Radial artery ultrasonic echo signal processing device and method based on CORDIC algorithm

Abstract

The invention relates to a radial artery ultrasonic echo signal processing device and method based on a CORDIC algorithm, belongs to the field of digital signal processing, and solves the problems of low operation speed, high resource consumption and low practicability when an existing CORDIC algorithm is applied to an ultrasonic signal processing device. The radial artery ultrasonic echo signal processing device comprises a dynamic filtering module based on a radix-4CORDIC algorithm, an orthogonal demodulation module based on a mixed coarse and fine tuning CORDIC algorithm and a logarithm compression module based on an improved extended hyperbolic CORDIC algorithm. And the processing module is used for extracting components with diagnostic value in the ultrasonic signal and filtering noise to obtain a filtered ultrasonic signal, orthogonally demodulating a frequency mixing signal formed by the carrier frequency signal and the filtered signal to obtain a demodulated ultrasonic signal, and compressing the demodulated ultrasonic signal to an imaging range which can be presented by a display. Based on the technical scheme of the invention, the ultrasonic signal processing device with high speed, high performance and practicability can be designed on the whole.

Description

Technical Field

[0001] The present invention relates to the technical field of digital signal processing, and in particular to a radial artery ultrasonic echo signal processing device and method based on a CORDIC algorithm. Background Art

[0002] The radial artery is rich in the most important physiological and pathological information of the human body. With the development of computer technology and microelectronics technology, ultrasonic diagnosis has emerged. Ultrasonic imaging has the advantages of low cost, safety, non-destructiveness, and high sensitivity. Therefore, the use of ultrasound to detect radial artery signals to assist in disease diagnosis has unique advantages. However, with the development of ultra-large array sensor technology, the effect of ultrasonic imaging is largely limited by the real-time nature of multi-channel signal acquisition and processing. Ultrasonic signal processing involves a large number of sine and cosine functions and square root and logarithmic solution operations. Traditional signal processing methods mainly include polynomial expansion method and table lookup method. However, due to the large hardware area of ​​polynomial expansion method and the difficulty in implementation, the storage resources occupied by table lookup method increase exponentially with the improvement of accuracy. The above methods cannot meet the current requirements for real-time processing of big data, and cannot fully utilize the parallel processing of FPGA to solve the problem of efficiency of solving key functions. The Coordinate Rotation Digital Computer (CORDIC) algorithm can decompose complex functions into shift, addition and subtraction operations by continuously rotating a series of angles related to the operation base to approximate the required rotation angle. It has the characteristics of high precision and low complexity in hardware implementation. Based on this advantage, the CORDIC algorithm has been widely used in the communication field, such as the design of FFT, NCO phase detector, filter, etc.

[0003] In the field of ultrasound, CORDIC algorithms are mostly used in beamforming and coordinate scanning transformation technologies, and few studies have applied the CORDIC algorithm throughout the entire medical ultrasound signal processing process. At the same time, the traditional CORDIC algorithm still has a large number of iterations and resource consumption, and is not suitable for use in situations where ultrasound system hardware and software resources are limited.

[0004] Therefore, the current signal processing devices based on the traditional CORDIC algorithm have at least the following problems: 1) The traditional CORDIC algorithm has too many iterations, slow data processing speed, consumes too many computing resources, and is difficult to implement in hardware circuits; 2) The algorithm is not accurate and difficult to implement, and is not practical; 3) The data range solved by the algorithm is too small, making it difficult to be applied in practice. In general, it is difficult to design a high-speed, high-performance, and practical ultrasonic signal processing device based on the traditional CORDIC algorithm. Summary of the invention

[0005] In view of the above analysis, the embodiments of the present invention aim to provide a radial artery ultrasonic echo signal processing device and method based on the CORDIC algorithm, so as to solve the problems that the existing CORDIC algorithm has slow data processing speed, consumes too many computing resources, has low algorithm accuracy and is difficult to implement, and the data range solved by the algorithm is too small, which makes it difficult to apply it in the ultrasonic signal processing process and it is impossible to design a high-speed, high-performance, and practical ultrasonic signal processing device.

[0006] On the one hand, an embodiment of the present invention provides a radial artery ultrasound echo signal processing device based on a CORDIC algorithm, comprising:

[0007] A dynamic filtering module based on the base-4 CORDIC algorithm is used to extract the diagnostically valuable components from the original ultrasound signal and filter out the noise to obtain the filtered ultrasound signal;

[0008] An orthogonal demodulation module based on a hybrid coarse and fine adjustment CORDIC algorithm is used to orthogonally demodulate a mixed signal formed by a carrier signal and a filtered signal to obtain a demodulated ultrasonic signal; wherein the hybrid coarse and fine adjustment CORDIC algorithm is used to obtain amplitude information of the demodulated ultrasonic signal;

[0009] The logarithmic compression module based on the improved extended hyperbolic CORDIC algorithm is used to compress the demodulated ultrasound signal and reduce its dynamic range to the imaging range that can be presented by the display.

