CN1410780A - Radar video frequency data real time compression and decompression transmission method - Google Patents

Abstract

The invention discloses a method for real-time compression, decompression and transmission of radar video data. Two parts, the encoding board and the decoding board respectively located on the radar vehicle and the command vehicle, are connected in the form of twisted pair through two RS485 ports. The radar vehicle, There is a main channel and a sub-channel between the command vehicles; the encoding board compresses the echo data and transmits it to the decoding board through the main channel for decoding, and the decoded echo data is directly sent to the graphic display terminal; the communication between the two vehicles is through The sub-channel is carried out, the radar vehicle transmits the beam attribute code including beam position, range, radar working status and other information to the command vehicle, the command vehicle sends the command control code to the radar vehicle, and at the same time, the sub-channel retransmits the error to the decoder board when the main channel fails Beam data; the main channel communication adopts simplex mode, only sending but not receiving; the transmission protocol is: logo + data length + beam data + checksum; the secondary channel adopts half-duplex communication, and the transmission protocol is: feature code + data length +transmitted content+checksum.

Description

Real-time compression and decompression transmission method for radar video data

1. Technical field

The invention relates to radar transmission, in particular to a real-time compression and decompression transmission method for radar video data.

2. Background technology

As an over-the-horizon sensor, radar plays an important role in national defense, land and resource exploration, etc. In the development of radar technology, radar networking and data fusion technology have become the main development trend, and the problem of real-time transmission of radar data has become increasingly prominent. The traditional multi-cable transmission method has disadvantages such as complex connection, low reliability, poor anti-interference ability, and large transmission loss ^[1][2] . Optical fiber transmission is high-speed, but in practical applications, there are problems such as the need for special tools for wiring, time-consuming and laborious, and high cost ^[3] .

3. Contents of the invention

According to the defects or deficiencies in the above-mentioned prior art, the purpose of the present invention is to provide a real-time compression and decompression transmission method for radar video data. The present invention uses military-used twisted-pair wires to complete the real-time transmission of radar video images, preferably The shortcomings of the existing methods are solved.

The radar video data real-time compression decompression transmission method of the present invention comprises the following steps:

1) Connect the two parts of the encoding board and the decoding board respectively located on the radar vehicle and the command vehicle through two RS485 ports in the form of twisted pairs, and there are main channels and secondary channels between the radar vehicle and the command vehicle;

2) The encoder board compresses the radar video data and transmits it to the decoder board for decoding through the main channel, and the decoded echo data is directly sent to the graphic display terminal;

3) The signaling communication between the two vehicles is carried out through the secondary channel. The radar vehicle transmits the beam attribute code containing important information such as beam position, range, and radar working status to the command vehicle, and the command vehicle sends a command control code to the radar vehicle. The channel retransmits the erroneous beam data to the decoder board when the main channel is in error;

4) The main channel communication adopts a simplex mode, only sending but not receiving; the transmission protocol is: logo + data length + beam data + checksum; different from the 10-bit data frame format of commonly used RS232, this channel proposes a unique 34-bit Data frame format, and realize the asynchronous transceiver driver based on this frame format with FPGA;

The flag is used to distinguish the first frame and the difference value, and the data length is the number of bytes after beam data compression; the sender is the encoder of the encoding board, and the receiver is the decoder of the decoding board; the receiver judges the transmission process according to the checksum Is there an error. If there is an error, the decoder immediately notifies the main communication controller on the decoding board in the form of an interruption, and the main communication controller requests the communication controller on the encoding board to resend the beam echo data; There is RAM for backing up beam data;

5) The sub-channel adopts half-duplex communication, and there are two communication controllers A and B respectively located on the decoding board and the encoding board, and the sending and receiving roles between the two coordinate the communication requests of the encoding board and the decoder through tokens;

Secondary channel transmission protocol: signature + data length + transmission content + checksum; the transmission content is determined by the signature;

Set up a separate receiver for the operation control code from the console; when the echo data received by the decoding board is wrong, communication controller A requests retransmission of data from communication controller B on the encoding board; the communication of the main channel does not And interrupt, the decoder discards the erroneous beam data, backs up the beam attribute data, and processes the beam data retransmitted through the sub-channel after processing the remaining beam data in a frame;

Secondary channel transmission content: control code from command vehicle, beam attribute code, error beam data.

