In recent years, video surveillance has been integrated with IP and IT technology, creating a new and unstoppable epic for intelligent video surveillance. As a leading security publication, A&S International hosted ¨Global Digital Surveillance Forum〃 (GDSF) in China for the second consecutive year from 28-29 August. Big names in the security industry were invited to share the latest and the most advanced in technologies and product concepts. Created as a platform for security players to interact, GDSF not only gives every attendee a chance to learn about the newest trends and applications, but it also allows for constructive exchange that should help bring the entire industry to the next level.
With increasing convergence of IP, IT and security technology, video surveillance is on a stable, if not explosive, growth curve. According to IMS Research, J.P. Freeman and Frost & Sullivan, the market for video surveillance is estimated to be US$6 billion and is expected to grow at 20 to 25 percent per year. It is estimated that 20 percent of video solutions would be based on some form of IP technology before 2008. By 2011, the revenue of network cameras and domes would reach up to $1.6 billion and 85 percent of video solutions would be IP-based. These figures show that digital surveillance is rapidly evolving and transforming, while research and development for IP applications and intelligent video surveillance (IVS) is on everyoneˇs radar.
¨GDSF China〃 was created to gather the elites of the security industry and unveil the latest developments in security. In this article, A&S International touches upon some of the highlights.
The Future is Bright for IVS and Application is the Key
The advancements in IP technology have revolutionized video surveillance by allowing cameras to transmit massive amounts of data quickly. The real challenge now lies in how to achieve an effective analysis. Additionally, while more and more cameras are put to use, there is not enough manpower to take charge of the back-end monitoring. Even if there is enough manpower to sit in front of monitors, it is still not feasible for guards to stay focused and react appropriately after a long period of idling. Without IVS, video monitoring would only be regarded as a tool for playback and retrieving images. This is why IVS holds the key to future developments in security.
What exactly is IVS? And what is the system structure of IVS? John Clark, Vice President of Business Development at ObjectVideo, explained that the main purposes of IVS are ¨to transform video signals into useful information and effectively respond to various kinds of events that IVS is intended to detect and alert.〃
According to Johan Lembre , Global Account Director at Axis Communications, there are four common IVS structures: ¨those that are embedded in IP cameras, those that use additional CPU or DSP processing power, those that use software on DVRs or PCs, and those that combine both front-end embedded functionality and back-end built-in software.〃
Key Criteria for IVSImage Quality
Regardless of system structure, the first priority of IVS is to have quality images for effective analysis. ¨High-resolution imagery is the key requirement for quality IVS analysis. With adequate images, IVS could work out a detailed analysis," said Peter MaKee, International Marketing Director at MOBOTIX.
Sony shares the same view on how image quality matters in IVS. Sony is currently working on advanced high-definition (HD) technology for video surveillance. Jun Wen, Product Manager of B2B Sales and Marketing Development at Sony, pointed out that working with HD is not enough in the pursuit of megapixel resolution. ¨It also requires other components such as sensors, lenses and processing chips. Sony has applied its next-generation CCD technology SuperExwave to new network cameras.〃
Distributed Intelligence St ructure Making IVS Easier
The quality of image determines the effectiveness of analysis. High-resolution quality requires strong processing capacity. ¨It requires powerful CPU processing capacity for centralized analysis systems like IVS. Therefore, the whole system should not incorporate too many cameras in order to bring about best functionality,〃 said Rei-mao Kao, Product Manager at Alpha Pricing. Michael Usami, Senior Manager of Product Planning Section at Sony, holds the same viewpoint: ¨Traditional CCTV systems send all video signals back to the back end for analysis, and a typical system requires an expensive and high-capacity server for all heavy processing loadings. But it came out to be ineffective and costly. On the other hand, using a distributed structure could significantly lower the cost and enhance the performance, no matter where IVS is actually executed.〃
1.IVS Embedded At the Front End
A distributed structure means that IVS is installed inside the camera, and this indicates an enhancement in the video analytics performance and system efficiency.
Because images are analyzed in the front end of network cameras even before they are compressed, this will significantly reduce server workload at the back end. This will also make searches and retrievals of video images easier and more economical. Furthermore, cameras are able to identify external factors that might trigger false alarms, such as shaking tree branches, raindrops and shadows. Most camera manufacturers, including Axis, Sony and Mobotix, have used distributed architecture in their IP camera designs.
Axis is embedding IVS functions into its network cameras. Providing video analysis not only lives up to most usersˇ demand, but also adds extra value for video applications. Axis has launched Application Development Program (ADP) and supplies its partners with SDKs, APIs and technical support.
