JP2016197775A - Signal processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve the problem in which it is difficult to suppress only pulse noises contaminated on a transmission line from a signal input from a sensor through the transmission line.SOLUTION: A signal processor includes a detection section and a suppression section. The detection section analyses a signal input from a sensor through a transmission line and detects pulse noises appeared outside an output frequency band of the sensor. The suppression section suppresses the pulse noises detected by the detection section from the signal.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a signal processing device, a signal processing method, and a program.

  Various proposals or practical use have been made of signal processing techniques for suppressing a part or all of noise from a signal in which a desired signal and noise are mixed and outputting a signal in which the desired signal is emphasized.

  For example, it has been proposed as a first related technique related to the present invention to estimate noise from an input signal and subtract the estimated noise from the input signal (see, for example, Patent Document 1). Also, the input signal is analyzed, noise information about the noise to be suppressed is mixed according to the analysis result of the input signal to generate mixed noise information, and noise is suppressed from the input signal using this mixed noise information Has been proposed as a second related technique related to the present invention (see, for example, Patent Document 2).

  On the other hand, in the passive sonar device, a signal processing method capable of detecting a transient signal even when a noise source having a large power is present outside the frequency band of the transient signal has been proposed as a third related technique related to the present invention ( For example, see Patent Document 3). However, this third related technique is a technique for extracting a desired transient signal from a signal input from an acoustic sensor, and is different from a technique for suppressing noise from an input signal. That is, the third related technique does not generate an input signal in which noise is suppressed. For this reason, a desired signal other than the transient signal included in the input signal, for example, a stationary sound that continues for a long time using an engine or the like as a source, is discarded in the process of signal processing. Moreover, an acoustic sensor capable of calculating the arrival direction of the acoustic signal is required.

JP-A-8-2221092 International publication WO2011 / 148860 JP 2014-32081 A Japanese Patent Laid-Open No. 10-177066

  By the way, when a signal detected by a sensor such as an acoustic sensor is input to the signal processing device through a wired or wireless transmission path, a new signal may be mixed as noise on the transmission path. Specifically, pulse noise such as static noise may be mixed. In order to correctly process the signal detected by the sensor, it is necessary to suppress pulse noise mixed on the transmission path. However, in the first and second related techniques related to the present invention, since it is impossible to determine which signal component of the input signal is a signal component mixed on the transmission line, the input signal mixed on the transmission line. It was difficult to suppress only the pulse noise.

  An object of the present invention is to provide a signal processing apparatus that solves the above-described problem, that is, it is difficult to suppress only pulse noise mixed on a transmission line from a signal input from a sensor through the transmission line. There is.

A signal processing apparatus according to an embodiment of the present invention includes:
Analyzing a signal input from a sensor through a transmission path, detecting a pulse noise appearing outside the output frequency band of the sensor; and
A suppressor that suppresses the detected pulse noise from the signal;
Have

A signal processing method according to another embodiment of the present invention includes:
Analyzing the signal input from the sensor through the transmission path, detecting pulse noise appearing outside the output frequency band of the sensor,
The detected pulse noise is suppressed from the signal.

A program according to another embodiment of the present invention is:
Computer
Analyzing a signal input from a sensor through a transmission path, detecting a pulse noise appearing outside the output frequency band of the sensor; and
A suppressor that suppresses the detected pulse noise from the signal;
And make it work.

  Since the present invention has the above-described configuration, it is possible to suppress only the pulse noise mixed on the transmission path from the signal input from the sensor through the transmission path.

1 is a block diagram of a signal processing device according to a first embodiment of the present invention. It is a block diagram which shows an example of the detection part in the signal processing apparatus which concerns on the 1st Embodiment of this invention. It is a block diagram which shows an example of the suppression part in the signal processing apparatus which concerns on the 1st Embodiment of this invention. It is a block diagram of the passive sonar apparatus which concerns on the 2nd Embodiment of this invention. It is an image figure of the sensitivity characteristic of the sensor which concerns on the noise determination process of the 2nd Embodiment of this invention, a signal, and the zone | band of a noise. It is an image figure of the process result of the noise determination process of the 2nd Embodiment of this invention.