Another feature of the present invention is that the relationship between the main channel and the secondary channel is:

(1) On the encoder board, there is a SRAM and a bus controller for storing backup beam data between the encoder and the second communication controller. The encoder continuously refreshes the contents of the SRAM, and the communication controller receives a request from the main communication controller After that, get backup beam data from SRAM. The two obtain control over the SRAM through the bus controller with a handshake signal;

(2) On the decoder board, in addition to the SRAM and the bus controller that store the backup beam data between the decoder and the communication controller A, the decoder can also notify the communication controller A to request to resend the wrong beam by means of hard interrupt Data; the decoder only detects whether A communication controller refreshes the SRAM after sending a request to resend the signal. If it finishes, it immediately reads the beam data from the SRAM by DMA, and obtains it through the bus controller with a handshake signal. Control over SRAM;

(3), A communication controller is also responsible for the task of reading control commands from the console. The receiver connected to the console provides data to the communication controller A in parallel data mode.

4. Description of drawings

Fig. 1 is the algorithm block diagram that the present invention adopts;

Figure 2 is a lossless image coding framework based on EZW;

Fig. 3 is a normalization diagram in arithmetic coding;

Fig. 4 is a hardware platform block diagram of an embodiment of the present invention;

FIG. 5 is a time sequence diagram of four DSPs working in time division according to the embodiment of the present invention.

5. Specific implementation

The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments given by the inventor.

The hardware system of the present invention includes two parts, an encoding board and a decoding board respectively located on the radar vehicle and the command vehicle, connected in the form of twisted-pair wires through two RS485 ports, and the shortest communication distance is 500m. There are two channels between the radar vehicle and the command vehicle (distinguished by the main channel and the secondary channel). The encoding board compresses the echo data and transmits it to the decoding board through the main channel for decoding, and the decoded echo data is directly sent to the graphic display terminal. The communication between the two vehicles is carried out through the secondary channel. The radar vehicle transmits beam attribute codes containing important information such as beam position, range, and radar working status to the command vehicle. The command vehicle sends command control codes to the radar vehicle. At the same time, the secondary channel is in the main channel When an error occurs, retransmit the error beam data to the decoder board.

The main channel communication adopts simplex mode, only sending but not receiving. The transmission protocol is: flag + data length + beam data + checksum. The flag is used to distinguish the first frame and the difference value, and the data length is the number of bytes after beam data compression. The sender is the encoder of the encoding board, and the receiver is the decoder of the decoding board. The receiver judges whether there is an error in the transmission process based on the checksum. If an error occurs, the decoder immediately notifies the main communication controller (located on the decoding board) in an interrupt mode, and the communication main controller requests the communication controller located on the encoding board to resend the beam echo data. There is a RAM between the encoder and the communication controller for backing up the beam data.

The sub-channel adopts half-duplex communication, and there are two communication controllers (AT89C55) located on the decoding board and the encoding board (distinguished by A and B communication controllers), and the sending and receiving roles between the two coordinate the encoding board through tokens and decoder communication requests. Set up the receiver separately for the operation control code from the console. When the echo data received by the decoding board is wrong, communication controller A requests retransmission of data from communication controller B located on the encoding board. The communication of the main channel is not interrupted because of this, the decoder discards the erroneous beam data, backs up the beam attribute data, and processes the beam data retransmitted through the sub-channel after processing the rest of the beam data in a frame. Secondary channel transmission protocol: signature + data length + transmission content + checksum. The transmission content is determined by the characteristic code. Secondary channel transmission content: control code from the command vehicle, beam attribute code, error beam data (transmitted only when the main channel is in error).

