Intelligent monitoring system, not only refers to has the first-class operation personnel and DVR equipment system, and the system should also can automatically realize the video content analysis, event detection, according to event generated alarms and send video, etc. In addition, this system also can record, store events and other video data, and the record of intelligent search and generate results analysis report. In general, this system is not a hybrid system, the whole IP video system.

From analog to IP

We are at the stage of transition from analog system into the IP system. Traditional simulation system is based on the simulation signal and the VCR as the center, is now gradually to the hybrid system (that is, the more centered on DVR system) shift, the shift is especially embodied in around 2005, and the next phase will be pure IP video system based on network.

Monitoring station through the network of DVR system, allowing the internal connection of arbitrary or traditional network browse on the Internet. In addition, the intelligent IP video solution analysis of the key components including digital equipment, equipment, compressed, local storage, decoding and video control equipment, etc.

In all of these devices, the camera is the most basic part. Image quality depends on many factors, not only the traditional monitoring is based on the CCD camera, and most of the IP is based on CMOS camera. The CCD camera can only provide a resolution of 750 x570 pixels of the traditional analog video, however, can improve the algorithm of maximum dynamic range and sensitivity of the camera is constantly.

Resolution is very important

CMOS can be progressive scanning equipment, its resolution can reach mega pixels, however, because of its sensitivity is limited, thus limiting the improvement of the algorithm of dynamic range, in addition, the sensor and processor is likely to be in harmony are an organic whole.

Fundamentally, the resolution is very important. The CIF format, for example, in the horizontal direction only 350 pixels; Use 2 cif format, can achieve 700 pixels; For two-thirds of cif format, horizontal pixel under 500 pixels; For the 4 cif and VGA format, horizontal pixel and up to 700 pixels. If you use the HDTV standards will be significantly improved, HD720 (1000 pixels), 1.2 Mega (more than 1300 pixels) and 3 Mega (2000 pixels).

At the same time, the bandwidth increases linearly increasing in the number of pixels. CIF, 2 CIF, two-thirds of CIF, 4 CIF and VGA generally are used less than 2 million per second, bandwidth, or in many cases is lower than 1 million per second. HD720, Mega 1.2 and 3 Mega use bandwidth are respectively 3 million, 5 million and 5 million per second. The increase of the bandwidth requirements, therefore, greater than the resolution increases the gain.

About the camera

Compared to traditional camera and mega pixel cameras, the former is characterized by: horizontal pixel can reach 756, for color and black and white images with good sensitivity, hybrid systems use interlaced scanning images, more commonly used in real time, 4 cif bandwidth is 300 million per second, lower integration costs, economic and practical.

And Mega pixel cameras generally horizontal pixel is 1350 (1.3 Mega) and 2000 (three Mega), the same sensitivity high, progressive scan avoiding the jagged edges of the image, and have real time effect. Typical MJPEG compression method, used for real-time video bandwidth is 10 million to 45 million per second. The camera integration, however, the high cost, can through the special lens and bandwidth to compensate. For example, HDD hard disk storage can be 10 times higher than that of using traditional cameras and encoder capacity.

Mega pixel cameras has its special position, they are a special camera for special occasions, especially need high resolution, for example in banking, retail stores, the cashier room, entertainment and other places. Bandwidth and storage capacity is not the main problem, that is to say, the mega pixel cameras in the accounts for only a small proportion of the overall system.

Compression technology

At present, the customer need higher quality compression technology, reducing the rate of data transmission and take up less transmission bandwidth and storage capacity. In different types of compression, JPEG and MJPEG based on single frame image, MJPEG using the finite difference algorithm, general transfer rate of 25 frames per second, bandwidth from 5 million to 12 million. In addition, the wavelet compression technology is suitable for the multi-channel video, video for 25 frames per second, the bandwidth from 4 million to 10 million. Mpeg-2 can provide video, DVD quality used dynamic detection, 50 frames per second in the bandwidth of the video from 2 million to 5 million, but there is no time delay model.

Mpeg-4 is one of the popular compression, image quality is dependent on the implementation, can provide delay model, the average compression ratio can reach 1 million to 4 million per second. Finally, there is the h. 264 and mpeg-4 10 Pt/AVC, higher than the efficiency of mpeg-4 10-40%, however, due to its high latency, therefore not suitable for application in some need PTZ camera to monitor the real-time scene, in general, the transmission rate is 0.5 million to 3 million per second.

In addition, the differential encoding is also very important. The whole frame is in the interval such as storage, and between the standard frame, store only a change of scene. I frame and P frame in the MPEG coding, I can see in real time and delay mode frame and P frame sequence.

bandwidth

The bandwidth of the importance of self-evident, network design is the key to minimizing the network load. Bandwidth requirements depending on the user needs to transmit video data, but also influenced by several other factors: such as resolution, including millions of pixels, D1, CIF, 2, 4 CIF and CIF. Compression techniques, such as Wavelet, MPEG; And the amount you need to focus on monitoring or storage of video streaming, etc. A system containing 25 cameras, for example, use full D1 mpeg-2 quality to monitor the busy scene, then the system needs of bandwidth will be 25 x 5 = 125 million/SEC, which requires a G bit Ethernet transmission.

