drawing out anomalies when required. They would also be able to experiment with different simulations and come up with possible scenarios where there could be potential threats.Faceless recognition systems would prove invaluable in situations such as the 2016 Berlin terrorist attacks where the attacker was caught on camera various times from angles where traditional biometric facial recognition systems failed.

Image Sharpening
Another huge benefit to using artificial intelligence in video surveillance systems is its ability to cross reference data with other neural networks or databases in order to fill the gaps in its own knowledge. The first artificial network enhance the details of the image in accordance to the layout provided in order to make it more feasible as an image.

This kind of image enhancing technology obviously has a huge amount of potential in video surveillance and security, from criminal and terrorist identification and tracking to staff monitoring and security access operations. This kind of technology could also allow remote IP cameras to stream enhanced video across a network that could then, in the case of a crime scene investigation, be remotely accessed by a forensics expert to assist in the investigation. The same could be true for security experts in video surveillance use cases whereby detectives can access image-enhanced video to investigate potential video recordings of wanted suspects.

Managing Massive deeper Data:
AI analyzes more and deeper data using neural networks that have many hidden layers. Building a fraud detection system with five hidden layers was almost impossible a few years ago. Surveillance is not just about data from a single source anymore. Cameras and sensors of diverse types are covering every nook and corner of locations to leave no room for error. This means that security solutions must deal with substantial amounts of data that require specialized software systems to manage.

AI systems could interpret data coming in via IP cameras and use it to automatically trigger alarms or notify security guards if it noticed something suspicious or out of place in the video data. Archive video footage could also be analyzed by AI systems using deep learning algorithms in order to place potentially new suspects at the scene of a crime automatically, with human agents merely updating a database the AI system had access to.

Video Summary
Video Summary is a fast and powerful application that reduces a long, archived video into a short, manageable video summary. Video Summary reduces the time required to analyze archived video by removing unnecessary video data such as static or empty scenes. Video Summary quickly analyzes the video and removes the static background data while retaining the relevant data. This allows security professionals to quickly review hours of archived video in minutes. Video Summary can receive video input from many different types of video archives recorded on DVRs, NVRs, or video management systems. It can also provide real-time video summaries on live video feeds from analog, digital, or IP/network cameras. With COM-SUR AI application help to Increase productivity and efficiency of security professionals.

Alarm Center
Alarm Center provides options to view, search, report and analyse all the alarms generated by AI analytics running on different systems across a LAN. It also provides heat map for multiple cameras and provides the paths of object or people movement. Reporting in PDF, JPG, Excel, Text file.

Unusual Motion Detection
UMD is an advanced artificial intelligence (AI) technology that brings a new level of automation to surveillance and is designed to reveal events that may have otherwise been missed. The analytic cameras will support either object detection or UMD but, not both at the same time.

The analytics definitely work and would be used for perimeter detection and alarming for sure but, if you aren't going to program the analytics and be proactive about it then UMD could be turned on and the camera will learn it's scene over the course of 1-2 weeks then when reviewing video instead of looking at lots of different motion events on the timeline you can filter the timeline to show just unusual events based on direction, speed, location.

"Today, most recorded video footage is never viewed, meaning critical events may go undetected. Avigilon is using artificial intelligence technology like UMD to help our customers effectively extract valuable insights from their video data," said Dr. Mahesh Saptharishi, Avigilon's CTO.

Edge Analytics
This is very cost effective solution without server requirement. Applying some supported cameras can provide Intrusion (Trespass, Boundary Crossing etc), Loitering, Crowding& Counting.

Enhanced Surveillance Monitoring
AI provide rapid detection of smoke even when smoke is formed in 10-15% of the viewfor both indoor and Outdoor Environments. It’s called Video based Smoke Detection(VSD).

Privacy Masking
Face and Skin area can be masked to protect the identity of the persons appearing in the camera view of surveillance system. While unmasked video is recorded & stored, masked video is viewed by operators during surveillance monitoring.

Thermal Sensing Camera
Thermal camera will allow you to measure the temperature of the particular object. It will install at a temperature sensitive place like industrial boiler etc. accordingly analysis object details through AI software.

Trigger Based Detection
The trigger can be obtained from devices like RFID, Traffic Light Signal. Triggers analytics for unauthorized people detection or object detection after people leave from the area of interest. So you can say this is apply for integration (BMS, ACS, Parking etc)

Real-Time Protection
Video surveillance solutions have always remained reactive and useful only after an incident has occurred or at least a threat is detected by human operators. This is all set to change with the arrival of AI systems that can detect a potential threat way before an incident occurs.

For instance,deep learning, neural network can detect unattended objects like bags in airports and public spaces, enabling authorities to act before it’s too late. Similarly, AI-enabled surveillance solutions can monitor the behavior of people in a store and identify those who might engage in shoplifting.

Demographic Analysis
It detects the gender for human face appearing in the camera field of view. Sometime is called Gender Detection.
This analysis can provide approximate age against human face appearing in the camera field of view.

Parking Management
Provides the parking occupancy level and free parking slot availability by detecting vehicle entry and exit by monitoring these transition points in parking lot areas.
Also detects the percentage of area occupied by vehicles and alerts against vehicle congestion as the vehicles occupy area beyond a threshold value.Illegal parking (no parking zone or restricted zone) of the vehicle in front of the entry/exit gates.

Beverage Management
Counts the number of cups / beverages passed over a line like in serving counter area.

Queue Management
Detects queue properties for waiting time analysis (Ingress – in flow, Egress – out flow. It’s applicable in service counters, ticket counters / travel desk, check in counters, flight boarding areas for waiting time analysis and improvement of service process.

Gesture Recognition
Detection of unusual gestures of persons by raising weapons (lathi/stick, rifle etc.) with intention to charge on others. Or a person slipping and falling on ground.Detection of object(s) removed / missing from the monitored zone in the camera view.

Appearance Search
Appearance Search technology will allow users to initiate a search for a person by selecting certain specific physical descriptions, including hair and clothing color, gender, and age. The new feature is designed to provide security operators enhanced speed and flexibility when conducting a search for a person. "While investigating critical events, time is of the essence, and security operators are often provided physical descriptions of the people involved," said Dr. Mahesh Saptharishi, Avigilon's CTO.

Crowding Detection
This technology allows you to know the density of the object in the crowd as the number of the object put at a particular place within the camera field of view / region of interest.
AI Analyses crowd movement patterns / direction in different way and marks it by different colours, and detects and movement of crowd in undesired direction.
Measures/counting the crowd level in terms of number of people occupying a specified region of interest in the camera field of view, provides live crowd count on screen.

PTZ Analytics
Speed dome camera / scan dome camera is called PTZ camera. Basically P – Pan, T – Tilt & Z indicate Zoom. Automatic tracking of object (Single or Multiple) using Pan-Tilt-Zoom (PTZ) camera.
Violation detected on any Fixed camera triggers PTZ camera to its view for auto tracking of the violator object.
Automatic one step PTZ operation to capture closer video of an agent on any pre-set rule violation such as crossing a virtual line, or entering a virtual area etc.
Different PTZ pre-set positions can be assigned to different regions and analytics can be run to monitor each of those pre-set positions.

Camera Tampering
Detection of camera tampering efforts by camera focus change or view obstruction or video cable cut.

ANPR
Number Plate Recognition Cameras allow you to recognize the number of vehicle and along with that it also stored the vehicle log. This camera is used in traffic management and traffic monitoring. It is also called automatic number plate reorganization (ANPR).

Most public spaces in India including railway stations, bus depots, metro stationand marketplaces are often under CCTV surveillance. New Delhi is all set to have one of the largest deployments in the country of CCTVs with the state government announcing plans to install 1.4 lakh CCTVs across Delhi. The India Railways is also setting aside Rs 3,000 crore in its 2018-19 budget to install CCTV systems across 11,000 trains and 8,500 stations, according to a news report.

AI performs frequent, high-volume, computerized tasks reliably and without fatigue” says Mr. Arindam Das Sarkar, Country Manager, Mirasys India. Mirasys is one of the leader in creating models using deep learning and Artificial intelligence based video analytics in India as well as global market.

Tamper Detection
Quickly get security officers to the damaged camera in areas with high crime rates, where there is a large potential for criminals purposely blocking the camera’s view.