Next, embodiments of the present invention will be described in detail with reference to the drawings.
[First embodiment]
Referring to FIG. 1, a signal processing apparatus 100 according to the first embodiment of the present invention is connected to a sensor 102 through a transmission path 101.

  The sensor 102 is, for example, an acoustic sensor when the signal processing device 100 is applied to a passive sonar device. However, application of the signal processing apparatus 100 is not limited to the passive sonar apparatus. For this reason, the sensor 102 is not limited to an acoustic sensor. For example, a vibration sensor that detects vibration, an optical sensor that detects an optical signal, a microphone that detects human conversation, music, or the like may be used. The sensor 102 transmits the detected signal 103 to the signal processing apparatus 100 through the transmission path 101.

  The signal processing apparatus 100 has a function of inputting the signal 103 from the sensor 102 through the transmission path 101 and suppressing only the pulse noise mixed on the transmission path 101 from the signal 103. For this purpose, the signal processing apparatus 100 includes a detection unit 110 and a suppression unit 120.

  The detection unit 110 has a function of analyzing the signal 103 and detecting pulse noise that appears outside the output frequency band of the sensor 102. The output frequency band of the sensor 102 is a frequency band that the sensor 102 can output. When the output frequency band of the sensor 102 is, for example, 0.1 kHz to 90 kHz, the detection unit 110 detects pulse noise that appears in a part or all of the frequency band of 90 kHz or more and the upper limit threshold or less. Here, the upper threshold is the upper limit of the frequency band of pulse noise. When detecting the pulse noise, the detection unit 110 outputs the detection information 111 to the suppression unit 120. The detection information 111 may be, for example, information (noise presence information) indicating whether or not pulse noise is present. For example, the detection information 111 is information indicating a predetermined value (for example, logical value 0) during a period when no pulse noise is present, and indicating a predetermined other value (for example, logical value 1) during a period when pulse noise is present. It may be.

  The suppression unit 120 has a function of suppressing only the pulse noise detected by the detection unit 110 from the signal 103. For example, the suppression unit 120 receives the signal 103 and the detection information 111, and suppresses the pulse noise from the signal 103 only during the period in which the pulse noise is present, which is indicated by the detection information 111. The suppression unit 120 outputs the signal 103 in which the pulse noise is suppressed as the signal 121. The signal 121 is then processed according to the purpose of use of the signal processing apparatus 100.

  Next, the operation of the signal processing apparatus 100 according to the present embodiment will be described.

  The detection unit 110 of the signal processing apparatus 100 analyzes the signal 103 input from the sensor 102 through the transmission path 101, detects pulse noise that appears outside the output frequency band of the sensor 102, and whether or not pulse noise exists. Is sent to the suppression unit 120.

  The suppression unit 120 receives the signal 103 and the detection information 111, suppresses only the pulse noise detected by the detection unit 110 from the signal 103, and outputs the signal 103 in which the pulse noise is suppressed as a signal 121.

  Thus, according to the present embodiment, it is possible to suppress only the pulse noise mixed on the transmission path 101 from the signal 103 input from the sensor 102 through the transmission path 101. The reason is that the signal input from the sensor 102 through the transmission path 101 is analyzed, and the detection unit 110 that detects the pulse noise that appears outside the output frequency band of the sensor 102, and the detected pulse noise is suppressed from the signal 103. This is because the suppression unit 120 is included.

  Next, a specific example of the detection unit 110 will be described. In addition, the structure of the detection part 110 is not limited to the structure illustrated below.

  Referring to FIG. 2, an example of the detection unit 110 includes a high pass filter 201 and a comparator 202. The high-pass filter 201 receives the signal 103, extracts a signal having a frequency component higher than the upper limit of the output frequency band of the sensor 102, and inputs the signal to the + input terminal of the comparator 202. A predetermined threshold value is input to the negative input terminal of the comparator 202. The comparator 202 sets the detection information 111 as an output to a logical value 1 if the input level of the + input terminal is higher than the input level of the −input terminal, and sets the logical value 0 otherwise.

