A lossless compression method for radar data
Through wavelet transformation and entropy coding technology, the characteristics of radar data are quantized and encoded, which solves the problem of low radar data compression efficiency in the existing technology, and achieves efficient lossless compression, reduces storage and transmission requirements, and reduces data length.
Patent Information
- Application Number
- CN202411178306.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-27
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-08-27
AI Technical Summary
Existing radar data compression technology cannot fully utilize the spatiotemporal correlation and signal characteristics of radar data, resulting in low compression efficiency, especially when processing high-frequency components, and fail to fully utilize compression for smooth areas of low-frequency components. Potential, leading to memory and computing resource bottlenecks, affecting storage and transmission efficiency.
The radar data is decomposed by wavelet transform, and the quantization process is carried out and entropy coding is combined with entropy coding. The optimal encoding is dynamically generated by the Hoffman entropy coding method. Different quantization steps are adopted for different frequency components, especially the quantization steps of the high-frequency components. And the encoded data is packaged and compressed through algorithms such as DEFLATE, LZMA and Brotli.
It significantly improves the compression efficiency and quality of radar data, reduces the storage space and transmission bandwidth requirements, maintains the losslessness of the data, and reduces the encoded data length.
Smart Images

Figure CN118713679B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of data compression, and in particular relates to a lossless compression method for radar data. Background Art
[0002] All lossless compression methods are based on the fact that when the original sample data has information redundancy or data randomness defects, it can be re-ordered by secondary encoding or statistical methods to remove information redundancy in the sample as much as possible to achieve the effect of information compression. The information redundancy process can also be called the degree of data chaos. If the data is completely randomly generated, it can hardly be compressed by extreme compression; on the contrary, if the data is more regular, the stronger the regularity, the greater the compression space.
[0003] In the field of radar data processing, with the continuous development of radar technology, the amount of data collected by radar systems has increased dramatically, especially for high-resolution, multi-channel, real-time processing radar systems, which generate even more massive amounts of data; these data not only contain key information such as the target's position, speed, shape, etc., but also contain a large amount of background noise and redundant information.
[0004] Existing radar data compression technologies are mainly traditional lossless compression algorithms such as ZIP and GZIP, which are not optimized for the characteristics of radar data. Radar data has unique spatiotemporal correlation and signal characteristics, and traditional algorithms often cannot fully utilize these characteristics, resulting in low compression efficiency. For high-frequency components in radar data, especially those containing a lot of detailed information, traditional compression methods often find it difficult to effectively remove redundant information. At the same time, for smooth areas in low-frequency components, although the redundancy is high, traditional methods may fail to fully utilize their compression potential due to overly conservative compression strategies. The open patent application CN117895950A shows a lossless compression method for radar data. The separated T1 and T2 data are stored separately as binary files, and a set of encoding formats are designed to re-encode the binary sequence. The re-encoded data is subjected to independent sign bit, forced encoding conversion of weight bit and data valid bit, so that each weight data is aligned to 8 bits, and compressed using the LZO algorithm. However, during the encoding process, due to data cutting and byte alignment, data expansion occurs, and this expansion affects the final compression efficiency. This method partially relies on the implementation of hardware platforms such as FPGA, which may increase the complexity of the implementation.
[0005] Traditional compression algorithms face bottlenecks in memory and computing resources when processing radar data, resulting in reduced compression efficiency or even failure to complete the compression task. Therefore, how to efficiently store and transmit these radar data has become an urgent problem to be solved. Summary of the invention
[0006] The purpose of the present invention is to provide a lossless compression method for radar data, which decomposes the echo data through wavelet transform and then quantizes it, and performs entropy coding on the quantized data, thereby significantly improving the efficiency and quality of data compression.
[0007] The specific technical solutions are as follows:
[0008] A lossless compression method for radar data, the method comprising the following steps:
[0009] Step S1, acquiring radar data, separating the data header file, T1 echo data and / or T2 echo data, where the echo data is a double-precision floating-point data sequence or a fixed-point data sequence.
