Smog frequency attack test monitoring

With the implementation of the "harmonious society" and "safe city" strategies implemented by all levels of government in China, monitoring technology has been effectively promoted and quality has been improved. In particular, the front-end data acquisition technology, as the most important part of the entire remote security monitoring network, collects and processes the original scene information to meet the application requirements.

Since autumn and winter, haze weather has been frequent in all provinces and cities, especially in northern China. Fog weather has caused inconvenience for people to go out and has had a serious impact on traffic monitoring. Haze and haze weather significantly reduces the quality of outdoor surveillance video. Not only are the images darker in color, the contrast is reduced, and the details of some important targets are even more difficult to identify. The impact of bad weather on video surveillance first stems from the impact on the camera. Needless to say, the camera needs to capture the light of the subject through the lens and further process it into a digital image. In the foggy weather, there are a large number of particles that contain dust and condensed water vapor. Once the density is too high, it will hinder the propagation of light. Not only can the naked eye not see the signal lights, ordinary lenses can only see blurry images, and many surveillance vehicles. A similar picture can be seen in the traffic news video.

To solve these problems, mainly rely on the camera's features and lens performance. Many cameras now have a fog-fighting function, so implementing these functions depends largely on the lens chosen. Ordinary lens, through the visible part, this part of the light wave can not bypass the dense mist particles in the air; and the infrared function of the lens can transmit infrared light, infrared light with a longer wavelength can be diffraction of these particles, and then be able to clear Imaging.

It can be said that hazy weather has brought business opportunities to front-end data acquisition technologies such as surveillance cameras and cameras. This has brought challenges. Then, in front of “big cake”, what are the trends of security front-end technology that users care about?

1. The camera technology will develop towards the big dynamics and low illumination. The camera is the executor of information collection and plays a key role in the monitoring system. Therefore, many manufacturers have invested a lot of energy in this area. For the conventional security monitoring system, it goes without saying that the focus of R&D is on improving the sharpness and sensitivity.

There are two major categories of current image sensors, namely CCD and CMOS sensors. In general, compared with CMOS, CCD has the advantages of high sensitivity, small noise and high resolution, but at the same time, it also has the disadvantages of large power consumption, high cost and complicated circuit. The power consumption of the CMOS is only about 1/10 of the CCD sensor. Moreover, since the CMOS technology utilizes a high degree of integration, many peripheral devices are integrated, which greatly reduces the complexity and reduces the size. Because of this, CMOS sensor technology is much faster than CCD technology.

Micron is a major supplier of CMOS sensors, as well as Omnivision and ST. Currently, the resolution of Micron's MT9T031 and MT9P031 CMOS sensors has reached 3 million and 5 million pixels respectively. According to a marketing manager of the company, the competitive landscape of CMOS and CCD technologies will continue, and 5 million pixels will be two. The demarcation point of this kind of technology. But an obvious trend is that CMOS's performance is gradually approaching CCD technology.

In addition, there are some related technological innovations worth mentioning. For example, Panasonic launched the third generation of super dynamic (SDG) technology. Makes the contrast dynamic range of the image more than 40 times. Samsung's SNR enhancement technology, as well as lens coating technology, can achieve ultra-low illumination.

2. Network video server will be the development direction of digital storage technology Since the digital era, digital storage technology has developed rapidly. Digital video recorders (DVRs) have fully embodied the advantages of digital storage and have been widely used in recent years.

The core function of the DVR is decompression. The methods are divided into three categories: hard decompression, soft decompression, and hardware and software integration. The soft decompression cost is relatively low, but software implementation is based on ordinary computers and software, and it also occupies the system resources of the computer, which affects the image processing capability, and the reliability is relatively poor. On the Other hand, the application is The requirements for image quality are getting higher and higher, and with the increase in the demand for remote transmission, MPEG4, which has a relatively high compression rate, will become more and more popular. Therefore, decompression of embedded hardware using dedicated chips will become the mainstream of DVR. In this architecture, the audio and video capture, decompression and decoding display functions are all integrated into one SoC chip, and then the bus structure is used to transfer the multiple streams directly to the graphics card, and the IDE bus displays and stores the images. The system CPU Only play a role in control and distribution, the remaining large amount of resources can be used for rich image processing. Typical representatives of this architecture are TI's Da Vinci and NXP's Nexperia.

