Video image compressed sensing coder-decoder system and coding-decoding method thereof

Abstract

The invention relates to a video image compressed sensing coder-decoder system. The system comprises a video acquisition subsystem, a processor module and the like. An FPGA module, a memory module DDR, a Flash and the like are in communication connection with the processor module. The processor module is in communication connection with an opposite end port and a computer control console through an Ethernet port and a UART interface respectively. An ARM kernel and a DSP kernel are arranged in the processor module, wherein the ARM kernel is used for a processing system to control and operate an operating system function, and the DSP kernel achieves voice and video coding and decoding functions through the CS technology. The DSP kernel and the ARM kernels operate in an interactive processing mode. The coding and decoding method includes the steps that a first frame of a video sequence serves as a first key frame of a first group; threshold value judgment is carried out; judgment is repeatedly operated, and finally image processing and output are carried out. The problem about operating efficiency of a CPU is solved, hardware and software resources of a DSP video system are saved, and the complexity of system operation is reduced.

Description

technical field [0001] The invention belongs to the field of signal and information processing in communication, and in particular relates to a video image compression sensing codec system and a codec method thereof. Background technique [0002] The device of the existing video compression encoder is realized based on the traditional Shannon sampling theorem. Under the requirements of the theorem, the sampling rate of the signal must be greater than twice the bandwidth of the input signal in order to achieve accurate reconstruction of the signal. Therefore, in order to restore video images, in hardware, the codec needs powerful CPU processing functions, more peripheral memory configurations, and higher cost; in software, the number of image processing samples increases and the amount of data is huge. Therefore, the main problems of current video image codecs are: [0003] (1) Encoders and decoders require high-speed CPU processing, requiring more auxiliary memory and peri...

AI Technical Summary

Problems solved by technology

[0003] (1) Encoders and decoders require high-speed CPU processing, requiring more auxiliary memory and peripheral expansion circuits

[0004] (2) In the process of video encoding, most of the coefficients are discarded after image transformation, resulting in a serious waste of data and system resources

[0005] (3) In a multi-user environment, it is difficult for this technology to achieve optimal codec of source and channel

[0007] The main difference between the CS decoder and the traditional decoder is that the traditional decoding process is the inverse process of encoding, the decoder is relatively simple, and the encoder is more complex, while decoding based on compressed sensing theory is no longer the inverse process of encoding, but a process of seeking The process of solving underdetermined linear equations, compared with the encoder, the decoder is more complicated

Images