Additionally, several technologies enable and support AI:
·         Graphical processing units are key to AI because they provide the heavy compute power that’s required for iterative processing. Training neural networks requires big data plus compute power.
·         The Internet of Things generates massive amounts of data from connected devices, most of it unanalyzed. Automating models with AI will allow us to use more of it.
·         Advanced algorithms are being developed and combined in new ways to analyze more data faster and at multiple levels. This intelligent processing is key to identifying and predicting rare events, understanding complex systems and optimizing unique scenarios.
·         APIs, or application processing interfaces, are portable packages of code that make it possible to add AI functionality to existing products and software packages. They can add image recognition capabilities to home security systems and Q&A capabilities that describe data, create captions and headlines, or call out interesting patterns and insights in data.
In Aug 2015, West Japan Railway Companymade a great example. After discovering that 60 percent of people hit by trains in Japan are intoxicated, they installed 46 security cameras that automatically search for and detect signs of intoxication. The AIvideo  looks for people who are napping on benches, stumbling, falling, or standing motionless for long periods of time and notifies human attendees who can then check to see if the person is in danger of harming themselves.

Using AI in video analytics, a number of systems will be able to communicate with each helping in taking decisions and readily catching suspicious activities or predicting them before they can happen.

With the advancement of AI, the future of video analytics is not limited to providing stagnant algorithms, but also to having the event and authorization-based alert systems. Wherein, the alerts will be sent only to an authentic person and the relevant department.

This Artical published in Safe secure magazine November 2018 edition.

Saturday, July 7, 2018

Choose ONVIF Prescribe ONVIF

Choose ONVIF Prescribe ONVIF

As an eSecurity Professional I always prefer to use ONVIF. ONVIF allows me to be truly independent of brands and recommend a way forward for customers that is not going to tie them in to a single, specific manufacturer. As I work as a System integrator in last 12yr found purchasers have probably all heard of ONVIF, but in my experience, the high level of competition in the CCTV market drives many manufacturers to attempt to differentiate themselves to understandably stay ahead of the pack. Unfortunately, I have seen many cases where the manufacturer tells the customer they make ONVIF-conformant products, but they then upsell additional features for their product and in many cases will say how that method is better than ONVIF. This doesn’t help many buyers of CCTV products today as they are not always experts in the field of CCTV or in the business of security, so they don’t see the strategic benefits of employing an ONVIF approach. We also see that the end users do not yet see the wider benefits of using an open standard and what that can offer them in the longer term.  They are instead more focused on the short term project, so there are many strategic influencers who are not yet aware of ONVIF.
Many projects I work which under government-funded customers that have long approval cycles and need to be able to have longer term plans. ONVIF allows this, as it has a stable roadmap and retains backwards compatibility in most things, so I am able to set a roadmap and a vision for a customers that they can plan around whilst having access to a large range of vendors and innovative new products. Government-funded long term projects will benefit the most from using ONVIF-conformant componentry. In India very few consultant prescribe ONVIF product, basically they still not clear about spec of ONVIF profiles. In UK, USA, Australia whether it is rail, roads, national or local projects, ONVIF has to be a key consideration in government-funded projects.

More systems will be ONVIF conformant and people buying products will be able to choose new and innovative solutions – like thermal cameras – as they become available, based purely on quality and cost instead of the manufacturer. That is a much simpler decision for them to make and the best products will rise to the top and remain in the market. There will also be more opportunities for systems to interoperate so that systems can be shared.

Each ONVIF profile has a fixed set of features that must be supported by a conformant device or client. Realizing a need for the standardization of communications between IP-based physical security products, Axis Communications, Bosch Security Systems and Sony Corp. in 2008 formed the Open Network Video Interface Forum (ONVIF).

ONVIF has surpassed 10,000 conformant products that meet its global interoperability standards, further broadening the organization’s offering and relevance in the IP-based physical security marketplace.
Membership is open to manufacturers, software developers, consultants, systems integrators, end users and other interested groups. The organization’s membership base covers six continents and offers more than 9,000 profile-conformant products.