[0010] Specifically, the dynamic filtering module based on the base-4 CORDIC algorithm includes: a frequency multiplier module, a filter coefficient generation module, and a parallel distributed FIR filtering module; wherein,

[0011] A frequency multiplication module is used to multiply the system clock clk_in of a specific frequency to obtain a multiplied clock clk_100 whose sampling frequency is twice the specific frequency;

[0012] The frequency multiplier clock clk_100 is used as the input clock signal of the filter coefficient generation module;

[0013] The filter coefficient generation module is used to calculate and read the filter coefficient, specifically including: a base-4 CORDIC algorithm operation module, a FIFO control module, and a multi-channel data switch module;

[0014] The detection depth dep, operation angle phi, passband cutoff frequency fp and sampling frequency fs are input into the base-4 CORDIC algorithm operation module to complete the dynamic solution of the filter coefficient;

[0015] The FIFO control module and the multi-channel data switch are used to read multiple filter coefficients generated by the radix-4 CORDIC algorithm operation module and send them to the distributed FIR filter module in sequence;

[0016] The FIR filter module solves the filter data of the original ultrasonic signal according to the filter coefficient and accumulates the final filter result.

[0017] The output y(l) of the FIR filter corresponding to the input time series x(l) is described by the following relationship:

[0018]

[0019] Wherein h(l) is the filter coefficient, N is the filter order, and h(m) is the mth filter coefficient of the filter coefficient generation module, which is related to the filter order and is calculated by the filter coefficient generation module.

[0020] Furthermore, the filter coefficient in the dynamic filtering module based on the base-4 CORDIC algorithm is designed using a Hamming window, and the filter coefficient h(l) is calculated as follows:

[0021]

[0022] ω c is the probe bandwidth, R N (l) is a rectangular window:

[0023]

[0024] Furthermore, the base-4 CORDIC algorithm operation module for calculating the filter coefficients takes 4 as the base, and the iteration formula is:

[0025]

[0026] where σ i ∈{-2,-1,0,1,2}.

[0027] Specifically, the orthogonal demodulation module based on the hybrid coarse and fine adjustment CORDIC algorithm includes: a direct digital synthesis module, a FIR low-pass filter module, and a hybrid coarse and fine adjustment CORDIC square root operation module; wherein,

[0028] The direct digital synthesis module is used to mix the carrier signal with the filtered signal. The mixed signal needs to contain two orthogonal signals. The generation process of the mixed signal is:

[0029] a. Input the reset signal rst_n and the clock signal clk_100 into the direct digital synthesis module to generate two sine and cosine local oscillator signals i and q;

[0030] b. Input the two signals i and q and the ultrasonic signal after filtering by the dynamic filtering module into the digital mixing modules mult_sin and mult_cos respectively, perform multiplication operation, and obtain the mixed signal;

[0031] The FIR low-pass filter module is used to process the above mixed signal to obtain an in-phase signal I and an orthogonal signal Q respectively;

[0032] The square sum of the I and Q signals is taken, and the signal amplitude is obtained by taking the square root of the square sum of the I and Q signals through the mixed coarse and fine adjustment CORDIC operation module.

[0033] Furthermore, the hybrid coarse-fine tuning CORDIC algorithm decomposes the rotation angle θ into coarse angles θ H and subdivision angle θ L Two parts, expressed as:

[0034] θ=θ H +θ L

[0035]

[0036]

[0037] The hybrid CORDIC algorithm is a combination of coarse and fine rotations. It uses a hierarchical pipeline method, with the CORDIC processor I module and the CORDIC processor II module performing coarse rotation phase and fine rotation phase on the required angle respectively. The specific steps are:

[0038] a. Roughly divide the angle θ H , the in-phase signal I and the quadrature signal Q are respectively used as the starting input values ​​x 0 and 0 Input CORDIC processor I module to perform coarse rotation phase and obtain the rough value x M and M , the execution operation is:

[0039]

[0040] b. Subdividing the angle θ H And the rough value x obtained in the previous step M and M As the input value of the CORDIC processor II module, a fine rotation phase is performed to obtain the final result x n and n , the execution operation is:

[0041]

[0042] Specifically, the logarithmic compression module based on the improved extended hyperbolic CORDIC algorithm includes: a preprocessing module, a positive iteration module, and an expansion module; wherein,

[0043] A preprocessing module is used to preprocess the demodulated signal data to obtain an initial iteration value;

[0044] Positive iteration module, the specific steps of iteration are:

[0045] a. Input the initial iteration value obtained by preprocessing into the pipeline iteration processing module;

[0046] b. According to the iteration formula of each level, each iteration is expanded into a pipeline unit of one level, including registers, shifters, and adders and subtractors;

[0047] c. Perform data post-processing operations, complete module correction, and obtain the logarithmic value of the input data;

[0048] The extension module expands the logarithmic convergence domain of the CORDIC algorithm by adding negative number iterations, and is used to quickly calculate input data beyond the original convergence domain.

