AVT digital high-definition video transmission technology promotes vehicle visual safety

    In-vehicle video surveillance has become an important application field covered by digital video surveillance technology. Due to various considerations such as automobile and driving safety, the video and visual systems involved must be stable and reliable, and there are more stringent requirements for video transmission. Digital video transmission has become an inevitable trend of development. Advanced Video Transfer (AVT), as a digital coaxial high-definition video transmission technology, fully supports visual lossless and real-time transmission of in-vehicle high-definition video equipment with its superior image performance and outstanding system characteristics, which is beneficial to expand from video recording to ADAS-related video applications.

　　The ideal of automotive vision becomes reality

　　After video surveillance became popular, transportation became an important area for the industry to expand its applications. In addition to implementing video surveillance on the road, on-board video equipment is also being gradually installed inside vehicles to collect video information in a timely manner and perform subsequent processing, which has opened the era of on-board monitoring. Early on-board video systems were mainly video recorders, whose purpose was to record video images of the surrounding area during the vehicle's driving. Often only cameras were installed in the front of the vehicle, and the collected videos were digitally compressed and stored. There were no high requirements for resolution and latency. In order to expand the scope of video surveillance, cameras were installed at the rear to increase recording channels. Multiple videos can be processed separately or transmitted to the front-end processor for synthesis, and there are no special requirements for the means and quality of signal transmission.

　　With the emergence of intelligent video processing technology, the video collected by the camera installed in the car can be analyzed and processed in real time, so as to provide early warning or warning for possible situations during driving. At this time, the vehicle-mounted equipment has certain visual functions, which can achieve the role of auxiliary driving. The vehicle-mounted video system can be divided into two categories according to the application. One is the 360° panoramic video driving system, also known as the surround view system; the other is the car rear view monitoring system. The following introduces the system composition and video transmission method involved.

　　As shown in Figure 1, the 360° panoramic driving system consists of a camera and a host on each side of the car. The four cameras use fisheye lenses to capture wide-angle images, and send them to the host for image stitching, trimming, etc. to generate a 360° panoramic image, which is then output to the LCD display. At present, the mainstream 360° panoramic products on the market are mainly standard definition solutions. The main reason is that when the pixels are low, the main processor capability of the panoramic host solution is not high. With the involvement of more and more host chip manufacturers, the processor performance is getting stronger and stronger, from dual-core to quad-core, and even to octa-core.

　　Therefore, the problem of processor capacity has been basically solved, and its performance is strong enough to support the load of high-definition video. Secondly, the resolution of the car screen used to be 1280*600, which is not high. With the popularization of 720P high-definition resolution car-level display screens, the demand for car-mounted high-definition monitoring has become more urgent. Each camera needs to transmit high-definition video signals to the host stably and reliably through cables, so it is imperative to provide high-capacity and high-reliability digital high-definition video transmission methods.

Figure 1 Composition of 360° panoramic driving system

　　The car rear-view monitoring system is a major turning point in the product form of the driving recorder. It has changed from satisfying all users with one product specification to satisfying the needs of more users with different specifications and configurations. Before the popularization of the 360° panoramic driving system, the demand for the car rear-view monitoring system was very high. It is reported that the market capacity and monthly shipments reached more than one million sets. Because the solution is inexpensive and the installation solution is simple, only a small camera needs to be installed at the rear license plate light. When the car starts to reverse, the screen automatically switches to the image of the reversing camera to assist the owner in reversing. Figure 2 shows the system block diagram of the solution. Similarly, with the popularization of high-definition screens, more and more car owners are more and more eager to demand high-definition cameras, so focusing on car rear-view applications, digital high-definition also faces great opportunities.

Figure 2 Composition of the vehicle rearview monitoring system

　　As the products gradually gain volume, more main control chip suppliers are joining the competition. For example, Allwinner mainly promotes T3 series and V series chips; MediaTek launches intelligent rear-view video chips; Meseda is the most representative solution provider using Leadcore 1860 chips; Rockchip and Intel jointly launched the code-named Sofia3G-R which will also be put into mass production; and there are also players waiting for Qualcomm platforms.
Innovative means of video transmission in the car

　　Innovative means of video transmission in the car

　　Digital high-definition video transmission technology is of extraordinary significance for improving the performance and quality of vehicle-mounted video equipment. It is not only used to improve the image quality and clarity of vehicle-mounted video sources, but also to provide visually lossless and real-time video transmission methods. In the vehicle-mounted market, there is demand for both pre-installed equipment and post-installed equipment, and both are looking forward to the use of more effective and reliable innovative methods in high-definition video transmission.

