KR20200000909A - Apparatus and method for suppressing reverberation signal - Google Patents

Abstract

According to one embodiment of the present invention, provided is an apparatus for removing a reverberation signal which comprises: a reception unit receiving a reception signal including an echo signal in which a radiation signal is reflected from a target object and a reverberation signal reflected from an object other than the target object; a spectrogram conversion unit performing Fourier transform on the reception signal to obtain a magnitude spectrogram of the reception signal; a base matrix estimation unit determining a frequency base matrix of the echo signal based on the radiation signal and estimating a time base matrix of the echo signal from the magnitude spectrogram of the reception signal using the determined frequency base matrix of the echo signal; and a reconstruction unit reconstructing the echo signal using the frequency base matrix of the echo signal and the time base matrix of the estimated echo signal.

Description

Reverberation signal removal device and method {APPARATUS AND METHOD FOR SUPPRESSING REVERBERATION SIGNAL}

The present invention relates to a reverberation signal removal device and method for reconstructing an echo signal by removing the reverberation signal among the signals received by the active sonar system.

SONAR (Sound Navigation and Ranging) refers to the method of detection based on various kinds of noise sources emitted from underwater targets. The sonar is divided into an active sonar and a passive sonar, depending on whether a radiation signal is generated.

An active sonar system refers to a system that detects a target by emitting a signal underwater and receiving the reflected sound. The active sonar system may receive a sound signal using a sensor and a beamformer, and then analyze the signal to estimate the distance to the target and the Doppler frequency.

It is important to remove unwanted signals, or noise, to get accurate reflections from the target using active sonar systems. In particular, the reverberation signal generated by the reflected sound reflected by the surrounding environment is similar to the reflection signal reflected by the target, and thus may be an obstacle to accurate target detection.

Korean Registered Patent 10-1558438 (October 07, 2015)

SUMMARY OF THE INVENTION An object of the present invention is to provide a reverberation signal removing device and method for reconstructing an echo signal using a frequency basis matrix of an echo signal and a time basis matrix of an echo signal.

However, the problem to be solved by the present invention is not limited to the above-mentioned, it is not mentioned but includes the purpose that can be clearly understood by those skilled in the art from the following description. can do.

In accordance with an aspect of the present invention, an apparatus for removing reverberation signal may include: a receiver configured to receive a reception signal including an echo signal reflected from an object and an echo signal reflected from an object other than the object; A spectrogram converter for performing Fourier transform on the received signal to obtain a magnitude spectrogram of the received signal; A base matrix estimator for determining a frequency basis matrix of the echo signal based on the radiated signal and estimating a time basis matrix of the echo signal from the magnitude spectrogram of the received signal using the determined frequency basis matrix of the echo signal ; And a reconstruction unit for restoring the echo signal using the frequency basis matrix of the echo signal and the time basis matrix of the estimated echo signal.

The base matrix estimator may estimate a time basis matrix of the echo signal based on the prediction length of the echo signal and the energy of the echo signal.

The base matrix estimator may further include a time basis matrix of the echo signal, a time basis matrix of the reverberation signal, and a frequency basis matrix of the reverberation signal from the magnitude spectrogram of the received signal using the determined frequency basis matrix of the echo signal. S may be sequentially estimated, and the time base matrix of the echo signal may be estimated by repeating the estimation process a predetermined number of times.

The base matrix estimator may determine the frequency basis matrix of the echo signal from the magnitude spectrogram of the radiated signal.

In addition, the basis matrix estimator may be configured to estimate a frequency basis vector of the radiation signal from the magnitude spectrogram of the radiation signal, wherein the frequency basis vector of the estimated radiation signal is composed of a plurality of vectors in parallel translation in the frequency axis direction. The frequency basis matrix of the echo signal can be determined.

The reconstruction unit may reconstruct the echo signal based on the row vector having the maximum energy among the temporal basis matrices of the echo signal and the column vector of the frequency basis matrix of the echo signal corresponding to the row vector.

The reconstructor may obtain a magnitude spectrogram of the echo signal using a row vector having the maximum energy among the temporal basis matrices of the echo signal and a column vector of the frequency basis matrix of the echo signal corresponding to the row vector. The echo signal may be restored by adding phase information of the received signal to the magnitude spectrogram of the acquired echo signal.

