FPGA-based ground penetrating radar real-time acquisition device and preprocessing method
By using an FPGA-based ground-penetrating radar real-time acquisition device and preprocessing method, the problem of insufficient real-time performance in real-time sampling and data processing was solved, achieving high-precision data acquisition and processing, improving data storage efficiency, and providing technical support for subsequent processing.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-30
- Publication Date
- 2026-03-31
AI Technical Summary
In existing ground-penetrating radar technology, real-time sampling places extremely high demands on the system, making it difficult to achieve. Furthermore, the real-time performance of data processing is insufficient, resulting in low acquisition accuracy and efficiency.
A ground-penetrating radar real-time acquisition device based on FPGA is adopted, including an acquisition timing control module, a high-speed parallel analog-to-digital conversion module, a pulse delay circuit module, a data transmission and storage module, and a power supply module. Combined with data preprocessing methods such as inter-channel compression, background denoising, DA distributed FIR filtering, AGC automatic gain control, and Hilbert-Huang transform, real-time acquisition and preprocessing are realized.
It achieves high-precision real-time acquisition and data processing, improves data storage efficiency, reduces redundancy, and provides technical support for subsequent data augmentation and interpretation.
Smart Images
Figure CN121763221A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of ground penetrating radar technology, and more specifically, to a real-time acquisition device and preprocessing method for ground penetrating radar based on FPGA. Background Technology
[0002] Ground-penetrating radar (GPR) detects the distribution of underground media by emitting high-frequency electromagnetic waves and is currently widely used for detecting underground targets. At present, GPR control methods, both domestically and internationally, are mainly based on FPGAs and employ an equivalent sampling method. This method involves the receiver collecting and plotting data from only one sampling point in a single transmitted pulse signal. By transmitting and receiving multiple times, data from different sampling points in each echo is collected, thus synthesizing a complete signal. Compared to equivalent sampling, real-time sampling can completely acquire the pulse signal from a single transmission. This allows for the elimination of random interference by superimposing multiple channels of data, resulting in higher acquisition accuracy for reproducing the original radar echo signal. However, real-time sampling requires extremely high acquisition rates and resolutions, making it difficult to implement in hardware. Furthermore, the current mainstream method for processing acquired data uses specialized host computer software, which suffers from insufficient real-time performance. Therefore, this invention proposes a real-time acquisition device and preprocessing method for GPR based on FPGAs. This method ultimately achieves higher acquisition accuracy for reproducing the original radar echo signal, improves data storage efficiency, reduces data redundancy, and provides technical support for subsequent rapid data augmentation and interpretation. Summary of the Invention
[0003] The technical problem to be solved by the present invention is to provide a ground penetrating radar real-time acquisition device and preprocessing method based on FPGA, with the aim of realizing real-time acquisition and preprocessing of ground penetrating radar.
[0004] This invention provides a real-time ground-penetrating radar acquisition device based on FPGA. The device mainly includes: an acquisition timing control module A, a high-speed parallel analog-to-digital conversion acquisition module B, a pulse delay circuit module C, a data transmission and storage module D, and a power supply module E.
[0005] Furthermore, the acquisition timing control module A generates corresponding working timing based on the control commands generated by the acquisition parameter settings issued by the USB3.0 in the data transmission module D, controls the high-speed parallel analog-to-digital conversion acquisition module B to start high-speed acquisition, and the pulse delay circuit module C generates corresponding trigger pulses to control the transmitter in the ground penetrating radar to work. The radar echo received by the receiving antenna is converted from analog to digital by the high-speed parallel analog-to-digital conversion acquisition module B and then transmitted to the acquisition timing control module A. The acquired radar echo is then locally stored and displayed by the data transmission and storage module D.
[0006] Furthermore, the FPGA-based ground-penetrating radar real-time acquisition device is connected to the radar antenna via a multi-core coaxial cable, which transmits trigger pulse signals and receives radar echo signals, respectively.
[0007] A real-time acquisition method for ground-penetrating radar based on FPGA mainly includes the following steps: ① After the radar antenna and the real-time acquisition device for ground-penetrating radar are connected, they start working and move synchronously to obtain real-time sampling data; ② Data preprocessing, removing redundant data through truncation and inter-channel compression processing methods; ③ Background denoising, filtering out background noise in the preprocessed real-time sampling data; ④ DA distributed FIR filtering, using a DA distributed algorithm to replace complex multiplication operations with table lookup operations to realize the function of an FIR finite-length unit impulse response filter, suppressing interference signals outside the effective frequency band; ⑤ AGC automatic gain control, automatically reducing the gain when the signal is strong and automatically increasing the gain when the signal is weak, thereby ensuring the uniformity of strong and weak signals and compensating for deep attenuation signals; ⑥ Hilbert-Huang transform, removing noise by performing empirical modulus decomposition on the data after automatic gain control and performing instantaneous phase analysis to obtain the real-time sampled data.