Bandwidth cost cost, this is why need distributed intelligent system, but it can reduce the load on the network? In fact, a distributed monitoring only need according to some relevant information, not for all of the data, and based on the selective record events and reduce the background of quality records. In addition, the user can also choose to not display the irrelevant video images, available only to operating personnel useful video, the video can be triggered by events or actions, the other VCA should be located in the camera or the position of the encoder, rather than focus on layout.

Recording method

As for the difference between concentrated and dispersed storage, the former can be stored in one of the most effective position, all of the stored data are transmitted in the form of data flow through the network, but need to ANR (automatic network routing) to ensure that the process of storage; In general, use RAID5 redundancy group, even a single HDD hard disk is damaged, also won't lost data, in addition, also can undertake maintenance center.

And stored separately on the other hand is a distributed system, storage without the network load. In general, this system is not redundant, HDD hard disk damage will cause the limited data loss. In addition, the maintenance is also distributed, storage integration in the encoder.

Distributed storage is a relatively low cost, the method of multiple channel video can be Shared a HDD hard disk. Eight channels, for example, 800 gigabytes sharing, each path is 100 g byte, the system can with 2 CIF, 5 FPS mode of image, record 21 days medium activity scene video data. Can also be used in single channel encoder flash memory, it can provide 4 million memory, every day can store 4 CIF, 1 FPS video image. However, since there is no storage redundancy and HDD drive, so memory failure will brings to the system records the loss of the video.

If you add directly on the encoder memory, will be a very suitable for small low cost solution of the system. General in this case, the HDD are connected to the USB hard drive, such as 250 gigabytes per channel, 14 days can be stored cif, 2 frames per second medium activity scene video data. This scheme also no storage redundancy and 24-7 HDD hard disk availability is limited, HDD hard disk failure will pose a threat to record video lost.

The virtual local storage

In addition, there are virtual local storage technology. I - SCSI technology is a new standard of medium or large storage solution, all the main storage device manufacturers support the standard, including Infotrend, EMC, NetApp, HP and Dell. SCSI cable through the network to send, can be directly through the crossover cable connected to the unit, also can be connected through the network, up to build 64 virtual local storage devices.

Virtual local storage has the advantage that it is the high cost performance of large system solutions, such as 4 t Shared 50 pathway, each path is 80 g, it can be stored for 30 days CIF, 5 FPS medium activity scene data. Use RAID5 redundancy, a single HDD hard disk failure will not bring the loss of stored data, and also can focus resources so as to effectively safeguard.

Virtual local storage can be used for large systems, a separate local network is used to store, so the backbone or wan will not bring any load. Many edge devices connected to a single disk array, which can provide RAID5 redundancy, also can focus resources to more effective maintenance.

In the aspect of network monitoring and control system for centralized storage, video recorder and memory is located in the most effective and most key position. NVR economical and practical for large systems, all the data stored in the form of data flow through the network to send. However, that leads to the larger network load, and if there is no use the NVR, in the network fault data will be lost when the records. But on the other hand, multiple NVR can be located in local site centralized management, provides the indicator, or network-attached storage, RAID redundancy and centralized maintenance.

Video content analysis: to make more intelligent monitoring system

Video content analysis is the simplest and the most common form is video dynamic detection, in fact it can be included in any more than one task, DVR, video server, and the camera, in addition, the ordinary video movement detector (VMD) include size, direction and speed. At the same time, enhanced video movement detector provides multi-level image pixels change analysis, qualitative analysis and motion analysis, excluding those invalid alarm, and can provide prevent damage detection, according to the actual direction, actual size, speed and position to detect and track objects.

Video content analysis also have different types, such as biometrics, incident detection and behavior detection. Among them, the biological recognition technology including face recognition, fingerprint recognition and iris scans, etc.; Event detection can be used in the analysis to identify the legacy or missing objects; Can be used to detect found that those who take more expensive items from the goods, the target group, in addition, it also will monitor the behavior of some irregular movements, such as people walk in the wrong direction, will be retrograde, through the fence to deliver goods, and vehicles parked in the wrong position, and so on.

Video content analysis for the license plate number recognition success rate can reach 98%, however, the success rate is depends on the specific site conditions, easily affected by many factors, including the speed of the vehicle movement, lighting (using infrared), license plate use of fonts and cleanliness, etc. Related applications could include: car park entrance guard system and instrument system; Used in transportation system, strengthen the management of vehicle speed measurement or vehicles through a red light; Vehicle positioning system, and through the checkpoint as the loss of the vehicle; The crime; In addition, you can like the London transport system used in road pricing system.

VCA

VCA Algorithm (VCA: Variable Complexity Algorithm and Variable Complexity of video coding Algorithm) can be placed inside a camera and network video encoder, and thus provide a simple, but advanced dynamic testing and detection on the basis of conventional event functions, these are usually based on T164 platforms. Embedded VCA application examples, including to protect the safety of high-risk equipment, real-time intelligence and virtual perimeter, this is a new trend of the development of the IP video.

In short, intelligent video monitoring system provides the function of distributed intelligence analysis and storage, thus improve the efficiency of the system, and improve the speed of response to emergencies, reduces the network load and the demand for video storage. In short, they are the best low cost monitoring solution.