According to an embodiment of the present invention, a method for removing reverberation signal may include: receiving a received signal including an echo signal reflected from an object and a reverberation signal reflected from an object other than the object; Fourier transforming the received signal to obtain a magnitude spectrogram of the received signal; Determining a frequency basis matrix of the echo signal based on the radiated signal; Estimating a time basis matrix of the echo signal from the magnitude spectrogram of the received signal using the determined frequency basis matrix of the echo signal; And reconstructing the echo signal using the frequency basis matrix of the echo signal and the time basis matrix of the estimated echo signal.

According to an embodiment of the present invention, a computer-readable recording medium includes: receiving a received signal including an echo signal reflected from an object and a reverberation signal reflected from an object other than the object; Fourier transforming the received signal to obtain a magnitude spectrogram of the received signal; Determining a frequency basis matrix of the echo signal based on the radiated signal; Estimating a time basis matrix of the echo signal from the magnitude spectrogram of the received signal using the determined frequency basis matrix of the echo signal; And reconstructing the echo signal using the frequency basis matrix of the echo signal and the time basis matrix of the estimated echo signal.

According to the embodiment of the present invention, since the reverberation signal from which the reverberation signal is removed from the received signal can be restored, it is possible to provide a more accurate target detection environment in the field of underwater target detection including an active sonar.

Effects obtained in the present invention are not limited to the above-mentioned effects, and other effects not mentioned above may be clearly understood by those skilled in the art from the following description. will be.

1A is a diagram illustrating an ideal underwater target detection environment, and FIG. 1B is a diagram illustrating an underwater target detection environment in which a plurality of reflectors exist.
2 is a control block diagram of an apparatus for removing reverberation signal according to an embodiment of the present invention.
3 is a view for explaining the operation of the spectrogram converter according to an embodiment of the present invention.
4A illustrates a magnitude spectrogram of a received signal according to an embodiment of the present invention, and FIG. 4B illustrates a magnitude spectrogram of an echo signal according to an embodiment of the present invention, and FIG. 4C illustrates an embodiment of the present invention. A diagram illustrating magnitude spectrograms of a reverberation signal, according to an exemplary embodiment.
5 is a view for explaining the operation of the restoration unit according to an embodiment of the present invention.
6 is a flowchart of a reverberation signal removing method according to an embodiment of the present invention.

Advantages and features of the present invention, and methods for achieving them will become apparent with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but can be implemented in various forms, only the embodiments are to make the disclosure of the present invention complete, and those skilled in the art to which the present invention pertains. It is provided to fully inform the scope of the invention, and the scope of the invention is defined only by the claims.

In describing the embodiments of the present invention, detailed descriptions of well-known functions or configurations will be omitted unless they are actually necessary in describing the embodiments of the present invention. In addition, terms to be described below are terms defined in consideration of functions in the embodiments of the present invention, which may vary according to intentions or customs of users and operators. Therefore, the definition should be made based on the contents throughout the specification.

Used below '… Wealth, The term 'herein' refers to a unit for processing at least one function or operation, which may be implemented by hardware or software, or a combination of hardware and software.

1A is a diagram illustrating an ideal underwater target (T) detection environment, and FIG. 1B is a diagram illustrating an underwater target (T) detection environment in which a plurality of reflectors exist.

 The active sonar system SN which emits a radiation signal in the water, which is a kind of an acoustic signal, may detect the target T by receiving the echo signal reflected by the target T and analyzing the reflected signal. Referring to FIG. 1A, the active sonar system SN traveling in the d direction according to the speed V emits a radiation signal toward the target T, and correspondingly, the echo signal reflected by the target T is the active sonar. It may be provided to the system SN. The active sonar system SN may detect the target T by checking the Doppler frequency and the distance to the target T based on the received echo signal.

However, in a real underwater environment, in addition to the target T, there may be a plurality of scatterers S _L reflecting the acoustic signal. As shown in FIG. 1B, a plurality of scatterers S _L may exist around the active sonar system SN.

As a result, the active sonar system SN in the actual underwater environment can receive the acoustic signal reflected by the plurality of scatterers S _L , that is, the reverberation signal together with the echo signal reflected by the target T. . In this case, the active sonar system SN analyzing the received signal may fail to detect the target T under the influence of the reverberation signal, or may mistake the scatterer S _L as the target T.

Accordingly, the active sonar system SN needs to remove the reverberation signal from the received signal and restore the echo signal for accurate target T detection.