[0010] Step S2, using wavelet transform to decompose the echo data, separate the high-frequency component and the low-frequency component, and quantize the separated high-frequency component and low-frequency component respectively.
[0011] Step S3, entropy coding is performed on the quantized data, and the coded data is packaged and compressed by a compression engine to complete lossless compression of the radar data.
[0012] Furthermore, in step S2, the wavelet transform is a discrete wavelet transform, and the decomposition model is expressed as:
[0013] ;
[0014] in, and Represent the low-frequency component and high-frequency component respectively. and are the filter coefficients of low-pass and high-pass filters respectively; when the wavelet basis function selects Haar wavelet, , ; is the input signal, i.e. the raw echo data without processing; is the index of the input signal, indicating the position of the input signal in the original signal. is the index of the output signal, indicating the position of the output signal in the decomposed signal.
[0015] The quantitative model is expressed as:
[0016] ;
[0017] and Respectively represent the quantized low-frequency component and high-frequency component, and Respectively represent the quantization step size of low-frequency components and high-frequency components; , Take a constant greater than 1 to amplify the quantization step of the high-frequency component. represents the basic quantization step size, Indicates that the result is rounded down.
[0018] Further, entropy coding is performed on the quantized data, including the following steps:
[0019] Step S31, quantize the radar signal to generate a quantized data set , each quantitative value in the statistical quantitative data set The frequency of occurrence of .
[0020] Step S32, according to the entropy coding formula , calculate the quantitative data set Information entropy , to determine the uncertainty of the data.
[0021] Step S33, based on the calculated probability distribution ,The quantized dataset is encoded using the Huffman entropy coding method.
[0022] Furthermore, the Huffman entropy coding method includes:
[0023] Each data value As a leaf node, construct a priority queue according to Sort the sizes of the nodes in the priority queue from small to large; select the two nodes with the smallest probability from the priority queue and construct a new node whose probability is the sum of the probabilities of these two nodes; insert the new node back into the priority queue; repeat the construction and insertion of the new node until there is only one node left in the queue, at which time the Huffman tree is constructed.
[0024] Starting from the root node of the Huffman tree, traverse each leaf node and generate the corresponding Huffman code, 0 represents the left branch and 1 represents the right branch; encode the quantized data according to the generated Huffman code to generate an entropy coded data stream.
[0025] Furthermore, the encoded data is packaged and compressed by a compression engine, including the following steps:
[0026] Step S41, merging the low-frequency component data stream and the high-frequency component data stream after entropy coding into a complete data stream to be compressed.
[0027] Step S42: Select a data compression algorithm to compress the merged data stream. The data compression algorithm includes: DEFLATE, LZMA and Brotli.
[0028] Step S43, packaging and storing the compressed data together with the original data header file to generate a final compressed file.
[0029] Furthermore, the lossless compression method further comprises the following steps:
[0030] Step S4, generating a check code for the compressed file. During the data decompression process, the correctness of the check code is first verified. If the check is successful, the decompression operation is performed; if the check fails, it is prompted that the data is damaged.
[0031] Compared with the prior art, the present invention has the following beneficial effects:
[0032] The present invention can efficiently remove redundant information in radar data through discrete wavelet transform and quantization methods, and realizes efficient lossless compression of radar data by combining entropy coding technology, greatly reducing the demand for storage space and transmission bandwidth; according to the characteristics of different frequency components, different quantization step sizes are adopted, especially the quantization step size of amplifying high-frequency components, so as to further improve the compression efficiency and maintain the losslessness of data; the Huffman entropy coding method is adopted to dynamically generate optimal coding according to the probability distribution of data, so as to effectively reduce the length of the encoded data. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Figure 1 The figure is a flow chart of a lossless compression method for radar data according to the present invention. DETAILED DESCRIPTION
[0034] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention is described clearly and completely below. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0035] like Figure 1 FIG. 1 is a flow chart of a lossless compression method for radar data according to the present invention, wherein the method comprises the following steps:
[0036] Step S1, acquiring radar data, separating the data header file, T1 echo data and / or T2 echo data, where the echo data is a double-precision floating-point data sequence or a fixed-point data sequence.