  Next, a specific example of the suppression unit 120 will be described. Note that the configuration of the suppression unit 120 is not limited to the configuration illustrated below.

  Referring to FIG. 3, an example of the suppression unit 120 includes an input terminal 301, a conversion unit 302, a noise suppression unit 303, an inverse conversion unit 304, an output terminal 305, an analysis unit 306, a mixing unit 307, a noise information storage unit 308, an input. A terminal 309 is provided. The suppressor 120 uses the second related technique related to the present invention. That is, the suppression unit 120 analyzes the input signal 103, generates premixed noise information by a mixing method according to the analysis result using noise information stored in advance, and further uses the mixed noise information to generate noise. To suppress.

  The signal 103 is input to the input terminal 301. The signal 103 input to the input terminal 301 is subjected to transformation such as Fourier transformation in the transformation unit 302 and divided into a plurality of frequency components. Amplitude spectra of a plurality of frequency components are supplied to the noise suppression unit 303, and the phase spectrum is transmitted to the inverse conversion unit 304. Here, the amplitude spectrum is supplied to the noise suppression unit 303, but a power spectrum corresponding to the square thereof may be supplied to the noise suppression unit 303.

  The noise information storage unit 308 includes a storage element such as a semiconductor memory, and stores information (noise information) related to characteristics of known noise as a suppression target. Known noise stored as a suppression target is, for example, pulse noise generated by static electricity or the like.

  On the other hand, the analysis unit 306 inputs the amplitude spectrum generated by the conversion unit 302 and performs analysis. The analysis unit 306 determines the characteristics of noise included in the signal 103 by analyzing the amplitude spectrum, and determines a mixing method of noise information according to the characteristics. Then, the analysis unit 306 passes the determined mixing method to the mixing unit 307. The mixing unit 307 generates mixed noise information from the noise information stored in the noise information storage unit 308 according to the mixing method received from the analysis unit 306.

  Detection information 111 is supplied from the input terminal 309 to the noise suppression unit 303. The noise suppression unit 303 uses only the amplitude spectrum supplied from the conversion unit 302 and the mixed noise information supplied from the mixing unit 207 for each period in which noise indicated by the detection information 111 exists. Noise is suppressed by frequency, and an amplitude spectrum as a noise suppression result is transmitted to the inverse transform unit 204. The noise suppression unit 303 transmits the amplitude spectrum supplied from the conversion unit 302 to the inverse conversion unit 204 as it is during the period when the noise indicated by the detection information 111 does not exist. The inverse conversion unit 304 performs inverse conversion by combining the amplitude spectrum supplied from the noise suppression unit 303 and the phase spectrum of the signal 103 supplied from the conversion unit 302, and supplies the signal 121 to the output terminal 305.

  Details of the conversion unit 302, the noise suppression unit 303, the inverse conversion unit 304, the analysis unit 306, the mixing unit 307, and the noise information storage unit 308 are described in Japanese Patent Application Laid-Open No. 2003-260688, and therefore the description thereof is saved.

  In the present embodiment, various additions and changes can be made based on the above-described configuration.

  For example, the detection unit 110 includes a storage unit that stores output frequency band information of the sensor 102, and a pulse that appears outside the output frequency band of the sensor 102 based on the output frequency band information of the sensor stored in the storage unit. May be configured to detect sexual noise. For example, when the output frequency band information of the sensor is 0.1 kHz to 90 kHz, it may be configured to detect pulse noise appearing in a band of 90 kHz + α or more and less than 90 kHz + α + β. Here, α is a constant that defines the distance from the upper limit of the output frequency band of the sensor, and β is a constant that defines the bandwidth for detecting pulse noise.