Hereinafter, the reverberation signal removal apparatus 1 provided in the active sonar system SN will be described in detail with reference to FIG. 2.

2 is a control block diagram of an apparatus for removing reverberation signal according to an embodiment of the present invention.

Referring to FIG. 2, an apparatus for removing reverberation signal 1 according to an embodiment includes: a receiver 100 receiving a received signal corresponding to a radiation signal; A spectrogram converter 200 for converting a received signal into a spectrogram; A base matrix estimator 300 for decomposing the magnitude spectrogram of the transformed received signal into a time basis matrix and a frequency basis matrix; And a reconstruction unit 400 for reconstructing the echo signal using the time basis matrix of the echo signal and the frequency basis matrix of the echo signal.

The receiver 100 may receive a reception signal corresponding to a radiation signal emitted from the active sonar system SN. In detail, the receiver 100 may receive a reception signal including an echo signal reflected from an object and an echo signal reflected from an object other than the object, for example, a scatterer S _L. The received signal x (t) received by the receiver 100 may be defined as in Equation 1.

[Equation 1]

Here, s (t) may be an echo signal reflected from the target T, r (t) may be a reverberation signal reflected from the surrounding environment, and n (t) may mean a background noise signal. In this case, the echo signal s (t) may be modeled by Equation 2.

[Equation 2]

Here, c (t) may mean a radiated signal emitted from the active sonar system SN, and f _d may mean a Doppler frequency. At this time, the radiation signal c (t) may be defined by equation (3).

[Equation 3]

Here, f ₀ may mean the center frequency of the radiation signal.

In order to receive the reception signal defined by Equations 1 to 3, the receiver 100 may be provided with a sensor array in which a plurality of sensors are arranged at equal intervals. The receiver 100 may receive a received signal through each of the plurality of sensors constituting the sensor array. Here, the sensor constituting the sensor array is a sensor that detects an acoustic signal or a radio wave signal and outputs an electrical signal corresponding thereto, and may include a microphone that detects an acoustic signal and / or an antenna that detects a radio wave signal.

The spectrogram converter 200 may convert the received signal received by the receiver 100 into a spectrogram. Here, the spectrogram means to represent a signal in the time-frequency domain. Hereinafter, the operation of the spectrogram converter 200 will be described with reference to FIGS. 3 and 4A, 4B, and 4C.

3 is a view for explaining the operation of the spectrogram converter according to an embodiment of the present invention, Figure 4a is a diagram showing the magnitude spectrogram of the received signal according to an embodiment of the present invention, Figure 4b is a view of the present invention FIG. 4C is a diagram illustrating magnitude spectrograms of echo signals according to an embodiment of the present invention, and FIG. 4C is a diagram illustrating magnitude spectrograms of reverberation signals according to an embodiment of the present invention.

로 표현될 수 있다. Referring to FIG. 3, the spectrogram converter 200 may obtain a spectrogram of the received signal by performing a short time Fourier transform (STFT) on the received signal received by the receiver 100. In detail, the spectrogram converter 200 performs a short-time Fourier transform by Fast Fourier Transform (FFT) every certain section along the time axis, and arranges the spectrogram of the received signal by arranging the result on the time axis. Can be obtained. At this time, the spectrogram of the received signal obtained is a complex matrix

It can be expressed as.

로부터 수신 신호의 크기 스펙트로그램 X를 획득할 수 있고, 일 실시예에 따른 수신 신호의 크기 스펙트로그램은 도 4a와 같다. 도 4a에서 파란 원은 반향 신호 정보가 포함되는 영역으로, 일정 시간 구간 내에서만 나타날 수 있다.In addition, the spectrogram converter 200 is a spectrogram of the received signal

The magnitude spectrogram X of the received signal can be obtained from FIG. 4A. The magnitude spectrogram of the received signal is shown in FIG. 4A. In FIG. 4A, a blue circle includes an echo signal information and may appear only within a predetermined time interval.

In addition, the magnitude spectrogram X of the received signal x (t) defined by Equation 1 may be defined by Equation 4.

[Equation 4]

Here, S represents the magnitude spectrogram of the echo signal s (t), R represents the magnitude spectrogram of the reverberation signal r (t), and N represents the magnitude spectrogram of the background noise signal n (t). In this case, when the number of frequency bins is K, and the number of time frames is A, the matrixes X, S, R, and N may each have a size of K × A.