[0049] Furthermore, the iterative formula of the CORDIC algorithm with the extended convergence domain is:

[0050] When i>0:

[0051]

[0052] z i+1 =z i -δ i tanh -1 (2 -i )

[0053] When i<0:

[0054]

[0055] z i+1 =z i -δ i tanh -1 (1-2 i-2 )

[0056] In vector mode:

[0057]

[0058] The expanded convergence region becomes:

[0059]

[0060] in:

[0061]

[0062] Furthermore, after the extended hyperbolic CORDIC algorithm optimizes the rotation angle θ as described below, the logarithmic operation of the CORDIC algorithm based on the extended convergence domain can be directly performed. The specific optimization steps are:

[0063] Calculate the inverse tangent function value of θ under positive iteration and negative iteration respectively, then convert the floating point number to a fixed point number and multiply θ by 2 16 .

[0064] On the other hand, the present invention provides a method for processing radial artery ultrasound echo signals based on a CORDIC algorithm, characterized in that the method comprises the following steps:

[0065] Step S1, firstly, the ultrasound signal is A / D converted, and the obtained digital ultrasound signal is transmitted to the entire radial artery ultrasound signal processing device under the control of the global clock;

[0066] Step S2: filtering, demodulating, and compressing the original ultrasonic signal using a radial artery ultrasonic signal processing device to complete subsequent ultrasonic imaging;

[0067] Specifically, processing the original ultrasonic signal using the radial artery ultrasonic signal processing device includes the following steps:

[0068] Step S201, using a dynamic filtering module based on a radix-4 CORDIC algorithm, extracting components with diagnostic value from the signal and filtering out noise to obtain a filtered ultrasonic signal;

[0069] Step S202, mixing the filtered ultrasonic signal generated in the previous step with the carrier signal, and orthogonally demodulating the mixed signal using an orthogonal demodulation module based on a hybrid coarse and fine adjustment CORDIC algorithm to obtain a demodulated ultrasonic signal; wherein the hybrid coarse and fine adjustment CORDIC algorithm is used to obtain amplitude information of the demodulated ultrasonic signal;

[0070] Step S203: The demodulated ultrasonic signal generated in the previous step is compressed by a logarithmic compression module of an improved extended hyperbolic CORDIC algorithm to reduce the signal dynamic range to an imaging range that can be presented by the display.

[0071] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects:

[0072] 1. In the dynamic filtering module for filtering the original ultrasonic signal, the filter coefficient is calculated based on the base-4 CORDIC algorithm, which can improve the trigonometric function calculation speed and reduce the output delay;

[0073] 2. In the orthogonal demodulation module that demodulates the mixed signal formed by the carrier signal and the filtered signal and extracts the amplitude by taking the square root of the square sum of the demodulated signal, a hybrid CORDIC algorithm including a combination of coarse and fine rotations is adopted, and the square root operation is performed in a hierarchical pipeline manner, so that the algorithm implementation difficulty is reduced and the calculation efficiency is improved;

[0074] 3. In the module for logarithmic compression of the demodulated ultrasonic signal, the improved extended hyperbolic CORDIC algorithm is used for logarithmic calculation. On the one hand, this method can still achieve fast logarithmic calculation when facing large input data, that is, compared with the traditional CORDIC algorithm, the algorithm described in the present invention has a larger solution range and is more suitable for actual ultrasonic signal processing; on the other hand, this method requires fewer iteration steps, which is convenient for hardware circuit implementation.

[0075] In summary, by adopting the technical solution provided by the present invention, a radial artery ultrasonic signal processing device with high speed, high performance, easy implementation and high practicality can be designed as a whole.

[0076] In the present invention, the above-mentioned technical solutions can also be combined with each other to achieve more preferred combination solutions. Other features and advantages of the present invention will be described in the subsequent description, and some advantages can become obvious from the description, or can be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained through the contents particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] The accompanying drawings are only used for the purpose of illustrating specific embodiments and are not to be considered as limiting the present invention. In the entire drawings, the same reference symbols represent the same components;

[0078] Figure 1 It is a radial artery ultrasound echo signal processing device and flow chart based on CORDIC algorithm;

[0079] Figure 2 Design block diagram and process flow chart for dynamic filtering module based on radix 4-CORDIC algorithm;

[0080] Figure 3 Design block diagram and process flow chart for the quadrature demodulation module based on hybrid coarse and fine tuning CORDIC algorithm;

[0081] Figure 4 It is a schematic diagram of the principle of the hybrid coarse and fine tuning CORDIC algorithm;

[0082] Figure 5 Design a block diagram for the logarithmic compression module based on the improved extended hyperbolic CORDIC algorithm. DETAILED DESCRIPTION

[0083] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0084] A specific embodiment of the present invention discloses a radial artery ultrasound echo signal processing device based on a CORDIC algorithm, such as Figure 1 The radial artery ultrasound signal processing module is shown, including:

[0085] A dynamic filtering module based on the base-4 CORDIC algorithm is used to extract the diagnostically valuable components from the original ultrasound signal and filter out the noise to obtain the filtered ultrasound signal;

[0086] An orthogonal demodulation module based on a hybrid coarse and fine adjustment CORDIC algorithm is used to orthogonally demodulate a mixed signal formed by a carrier signal and a filtered signal to obtain a demodulated ultrasonic signal; wherein the hybrid coarse and fine adjustment CORDIC algorithm is used to obtain amplitude information of the demodulated ultrasonic signal;

[0087] The logarithmic compression module based on the improved extended hyperbolic CORDIC algorithm is used to compress the demodulated ultrasound signal and reduce its dynamic range to the imaging range that can be presented by the display.