  The detection unit 110 may have a learning function for output frequency band information of the sensor 102. For example, the detection unit 110 is configured to observe the output frequency band of the signal 103 a plurality of times at regular intervals, calculate an average thereof, and store the calculated average main frequency band in the storage unit. It's okay.

  The detection unit 110 and the suppression unit 120 can be realized by a computer and a program that configure the signal processing device 100. The program is recorded on a recording medium such as a semiconductor memory, read by the computer when the computer is started, and controls the operation of the computer, thereby realizing the detection unit 110 and the suppression unit 120 on the computer.

[Second Embodiment]
Next, a second embodiment of the present invention will be described. In this embodiment, an application in a sonar device will be described as an example of a signal processing device in the present invention.

  The sonar device is roughly classified into a passive sonar device and an active sonar device. A passive sonar device is a device that captures sound waves generated by a target (for example, a submarine or an underwater mine) existing in water (including the water surface) and calculates the presence and target position (orientation) of the target. In contrast, an active sonar device radiates sound waves into the sea, detects the reflected sound from the target, and measures the position and distance of the target from the time difference between the emitted sound and the reflected sound and the detection direction of the reflected sound. Refers to the device to be used.

  Among the sounds emitted by the target, there are stationary sounds that are generated for a long time from the engine, etc., and transient sounds that are not continuous in time or that last only for a short time, such as when steering a ship. is there. Hereinafter, a signal based on a received sound wave is referred to as a received signal, and a signal based on a transient sound is referred to as a transient signal.

  These received signals are processed by the sonar device, and the operator visually observes the sound and the screen display of the signal analysis result processed by the sonar device, and integrates the information to determine the presence or absence of the signal. Two typical techniques are known for signal analysis. One method is a method of performing frequency analysis by a method such as Fourier transform (FFT). A method called short-time FFT is generally used in which time resolution is set in accordance with the duration of a signal and calculation is periodically repeated while shifting an analysis interval. The other one is a method of detecting the received signal and analyzing the power, and is often used as a means for extracting a transient signal.

  Patent Document 4 discloses a technique for detecting a transient signal using power analysis. In the power analysis, there is a drawback that the signal cannot be recognized when there is a large power noise outside the band of the transient signal.

  According to the third related technique related to the present invention, it is possible to display a transient signal that does not depend on the band characteristics representing the amount of change in the level by using the sum value of the normalized levels for each frequency. . However, in the third related technique related to the present invention, since an input signal with suppressed noise is not generated, a desired signal other than the transient signal included in the input signal, for example, a steady state that continues for a long time using an engine or the like as a generation source. Sounds are thrown away in the process of signal processing. Moreover, an acoustic sensor capable of calculating the arrival direction of the acoustic signal is required.

<Problem to be solved by this embodiment>
The operator must find an acoustic signal generated by the target from a myriad of noises such as noise generated by a sound source other than the target in water and electrical noise mixed in the transmission line of the sonar device and the sonar device. These noises may cause target sound quality deterioration when listening to the signal, deterioration of target signal visibility in the display of signal analysis results, decrease in detection probability in automatic detection processing, increase in false alarm probability, etc. This is a major hindrance when finding signals. In particular, noise such as static noise, which is non-stationary like the transient signal, is a major factor in increasing the false alarm probability in detecting the transient signal using power analysis.

  Although the impulsive noise such as the static noise can be removed by using the second related technique related to the present invention, the passive sonar device removes the transient noise generated by the target together with the static noise. Therefore, the detection probability may be reduced. Even in a radio or the like, there is a possibility that a desired pulsed sound (such as a percussion instrument) other than sound may be suppressed.

  The main purpose of the present embodiment is to remove non-stationary pulse noise such as static noise from the acoustic signal and frequency analysis result without suppressing the desired sound, and to visually and audibly hear the target signal in the signal processing result. It is to provide a signal processing device, a signal processing method, and a signal processing program thereof that improve the false alarm probability in detection recognition and automatic detection by the computer.