As such, the magnitude spectrogram X of the received signal may include the magnitude spectrogram R of the reverberation signal of FIG. 4C in addition to the magnitude spectrogram S of the echo signal of FIG. 4B. Therefore, it is necessary to restore the magnitude spectrogram S of the echo signal by removing the magnitude spectrogram R of the reverberation signal from the magnitude spectrogram X of the received signal.

To this end, the base matrix estimator 300 may estimate the time basis matrix of the echo signal and the time basis matrix of the reverberation signal from the magnitude spectrogram of the received signal. The base matrix estimator 300 estimates the temporal basis matrix of the echo signal and the temporal basis matrix of the reverberation signal from the magnitude spectrogram of the received signal by using a nonnegative matrix decomposition technique.

Here, the non-negative matrix decomposition technique is a technique of estimating and decomposing each matrix after representing a non-negative matrix of size K × A as a product of a non-negative matrix of size K × B and a non-negative matrix of size B × A. The base matrix estimator 300 applies the magnitude spectrum X of the received signal as a non-negative matrix of size K × A, the frequency basis matrix W and the non-negative matrix of size B × A as the non-negative matrix of size K × B. We can set the time basis matrix H as.

Prior to this, the base matrix estimator 300 may determine the frequency basis matrix W _p of the echo signal as an initial value. Since the echo signal has the same frequency structure as the radiation signal, the base matrix estimator 300 determines the frequency basis matrix W _P of the echo signal by obtaining the frequency basis vector w _P and the time basis vector h _P of the radiation signal. Can be. The basis matrix estimator 300 may obtain a frequency basis vector w _P (K × 1 column vector) of the radiation signal and a time basis vector h _P (1 × A row vector) of the radiation signal according to Equation 5.

[Equation 5]

Here, P may mean a magnitude spectrogram (K × A matrix) of the radiation signal, and '. ×' may mean a multiplication between matrix elements in the Hadmard product.

The base matrix estimator 300 repeats Equation 5 a predetermined number of times so that the product of w _P and h _P is close to _P, and obtains the frequency basis vector w _P of the radiated signal, from which the frequency of the echo signal is obtained. We can determine the base matrix W _P. In detail, the base matrix estimator 300 may determine the frequency basis matrix W _P of the echo signal according to Equation 6.

[Equation 6]

은 방사 신호의 주파수 기저 벡터 w_P를 주파수 축 방향으로 r 만큼 평행 이동시킨 값을 의미할 수 있다.here,

May mean a value obtained by moving the frequency basis vector w _P of the radiation signal in parallel in the frequency axis direction by r.

Since the Doppler effect may occur between the radiated signal and the received signal when the active sonar system SN moves, the base matrix estimator 300 moves the frequency basis vector w _P of the radiated signal in parallel in the frequency axis direction by r. The frequency basis matrix W _P of the echo signal can be constructed.

When the frequency basis matrix W _P of the echo signal is determined, the base matrix estimator 300 determines the time basis matrix H _P of the echo signal and the time basis matrix H _R of the reverberation signal, and the reverberation signal from the magnitude spectrogram X of the received signal. The frequency basis matrix W _R can be estimated. The base matrix estimator 300 may sequentially estimate each base matrix, that is, the frequency base matrix W and the time basis matrix H, in a nonnegative matrix decomposition model by a multiplicative update method. have.

In detail, the base matrix estimator 300 may estimate the time basis matrix H _P of the echo signal according to Equation (7).

[Equation 7]

Where W _P is the frequency basis matrix of the echo signal, X is the magnitude spectrum of the received signal, h _P (r, n) is the element of the time basis matrix H _P of the echo signal, and N is It may mean the number of time frames of data, and α may be any parameter value. In addition, '. ×' is a Hadmard product and may mean multiplication between matrix elements.

Referring to Equation (7), means for determining the portion closer to the W H _{P P} H _P of X and multiplication according to the update method by the left side of the expression is the multiplication parameter α. On the other hand, the expression on the right side of the parameter α is a limiting condition for spacing the energy distribution of H _P , and means a part of minimizing the base number by reducing the value of the activated element.

In this case, since the reverberation signal has a limitation in length compared to the reverberation signal, the base matrix estimator 300 may apply a time restriction condition to h _P (r, n) which is an element of the time basis matrix H _P of the reverberation signal. Can be. To obtain h _P (r, n) according to the time constraints is given by Equation 8.