How important is it for devices to be able to openly communicate, with some sort of standardization in effect? The technology researchers at the Gartner Group predict that by 2020 there will be 50 billion devices connected over IP networks, or what’s known now as the Internet of Things (IoT).
As ONVIF standards have evolved they have been designed around a list of performance profiles. Each ONVIF profile has a fixed set of features that must be supported by a conformant device or client. An example would be that an ONVIF Profile S device must work with other ONVIF Profile S devices.
Within each profile are product features listed by ONVIF and classified as mandatory (M), conditional (C) or optional (O). You can check online at the ONVIF Conformant Product List page and search for compliance confirmation. Do remember smarter video systems play a clear role in generating data.
To get an idea where products for your next project stand, here’s a look at current ONVIF profile specifications.
ONVIF Profile Specifications
For electronic access control systems:-
  • Profile C conformant devices and clients support site information, IP door access control and event and alarm management.
  • Profile A devices can retrieve information, status and event, and configure entities such as access rules, credentials and schedules. Conformant clients can provide configurations of access rules, credentials and schedules. The client can also retrieve and receive standardized access control-related events.
For IP-based video systems:-
  • Profile S could be a camera or video encoder that can send video data over an IP network to a Profile S client. A client might be video management software (VMS). This profile also covers pan/tilt/zoom control, audio, multicasting and relay outputs for conformant devices and clients that support such features. So it addresses common functionalities of IP video systems, such as video and audio streaming, PTZ controls, and relay activation.
  • Profile G products such as a camera or video encoder device can record video over an IP network or on the device itself. A Profile G VMS client is one that configure, request and control recording of video data over an IP network. Profile G also includes support for receiving audio and metadata streams. So it addresses video storage, recording, search, and retrieval.
  • Profile Q aims to provide quick discovery and basic configuration for Profile Q-conformant products. Profile Q also covers specs for TLS (transport layer security) configuration. TLS is a secure communication protocol to protect against tampering and eavesdropping.
  • Profile T is a release candidate for IP-based video systems. It supports video streaming such as the use of H.264 and H.265 encoding formats, imaging settings, and alarm events such as motion and tampering detection. This profile is designed to work with Profile S and not replace it.
The most majority IP-based security products claim to compliant with ONVIF standard. However, when end-users try to install these devices marked with this standard might encounter some problems, such as network video recorder can't discover the IP cameras in the same LAN network, motion detection doesn't work. Why is this happening?

First, you need to verify that your installed devices are truly compliant with ONVIF standard. Some manufacturers often mark their products as compatible with ONVIF, although this is not true. It's better to buy/use the IP-based security products from members of ONVIF. You can find the full list of members/manufacturers who are verified to offer ONVIF compliant products.

Second, it's not possible to be ONVIF compliant only; the IP-based security products need to be compliant to a Profile as well. You can be ONVIF conformant only, but if you are it is not guaranteed that you would get interoperability between other ONVIF devices and clients. To truly reap the benefits of ONVIF, you need to be Profile S compliant as it dramatically increases the chance of things like video and audio streaming or video configuration and multicast actually working well together.
One may encounter cases within a set of devices with Profile S, having no problems with connectivity and video streaming, but with some functions (e.g. motion detection). It is usually caused by errors of the standard or, more often, by an incorrect implementation of the standard/profile made by the manufacturer. Such problems should be immediately reported to the manufacturer that, in turn, should release improved software. ONVIF forum is heavily working on new, improved versions of the standard, which will be debugged and more precise in terms of terminology.

ONVIF has suspended Dahua & HikVision on October 9, 2019.

ONVIF Benefits to End Users:
  • Greater flexibility
  • Increased product choice
  • Future proof

ONVIF gives you the flexibility to pick and choose optimal products for your needs without being locked into a specific brand. An ONVIF conformant video management software, for instance, will allow you to integrate ONVIF conformant devices from different hardware manufacturers. With software that supports both ONVIF and proprietary application programming interfaces (APIs), you can choose to use the ONVIF interfaces for certain functionalities and the proprietary API for other features at the same time. Having ONVIF conformant products is like having an insurance policy that future-proofs your system for a more secure investment.

ONVIF offers thousands of conformant products from different vendors that support one or more.

ONVIF Benefits for System Integrators/Specifiers:
  • Greater product choice
  • Assured interoperability
  • Easier integration

ONVIF conformant products allow you to create a flexible, cost-effective and future-proof system. ONVIF enables you to choose optimal products from different vendors. The conformant products allow you to be confident in their interoperability and makes integration of products from different vendors easier. The interoperable functionalities are determined by a conformant product’s ONVIF profile(s). 

Please share this handy guide to all consultant, System integrator & customers. After reading this guide if you ask any of them, what would make your life easier? They’ll likely reply with open source integration of systems. Well, it’s been some time coming, but it looks like that wish is becoming more of a reality these days.