Panasonic System Solutions Company discussed its innovative i-Pro IP surveillance solution at the Global Digital Surveillance Forum, part of SecuTech Expo 2008 held in Taipei, Taiwan.
Panasonic's i-Pro series of network surveillance products were launched three years ago, signifying its dedication to leading the way in IP technology. The i-Pro series is constantly growing, with new innovations based on intensive research and development. Panasonic has a total product lineup ranging from box cameras to various types of PTZ and fixed dome cameras, network disk recorders, encoders and management software. This year, the i-Pro Network Surveillance System will be further reinforced with its core competency for superior images and total system quality.
Panasonic has been working on video intelligence for many years, mainly for Japanese domestic projects, including traffic monitoring for railways and highways. Installing many of these systems raised overall security awareness and reduced manpower. Also, for logistics and industrial usage, Panasonic installed systems for high-level security requirements.
Through these project-based experiences, Panasonic has strengthened its video analytic algorithms to reduce false alarms. These algorithms have been improved to be more general and standard, which have then been embedded into a product, the NT314. The NT314 is capable of object detection, tripwire detection and auto tracking.
This system was installed at a public park in Tokyo to detect intruders, theft, violations, threats and so on. The environment has water and becomes dark at night, making video analytics critical. Motion detection algorithms work with low contrast for nighttime situations with dynamic sensitivity adjustment, calculating the standard deviation of the histogram. With the three-dimensional object size measurement algorithms, the system can determine whether it was a person or another object, such as cats, dogs, trees, leaves or waves. With this algorithm, the NT314 can detect people moving and a boat crossing, without detecting the waves. Even at night in the dark shadows with low contrast, it can detect a person walking around.
Even with these algorithms, there were some remaining issues. The first one is the fact that the quality of the input picture determines the preciseness of video intelligence. Even with a superior intelligence algorithm, if the input picture quality from the camera is not good enough, users only get many false alarms. This results in the system being turned off.
Also, fine adjustment is required for certain environments. Even though the dynamic sensitivity adjustment and three-dimensional object measurement algorithms simplify the calibration for setup, in some situations, the setup needs to be more accurate and optimized.
In an IP system, generating intelligent data from video streams requires real-time, simultaneous decoding and signal detection from multiple camera data streams. All this is performed by one PC with video management software. This means that generating intelligent data consumes much of the CPU power. Even with a high-powered dual CPU, the maximum number of cameras that could be processed would be four to five units. For demonstration purposes, this could be enough; but for a real system with 50 to 100 cameras, this approach is impractical. Therefore, intelligent engines need to be distributed, to be placed in either the camera or the encoder, as an edge device.
Usually, when embedding intelligent software into edge devices, standard DSPs (digital signal processors) such as the Texas Instruments DM642 or DaVinci devices are used. Panasonic sees this kind of approach from several manufacturers. However, additional DSPs become an additional cost to the camera, making the camera expensive and limiting applications.
Also, power consumption and heat may become an issue for development and reliability, as cameras have to be designed small. Judging from history, real technical breakthroughs provide additional value without increasing cost sometimes even helping to lower costs. As they become more widely adopted, they help make that value into standard. In this scenario, intelligence needs to be embedded into a new camera DSP, which can handle not only video signal processing and encoding, but can also generate intelligent data. When this type of DSP becomes available, intelligence can be provided without additional cost and be widely used.
To get better results from video intelligence, it requires better image quality and better resolution. Therefore, Panasonic has developed a new series of megapixel cameras, designed as leading models with reasonable prices. The benefit of IP systems is that they exceed NTSC television standards with higher resolution.
The key developments required for intelligence are:
1. Megapixel cameras capable of providing high-definition quality images.
2. Fine adjustment to suit certain environments. Technology is improving rapidly for video intelligence. But it is difficult to get the best results for many situations. Usually there are many settings that have to be optimized to meet customer requirements. Therefore, feedback from the field is required, which Panasonic is continuously improving on.
3. Single application-specific integrated circuit (ASIC) to handle both video signal processing and intelligence processing.
The direction of IP technology is toward open-system architecture. That means a multivendor solution for IP video surveillance. The advantage of multivendor solutions is the flexibility of product selection for performance, feature and cost.
On the other hand, the disadvantage may be the responsibility for full system operation. It is difficult and unclear to guarantee the whole system will work. It is difficult for integrators to test all the functions in advance. Another disadvantage is the inconsistency for continuous connectivity and upgrading for the system. As there is no standard in the security industry for IP interface, supporting many vendors requires a great deal of support and testing. It is becoming increasingly difficult to keep supporting connectivity with many other systems.
For closed systems in other words, a single-vendor system the advantage is that the end-to-end system is assured by a single vendor. Also, there is consistency for continuous connectivity and upgrading. The disadvantage may be the limited flexibility, because everything will rely completely on what the manufacturer can or cannot offer.
The industry will continue to try to overcome these issues, improving intelligent technology and connectivity to open IP systems. And Panasonic will continue to be part of this process and transition.