[Equation 8]

Here, n denotes a frame number, and l _P denotes a prediction length of an echo signal, which may be determined by an external input or an internal operation.

Referring to Equation 8, the base matrix estimator 300 may calculate a moving average on the time axis of each estimated time basis vector according to Equation (1). Next, the base matrix estimator 300 may find a frame at which the moving average becomes the maximum according to equation (2). After finding the frame with the maximum moving average, the basis matrix estimator 300 calculates h _c (r, n) in which h other than the frame is 0 in h _P (r, n) according to equation (3). Can be obtained. Finally, the base matrix estimator 300 may obtain h _P (r, n) through attenuation of the remaining frames except the frame having the maximum energy, according to Equation (4). Here, γ may mean an arbitrary parameter that attenuates as close to 0 as the frame other than the maximum energy does not attenuate.

Next, the base matrix estimator 300 may estimate the time basis matrix H _R of the reverberation signal according to Equation (9).

[Equation 9]

Here, W _R means the frequency basis matrix of the reverberation signal, X means the magnitude spectrum of the received signal, h _R (r, n) means the element of the time basis matrix H _R of the reverberation signal, N is It may mean the number of time frames of data, and α may be any parameter value. In addition, '. ×' is a Hadmard product and may mean multiplication between matrix elements.

Referring to Equation (9) means a part for obtaining a H _R closer to the multiplication of X and W _R H _R in accordance with the update method of the left side of the expression is the multiplication parameter α. On the other hand, the expression on the right side of the parameter α is a constraint condition for spacing the energy distribution of H _R , and means a part of minimizing the base number by reducing the value of the activated element.

In addition, the base matrix estimator 300 may estimate the frequency basis matrix W _R of the reverberation signal according to Equation 10.

[Equation 10]

Here, H _R may mean a time basis matrix of the reverberation signal, and X may mean a magnitude spectrum of the received signal.

Referring to Equation 10, the base matrix estimator 300 may obtain W _R , which is a product of W _R and H _R close to X according to an update method by multiplication.

The base matrix estimator 300 repeatedly performs the above Equations 7 to 10 for a predetermined number of times, thereby updating and estimating the echo time basis matrix H _P , the reverberation time basis matrix H _R , and the reverberation frequency basis matrix W _R. can do. In contrast, the base matrix estimator 300 calculates the above-described equations 7 to 10 until the echo time basis matrix H _P , the reverberation time basis matrix H _R , and the reverberation frequency basis matrix W _R converge to a predetermined convergence value. It may be performed repeatedly.

When the estimation of the echo time base matrix H _P is completed by the basis matrix estimator 300, the reconstructor 400 estimates the frequency basis matrix W _P of the echo signal predetermined as an estimated echo time basis matrix H _P and an initial value. Can be used to recover the echo signal.

To this end, the reconstruction unit 400 may first reconstruct the magnitude spectrogram S of the echo signal according to Equation (11).

[Equation 11]

Here, r _max means the row having the largest energy among the time basis matrix H _P of the echo signal, and h (r _max ) represents the r _max th which is the basis vector having the largest energy among the time basis matrix H _P of the echo signal. The row vector, w _P (r _max ) may refer to the r _max th column vector, which is the basis vector having the largest energy among the frequency basis matrix W _P of the echo signal.

중 위상 값을 반향 신호의 크기 스펙트로그램 S에 적용하여 반향 신호의 스펙트로그램

를 획득할 수 있으며, 이는 수학식 12를 따른다.The magnitude spectrogram S of the echo signal thus obtained does not include phase information of the echo signal. Therefore, the restoration unit 400 is a spectrogram of the received signal

Spectrogram of echo signal by applying middle phase value to magnitude spectrogram S of echo signal

It can be obtained, which follows the equation (12).

[Equation 12]

는 수신 신호의 스펙트로그램을 의미한다. 또한, *는 행렬의 원소 단위 곱셈 연산자를 의미한다.Here, S means the magnitude spectrogram of the echo signal,

Means the spectrogram of the received signal. In addition, * means the element multiplication operator of the matrix.

로부터 반향 신호 s(t)를 복원할 수 있다. 이하에서는 도 5를 참조하여 반향 신호의 스펙트로그램

로부터 반향 신호 s(t)를 복원하는 방법을 설명한다.Finally, the restoration unit 400 spectrogram of the acquired echo signal

The echo signal s (t) can be recovered from. Hereinafter, the spectrogram of the echo signal with reference to FIG.