[0037] For example: obtain radar data example1, data header file: header="RadarDataHeader"; original echo data x=[0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0]. The compression method of the present invention is explained below through an actual calculation process.
[0038] Step S2, using wavelet transform to decompose the echo data, separate the high-frequency component and the low-frequency component, and quantize the separated high-frequency component and low-frequency component respectively.
[0039] The wavelet transform is a discrete wavelet transform, and the decomposition model is expressed as:
[0040] ;
[0041] in, and Represent the low-frequency component and high-frequency component respectively. and are the filter coefficients of low-pass and high-pass filters respectively; when the wavelet basis function selects Haar wavelet, , ; is the input signal, i.e. the raw echo data without processing; is the index of the input signal, indicating the position of the input signal in the original signal. is the index of the output signal, indicating the position of the output signal in the decomposed signal.
[0042] After example1 is decomposed using Haar wavelet, the low-frequency component is:
[0043]
[0044] The low-frequency component after calculation is: ; Using the same calculation method, the high frequency component is: .
[0045] The quantitative model is expressed as:
[0046] ;
[0047] and Respectively represent the quantized low-frequency component and high-frequency component, and Respectively represent the quantization step size of low-frequency components and high-frequency components; , Take a constant greater than 1 to amplify the quantization step of the high-frequency component. represents the basic quantization step size, Indicates that the result is rounded down.
[0048] Set the base quantization step size ; Then the low-frequency quantization of example1: , , , ; The quantized low-frequency component is: ; The quantized high-frequency component is: .
[0049] Step S3, entropy coding is performed on the quantized data, and the coded data is packaged and compressed by a compression engine to complete lossless compression of the radar data.
[0050] The entropy encoding of the quantized data includes the following steps:
[0051] Step S31, quantize the radar signal to generate a quantized data set , each quantitative value in the statistical quantitative data set The frequency of occurrence of .
[0052] Example1 quantized dataset: , frequency statistics: 2 appears once, 3 appears once, 4 appears once, 6 appears once, -1 appears 4 times; then:
[0053] .
[0054] Step S32, according to the entropy coding formula , calculate the quantitative data set Information entropy , to determine the uncertainty of the data.
[0055] Step S33, based on the calculated probability distribution , the quantized data set is encoded using the Huffman entropy coding method. Low-probability events are assigned longer codes, thereby achieving effective data compression, and the compressed data size does not exceed the theoretical limit indicated by the information entropy H(Q); the information entropy of example 1 .
[0056] The Huffman entropy coding method comprises:
[0057] Each data value As a leaf node, construct a priority queue according to Sort the sizes of the nodes in the priority queue from small to large; select the two nodes with the smallest probability from the priority queue and construct a new node whose probability is the sum of the probabilities of these two nodes; insert the new node back into the priority queue; repeat the construction and insertion of the new node until there is only one node left in the queue, at which time the Huffman tree is constructed.
[0058] Starting from the root node of the Huffman tree, traverse each leaf node to generate the corresponding Huffman code, 0 represents the left branch, and 1 represents the right branch; encode the quantized data according to the generated Huffman code to generate the entropy coded low-frequency component data stream and high-frequency component data stream.
[0059] Set each data value of example1 As a leaf node: (2,1 / 8),(3,1 / 8),(4,1 / 8),(6,1 / 8),(-1,4 / 8),
[0060] Combine the two smallest nodes to construct a new node: combine (2,1 / 8) and (3,1 / 8) to form a new node (2,3,2 / 8); combine (4,1 / 8) and (6,1 / 8) to form a new node (4,6,2 / 8); combine (2,3,2 / 8) and (-1,4 / 8) to form a new node (2,3,-1,6 / 8); combine (4,6,2 / 8) and (2,3,-1,6 / 8) to form the final node.
[0061] Generate Huffman code: -1->0; 2->10; 3->110; 4->1110; 6->1111.
[0062] The encoded low-frequency component data stream and high-frequency component data stream are: and .