[0088] The process of filtering, demodulating and compressing the ultrasonic echo signal using the above device includes:

[0089] Step S1, firstly, the ultrasound signal is A / D converted, and the obtained digital ultrasound signal is transmitted to the entire radial artery ultrasound signal processing device under the control of the global clock;

[0090] Step S2: filtering, demodulating and compressing the input digital ultrasound signal using the radial artery ultrasound signal processing device to complete subsequent ultrasound imaging;

[0091] Specifically, the process of processing the input digital ultrasound signal using the radial artery ultrasound signal processing device includes the following steps:

[0092] Step S201, using a dynamic filtering module based on a radix-4 CORDIC algorithm, extracting components with diagnostic value from the signal and filtering out noise to obtain a filtered ultrasonic signal;

[0093] Step S202, mixing the filtered ultrasonic signal generated in the previous step with the carrier signal, and orthogonally demodulating the mixed signal using an orthogonal demodulation module based on a hybrid coarse and fine adjustment CORDIC algorithm to obtain a demodulated ultrasonic signal; wherein the hybrid coarse and fine adjustment CORDIC algorithm is used to obtain amplitude information of the demodulated ultrasonic signal;

[0094] Step S203: The demodulated ultrasonic signal generated in the previous step is compressed by the logarithmic compression module of the improved extended hyperbolic CORDIC algorithm to reduce the signal dynamic range to the imaging range that can be presented by the display, thus laying a foundation for subsequent ultrasonic imaging.

[0095] Specifically, the dynamic filtering module based on the base-4 CORDIC algorithm is used to filter the original ultrasound signal, extract components with diagnostic value, and filter out noise. Figure 2 As shown, the module mainly includes: a frequency multiplier module, a filter coefficient generation module, and a parallel distributed FIR filter module.

[0096] First, the frequency multiplication module is used to multiply the system clock clk_in of a specific frequency to obtain a multiplied clock clk_100 with a sampling frequency twice the specific frequency. According to the sampling theorem, the center frequency of the probe is 3.5MHz, and the sampling frequency should be greater than 7MHz. In order to reduce the impact of quantization error on the ultrasonic sound beam, the sampling frequency should be greater than or equal to 8 times the center frequency of the probe. To meet this condition, the present invention adopts a 50MHz clock input, inputs the reset signal rst_n and the 50MHz system clock clk_in into the frequency multiplication module, and obtains the sampling frequency f after multiplication. s The multiplied clock clk_100 is 100MHz.

[0097] The frequency multiplied clock clk_100 is an input clock signal of the filter coefficient generation module and the parallel distributed FIR filter module.

[0098] The output y(l) of the FIR filter corresponding to the input time series x(l) is described by the following relationship:

[0099]

[0100] Wherein h(l) is the filter coefficient, N is the filter order, and h(m) is the mth filter coefficient of the filter coefficient generation module, which is related to the filter order and is calculated by the filter coefficient generation module.

[0101] Next, the filter coefficient generation module is used to calculate the filter coefficient, including the base-4 CORDIC algorithm operation module, the FIFO control module, and the multi-channel data switch module. The filter coefficient is designed using the Hamming window, and the coefficient calculation formula is:

[0102]

[0103] ω c is the probe bandwidth, R N (l) is a rectangular window:

[0104]

[0105] As can be seen from the above formula, the FIR filter module involves a large number of sine and cosine function calculations. Therefore, the CORDIC algorithm can be used for fast calculation to improve the data processing speed.

[0106] The present invention adopts a base-4 CORDIC algorithm operation module, which is different from the traditional base-2 CORDIC algorithm. It uses 4 as the base and the iterative formula is:

[0107]

[0108] where σ i ∈{-2,-1,0,1,2}.

[0109] Assuming the output precision is n, the number of iterations required for the above algorithm is Compared with the traditional radix-2 CORDIC algorithm, the number of iterations is reduced by half. Therefore, the use of the radix-4 CORDIC algorithm can increase the calculation speed of the sine and cosine functions and reduce the output delay.