　　The wiring inside the car is very complicated and the reliability requirements are very high, but there are not many options for video connections. Ordinary AV audio and video cables are obviously not suitable for use in the car, so you need to consider choosing coaxial cables or Ethernet cables. For high-definition video transmission, the coaxial high-definition transmission system is currently more popular, among which the analog coaxial high-definition transmission system is mainly used in high-definition video surveillance cameras, such as AHD, HD-TVI and HD-CVI, and the digital coaxial high-definition system must be AVT.

　　HD-TVI, HD-CVI, and AHD are high-definition video transmission specifications based on coaxial cables. They belong to analog high-definition solutions and are implemented through intra-frame frequency addition. The only difference is the size of the frequency increase, while the implementation means and methods are basically the same. The advantage of this implementation method is that the solution cost is low, and the image quality can be improved to 1 million or 2 million pixels relative to the standard definition effect, which is the so-called analog high definition. However, because the pixel points of the image are improved by intra-frame frequency addition, the frequency within the frame increases, and the image brightness interference caused is difficult to avoid. Although ISP processing can make up for some defects that the human eye can perceive, making this type of solution popular in the industrial monitoring market, for the automotive market, in addition to satisfying the viewing of the human eye, the main control chip platform also needs to collect video for intelligent analysis and processing to achieve the effect of machine vision. For example, automatic parking technology and more assisted driving, as well as future unmanned driving technology, all require accurate video collection. To ensure the accurate transmission and reception of these videos, complete digital transmission must be achieved. If the analog signal method is still used during the transmission process, it will easily lead to video loss or even loss in the complex application environment inside the car, which will greatly reduce the ability of these main control chips to recognize high-definition video. This is indeed unacceptable for the automation and safety of the in-vehicle video system.

　　As a new digital coaxial transmission technology, AVT technology combines the characteristics of full digital architecture and digital signal transmission. In order to overcome the transmission distance limitation caused by the attenuation of digital signals on the transmission cable, it adopts a new generation of digital video compression algorithm, which is more effective than the standard VC-2 LD compression algorithm, and the video delay is as small as one frame image cycle, less than 3 milliseconds. Therefore, AVT is truly visually lossless, and the loss and delay of high-definition video signals are completely imperceptible to the human eye, which can be regarded as a real-time effect.

　　The visual lossless compression technology used by AVT is less than -40dB different from the original image, and the image quality is comparable to digital uncompressed transmission, and the image quality does not change with distance, wire material, temperature, etc. In contrast, analog coaxial HD needs to use DAC to convert to analog signal before transmission, and ADC needs to be used to convert back to digital signal at the receiving end. The two conversions inevitably cause video loss and deteriorate the image quality. In addition, analog video signals will cause nonlinear attenuation during transmission, resulting in more image quality loss. Another serious problem with analog coaxial HD is that the sampling clock recovered from the receiving end has a lot of jitter, which will cause inaccurate sampling position of the ADC at the receiving end, resulting in poor image quality after analog coaxial HD transmission, and the image quality between different machines is also different, with large dispersion and poor consistency, and changes with temperature. Therefore, it is not suitable for application scenarios such as in-vehicle video that have high requirements for image quality and large temperature range changes.

　　AVT also has innovative rate adaptive transmission technology, which can adaptively adjust the digital compression ratio according to the transmission distance and signal attenuation to achieve the best video image effect. Therefore, high-definition video transmission is no longer affected by the external environment, and it is also conducive to future product upgrades. In addition, AVT can also realize bidirectional data transmission. It supports UART, IR, I²C (master/slave) bidirectional transparent transmission and SPI forward transparent transmission on the control interface; on the audio interface, AVT supports SPDIF, I²S 8/16/24/32/44.1/48/96Kbps bidirectional transmission of sampling rates; in terms of wire use, AVT supports Cat 5 Ethernet cable transmission. In fact, as long as one pair of twisted pair cables is used, one channel of high-definition video can be transmitted. If an Ethernet cable is used, one cable can transmit four channels of high-definition video, so it is very suitable for in-vehicle wiring of vehicle-mounted video equipment, facilitating the multi-functional needs of vehicle-mounted systems and the development of ADAS systems.