The relationship between the main channel and the secondary channel is relatively close:

(1) On the encoder board, there is a SRAM and a bus controller for storing backup beam data between the encoder and the second communication controller. The encoder continuously refreshes the contents of the SRAM, and the communication controller receives a request from the main communication controller After that, get backup beam data from SRAM. The two obtain control over the SRAM through the bus controller with a handshake signal.

(2) On the decoder board, in addition to the SRAM and the bus controller that store the backup beam data between the decoder and the communication controller A, the decoder can also notify the communication controller A to request to resend the wrong beam by means of hard interrupt data. The decoder only detects whether communication controller A finishes refreshing the SRAM after sending out a retransmission request signal. If it finishes, it immediately reads the beam data from the SRAM by DMA. The two obtain control over the SRAM through the bus controller with a handshake signal.

(3), A communication controller is also responsible for the task of reading control commands from the console. The receiver connected to the console provides data to the communication controller A in parallel data mode.

The basic principle of the algorithm:

The radar echo signal is a non-stationary random process, and its inherent information entropy is very large. The compression ratio of more than 2 times cannot be achieved by using the lossless compression algorithm alone. Therefore, the reversible integer wavelet transform is introduced, and the lossy Compression scheme. See Figure 1 for the algorithm block diagram.

The algorithm scheme introduces the reversible wavelet transform, which decorrelates the echo data, makes the energy more concentrated, and provides a compact representation of multi-resolution data. Before performing wavelet transformation, symmetrical expansion is required, which is to meet the needs of wavelet filter for convolution operation. The choice of wavelet filter has a crucial impact on the performance of the compression algorithm. This algorithm chooses the 5-3 filter with good computational complexity and good reconfigurability. Embedded wavelet zerotree coding encodes wavelet coefficients according to their order of importance, where the importance of wavelet coefficients is measured by their contribution to image reproduction. Finally, entropy coding makes full use of the redundancy that still exists in the same frequency band and between different frequency bands after wavelet transform, and further improves the compression efficiency.

1). Embedded wavelet zerotree coding

The embedded wavelet zerotree framework provides an excellent solution to the image compression problem. The Embedded Wavelet Zero-Tree Coding (EZW) proposed by Shapiro and the Hierarchical Tree Set Partitioning Method (SPIHT) proposed by Said and Pearlman, both show high compression ratio and low computational complexity based on the wavelet compression scheme .

The lossless image coding framework based on EZW consists of three parts: 1) Reversible discrete wavelet change; 2) Hierarchical classification and selection of wavelet coefficients; 3) context-modeling-based. (arithmetic) entropy coding; as shown in Figure 2:

It is not difficult to see from the block diagram in Figure 2 that a choice needs to be made in each part, 1) the first step: select the wavelet filter; 2) the second step: select the appropriate wavelet coefficient classification method; 3) the third step: for the entropy The encoder selects the context model. Using a suitable wavelet filter can minimize the correlation of the data, properly sorting and classifying the generated wavelet coefficients and choosing a suitable context model for arithmetic coding will improve the compression efficiency. For image compression applications, the performance of the three-step model depends on all three modules. Each step must be well combined with the other two steps to produce an optimal, compact and embeddable bitstream.

In the first step, the wavelet transform decorrelates the image data, which concentrates the energy and provides a compact representation of the image at multiple resolutions. In this step, the choice of wavelet filter has a crucial impact on the performance of the compression algorithm. For lossless image compression, the reversible integer wavelet transform is widely used because it can complete the mapping from integer to integer. Experimental results show that (5,3) filter has the best overall performance among the selected filters.

In the second step, the wavelet coefficients are encoded according to their order of importance, where the importance of wavelet coefficients is measured by their contribution to image reproduction. Both EZW and SPIHT determine their importance according to the absolute value of wavelet coefficients. In progressive transfer applications, the selection criterion determines the bit allocation strategy for successive approximation and thus plays a very important role in the fidelity of the reconstructed image. In EZW, Shapiro defines importance maps. In the importance map, each binary value indicates whether the coefficient at its corresponding position is important for a given threshold. The successive approximation of the wavelet coefficients is realized by changing the threshold to generate the importance map.