[0063] The encoded data is packaged and compressed by the compression engine, including the following steps:
[0064] Step S41, merging the low-frequency component data stream and the high-frequency component data stream after entropy coding into a complete data stream to be compressed.
[0065] Then the final complete data stream to be compressed in example1 is: .
[0066] Step S42: Select a data compression algorithm to compress the merged data stream. The data compression algorithm includes: DEFLATE, LZMA and Brotli.
[0067] Step S43, packaging and storing the compressed data together with the original data header file to generate a final compressed file.
[0068] The lossless compression method also includes the following steps:
[0069] Step S4, generating a check code for the compressed file. During the data decompression process, the correctness of the check code is first verified. If the check is successful, the decompression operation is performed; if the check fails, it is prompted that the data is damaged.
[0070] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A lossless compression method for radar data, characterized in that: The method comprises the following steps: Step S1, acquiring radar data, separating the data header file, T1 echo data and / or T2 echo data, where the echo data is a double-precision floating-point data sequence or a fixed-point data sequence; Step S2, decomposing the echo data by wavelet transform, separating high-frequency components and low-frequency components, and quantizing the separated high-frequency components and low-frequency components respectively; The wavelet transform is a discrete wavelet transform, and the decomposition model is expressed as: ; in, and Represent the low-frequency component and high-frequency component respectively. and are the filter coefficients of low-pass and high-pass filters respectively; when the wavelet basis function selects Haar wavelet, , ; is the input signal, i.e. the raw echo data without processing; is the index of the input signal, indicating the position of the input signal in the original signal. is the index of the output signal, indicating the position of the output signal in the decomposed signal; The quantitative model is expressed as: ; and Respectively represent the quantized low-frequency component and high-frequency component, and Respectively represent the quantization step size of low-frequency components and high-frequency components; , Take a constant greater than 1 to amplify the quantization step of the high-frequency component. represents the basic quantization step size, Indicates that the result is rounded down; Step S3, entropy coding the quantized data, and packing and compressing the coded data through a compression engine to complete lossless compression of the radar data; The entropy encoding of the quantized data includes the following steps: Step S31, quantize the radar signal to generate a quantized data set , each quantitative value in the statistical quantitative data set The frequency of occurrence of ; Step S32: according to the entropy coding formula , calculate the quantitative data set Information entropy , to determine the uncertainty of the data; Step S33, based on the calculated probability distribution , the quantized data set is encoded using the Huffman entropy coding method; Step S4, generating a check code for the compressed file. During the data decompression process, the correctness of the check code is first verified. If the check is successful, the decompression operation is performed; if the check fails, it is prompted that the data is damaged.
2. A lossless compression method for radar data according to claim 1, characterized in that: The Huffman entropy coding method comprises: Each data value As a leaf node, construct a priority queue according to Sort the nodes from small to large; select two nodes with the smallest probability from the priority queue, construct a new node whose probability is the sum of the probabilities of these two nodes; insert the new node back into the priority queue; repeat the construction and insertion of the new node until there is only one node left in the queue, at which point the Huffman tree is constructed; Starting from the root node of the Huffman tree, traverse each leaf node and generate the corresponding Huffman code, 0 represents the left branch and 1 represents the right branch; encode the quantized data according to the generated Huffman code to generate an entropy coded data stream.
3. A lossless compression method for radar data according to claim 2, characterized in that: The encoded data is packaged and compressed by the compression engine, including the following steps: Step S41, merging the low-frequency component data stream and the high-frequency component data stream after entropy coding into a complete data stream to be compressed; Step S42, selecting a data compression algorithm to compress the merged data stream, the data compression algorithm includes: DEFLATE, LZMA and Brotli; Step S43: Pack and store the compressed data together with the original data header file to generate a final compressed file.
Citation Information
Patent Citations
Data compression method based on radar algorithm
CN117895950A
Image compression coding method based on combination of wavelet packets and vector quantization
CN103546759A
Double-path remote sensing image compression method based on frequency domain characteristics
CN117336494A
Improved Huffman data compression method
CN1478255A
Compressing encoder, compressing encoding method and program
JP2007104645A