[0008] Furthermore, in step ②, data preprocessing first involves extracting the first n points of each channel of the collected data as valid data, then determining the compression factor and performing inter-channel mean compression transformation to complete the data preprocessing.
[0009] Furthermore, in step ③, background denoising first uses a data buffer (FIFO) to cache the data for different channels and calculate their mean. Then, each channel's data is subtracted from the mean to obtain the real-time sampled data after background denoising.
[0010] Furthermore, in step ④, the DA distributed FIR filtering method first designs the relevant attributes and parameters of the ground penetrating radar FIR filter, including: filtering method, sampling frequency, high-frequency and low-frequency cutoff frequencies, and filter order. The filter coefficients are obtained from the above attributes and parameters. Then, the corresponding filter coefficients and binary address combination results are stored in a lookup table. Then, the address obtained by splicing the background denoising data is searched in the table, and the corresponding lookup results are output and superimposed to obtain the data after DA distributed FIR filtering.
[0011] Furthermore, in step ⑤, the AGC automatic gain method first performs a square operation on the signal of the data after preprocessing, background denoising, and DA distributed FIR filtering to obtain the signal power value, determines the segment length, and calculates the average value of the sum of signal power values at that length. Then, it calculates the reciprocal of the average value of the sum of the squares of the signal as the gain weight, draws a table of gain coefficients under different gradients, inputs the calculated sum of the squares of the signal at the segment length, finds the corresponding gain coefficient, and finally adjusts the gain of the signal according to the obtained gain control value to obtain the data after automatic gain.
[0012] Further, in step ⑥, the Hilbert-Huang transform first reads the data signal after AGC (Automatic Gain Control) and stores it in registers a, b, and c respectively. Then, it compares whether the value in register b is an extreme value among the three registers, buffering the maximum and minimum values using different FIFOs. It also identifies and buffers the zero point of the data in register b. After the clock cycle ends, the register values are sequentially passed from b to a, then from c to b, and register c reads the new value. This process is repeated until the original signal is completely read. Then, cubic spline interpolation is used to calculate the interpolation points. After determining whether the IMF condition is met, the average value of the upper and lower envelopes is calculated, and the average value is subtracted from the input signal. This yields the IMF components of each intrinsic mode function after empirical mode decomposition. Finally, noise is removed by selectively reconstructing each IMF component. The reconstructed data is then subjected to a Hilbert transform using an FIR Hilbert filter to obtain the imaginary part data v(t). The reconstructed data is used as the real part data u(t). According to the instantaneous phase formula: Calculate the instantaneous phase information of the data to obtain the data after real-time sampling and processing. Attached Figure Description
[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0014] Figure 1 This is a block diagram of the sampling device and method disclosed in Embodiment 1 of the present invention. Detailed Implementation
[0015] To clearly illustrate the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:
[0016] Example 1:
[0017] This example provides a real-time data acquisition device for ground-penetrating radar based on FPGA, such as... Figure 1 As shown, the device mainly includes: a timing control module A, a high-speed parallel analog-to-digital conversion acquisition module B, a pulse delay circuit module C, a data transmission and storage module D, and a power supply module E.
[0018] like Figure 1 As shown, the acquisition timing control module A generates corresponding working timing based on the control commands generated by the acquisition parameter settings issued by the USB3.0 in the data transmission module D. It controls the high-speed parallel analog-to-digital conversion acquisition module B to start high-speed acquisition. The module connects to the radar antenna through a multi-core coaxial cable, transmits the corresponding trigger pulse generated by the pulse delay circuit module C to control the transmitter in the ground penetrating radar, and receives the radar echo received by the antenna. Then, the high-speed parallel analog-to-digital conversion acquisition module B performs analog-to-digital conversion and transmits the data to the acquisition timing control module A. Finally, the data transmission and storage module D stores and displays the acquired radar echo locally.