[0110] use Figure 2 The process of calculating the filter coefficient by the filter coefficient generation module is as follows:

[0111] The multiplied clock clk_100 with a sampling frequency of 100MHz is used as the input clock signal to input the radix-4CORDIC algorithm operation module and the FIFO control module respectively. At the same time, the parameter detection depth dep, operation angle phi, cutoff frequency fp and sampling frequency fs are input into the radix-4CORDIC algorithm operation module, so that the filter coefficients can be dynamically solved; in the solution process, the FIFO control module is called to control the filter data, and the configuration parameters are controlled through the multi-way data switch; finally, the multiple filter coefficients h(m) (coe1~coe17) generated by the radix-4CORDIC algorithm operation module are read and sent to the subsequent distributed FIR filter module in sequence.

[0112] Finally, the distributed FIR filter module solves the filter data of the original ultrasonic signal according to the above filter coefficients, accumulates and outputs the final ultrasonic filter result dout.

[0113] Exemplarily, the present invention designs a set of 33-order dynamic FIR filters with a sampling frequency of 100 MHz and a probe bandwidth of ω c is 1MHz, the center frequency of the initial ultrasonic echo signal ω 0The input parameters of detection depth dep, operation angle phi, cutoff frequency fp and sampling frequency fs are 16-bit fixed-point data, which are implemented using an 8-stage pipeline structure. The FIFO control module and the multi-channel data switch can be used as filter coefficient reading controllers to read the 17 filter coefficients coe1 to coe17 generated by the base-4 CORDIC algorithm operation module. The distributed FIR filter module is composed of 4 small lookup tables, which solve 17 filter data for the original ultrasonic signal respectively and accumulate them to obtain the filtered ultrasonic signal.

[0114] The signal after dynamic filtering not only reflects the tissue impedance characteristics, but also carries the delay of different echo phases. The present invention adopts orthogonal demodulation method for detection, divides the signal into I and Q paths, mixes and superimposes them with local oscillator signals with a phase difference of 90°, thereby eliminating the influence of echo phase.

[0115] Specifically, the orthogonal demodulation module based on the hybrid coarse and fine adjustment CORDIC algorithm is as follows: Figure 3 It mainly includes: direct digital synthesis module, FIR low-pass filter module, mixed coarse and fine adjustment CORDIC square root operation module.

[0116] First, the direct digital synthesis module is used to mix the dynamically filtered signal with the carrier signal. The specific process is as follows: Figure 3 As shown:

[0117] The reset signal rst_n and the clock signal clk_100 are input into the direct digital synthesis module to obtain two local oscillator signals i and q; they are respectively input into the digital mixer modules mult_sin and mult_cos with the dynamically filtered data data_in, and multiplied respectively to obtain two mixed signals.

[0118] The mixing signal needs to contain two orthogonal signals. For example, a DDS signal generator can be used to generate Two local oscillator signals.

[0119] Next, the two mixed signals are processed by the FIR low-pass filter module to obtain the in-phase signal I and the quadrature signal Q. The specific operation is: the two mixed signals are used as the input values ​​din_a and din_b of the low-pass filters FIR_1 and FIR_2 respectively to obtain the in-phase signal I and the quadrature signal Q.

[0120] Exemplarily, the entire orthogonal demodulation process can be described by the following derivation:

[0121] Assume the echo signal is:

[0122] s(t)=A(t)cos[ω 0 t+φ(t)]

[0123] Where A(t),ω 0 and φ(t) are the echo signal amplitude, carrier frequency, and phase respectively. The signal obtained after sampling and filtering is:

[0124] s(nT s )=A(nT s )cos[ω 0 nT s +φ(nT s )]

[0125] =A(nT s )cosφ(nT s )cosω 0 nT s -A(nT s )sinφ(nT s )sinω 0 nT s

[0126] =A I (nT s )cosω 0 nT s -A Q (nT s )sinω 0 nT s

[0127] The sampling interval

[0128] The in-phase component is:

[0129] A I (nT s )=A(nT s )cosφ(nT s )

[0130] The orthogonal components are:

[0131] A Q (nT s )=A(nT s )sinφ(nT s )

[0132] After FIR low-pass filtering, the in-phase signal I is:

[0133] I=LPF(s(nT s )×2cos(ω 0 nT s ))

[0134] =A I cos((ω0 -ω c )nT s )-A Q sin((ω 0 -ω c )nT s )

[0135] Similarly, the orthogonal signal Q is:

[0136] Q=A I cos((ω 0 -ω c )nT s )+A Q sin((ω 0 -ω c )nT s )

[0137] Taking the modulus gives the amplitude:

[0138]

[0139] The key operation of the orthogonal demodulation module is the square root operation. The present invention obtains the amplitude value of the demodulated signal through the hybrid coarse and fine adjustment CORDIC square root operation module. The block diagram of the hybrid coarse and fine adjustment CORDIC algorithm is shown in FIG. Figure 4 As shown, it contains a coarse and fine rotation combination, which is composed of a CORDIC processor I module and a CORDIC processor II module, which respectively perform coarse rotation phase and fine rotation phase. The calculation principle and process are as follows:

[0140] The hybrid coarse and fine adjustment CORDIC algorithm decomposes the required angle θ into coarse angles θ H and subdivision angle θ L Two parts, expressed as:

[0141] θ=θ H +θ L

[0142]

[0143]

[0144] In order to improve the efficiency of hybrid CORDIC solution, a hierarchical pipeline method is adopted. H , in-phase signal I and quadrature signal Q as the starting input value x 0 and 0 Input CORDIC processor I module to perform coarse rotation phase and obtain the rough value x M and M :

[0145]

[0146] The above intermediate values ​​are used as the input values ​​of the CORDIC processor II module, and the fine rotation phase is performed to obtain the final result x n and n :

[0147]

[0148] Obviously, coarse calculations and fine calculations can be performed simultaneously, the algorithm is easy to implement, and the computational efficiency of the CORDIC algorithm is improved.

