Dear Friends,
Here’s a article on Video Analytic & Face Detection System. The most important thing which requires in video analytic is a Camera & regarding (IP base or Stand alone) Software.
What is Video Analytics?
Video Analytics, also known as IVS (Intelligent Video Surveillance) is a new emerging market for security allowing its users to easily monitor and secure areas with security cameras. With this new state of the art technology, businesses can easily monitor places of interest with sophisticated software that makes detecting threats or unwanted visitors simple and effective.
Intelligent Video Surveillance consists of algorithms that detect movement or changes in live and recorded video to see whether the movement or changes mean a possible threat is about to occur or occurring. These algorithms work by examining each pixel of the video and putting together all the pixel changes. If many pixels are changing in one area and that area is moving in a direction, the software considers this to be motion. Depending on the policies and alerts you have setup, you will be notified of this motion or other actions can be automatically taken by the software such as motion tracking which follows the motion until it is no longer detected.
Intelligent Video Surveillance consists of algorithms that detect movement or changes in live and recorded video to see whether the movement or changes mean a possible threat is about to occur or occurring. These algorithms work by examining each pixel of the video and putting together all the pixel changes. If many pixels are changing in one area and that area is moving in a direction, the software considers this to be motion. Depending on the policies and alerts you have setup, you will be notified of this motion or other actions can be automatically taken by the software such as motion tracking which follows the motion until it is no longer detected.
Object Recognition
You can program the recording software to distinguish certain objects within the video feed. Should this object change, be moved or removed, the software will immediately pick up this change.
Advanced Motion Detection/Tracking
Pick up even the slightest movement within the monitored area from any angle, and program the camera to follow these movements. The software can even be used to track speed, direction of travel and more.
License Plate Recognition
Video analytics software can even be programmed to detect license plates and take special images of these plates for retrieval. This sort of software is commonly used in places such as gas stations and toll booths to keep a track of traffic passing through.
Centralized Video Analytics Control
Another important element of video analytics is that all of these uses can be controlled from a central point. Video analytics software is connected to the each individual camera through a network, allowing control over any number of different feeds and cameras from a single point. It can also be used to send alerts, notices, etc. in cases of emergency.
Video analytics can be used with either digital network cameras or analog CCTV (Closed Circuit TV) cameras. However, the software requires a digital image to work with. In order to incorporate video analytics into an analog system you must first be able to convert the video feeds into a digital format. This is required in order for the software to analyze the video in such a precise way.
The increasing use of video is driving enterprises to look for better ways to manage their video platforms. One approach is to use "intelligent" video components with built-in video analytics (VA). An intelligent video surveillance platform can analyze scenes in real time for suspicious situations and immediately share the information with appropriate decision-makers. IP-based video systems can integrate with motion detection and alarm management applications so the system decides when to send video, at what frame rate and resolution, and when to alert a specific operator for monitoring and response.
Video Analytics
There are many ways to implement VA, ranging from general purpose dedicated VA systems to VA-enabled components that tackle a specific problem. Though a dedicated VA system can satisfy many pressing organizational requirements, there are drawbacks to the approach, especially when the enterprise has diverse VA goals, mixed environments and scalability requirements to satisfy
An alternative is to deploy an IP-based open systems video management platform that can support diverse "best of breed" VA components. This approach allows a company to "mix and match" VA solutions to specific needs. In addition to being more cost effective, an open systems video management platform ensures flexibility and scalability. As the VA needs of the organization change and as VA technology improves, an open video management platform ensures that the organization will be able to shift focus and keep pace with new technology
If designed wisely and used to their full potential, VA solutions can provide an ROI that surpasses many other types of video or surveillance applications.
Overview of Video Analytics
The main purpose of VA is to provide better information faster. Instead of relying solely on human scrutiny of video monitors to detect anomalies in the environment, VA software uses predefined rules and algorithms to analyze the video and extract information. These algorithms detect the outlines or edges of objects or in some cases identify the object or the pattern of its movement. VA then makes decisions about the images based on whether a preset protocol is confirmed or violated.
A standard CCTV design differs in fundamental ways from a video design that incorporates VA. In the former situation, guards interact with pan/tilt/zoom cameras that are capable of high resolution, requiring enough cameras to enable close-ups of images with enough detail to be recognizable. These cameras in the required quantities are typically very costly. The cost of such a system is increased by the number of guards that might be necessary to monitor a large number of cameras, such as would be the case for securing wide areas.