[0019] A real-time data acquisition method for ground-penetrating radar based on FPGA, such as Figure 1 As shown, this method mainly includes: ① After the radar antenna and the ground penetrating radar real-time acquisition device are connected, they start working and move synchronously to obtain real-time sampling data; ② Data preprocessing, removing redundant data through truncation and inter-channel compression processing methods; ③ Background denoising, filtering out background noise in the preprocessed real-time sampling data; ④ DA distributed FIR filtering, using the DA distributed algorithm to replace complex multiplication operations with table lookup operations to realize the function of FIR finite-length unit impulse response filter, suppressing interference signals outside the effective frequency band; ⑤ AGC automatic gain, that is, automatically reducing the gain when the signal is strong and automatically increasing the gain when the signal is weak, thereby ensuring the uniformity of strong and weak signals and compensating for deep attenuation signals; ⑥ Hilbert-Huang transform, removing noise by performing empirical modulus decomposition on the data after automatic gain and performing instantaneous phase analysis, thereby obtaining the real-time sampled data.
[0020] The specific calculation method for the real-time sampling and processing data of the FPGA-based ground penetrating radar real-time acquisition method is as follows: For the data preprocessing in step ②, data preprocessing is first performed. The ground penetrating radar real-time acquisition device actually collects 8400 sampling points per channel. If the required time window depth tns is required in actual use, the total time window length tns is divided by the time interval of collecting a single point 1 / 10.4ns to calculate the required number of points n. That is, the first n points of each channel of the 8400 collected data are taken as valid data. Then, the compression factor is determined and the inter-channel mean compression transformation is completed to complete the data preprocessing.
[0021] For background denoising in step ③, firstly, a data buffer (FIFO) is used to buffer the data of different channels and calculate the mean. Then, each channel's data is subtracted from the mean to obtain the real-time sampled data after background denoising.
[0022] For the DA distributed FIR filtering method in step ④, the relevant attributes and parameters of the ground penetrating radar FIR filter are first designed, including: filtering mode, sampling frequency, high-frequency and low-frequency cutoff frequencies, and filter order. The quantized filter coefficients are obtained from the above attributes and parameters. Then, the corresponding filter coefficients and binary address combinations are stored in a lookup table. The address obtained by splicing the background denoising data is then searched in the table. The corresponding search results are output and superimposed to obtain the data after DA distributed FIR filtering.
[0023] For the AGC automatic gain method in step ⑤, firstly, the signal power value is obtained by squaring the signal after preprocessing, background denoising, and DA distributed FIR filtering. Then, the segment length is determined and the average value of the sum of signal power values at that length is calculated. Next, the reciprocal of the average value of the sum of signal squares is calculated as the gain weight. A table of gain coefficients under different gradients is drawn. After inputting the sum of squares of the signal at the calculated segment length, the corresponding gain coefficient is found. Finally, the gain is adjusted according to the obtained gain control value to obtain the data after automatic gain.
[0024] For the Hilbert-Huang transform in step ⑥, firstly, the data signal after AGC (Automatic Gain Control) is read and stored in registers a, b, and c respectively. Secondly, the value in register b is compared to the extreme values of the three registers. The maximum and minimum values are buffered using different FIFOs, and the zero point of the data in register b is identified and buffered. Then, after the clock cycle ends, the register value is passed sequentially from b to a, and from c to b. Register c reads the new value. The above steps are repeated until the original signal is read. Then, the interpolation point is calculated using cubic spline interpolation. After determining whether the IMF condition is met, the average value of the upper and lower envelopes is calculated, and the average value is subtracted from the input signal. Thus, the IMF components of each intrinsic mode function after empirical mode decomposition are obtained. Finally, the noise part is removed by selective reconstruction of each IMF component. The Hilbert transform of the reconstructed data is performed through an FIR Hilbert filter to obtain the imaginary part data v(t). The reconstructed data is used as the real part data u(t). According to the instantaneous phase formula: Calculate the instantaneous phase information of the data to obtain the data after real-time sampling and processing.
[0025] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. An FPGA-based ground penetrating radar real-time acquisition device, characterized in that: The device comprises a collection timing control module A, a high-speed parallel analog-digital conversion collection module B, a pulse delay circuit module C, a data transmission and storage module D and a power supply module E. The collection timing control module A generates corresponding working timing according to the control instruction generated by the collection parameter setting of the USB3.0 in the data transmission module D; the high-speed parallel analog-digital conversion collection module B receives the trigger signal of A and starts high-speed collection; the corresponding trigger pulse is generated by the pulse delay circuit module C to control the transmitter in the ground penetrating radar to work; the radar echo received by the receiving antenna is subjected to analog-digital conversion by the high-speed parallel analog-digital conversion collection module B and then transmitted to the collection timing control module A; and the radar echo collected is stored and displayed locally by the data transmission and storage module D. 2.The FPGA-based ground penetrating radar real-time acquisition device of claim 1, wherein, The connection with the radar antenna is through a multi-core coaxial cable, which respectively sends a trigger pulse signal and receives a radar echo signal.