Finally, entropy coding makes full use of the relationship between the same frequency band and between different frequency bands after wavelet transformation. This relationship manifests itself as a statistical dependence between the importance values of adjacent coefficients and between parent and child importance values. Arithmetic coding schemes based on the foregoing use neighborhood contexts to exploit high-order entropy.

2). Entropy encoder

The applicant adopts the arithmetic coding algorithm as the entropy coder, and its basic principle is as follows: the state of the coder is recorded with two variables L (the low endpoint of the bounded interval) and R (the width of the interval). For the convenience of discussion, it is assumed that L and R is a real number between 0 and 1. In the specific implementation, it can be an integer, which is constantly scaled by the power of 2.

①Initial L=0, R=1.

Assuming that P=[p _i ] is a normalized probability distribution, we have ${\mathrm{\Σ}}_{i=1}^{\mathrm{no}}{p}_t=1.$ definition $\mathrm{low}\left[i\right]={\mathrm{\Σ}}_{j=1}^{i-1}{p}_j,$ The cumulative probability of all letters preceding i in the source alphabet. Also define low[n+1]=1.

②L←L+R×low[i]

③R←R×p _i

④In order to solve the problem of precision and overflow, this step must be added (renormalization)

while R＜0.25do

(a)if L+R＜0.5 then bit_plus_follow(0)

(b) else if 0.5≤L then bit_plus_follow(1)

Set L←L-0.5

(c)Else

Set bits_outstanding← bits_outstanding+1

Set L←L-0.25

(d) Then, for all cases, Set L←2×L and R←2×R⑤To peeform bit_plus_follow(b)

(a) write_one_bit (b)

(b)Use write_one_bit(1-b)to output bits_outstanding bits of the opposite polarity

(c) Set bits_outstanding←0.

This maintenance of R ≥ 0.25 before encoding one symbol at a time makes the precision of L and R only at most two bits more than the precision of the probability _pi .

Figure 3 corresponds to (a), (b), and (c) in step 4. In Figure (a), the output bit is obviously 0, and L and R are adjusted accordingly. In the second case (Figure (b)), output 1. The special case is the third case, when R<0.25, while L and L+R are located on both sides of 0.5, the bit to be output cannot be determined, it depends on the next symbol to be input, but we can know that the next The bit next to the currently output bit must be the opposite of the currently output bit. So in the third case, it does not output, but still expands L and R, and just records this case (using bits_outstanding) to additionally output the number of bits opposite to the output bit next time.

Decoders are the inverse of this process. Given a code c, the decoder must determine the sequence of m symbols that produce code c. Assume V is the current window into c, with the same precision as L and R. Boundaries L and R are reinitialized to 0 and 1 respectively, and V must be initialized to the first few bits of bitstream c before the first symbol.

① Determine i so that low[i]≤(V-L)/R<low[i+1]

②L←L+R×low[i]

③R←R×p _i

④ In addition to L multiplication, V is treated the same as L in the encoder. The multiplication of V reads 1 bit from the encoded bit stream, shifts V to the left and adds it to the lowest bit of V.

⑤ Output the symbol i.

The selected hardware structure of the present invention

Due to the high complexity of the algorithm, a single DSP is no longer competent for the calculation work. At the same time, considering the redundancy of calculation in the future, a hardware platform with four DSPs is designed and a pipeline mechanism is adopted. The encoding board and the decoding board have exactly the same hardware structure, which reduces the workload of debugging and plate making. Since the decoding board and the encoding board have the same structure, they are not drawn separately.

The hardware platform is shown in Figure 4. Four DSP chips are interconnected through address bus and data bus, and can share on-chip memory with each other. The input data of the system enters the dual-port memory through the buffer (BUF). FPGA1 completes the addressing and writing signals of the dual-port memory according to the timing signal, and acts as a timer, sending four trigger signals that determine the DSP to start working, realizing the system. Pipeline works. The DSP receiving the work order obtains the echo data to be processed from the dual-port memory and starts encoding.