Sony uses DEPA, or Distributed Enhanced Processing Architectureenabled model, for its IP cameras. DEPA can analyze as many as 15 frames and identify differences between the frames to distinguish shadows and to filter out redundant, overlapping movements. DEPA can also lessen camera vibrations and the end result is that false alarms caused by environmental noise can be minimized. Different applications of DEPA include perimeter protection and people counting.
¨Incorporating IVS into cameras provides every camera the ability to analyze different parameters. Under this structure, even if external networks are disrupted, cameras can still function via wireless transmission,〃 said McKee. ¨The structure can shift most of the analysis to the front end. This way, a simple PC-based DVR can work well with the whole system too, and the back end of the system would accommodate more cameras than before.〃
2. IVS At Back End
In a distributed structure, while one common approach is to implement IVS at the front end (IP camera), another is to implement IVS in the back-end hardware (DVR). ObjectVideo is an IVS software developer who provides IVS software to manufacturers of back-end hardware.
When addressing what the real IVS should be able to do, Clark explained: ¨The real IVS technologies need to have three different aspects. The first is to be able to extract target objects from the background and turn them into independent detailed datathe so-called video metadata. Metadata can also single out specific events through pre-programmed parameters, providing specific applications for system integrators.〃
IVS technology has been adopted in many different security applications, including airports, banks, schools, corporations and traffic surveillance. Is IVS really this magical that no deficiencies exist? Axis shared its view.
As promising as IVS may sound, the technology has limitations. There is a great deal of confusion between what is expected by end users and what can actually be done. Tasks that are difficult to achieve are people counting in crowds and object detection in extremely busy environments. Also, identification of unattended baggage in a busy airport is not something the current IVS can do. Some vendors are simply over-selling and many users have unrealistic expectations.
Axis recognizes the limitations in IVS and takes a different approach. Axis embeds IVS within the camera; when the camera is moved and the scene is changed, IVS will immediately detect the change and issue a warning. Such a design provides an extra layer of protection for the customers.
Qinghai-Tibet Railway The World's Highest Video Surveillance Application
Sprawling over 1043 kilometers, Qinghai-Tibet Railway is the highestelevated railway in China. A vast section of the railway (600 km) is covered by tundra and extreme environment makes it difficult for railway construction. In practice, many of the 30-plus railway stations are not staffed at all. Extreme climate conditions such as sand and snow storms also add to the challenges of the daily maintenance of the railway. China Railway Signal & Communication Corp. (CRSC) is responsible for the system integration of Qinghai-Tibet Railway and NICE products were installed.
Overall, 223 video stations and as many as 1,013 channels have been set up along the railways. The cameras include long-, mid- and short-range cameras, as well as thermal imaging and night vision cameras. According to CRSC, the system consists of technologies such as intelligent video analytics, networked digital video transmission and thermal infrared.
For critical areas like railway stations and bridges, cameras and video servers have been put in place. The main purpose of such a surveillance system is to monitor train operations and facility construction, and to respond to emergencies. Since many of the establishments along the railway are unstaffed, part of railway management is undertaken by said video surveillance system as well. Many cable lines and wires need to be monitored for situations that require immediate responsesfor example, when the railways are covered by sand or snow.
A system used for severe conditions like these requires comprehensive functionality and ultra-high performance specifications. The following is a brief introduction of the surveillance system used by Qinghai-Tibet Railway.
1.Video Image Analysis
A total of 1,013 video inputs are deployed in the surveillance system. One control center can manage to monitor 16 channels at the same time. It is normal for security staff to lose their concentration after staring at the monitoring screen for a long period of time and an unusual event could easily occur unnoticed; but with the aid of video analysis, alerts would be sent out immediately to patrolling security personnel.
2. Central Video Surveillance Platform
Qinghai-Tibet Railway has developed and used its own management platform so it can be more flexible when integrating systems from different vendors in the future.
3. Image Viewing and Retrieval Permit
Management of user authorization is pivotal in the overall security surveillance process. An authorization hierarchy should be set up for the the viewing and retrieval of video images. Those with lower clearance should only have limited access to their own images while those higher up can have full access to all 1,013 channels.
With more than 1,000 video channels implemented, Qinghai-Tibet Railway is considered the largest application of video analysis in the world. The most unique feature of this project is its adoption of thermal infrared imaging and night vision technology. Unmanned bridges and tunnels can be monitored through remote monitoring 24/7.
Qinghai-Tibet Railway is located in a rugged terrain and in high elevation; therefore, it has specific demands on hard-disk drives and encoders. In addition, because of the lower oxygen level, products may need to have lower heat conductivity or their life expectancy may be affected.
New technologies and product concepts were introduced in this GDSF event; more importantly, the myth of intelligent video was dispelled for many industry players. There is no doubt that the next GDSF event will continue to serve as the indicator of security industry and market trends.