<Means for solving the problem>
In order to solve the above-described problem, in this embodiment, a target signal and noise determination unit based on sensor sensitivity information is provided before the pulse noise suppression process (sensor characteristic information 3 in FIG. 4 and impulse noise determination). Processing unit 4). In addition, when the sensor characteristic information 3 cannot be obtained in advance, the impulse noise determination processing unit 4 accumulates the signal output from the reception processing unit 2 for a certain period of time and performs an averaging process to obtain the sensor characteristic information 3 by itself. It may be generated.

  Usually, an acoustic sensor has a sensitivity characteristic with respect to the frequency of an input acoustic signal (circles 1 and 2 in FIG. 5; hereinafter, the numbers with circles are as in (1) in the specification. (Shown in parenthesized numbers). Therefore, the acoustic signal received by the acoustic sensor is transmitted to the processing unit in a state where the sensitivity characteristic of the acoustic sensor is convoluted ((3) in FIG. 5), and the frequency analysis processing by FFT or the like is performed. On the other hand, noise mixed in a wired or wireless transmission path behind the acoustic sensor or inside the signal processing unit is transmitted to the signal processing unit without convolution of the sensitivity characteristics of the signal of the acoustic sensor, and frequency analysis processing is performed. . Therefore, in the frequency analysis results, noise that spreads over a wide band, such as static noise mixed in the transmission path, appears in its frequency component even in a band where the sensitivity of the acoustic sensor is originally low ((4) in FIG. 5).

  In the present embodiment, pulse noise that appears in a band different from the sensitivity characteristic of the acoustic sensor is detected and determined, and the pulse noise in the reception band of the acoustic sensor is suppressed.

<Effect of this embodiment>
According to the present embodiment, it is possible to suppress pulsating noise input from outside the acoustic sensor, such as static noise, without suppressing pulsating desired sound. In the sonar device, by suppressing noise, false detection can be reduced and the false alarm probability can be reduced.

  FIG. 6 shows an image of the processing result according to this embodiment. In the method according to the present embodiment, not only the acoustic processing device having the azimuth calculation sensor but also a single sensor without the azimuth calculation sensor can suppress the pulse noise without suppressing the desired sound. is there. In addition to the sonar device, this embodiment can be applied to an acoustic processing device having an acoustic sensor such as a radio and a transceiver, and an acoustic communication device, without erroneously suppressing a desired pulse-like sound such as a percussion instrument. The sound quality can be improved by suppressing the pulsating noise mixed in the transmission path.

<Configuration of this embodiment>
FIG. 4 is a block diagram of the passive sonar device according to the present embodiment. The passive sonar device includes an acoustic sensor unit 1, a display unit 9, and a processing unit 10. The processing unit 10 includes a reception processing unit 2, sensor characteristic information 3, impulse noise determination processing unit 4, impulse noise suppression processing unit 5, frequency analysis processing unit 6, power analysis processing unit 7, and signal detection processing unit 8. Have.

  The acoustic sensor unit 1 is installed in water and receives an underwater acoustic signal. Acoustic signals include stationary signals from the target, transient sounds, and stationary and non-stationary noise from noise sources. The output signal from the acoustic sensor unit 1 is in a state where the sensitivity characteristic of the acoustic sensor is convoluted ((1), (2), (3) in FIG. 5). The acoustic sensor unit 1 modulates an acoustic signal in which its own sensitivity characteristic is convoluted as necessary, and transmits it through a wired or wireless transmission path. At this time, the signal is in a state where noise such as static noise is added in the transmission line ((4) in FIG. 5). The reception processing unit 2 receives the transmitted signal, performs demodulation processing as necessary, takes out acoustic information, and outputs it to the impulse noise determination processing unit 4.

  The impulse noise determination processing unit 4 detects pulsed noise in the transmission noise band outside the sensor output signal band based on the sensor characteristic information 3 with respect to the received signal. When the sensor characteristic information 3 is not provided in advance, the impulse noise determination processing unit 4 accumulates input signal information for a predetermined time and generates sensor characteristic information 3 used for transmission noise band determination by performing an averaging process or the like. May be. The impulse noise determination processing unit 4 outputs the input acoustic signal and the detection result of the wideband signal in the transmission noise band to the noise suppression processing unit 5.

