Industry News
    Home > News
  • Several techniques that affect the development of intelligent video analysis

        From the technology trends, IP video monitoring will gradually become the mainstream mode; unquestionable trend. With the development of network monitoring, intelligent video analysis has become an important direction of the development of IP video monitoring. However, for most chip manufacturers and system vendors, IP video surveillance gradually to the development of high-definition, intelligent direction, brought to them in addition to the huge market opportunities, as well as new challenges. The major technology trends of the design of a digital video surveillance system is needed to achieve the high-definition (1080p resolution), video analysis, intelligence management, IP, H.264 video compression standard, the higher the clarity of the image signal processing, the device features for application and with the project algorithm. These new trends so that developers have to be the face of low-power, low-cost, and the environmental aspects of the challenge.
        Used in the monitoring market, Digital Video Recorder (DVR) occupy half of the video surveillance market, however, with the development of network monitoring, network cameras (IPC) and network hard disk video recorder (NVR) to catch up. Which intelligent video analysis is becoming a hotspot of digital video surveillance, the DSP has become the driving force of this technology. There is no international standard on intelligent digital video analysis, so the the only programmable high-performance DSP and video processor that makes the realization of more complex monitoring algorithm becomes possible.
        Powerful SoC system
        TI´s digital signal processor (DSP) has a strong impetus to the development of digital video surveillance technology, which has a milestone in the the TMS320DM642 universal digital media processor. After the launch of TI digital media processing system-on-a-chip Da Vinci (DavinciTM) provides a powerful platform for digital video surveillance. Da Vinci various series platform can be completely covered by the digital video surveillance applications direction, the master processor of ARM and DSP-based processor two main development direction. The common feature is that you can support a variety of video compression formats, including H.264/MPEG-4 all DSP core has the ability to support the VCA.
        DSP SoC-based program in monitoring the market are dominant. Meanwhile, some FPGA vendors hoping to show their strengths in the differentiation and customization. Advantage of the inherent flexibility of the programmable coprocessor FPGA devices for monitoring applications and MPU / DSP combination, the designer is able, but also to take advantage of the FPGA to concentrate on developing algorithms to take full advantage of the MPU / DSP processing capabilities. Some manufacturers using its multi-channel video surveillance system hardware platform DSP + FPGA way to use TI´s TMS320DM6437, respectively, only one digital video input interface, connect 8-channel video decoder, video decoder increase the coupling between the video input interface and DSP logic, programmable FPGA can facilitate the achievement of the coupling between the two logic.
        Greater level of flexibility is a must in order to meet the market needs of the future digital video surveillance, and video compression technology to intelligent codec development, the traditional monitoring to the development of real-time intelligent analysis, optical system will be digitized electronic systems be more closely integrated.
        DSP processor to enhance intelligent analysis and processing capacity
        After generations of technology evolution, the video surveillance system is moving forward to the network (IP) and intelligent direction. In addition, the coexistence of a mix of analog and digital surveillance systems will also be the characteristics of China´s security market in the next few years. Selected as a high-end program, DSP the surveillance market mainstream processor. ASIC SoC latecomers have some impact on the market. In addition, FPGA is also trying to enter the market, but there are no obvious examples of success. Because at this stage there is no standard for intelligent analysis and processing algorithms, intelligent video analysis and processing be DSP MIPS requirements are very different based on the different compression formats, optimization of the algorithm is very critical. Since both flexibility and high performance DSP is more good intelligent analysis and processing work.
        The the ideal DSP processor for image processing applications must have the following characteristics: kernel processing capabilities; specifically for image processing instruction set; bulk data transmission of low-power hardware architecture; highly integrated; rich software module library ; powerful development tool. For video surveillance systems, the most suitable for the control and treatment should be convergence control functions and digital signal processing functions into one processor.
        Video surveillance products in China early mostly DSP, multi-channel, high-resolution video codec applications, DSP software is difficult to achieve the requirements of the system. FPGA hardware acceleration, real-time multi-channel compression. The advantages of the FPGA is a programmable, very suitable for high-need flexibility to provide differentiated application design. At present, the the FPGA chip video surveillance products production.
        H.264 video codec technology
        In the history of video surveillance, has a variety of video compression standards. Dedicated to low bit rate video telephony standard defined by the International Telecommunication Union ITU, including H.261/2/3; by the International Organization for Standardization ISO for consumer applications, sports image compression defined in the MPEG standard, including MPEG1 / 2 / 4; H.264 codec technology is based on the cooperation of the target application (Joint Video Team JVT) defined by the ITU and ISO, some standards still coexist. Mainly in the application of the M-JPEG, MPEG4 and H.264 three, we compare these standards through the following.
        Implementation costs: including algorithm complexity and degree of possession of resources, the level of difficulty and cost of technology acquisition, and ultimately reflected in the product cost;
        Encoding performance: mainly embodied in the equivalent main objective visual quality case, the size of the compressed stream obtained, which are very critical for the network transmission and storage;
        Network support: network support is good or bad depends on the adaptive capacity of the network packet loss and error fault tolerance and support network;
        Range of application support: the application needs of the ever-changing field of video surveillance, the ability to support a platform for different needs is also very important;
        Future potential: a development point of view, a standard potential size is reflected in the extent of future industry accepted, the smoothness of the transformation in the relevant application, and the degree of support for interoperability.
        Video surveillance products, the encoder is always the focus of the program, a suitable encoder programs tend to be able to determine the success or failure of a product. H.264 coding scheme is becoming the mainstream of the market, has been widely recognized by the market for its superior coding performance and good network adaptability. It not only can support the broad application environment, but also to meet the needs of different resolutions, bit rate, and transmission and storage environment.
        H.264 is an advanced video processing technology, the industry recognized, under conditions of the same quality of video playback, H.264 MPEG4 about saving nearly 40% of the transmission bandwidth or hard disk storage space, this is undoubtedly the promotion and application of remote monitoring from to a very important role. H.264 coding efficiency with its far more the MPEG4 complex computing the amount of the cost of the use of the most advanced frame between the prediction mode, including the complexity of the motion estimation, 1/2 and 1/4 pixel prediction; far MPEG4 advanced frame prediction modes, including as many as 13 kinds Zhen intra prediction mode; H.264 introduction of the new loop filter (in-loopfiltering) technology, the contribution to the substantial increase of the quality of video playback huge.
        However, the application of new technologies require a lot of computational processing, video codec processing platform also proposed new requirements. Due to capacity constraints, the software in the form of video codecs, such as codecs, general-purpose processor with a DSP, etc. can only give up a lot of H.264, the most advanced technology (such as Zhen between coding prediction mode, the ring Road filtering, etc.), such a result is the coding effect to return to the level of the MPEG4. More serious is this encoder the output code stream does not conform to the H.264 standard, all the universal decoder can not be its supporting composed of an open system. Selection of a non-standard encoder, the user can only limit the selection of its supporting non-standard decoder, and thereby becomes a closed system.

        Apexis H.264 IP camera recommend: