In recent years, with the development of embedded technology, the remote network video surveillance equipment based on embedded video server technology has appeared. The main working principle of embedded video server is: Analog video signal through A/D , using MPEG-2 or MPEG-4 Protocol compression, The compressed video data is then packaged and routed to the network. Users can directly use the browser or client-specific software to view the image on the video server, authorized users can also control the camera PTZ lens action or configuration of the embedded system. The compression section can be implemented by a dedicated hardware chip or by the embedded CPU for soft compression. If the latter is used, the CPU is required to have a strong computing function.
Embedded Video network server can be used in metropolitan Education Network, remote online training, network TV / Radio, real-time digital monitoring, remote video conferencing and other fields.
existing video server with cpu soft compression mode. 4 The analog video data is connected to the video decoder, and the decoder transmits the digitized video stream data to cpu--pnx1300 , mpeg The compression protocol is stored in the flash pnx1300 After compressing the video data, output to the Ethernet control chip, then the video data through the rj-45 port is uploaded to Ethernet.
PNX1300 is a Philips the company's high-performance, low-power chip, optimized for multimedia applications, is designed for + -bit Universal VLIW Structure of DSP . It can realize the algorithm of audio-video compression and decompression better.
but the soft compression speed is slower than hard compression, and to achieve Mpeg-x compression protocol, the workload is very large. In addition, the cost of protocol upgrades is significant.
The multi-channel Data packager integrates the data into a dedicated hardware compression chip for processing. Taking AT2021 ( integrated MPEG-2 Protocol ) chip as an example, the hardware compression algorithm is composed of several parts.
Video Input interface: Provides a video signal input interface, and some pre-compression pretreatment to improve the compression ratio.
Motion Prediction Compensation Unit: according to MPEG Recommended for BMA the algorithm makes the motion prediction compensation compression of the image.
dct/ quantization unit: For each image of the macro block DCT and Inverse DCT and then the quantization matrix to be obtained from the experiment DCT coefficients are quantified.
VLC and Entropy Coding unit: to quantify the DCT the DC coefficients are DPCM encoding, as well as the quantification of DCT The AC coefficients are encoded with variable length, and then the alternating current coefficients are entropy coded.
after multi-channel video data compression, you need to enter CPU package to send to the network.
CPU Select Samsung's 32 bit risc chip s3c4510 , The chip is specifically developed for embedded Ethernet applications, and the kernel is arm7tdmi , which supports high code density thumb instruction set. Through its peripherals i2c bus, you can control the front-end video capture, compression module chip operating mode, If you can set the brightness, chroma, saturation control register when the video decoder is decoded, s3c4510 support uart , add-on rs-232 port, you can connect the video server to the host for data exchange between the host and the server.
s3c4510 support 10m , 100m simplex and duplex mode Ethernet transmission. In single-mode, it supports csma/cd protocol, which in duplex mode supports Span style= "COLOR: #454545" >ieee802.3 mac layer. Therefore, you also need to give s3c4510 Connect the chip that controls the physical layer of the network, such as intel lxt970a
compression chip to AT2021 , CPU in s3c4510 For example, the compression chip and CPU the interface.
s3c4510 by address bus la[0..3] Addressing AT2021 related registers, via the data bus ld[0..15] Read and write data to set AT2021 The working parameters, monitoring AT2021 the operating state of the kernel and gets the compressed data.
For example, you can set the compressed video output stream to the format IPP or IBBP .
s3c4510 after you have taken the data to memory, follow each video data before the ID Data , then package the data in chunks, set its IP The destination address is then routed to the network, and the remote can receive it.
Of course, the packaged data can also be stored directly on the local hard disk for later query, that is, the DVR (DVR) .
The use of dedicated compression chip for data compression, not only reduces the development of video server difficult and cost, but also easy to compress protocol upgrades. With the rapid development and popularization of embedded technology, embedded devices will increasingly replace traditional PCsin the entire post-PC era, penetrating into industrial production and people's daily life.
Embedded Video Communication Network Service