In a design using VA for wide-area surveillance, the number of cameras and guards is reduced. Fixed cameras generally cover the fence line or property boundaries and are connected to computers running VA software for detection, with a smaller number of PTZ cameras used for tracking by guards. This reduces the number of cameras, thus reducing the cost of the infrastructure. Some VA products also offer target hand-off from one PTZ to the next as the target moves from one camera’s field of view to another.
VA may also relieve the burden on the infrastructure. If the VA algorithm determines that the current video data is not of high priority, it might lower resolution to conserve bandwidth or storage space. Conversely, if the data is of high interest, it might automatically remove noise or improve resolution to enhance the image.
VA also provides capabilities that people simply cannot perform well such as counting and classifying many objects rapidly. Even when used for very complex operations such as facial recognition (comparing video images of faces against a database of criminal mug shots), many VA programs attain accuracy levels between 80 and 99 per cent.
Motion Detection and Beyond
Research indicates that people cannot remain attentive when required to monitor video for an extended period of time, especially when there are many cameras to watch. For this reason, most VA programs generate an alert when movement is detected within the camera's field of view. This motion detection capability reduces the amount of visual information that security personnel have to view at one time, allowing them to focus on more critical tasks.
At the most basic level, VA incorporates motion detection and event handling "rules" that make decisions about when to record and send video, at what frame rate and resolution, and when and whom to alert about the event. Early motion detection VA led to an unacceptable level of "false alarms," but most VA software today can be tuned to be more discerning or less discerning, depending on the asset you are protecting. Further, VA software can “learn” the scene, further reducing false alarms by understanding whether a moving object is truly an “object of interest”. If the asset is extremely costly, the system can be tuned to be more conservative. In this situation more kinds of movement will trigger an alert.
General Purpose versus Specialized Video Analytics Applications
Today VA extends far beyond motion detection. Despite many solutions that purport to be "general purpose," the most effective VA solutions are designed for a specific purpose -- such as facial recognition, behavior recognition, wide-area perimeter detection or people counting --with little overlap between applications.
Advanced analytics are currently used in defense and intelligence applications, and some of these techniques are beginning to reach the commercial market. For example, vendors are offering face recognition software that can be integrated into video surveillance systems. Other software is being developed that can recognize specific objects like weapons. And motion detection is being extended to include detection of specific activities (directional motion, atypical motion, or no motion) that might be considered suspicious in certain environments. Most of these applications are best satisfied by high specialized "best of breed" solutions.
As technology improves, the uses and capabilities of VA will expand and become even more specialized.
VA has many uses, including:
- Motion detection and/or directional motion
- Behavior recognition
- Facial recognition
- License plate recognition
- Tracking and classifying objects
- Smoke and fire sentries
- Wide area perimeter security
- Detection of objects left behind such as unattended baggage, abandoned vehicles
Common Video Analytic Methodologies
Some VA applications operate at the pixel level of an image, and within the software is an algorithm that describes the arrangement of pixels or changes within the pixels within the camera's field of view. This "pixel recognition" algorithm detects certain things or events by comparing objects in view with reference images. Others rely on pre-processing a scene at a “macro” level, to reduce the possible targets for observation, then analyzing shapes and forms and their interaction with each other.
At the other end of the VA scenario is the user. The user sets up rules that would define an “object of interest” -- specific security rules relating to the object or event. Common rules include parameters describing speed, direction, time of day, size of object, etc. When the object or event is detected by the pixel recognition algorithm that violate a rule, an alert or alarm is triggered.
For example, most perimeter security VA applications are designed to detect someone crossing a predetermined line. When a change is noted in the pixels within a camera's FOV, the VA software analyzes the difference in pixel count along with the size, location and the speed of the change. These measurements help classify the most likely cause of the change. The VA program subsequently consults its list of user-defined rules and performs a pre-specified action such as triggering an alarm or moving a camera.
Pixels on Target
The term "pixels on target" is a measurement of the number of pixels required for a given VA algorithm to be able to detect an object within an image. Some VA software can detect an intrusion with only four pixels on target, which represents a very small area of the scene. (Twelve pixels, for example, represent 1/20,000 of a scene).
Some VA solutions require only ten pixels to accurately detect and classify a human, but would require the camera’s to be much closer to the targeted area. The lower the number of "pixels on target" the farther away targets can be detected and the fewer cameras you will need. The chart below shows the pixels-on-target for a camera with a 30-degree field of view digitized at a 640x480 resolution. If 32,000 pixels-on-target are required for detection of a given object or event, the maximum distance of the target is 75 feet. If 8,000 pixels-on-target are required, 38 feet is the maximum distance. Note that algorithms that require only 8 pixels-on-target enable the same camera to detect targets at the much longer distance of 1,200 feet.
Integration Issues
Video analytics solutions are often used with other applications such as intrusion detection sensors, access control, card systems used for security purposes, and even business database software. Since any given VA application feeds into other systems, integration is a significant issue that must always be taken into account.
When deploying video analytics, it is critical to ensure that the video platform can accommodate a wide variety of hardware and software. An open systems platform designed for interoperability ensures that organizations and integrators alike will be able to select the best solution for today while retaining the flexibility to change solutions if technology or requirements change. This approach minimizes risk and future proofs the core solution to other network applications such as email servers or alarm systems.
Benefits:
Real-Time Video Monitoring
IVS software can display live video in real time. Also in real time are alerts for security policies you setup. These alerts will notify you immediately when there is a threat. Various alerts can be setup. These include email notification SMS messaging, on-screen alerts, alarms and triggers and even contacting the proper authorities.
Improves Quality of Surveillance
With IVS software you have the ability to be notified immediately when unusual activity is detected. This allows any available personnel to react upon alerts from the IVS software.
Accurate Detection
IVS software has the ability to detect specific behaviors. This means if someone is heading towards somewhere they should not be, you will be warned. The software also has the ability to be used indoors and outdoors, even in low light situations. Imagine being able to detect a car thief trying to break into a car in your parking lot in the middle of the night.
Ease of Implementation
IVS software can be integrated with existing CCTV/Analog systems or implemented with new state of the art IP network cameras. With a wide range of compatible cameras, IVS software can be implemented in any solution using security cameras.
Decreases Labor Costs
Due to the IVS software handling all alerts from possible threats, less personnel is needed to view video. With standard video systems, someone must always be watching for unwanted visitors but with the IVS software watching for you, more video can be watched by a lower number of people. Video no longer needs to be watched by a live person 24 hours a day, 7 days a week. This decreases labor costs and increases productivity.
IVS software can display live video in real time. Also in real time are alerts for security policies you setup. These alerts will notify you immediately when there is a threat. Various alerts can be setup. These include email notification SMS messaging, on-screen alerts, alarms and triggers and even contacting the proper authorities.
Improves Quality of Surveillance
With IVS software you have the ability to be notified immediately when unusual activity is detected. This allows any available personnel to react upon alerts from the IVS software.
Accurate Detection
IVS software has the ability to detect specific behaviors. This means if someone is heading towards somewhere they should not be, you will be warned. The software also has the ability to be used indoors and outdoors, even in low light situations. Imagine being able to detect a car thief trying to break into a car in your parking lot in the middle of the night.
Ease of Implementation
IVS software can be integrated with existing CCTV/Analog systems or implemented with new state of the art IP network cameras. With a wide range of compatible cameras, IVS software can be implemented in any solution using security cameras.
Decreases Labor Costs
Due to the IVS software handling all alerts from possible threats, less personnel is needed to view video. With standard video systems, someone must always be watching for unwanted visitors but with the IVS software watching for you, more video can be watched by a lower number of people. Video no longer needs to be watched by a live person 24 hours a day, 7 days a week. This decreases labor costs and increases productivity.
Conclusion
No one video analytics solution does everything. As with any technology decision, it is important to choose a solution that fulfills your organization's current requirements while providing a foundation for any and all later possible combinations of VA solutions and other technologies.
No one video analytics solution does everything. As with any technology decision, it is important to choose a solution that fulfills your organization's current requirements while providing a foundation for any and all later possible combinations of VA solutions and other technologies.
Hope the readers will find all the above information helpful to understand the topic & clear the concept of Video Analytic.
Click the below link to download the PPT on Video Analytic.
Thanks & regards.
Jayesh Ahire.
No comments:
Post a Comment