A method of recovering the echo signal s (t) from the following will be described.

5 is a view for explaining the operation of the restoration unit according to an embodiment of the present invention.

을 역(Inverse) 단시간 푸리에 변환하여 시간 영역에서의 반향 신호 s(t)를 복원할 수 있다. 구체적으로, 복원부(400)는 반향 신호의 스펙트로그램

의 시간 축을 따라 일정 구간마다 역(Inverse) 고속 푸리에 변환함으로써 역(Inverse) 단시간 푸리에 변환을 수행하고, 그 결과를 시간 축 상에 배열함으로써 시간 영역에서의 반향 신호 s(t)를 복원할 수 있다. Referring to FIG. 5, the reconstruction unit 400 may spectrogram the acquired echo signal.

Inverse short-time Fourier transform may be used to restore the echo signal s (t) in the time domain. Specifically, the restoration unit 400 is a spectrogram of the echo signal

The inverse short-time Fourier transform is performed by inverse fast Fourier transforms at predetermined intervals along the time axis of, and the echo signal s (t) in the time domain can be restored by arranging the result on the time axis. .

As a result, the reconstructed echo signal s (t) may mean the result of the reverberation signal r (t) being removed from the received signal x (t). The active sonar system SN may enable more accurate detection of the target T by using the echo signal s (t) from which the reverberation signal r (t) is removed.

6 is a flowchart of a reverberation signal removing method according to an embodiment of the present invention.

First, the reverberation signal removing apparatus 1 may receive a reception signal including an echo signal and a reverberation signal (S100). Here, the echo signal may mean an acoustic signal in which the radiation signal is reflected from the object, and the reverberation signal may mean an acoustic signal in which the radiation signal is reflected from an object other than the object.

Next, the reverberation signal removing apparatus 1 may obtain a magnitude spectrogram of the received signal (S110). Specifically, the reverberation signal removing apparatus 1 may obtain a spectrogram representing the time-frequency domain by short-time Fourier transforming the received signal, and obtain the magnitude spectrogram of the received signal including only the magnitude component from the obtained spectrogram. Can be obtained separately.

Apart from the above-described steps, the reverberation signal removing apparatus 1 may determine the frequency basis matrix of the echo signal (S120). Since the echo signal has the same frequency structure as the radiation signal, the reverberation signal removing apparatus 1 may determine the frequency basis matrix of the echo signal by obtaining the frequency basis vector of the radiation signal and the time basis vector of the radiation signal.

When the magnitude spectrogram of the received signal is obtained and the frequency basis matrix of the echo signal is determined, the reverberation signal removing apparatus 1 may estimate the time basis matrix of the echo signal from the magnitude spectrogram of the received signal (S130). To this end, the reverberation signal removing apparatus 1 may apply a non-negative matrix decomposition technique. In particular, the reverberation signal removal apparatus 1 may use the predicted length of the echo signal and the energy of the echo signal to estimate the time basis matrix of the echo signal.

Finally, the reverberation signal removing apparatus 1 may restore the echo signal using the frequency basis matrix of the echo signal and the time basis matrix of the estimated echo signal (S140). In detail, the reverberation signal removing apparatus 1 may restore the echo signal based on the column vector having the maximum energy among the temporal basis matrices of the echo signal and the row vector of the frequency basis matrix of the echo signal corresponding to the column vector.

The reverberation signal removing apparatus 1 and the method described above can restore the reverberation signal from which the reverberation signal has been removed from the received signal, and thus have a more accurate target detection environment in the underwater target (T) detection field including an active sonar. Can be provided.

Combinations of each block of the block diagrams and each step of the flowcharts attached herein may be performed by computer program instructions. These computer program instructions may be mounted on a processor of a general purpose computer, special purpose computer, or other programmable data processing equipment such that instructions executed through the processor of the computer or other programmable data processing equipment may not be included in each block or flowchart of the block diagram. It will create means for performing the functions described in each step. These computer program instructions may be stored in a computer usable or computer readable memory that can be directed to a computer or other programmable data processing equipment to implement functionality in a particular manner, and thus the computer usable or computer readable memory. It is also possible for the instructions stored in to produce an article of manufacture containing instruction means for performing the functions described in each block or flowchart of each step of the block diagram. Computer program instructions may also be mounted on a computer or other programmable data processing equipment, such that a series of operating steps may be performed on the computer or other programmable data processing equipment to create a computer-implemented process to create a computer or other programmable data. Instructions that perform processing equipment may also provide steps for performing the functions described in each block of the block diagram and in each step of the flowchart.