The working sequence of the four processors is shown in Figure 5. The DSP that completes the encoding task sends the data into FPGA2 through its link port, and FPGA2 completes the parallel-to-serial conversion of the data, and sends the data out in an asynchronous serial manner. The working mode of the decoding board is the same as that of the encoding board, so I won't go into details here.

,references

[1] Tian Erwen, Radar Signal Transmission and Processing and Its Modular Application, Microelectronics, Issue 01, 1994, Volume 24, Issue 1, p75-81.

[2] Yang Mei, Analysis of software interface between 1553B bus driver layer and transport layer of airborne PD fire control radar system, Modern Radar, April 1994, No. 2. P50-56.

[3] Research on high-speed data transmission of 8mm radar based on parallel port, System Engineering and Electronic Technology, No. 2, 2001. Volume 23 Issue 2.

Claims (2)

1. the compressed and decompressed transmission method of radar video data in real time is characterized in that, may further comprise the steps:

1) will lay respectively at encoding board, decoding deck two parts on radar truck, the command car,, connect, be provided with main channel and subchannel between radar truck, the command car with the twisted-pair feeder form by two RS485 mouths;

2) send decoding deck to by main channel after encoding board compresses echo data and decode, decoded echo data is directly given graphic display terminal;

3) communication between two cars is undertaken by subchannel, radar truck transmits the beam properties sign indicating number that comprises important informations such as beam. position, range, radar duty to command car, command car sends the order control code to radar truck, and subchannel retransmits the beam data of makeing mistakes to decoding deck when main channel is made mistakes simultaneously;

4) simplex mode is adopted in the main channel communication, only sends out and does not receive; Host-host protocol is: sign+data length+beam data+verification and; Adopt 34 bit data frames forms, and realize asynchronous receiving-transmitting driver based on this kind frame format with FPGA;

Sign is used to distinguish first frame and difference, and data length is the byte number after the beam data compression; Transmit leg is the scrambler of encoding board, and reciever is the demoder of decoding deck; Reciever is according to verification and judge whether transmission course makes mistakes, if make mistakes, demoder is positioned at the main communication controller of decoding deck immediately with the interrupt mode notice, and main communication controller is to this wave beam echo data of communication controller request repeat that is positioned on the encoding board; Be provided with RAM between scrambler and the communication controller, be used to back up beam data;

5) subchannel adopts the half-duplex mode communication, has two communication controllers of first, second to lay respectively on decoding deck and the encoding board, receives and dispatches the role between the two and coordinates encoding board and demoder communication request by token;

The subchannel host-host protocol: condition code+data length+transmission content+verification and; Determine the transmission content by condition code;

Operation control code from operation bench is set up reciever separately; When the echo data that decoding deck is received was made mistakes, the first communication controller was to the second communication controller request repeat data that are positioned on the encoding board; Therefore the communication of main channel is not interrupted, and demoder is abandoned the beam data of makeing mistakes, and backs up this beam properties data, after all the other beam data are handled in the frame, handles the beam data that retransmits through subchannel;

Subchannel transmission content: from the sign indicating number of controlling of command car, beam properties sign indicating number, the beam data of makeing mistakes.

2. the compressed and decompressed transmission method of radar video data in real time as claimed in claim 1 is characterized in that the relation between described main channel and the subchannel is:

(1), on encoding board, be provided with SRAM and the bus controller that is used for the storage backup beam data between scrambler and the second communication controller, scrambler constantly refreshes the SRAM content, and communication controller obtains the backup beam data from SRAM after receiving the request of autonomous communication controller.Obtain control with handshake by bus controller between the two to SRAM;

(2), on decoding deck, except the SRAM and bus controller of storage backup beam data, demoder can also hard interrupt mode notice first communication controller between demoder and the first communication controller, the request repeat beam data of makeing mistakes; Demoder only just detects the first communication controller and whether SRAM is refreshed end after sending request repeat signal, if finish, read beam data with dma mode from SRAM immediately, obtains control to SRAM with handshake by bus controller between the two;

(3), the first communication controller is also undertaken from operation bench and is read the task of controlling order; The reciever that links to each other with operation bench provides data in the parallel data mode to the first communication controller.

Images