         Combined with the application of 360° panoramic driving system

　　Based on AVT technology, the designed chip NS2520 is a transmitter (Tx), and the NS2521 chip is a receiver (Rx), which can support dual channels. NS2520 serially encodes the parallel signals output by the camera ISP. The video input interface is compatible with 10/20bit bit width, and has multiple video formats such as BT656/1120, CEA-861, CPI, DVP, etc., and supports up to 1080P@30fps. It also supports SPDIF/I2S audio embedding. One NS2521 can be used with two NS2520 chips to fully support the above formats.

Figure 3. Block diagram of AVT in panoramic parking system

　　The future of assisted driving is bright

　　With the development of intelligent automobile industry, Advanced Driver Assistant System (ADAS) will be popular in the market, and various ADAS-related video systems will continue to emerge. In addition to the panoramic parking and rear-view monitoring systems introduced, there will also be adaptive cruise control systems, lane keeping systems, forward collision warnings, lane departure warnings, brake assist systems, rear monitoring systems, night vision systems, driver anti-fatigue monitoring systems, and traffic signal recognition systems. All of these require high-definition real-time video as a guarantee to improve the image quality and clarity of the on-board video source, and real-time large-capacity reliable video transmission is crucial for ADAS applications. Traditional video high-speed transmission uses analog coaxial cables, but due to the presence of various electromagnetic radiation in the car, it is easy to cause various interferences. For high-definition video with a resolution of 1080P, the bandwidth will be as high as three megabits. Under the complex conditions of the internal lines of the vehicle, high reliability requirements must be met. Digitalization and networking can compress the video and transmit it through Ethernet cables, but it will inevitably bring network delays. As an advanced video transmission technology, AVT adopts digital coaxial high-definition transmission, but can effectively transmit visually lossless and extremely low-latency high-definition video signals with the help of general Ethernet cables. Therefore, the application of AVT digital high-definition transmission solutions will help promote the automotive vision market.

　　As shown in Figure 4, the application of AVT high-definition cameras in ADAS systems, in which the AVT chip is used to realize a new type of AVT high-definition camera, in which the high-definition video acquisition part composed of the CMOS sensor and the ISP processing unit is highly coupled with the AVT sending chip NS2520, completing the seamless video connection, directly outputting the digital high-definition video signal, and transmitting it through the Ethernet cable. The transmission signal in the car can be received by the host through NS2521, and restored to a visually lossless and extremely low-latency digital high-definition video signal, and then processed in real time by the main processor. The processing results are output on the screen display on the one hand, and are also output to the ADAS control unit to adjust the situation encountered during the driving of the car. It should be noted that two AVT cameras can share one NS2521 receiving unit. If more AVT cameras need to be connected, more NS2521 chips need to be added in advance to the design of the host.

Figure 4 Application of AVT HD Camera in ADAS System

　　In recent years, the in-vehicle video market, dominated by smart rearview mirrors and smart driving recorders, has shown a blowout trend, with shipments starting to reach tens of millions. The integration of in-vehicle video products has already appeared, and various devices combining vehicle navigation, in-vehicle multimedia, and in-vehicle entertainment are undergoing a transformation from standard definition to high definition, from analog to digital, and from single function to intelligent multi-function, among which digital high-definition video transmission technology will play an extremely important role.

　　This year is the first year of China's intelligent driving assistance system market. ADAS will be mainly used for monitoring, warning, braking and guiding tasks. Its demand will maintain a high growth rate in the next decade. It is a safety measure to protect drivers and effectively reduce accidents. It has both regulatory requirements and consumer attention. From an international perspective, both the EU and the United States have mandatory requirements for all motor vehicles to be equipped with autonomous emergency braking systems and forward collision warning systems by 2020. A recent report from an international survey agency shows that car buyers are increasingly interested in ADAS systems that provide functions such as assisted parking systems or rear blind spot monitoring, and that balance comfort and economy.

　　As processors and sensors are expected to account for the majority of sales, semiconductor companies are considering competing in these areas by offering distinctive products. With the improvement of video vision processors, combined with complete system solutions, they will not only occupy a place in the aftermarket, but also in the original equipment market. Therefore, the AVT series chips have undergone rigorous design and strict production and manufacturing processes and testing to meet the requirements of automotive regulations. AVT chips supporting 4K x 2K will be launched in the second half of the year, which will fully support ultra-high-definition ADAS requirements.