3. A real-time acquisition method of a ground penetrating radar based on FPGA, characterized in that: The method comprises the following steps: ①The radar antenna and the ground penetrating radar real-time collection device are connected and start to work and move synchronously to obtain real-time sampling data; ②Data preprocessing, redundant data is removed through interception and inter-channel compression processing method; ③background noise removal, background noise in the preprocessed real-time sampling data is filtered out; ④DA distributed FIR filtering, the DA distributed algorithm is used to replace complex multiplication operation with table lookup operation to realize the function of FIR finite-length unit impulse response filter and suppress interference signals outside the effective frequency band; ⑤AGC automatic gain, that is, when the signal is strong, the gain is automatically reduced, and when the signal is weak, the gain is automatically increased to ensure the uniformity of strong and weak signals and make up for deep attenuation signals; ⑥Hilbert-Huang transform, the data after automatic gain are subjected to empirical mode decomposition to remove noise and perform instantaneous phase analysis to obtain the data after real-time sampling processing.
4. The FPGA-based ground penetrating radar real-time acquisition method of claim 3, wherein: The ground penetrating radar real-time collection device can realize real-time processing of the collected high-level data and improve the processing efficiency of the real-time sampling data.
5. The FPGA-based ground penetrating radar real-time acquisition method according to claim 3, characterized in that: In the data preprocessing of step ②, the first n points of each channel data of the collected data are intercepted as effective data, then the compression multiple is determined and the inter-channel mean compression transformation is completed to complete the data preprocessing.
6. The FPGA-based ground penetrating radar real-time acquisition method according to claim 3, characterized in that: In the background noise removal of step ③, the data of different channels are buffered by using the data buffer FIFO and the mean value is calculated, the mean value is subtracted from each channel data to obtain the real-time sampling data after background noise removal.
7. The FPGA-based ground penetrating radar real-time acquisition method of claim 3, wherein the DA distributed FIR filtering method in step ④, first designs the relevant properties and parameters of the ground penetrating radar FIR filter, including: The filter mode, sampling frequency, high frequency and low frequency cutoff frequency and filter order are used to obtain the filter coefficients according to the above attributes and parameters, then the corresponding filter coefficients and binary address combination results are stored in the lookup table, then the address obtained by splicing the data after background noise removal is looked up in the table, the corresponding lookup result is output and superimposed to obtain the data after DA distributed FIR filtering.
8. The FPGA-based ground penetrating radar real-time acquisition method of claim 3, wherein: The AGC automatic gain method in the step 5 first carries out square operation on the signal of the data after the pre-processing, background denoising and DA distributed FIR filtering to obtain signal power value, determines the segment length and calculates the average of the sum of the signal power values under the length, then calculates the reciprocal of the signal square sum average as the gain weight, draws the gain coefficient table under different gradients, inputs the calculated square sum of the signal of the segment length, finds the corresponding gain coefficient, and finally adjusts the signal according to the obtained gain control value to obtain the data after automatic gain.
9. The FPGA-based ground penetrating radar real-time acquisition method of claim 3, wherein: The step ⑥ of Hilbert transform, first reads the data signal after AGC automatic gain and stores in register a, b, c, second compares the value in register b whether is the extreme value in three registers, the extreme value and minimum value are respectively cached by different FIFO, and the zero point of register b data is identified and cached, then at the end of clock cycle, the register value is transferred from b to a, c to b, register c reads new value, repeats the above steps until the original signal is read, then the interpolation point is calculated by cubic spline interpolation, the average of upper and lower envelope is calculated after determining whether the IMF condition is met, and the mean value is subtracted from the input signal, thus obtaining the intrinsic mode function IMF component after empirical mode decomposition, finally, the noise part is removed by selective reconstruction of each IMF component, the virtual part data v(t) is obtained by Hilbert transform of the reconstructed data through FIR Hilbert filter, and the reconstructed data is taken as the real part data u(t), and the instantaneous phase formula is: The data instantaneous phase information is calculated, and the data after real-time sampling processing is obtained.