[0149] The demodulated ultrasonic signal needs to be logarithmically compressed to reduce its dynamic range. At the same time, the accuracy and speed of the algorithm implementation need to be considered, and the consumption of system resources should be minimized as much as possible.

[0150] The present invention proposes an improved extended hyperbolic CORDIC algorithm for logarithmic operations. Specifically, the logarithmic compression module based on the improved extended hyperbolic CORDIC algorithm is as follows: Figure 5 As shown, it includes a preprocessing module, a positive number iteration module, and an extension module.

[0151] The preprocessing module is used to generate an initial iteration value. For example, the input data x, that is, the demodulated ultrasonic signal, is added by 1 and subtracted by 1 to obtain the initial iteration value x. 0 ,y 0 , and expand the data by 2 16 to avoid floating point operations.

[0152] The positive iteration module adopts the iterative steps of the existing hyperbolic CORDIC algorithm. First, the input data is preprocessed, including adjusting the data quadrant and taking measures to improve the accuracy, such as multiplying by a power of 2 to expand the multiple; then, the data is input into the pipeline iteration processing module, and each iteration is expanded into a first-level pipeline unit according to the iteration formula of each level, which is mainly composed of registers, shifters and adders and subtractors; finally, the data post-processing operation is performed to complete the modulus correction and obtain the logarithm of the input data. The iterative process requires that when the iteration number is 4, 13, 40, etc., and satisfies i=3k+1, the iteration needs to be repeated.

[0153] In the hyperbolic coordinate system, the traditional CORDIC algorithm iteration formula is:

[0154]

[0155] z i+1 =z i -δ i θ i

[0156] θi =tanh -1 (2 -i )

[0157] in:

[0158]

[0159] The convergence range of the CORDIC algorithm in the hyperbolic coordinate system is:

[0160]

[0161] Obviously, in the hyperbolic coordinate system, the solution range of the traditional CORDIC algorithm is very small, which does not completely include the definition domain of the independent variable in the natural logarithm, and is of little significance in practical ultrasound applications. Therefore, it is necessary to expand the convergence domain of the traditional CORDIC.

[0162] Specifically, the improved extended hyperbolic CORDIC algorithm of the present invention can realize fast calculation of logarithms even when the input data is large by selecting appropriate extended convergence domain conditions and adopting an iterative module in which i is a negative number. In addition, the method requires fewer iteration steps, which is convenient for hardware circuit implementation.

[0163] The iterative formula of the CORDIC algorithm with extended convergence domain is:

[0164] When i>0:

[0165]

[0166] z i+1 =z i -δ i tanh -1 (2 -i )

[0167] When i<0:

[0168]

[0169] z i+1 =z i -δ i tanh -1 (1-2 i-2 )

[0170] in:

[0171]

[0172] The expanded convergence region becomes:

[0173]

[0174] in:

[0175]

[0176] The relationship between the domain Φ and θ of the natural logarithm in the extended convergence domain CORDIC algorithm is:

[0177]

[0178] From the properties of the hyperbolic tangent function, we know that Φ is an increasing function within its domain. Define Φ max is the maximum value of Φ in its domain. max The changes with M are shown in Table 1.

[0179] For example, in this ultrasound system design, the ADC acquisition accuracy is 12 bits, and 32 channels are used for data sampling. The data bit width entering the logarithmic compression link is 16 bits, that is, the maximum data volume is 65535. Considering the real-time performance of the entire system in the design, the FPGA circuit of logarithmic compression is designed with a pipeline architecture to improve system efficiency. This method is to sacrifice hardware resources in exchange for a higher working speed. In order to save system resources, M=3 is preferably used in the design.

[0180] Table 1 Φ in the natural logarithm convergence domain max With the change of M

[0181]

[0182] The formula for calculating logarithms in the extended hyperbolic CORDIC algorithm is:

[0183]

[0184] Obviously, the logarithm calculation process in the above formula also includes multiplication and division operations, which should be avoided as much as possible in hardware design.

[0185] In order to avoid floating point numbers, multiplication and division operations in solving logarithms, the extended hyperbolic CORDIC algorithm optimizes the rotation angle θ so that the logarithmic operation of the CORDIC algorithm based on the extended convergence domain can be directly performed.