  The impulse noise suppression processing unit 5 performs an impulse noise suppression process only on a signal that has been detected based on the detection result of the wideband signal in the received transmission noise band. By performing the impulse noise suppression process based on the detection result of the broadband signal in the transmission noise band, it is possible to suppress the noise mixed in the signal transmission path without suppressing the transient sound from the target. The impulse noise suppression processing unit 5 outputs to the frequency analysis processing unit 6 and the power analysis processing unit 7 an acoustic signal that has been subjected to impulse noise suppression processing only on the detected signal. A pulse noise suppression technique described in Patent Document 2 can be applied to the impulse noise suppression processing unit 5.

  The frequency analysis processing unit 6 performs frequency analysis processing such as FFT or wavelet transform and outputs the result to the signal detection processing unit 8. Further, the power analysis processing unit 7 calculates the power of the signal and outputs it to the signal detection processing unit 8. The signal detection processing unit 8 performs threshold determination on the input frequency analysis result and power analysis processing result, and outputs the result to the display unit 9. Based on the received signal, the display unit 9 displays information such as the frequency analysis result, power analysis result, and detection result of the signal on the screen.

  As described above, according to this embodiment, it is possible to suppress the noise mixed in the signal transmission path without suppressing the transient sound from the target and present the detection result of the state in which the false alarm is reduced. Become.

  The present invention can be applied to a signal processing apparatus that inputs a signal from a sensor through a transmission line and suppresses only the pulse noise mixed on the transmission line from the signal.

DESCRIPTION OF SYMBOLS 1 ... Acoustic sensor part 2 ... Reception processing part 3 ... Sensor characteristic information 4 ... Impulse noise determination processing part 5 ... Impulse noise suppression processing part 6 ... Frequency analysis processing part 7 ... Power analysis processing part 8 ... Signal detection processing part 9 ... Display Unit 10 processing unit 100 signal processing apparatus 101 transmission path 102 sensor 103 signal 104 signal 110 detection unit 111 detection information 120 suppression unit 121 signal 201 high-pass filter 202 comparator 301 input terminal 302 ... Conversion unit 303 ... Noise suppression unit 304 ... Inverse conversion unit 305 ... Output terminal 306 ... Analysis unit 307 ... Mixing unit 308 ... Noise information storage unit 309 ... Input terminal

Claims (7)

Analyzing a signal input from a sensor through a transmission path, detecting a pulse noise appearing outside the output frequency band of the sensor; and
A suppressor that suppresses the detected pulse noise from the signal;
A signal processing apparatus. A storage unit for storing output frequency band information of the sensor;
The detection unit detects pulse noise that appears outside the output frequency band of the sensor based on the output frequency band information of the sensor stored in the storage unit.
The signal processing apparatus according to claim 1. The detection unit learns output frequency band information of the sensor from the signal and stores it in the storage unit.
The signal processing apparatus according to claim 2. The suppression unit analyzes the signal, generates noise information by mixing noise information related to noise to be suppressed according to the analysis result of the signal, and detects the noise using the mixed noise information. To suppress pulse noise,
The signal processing apparatus according to claim 1. The detection unit notifies the suppression unit of detection information indicating whether or not the pulse noise is present,
The suppression unit suppresses only the pulse noise detected by the detection unit from the signal based on the detection information.
The signal processing device according to claim 1. Analyzing the signal input from the sensor through the transmission path, detecting pulse noise appearing outside the output frequency band of the sensor,
Suppressing the detected pulse noise from the signal;
Signal processing method. Computer
Analyzing a signal input from a sensor through a transmission path, detecting a pulse noise appearing outside the output frequency band of the sensor; and
A suppressor that suppresses the detected pulse noise from the signal;
Program to make it function.

Images