In addition, each block or step may represent a portion of a module, segment or code that includes one or more executable instructions for executing a specified logical function (s). It should also be noted that in some alternative embodiments, the functions noted in the blocks or steps may occur out of order. For example, the two blocks or steps shown in succession may in fact be executed substantially concurrently or the blocks or steps may sometimes be performed in the reverse order, depending on the functionality involved.

The above description is merely illustrative of the technical idea of the present invention, and those skilled in the art to which the present invention pertains may make various modifications and changes without departing from the essential quality of the present invention. Therefore, the embodiments disclosed herein are not intended to limit the technical spirit of the present invention but to explain, and the scope of the technical spirit of the present invention is not limited by these embodiments. The scope of protection of the present invention should be interpreted by the following claims, and all technical ideas that fall within the scope of equivalents thereof should be construed as being included in the scope of the present invention.

1: Reverberation signal canceller
100: receiver
200: spectrogram converter
300: base matrix estimator
400: restoration unit

Claims (9)

A receiver configured to receive a received signal including an echo signal reflected from an object and a reverberation signal reflected from an object other than the object;
A spectrogram converter for performing Fourier transform on the received signal to obtain a magnitude spectrogram of the received signal;
A base matrix estimator for determining a frequency basis matrix of the echo signal based on the radiated signal and estimating a time basis matrix of the echo signal from the magnitude spectrogram of the received signal using the determined frequency basis matrix of the echo signal ; And
And a reconstruction unit for recovering the echo signal using the frequency basis matrix of the echo signal and the estimated time basis matrix of the echo signal. The method of claim 1,
The basis matrix estimator,
And estimating a time basis matrix of the echo signal based on the predicted length of the echo signal and the energy of the echo signal. The method of claim 1,
The basis matrix estimator,
The time basis matrix of the echo signal, the time basis matrix of the reverberation signal, and the frequency basis matrix of the reverberation signal are sequentially estimated from the magnitude spectrogram of the received signal by using the determined frequency basis matrix of the echo signal.
And estimating the temporal basis of the echo signal by repeating the estimation process a predetermined number of times. The method of claim 1,
The basis matrix estimator,
And a reverberation signal canceller for determining the frequency basis matrix of the echo signal from the magnitude spectrogram of the radiated signal. The method of claim 4, wherein
The basis matrix estimator,
Estimate a frequency basis vector of the radiation signal from the magnitude spectrogram of the radiation signal, and determine a frequency basis matrix of the echo signal composed of a plurality of vectors in which the frequency basis vector of the estimated radiation signal is parallelly shifted in the frequency axis direction Reverberation signal cancellation device. The method of claim 1,
The restoration unit,
And reverberating the echo signal based on the row vector having the maximum energy among the temporal basis matrices of the echo signal and the column vector of the frequency basis matrix of the echo signal corresponding to the row vector. The method of claim 6,
The restoration unit,
The magnitude spectrogram of the echo signal is obtained using a row vector having the maximum energy among the temporal basis matrices of the echo signal and a column vector of the frequency basis matrix of the echo signal corresponding to the row vector, and the magnitude of the obtained echo signal is obtained. The reverberation signal removing device restores the echo signal by adding phase information of the received signal to a spectrogram. Receiving a received signal including an echo signal reflected from an object and a reverberation signal reflected from an object other than the object;
Fourier transforming the received signal to obtain a magnitude spectrogram of the received signal;
Determining a frequency basis matrix of the echo signal based on the radiated signal;
Estimating a time basis matrix of the echo signal from the magnitude spectrogram of the received signal using the determined frequency basis matrix of the echo signal; And
And reconstructing the echo signal using the frequency basis matrix of the echo signal and the time basis matrix of the estimated echo signal. Receiving a received signal including an echo signal reflected from an object and a reverberation signal reflected from an object other than the object;
Fourier transforming the received signal to obtain a magnitude spectrogram of the received signal;
Determining a frequency basis matrix of the echo signal based on the radiated signal;
Estimating a time basis matrix of the echo signal from the magnitude spectrogram of the received signal using the determined frequency basis matrix of the echo signal; And
And recovering the echo signal using the frequency basis matrix of the echo signal and the temporal basis matrix of the estimated echo signal.

Images