[0186] The optimization steps are: first calculate the inverse tangent function value corresponding to θ under positive iteration and negative iteration respectively, then realize the conversion from floating point to fixed point, and multiply θ by 2 16 , which can be achieved by left shifting 16 bits in hardware. Specifically:

[0187] Assume z in The initial value of is 0, and after multiple iterations ln(x) = 2z n , in the extended CORDIC algorithm:

[0188]

[0189]

[0190] To ensure iteration accuracy, the improved extended hyperbolic CORDIC algorithm adopts 16 steps of forward iteration.

[0191] To achieve the conversion from floating point to integer, the optimized rotation angle θ is first multiplied by 2 16 Then round up.

[0192] Therefore, in hardware implementation, the logarithmic operation can be expressed as:

[0193]

[0194] The optimized rotation angle is expressed as:

[0195]

[0196] Assume z in The initial value is 0, and the optimized iteration expression is updated to:

[0197]

[0198] It can be seen that for z′ n , divided by 2 16 , just right shift 16 bits to get the logarithmic compressed value.

[0199] In summary, compared with the prior art, the radial artery ultrasound signal processing device and method based on the CORDIC algorithm provided in the embodiment of the present invention can produce at least the following beneficial effects by introducing a high-performance CORDIC algorithm in the filtering, demodulation and compression stages of signal processing:

[0200] 1) In the stage of filtering the original ultrasonic signal, the filter coefficients are calculated by using the base-4 CORDIC algorithm, which improves the calculation speed of trigonometric functions, reduces output delay, reduces the number of iterations, and reduces the consumption of computing resources;

[0201] 2) In the demodulation stage of the filtered signal, a hybrid CORDIC algorithm including a combination of coarse and fine rotations is used, and a hierarchical pipeline is used to perform square root operations, which reduces the difficulty of algorithm implementation and improves computational efficiency;

[0202] 3) In the stage of logarithmic compression of the demodulated ultrasonic signal, the improved extended hyperbolic CORDIC algorithm is used for logarithmic calculation. On the one hand, this method can still achieve fast logarithmic calculation when facing large input data, that is, compared with the traditional CORDIC algorithm, the algorithm described in the present invention has a larger solution range and is more suitable for actual ultrasonic signal processing; on the other hand, this method requires fewer iterations, has a fast calculation speed, and is easy to implement in hardware circuits.

[0203] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0204] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A radial artery ultrasound echo signal processing device based on CORDIC algorithm, characterized in that: include: A dynamic filtering module based on the base-4 CORDIC algorithm is used to extract the diagnostically valuable components from the original ultrasound signal and filter out the noise to obtain the filtered ultrasound signal; An orthogonal demodulation module based on a hybrid coarse and fine adjustment CORDIC algorithm is used to orthogonally demodulate a mixed signal formed by a carrier signal and a filtered signal to obtain a demodulated ultrasonic signal; wherein the hybrid coarse and fine adjustment CORDIC algorithm is used to obtain amplitude information of the demodulated ultrasonic signal; The logarithmic compression module based on the improved extended hyperbolic CORDIC algorithm is used to compress the demodulated ultrasound signal and reduce its dynamic range to the imaging range that can be presented by the display.

2. The radial artery ultrasound echo signal processing device based on the CORDIC algorithm according to claim 1, characterized in that: The dynamic filtering module based on the base-4 CORDIC algorithm includes: a frequency multiplier module, a filter coefficient generation module, and a parallel distributed FIR filtering module; among them, A frequency multiplication module is used to multiply the system clock clk_in of a specific frequency to obtain a multiplied clock clk_100 whose sampling frequency is twice the specific frequency; The frequency multiplier clock clk_100 is used as the input clock signal of the filter coefficient generation module; The filter coefficient generation module is used to calculate and read the filter coefficient, specifically including: a base-4 CORDIC algorithm operation module, a FIFO control module, and a multi-channel data switch module; The detection depth dep, operation angle phi, passband cutoff frequency fp and sampling frequency fs are input into the base-4 CORDIC algorithm operation module to complete the dynamic solution of the filter coefficient; The FIFO control module and the multi-channel data switch are used to read multiple filter coefficients generated by the radix-4 CORDIC algorithm operation module and send them to the distributed FIR filter module in sequence; The FIR filter module solves the filter data of the original ultrasonic signal according to the filter coefficient and accumulates the final filter result.

3. The radial artery ultrasound echo signal processing device based on the CORDIC algorithm according to claim 1, characterized in that: The orthogonal demodulation module based on the hybrid coarse and fine adjustment CORDIC algorithm includes: a direct digital synthesis module, a FIR low-pass filter module, and a hybrid coarse and fine adjustment CORDIC square root operation module; The direct digital synthesis module is used to mix the carrier signal with the filtered ultrasonic signal. The mixed signal contains two orthogonal signals. The generation process of the mixed signal is as follows: a. Input the reset signal rst_n and the clock signal clk_100 into the direct digital synthesis module to generate two orthogonal signals i and q; b. Input the i and q signals and the ultrasonic signals filtered by the dynamic filtering module into the digital mixing modules mult_sin and mult_cos respectively, and perform multiplication operations to obtain the sine and cosine local oscillator signals, i.e., the mixing signals; The FIR low-pass filter module is used to process the above mixed signal to obtain an in-phase signal I and an orthogonal signal Q respectively; The square sum of the I and Q signals is taken, and the signal amplitude information is obtained by taking the square root of the square sum of the I and Q signals through the mixed coarse and fine adjustment CORDIC operation module.