However, due to the large scale of the security monitoring network in the "Safe City", the interconnection of information among various sub-networks is very important, which inevitably needs to consider the data transmission problem. Based on this, there is a certain limit to this kind of target for the full hardware decompression scheme of localized monitoring. Because this program emphasizes the decompression speed, image processing capabilities and large-capacity storage. In a large network such as the "Safe City", the number of front-end sites for monitoring systems is so many that there must be a cascade of multiple machines, as well as access problems between internal and external networks, all of which make the control during remote monitoring complicated. Makes the hardware-based decompression method based on embedded hardware powerless.

Of course, the higher integration of super-capacity and ultra-multi-channel DVR can alleviate this problem, but it is not realistic to achieve. Therefore, a new solution is the network video server solution. In this scheme, storage uses general-purpose storage technology in the computer field. The solution aims at network video applications and adopts streaming media technology suitable for network transmission. It will become a development direction of the future "Safe City" monitoring program. Hou Yulin of Hitho Technology agrees with this. He said: “Video servers and network cameras do have certain advantages in realizing remote monitoring. However, there are still many problems to be solved before scale application, such as poor real-time performance. Delay is too large, the screen is not smooth, etc."

Another development trend of DVR is multi-interface technology. In the future, DVR will integrate interfaces from the traditional 485 and 232 to 1394, USB and Ethernet. In addition, the ever-increasing search and playback speed ratio is also a development trend. Of course, supporting larger and larger hard disk storage capacity is a goal DVR has always pursued.

3, two kinds of codec processing implementation will be leading the current mainstream video surveillance system compression format is MPEG4, although its processing delay is slightly larger, but it also takes into account the ability of all aspects, is a better compromise .

For the resolution format, the type is from the traditional CIF format to the current pixel with a 720×576 D1 format, and there are Half-D1, DCIF and so on. And Vweb Company introduced the full D1 resolution MPEG-1, 2, 4 and H.263 audio and video encoders, whether it is the memory capacity or bandwidth seems to be no problem for Chinese users. However, after a cool analysis, it is not difficult to find whether D1 can enter China's audio and video encoding and decoding national standards, based on the amount of D1 storage and cost factors, feeling that the promotion of large areas of D1 has not yet reached the heat, but the market is more emphasis on pixels 528 ×384 DCIF. Xie Yongqiang, an engineer of Shenzhen Haotian Technology Co., Ltd., also holds this view. He said: “The DCIF format seems to be more modest both in terms of bit rate and resource requirements, and it appears to be more balanced in all directions. Therefore, it can be effectively solved. The conflict between clarity and transfer rate, storage capacity, and cost."

In response to various resolution formats and compression formats, codec technology and chip implementation methods present a variety of features, but they can basically be summed up in two major categories, namely DSP and ASIC.

DSP is represented by TI's products. The advantage is that it has a very flexible processing method and the graphics processing function is powerful. However, such devices are uploaded from the previous dedicated signal processing architecture, and therefore have certain restrictions in certain aspects. For example, if there is a large amount of system resources, it seems that there is still lack of competitiveness for some applications with high control requirements and high bit rates.

However, TI has further optimized this point and won many orders. Hikvision, China's largest manufacturer of surveillance products, has used 3 million TI's TVP5150AM1 ultra-low-power video decoder chips in its video DVRs and video servers. The chip consumes only 113mW in typical operating conditions. The power consumption in the power-down mode is less than 1mW. The product requires only one crystal to support all standards, with analog and digital supply voltages of 1.8V and I/O supply voltages of 3.3V, meeting market demand.

The ASIC is mainly dedicated to the application-specific SoC. It can use the most advanced multi-core technology, adopt a standard bus architecture, match a certain storage resource, and expand the interface with strong performance. Based on these advantages, the ASICSoC chip is more applicable. High-end applications that can be applied to video processing.

These ASIC devices are less constrained in design time and can be specifically optimized for video codec, digital encryption, dynamic adjustment of resolution, and various remote control interfaces, thereby solving the image processing and complex control in network monitoring. And interface constraints bottlenecks.

As far as these two types of processing technologies are concerned, they have their own living space and will develop in parallel around the right applications.

Conclusion The security industry has always been a professional and closed market, especially in the age of digitization and networking. In the operation of mainstream IP cameras, professional technical support is required for installation, debugging, and subsequent system maintenance. The knowledge of IT networks has led to a long period of slowness in the promotion of network monitoring in the domestic market. The development of front-end technologies has also been subject to certain restrictions. It is believed that under the massive demand of the civilian market, security products will move closer to consumer electronics products, and front-end technology will continue to advance!