4. The radial artery ultrasound echo signal processing device based on the CORDIC algorithm according to claim 1, characterized in that: The logarithmic compression module based on the improved extended hyperbolic CORDIC algorithm includes: a preprocessing module, a positive iteration module, and an expansion module; among them, A preprocessing module is used to preprocess the demodulated signal data to obtain an initial iteration value; Positive iteration module, the specific steps of iteration are: a. Input the initial iteration value obtained by preprocessing into the pipeline iteration processing module; b. According to the iteration formula of each level, each iteration is expanded into a pipeline unit of one level, including registers, shifters, and adders and subtractors; c. Perform data post-processing operations, complete module correction, and obtain the logarithmic value of the input data; The extension module expands the logarithmic convergence domain of the CORDIC algorithm by adding negative number iterations, and is used to quickly calculate input data beyond the original convergence domain.

5. The radial artery ultrasound echo signal processing device based on the CORDIC algorithm according to claim 1 or 2, characterized in that: The filter coefficient is designed using a Hamming window, and the filter coefficient h(l) is calculated as follows: where ω c is the probe frequency, N is the filter order, R N (l) is a rectangular window:

6. The radial artery ultrasound echo signal processing device based on the CORDIC algorithm according to claim 5, characterized in that: The base-4 CORDIC algorithm operation module for calculating the filter coefficient takes 4 as the base, and the iteration formula is: where σ i ∈{-2,-1,0,1,2}.

7. The radial artery ultrasound echo signal processing device based on the CORDIC algorithm according to claim 3, characterized in that: The hybrid coarse and fine tuning CORDIC algorithm decomposes the rotation angle θ into coarse angles θ H and subdivision angle θ L Two parts, expressed as: θ=θ H +θ L The hybrid CORDIC algorithm is a combination of coarse and fine rotations. It uses a hierarchical pipeline method, with the CORDIC processor I module and the CORDIC processor II module performing coarse rotation phase and fine rotation phase on the required angle respectively. The specific steps are: a. Roughly divide the angle θ H The in-phase signal I and the quadrature signal Q are respectively used as the starting input values ​​x0 and y0 to be input into the CORDIC processor I module to perform a coarse rotation phase and obtain a coarse value x M and M , the execution operation is: b. Subdividing the angle θ H And the rough value x obtained in the previous step M and M As the input value of the CORDIC processor II module, a fine rotation phase is performed to obtain the final result x n and n , the execution operation is:

8. The radial artery ultrasound echo signal processing device based on the CORDIC algorithm according to claim 4, characterized in that: The iterative formula of the CORDIC algorithm with extended convergence domain is: When i>0: z i+1 =z i -δ i tanh -1 (2 -i ) When i<0: z i+1 =z i -δ i tanh -1 (1-2 i-2 ) In vector mode: The expanded convergence region becomes: in:

9. The radial artery ultrasound echo signal processing device based on the CORDIC algorithm according to claim 8, characterized in that: The extended hyperbolic CORDIC algorithm can directly perform logarithmic operations based on the extended convergence domain of the CORDIC algorithm after optimizing the rotation angle as described below. The optimization steps are as follows: Calculate the inverse tangent function value of θ under positive iteration and negative iteration respectively, then convert the floating point number to a fixed point number and multiply θ by 2 16 .

10. A radial artery ultrasound echo signal processing method based on CORDIC algorithm, characterized in that: The following steps are involved: Step S1, firstly, the ultrasound signal is A / D converted, and the obtained digital ultrasound signal is transmitted to the entire radial artery ultrasound signal processing device under the control of the global clock; Step S2: using a radial artery ultrasound signal processing device to process the input digital ultrasound signal to complete subsequent ultrasound imaging; The process of processing the input digital ultrasonic signal by using the radial artery ultrasonic signal processing device comprises the following steps: Step S201, using a dynamic filtering module based on a radix-4 CORDIC algorithm, extracting components with diagnostic value from the signal and filtering out noise to obtain a filtered ultrasonic signal; Step S202, mixing the filtered ultrasonic signal generated in the previous step with the carrier signal, and orthogonally demodulating the mixed signal using an orthogonal demodulation module based on a hybrid coarse and fine adjustment CORDIC algorithm to obtain a demodulated ultrasonic signal; wherein the hybrid coarse and fine adjustment CORDIC algorithm is used to obtain amplitude information of the demodulated ultrasonic signal; Step S203: The demodulated ultrasonic signal generated in the previous step is compressed by a logarithmic compression module of an improved extended hyperbolic CORDIC algorithm to reduce the signal dynamic range to an imaging range that can be presented by the display.

Images