JP6236755B2 - Passive sonar device, transient signal processing method and signal processing program thereof - Google Patents

Description

  The present invention relates to a passive sonar device, a transient signal processing method, and a signal processing program thereof.

  A passive sonar is a device that captures sound waves generated by a target existing in water (including the water surface) and calculates the presence and target position (orientation) of the target. Such a received sound wave includes a transient sound that is not continuous in time among sounds emitted by a target or continues for a short time. 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.

  Two typical techniques are known for analyzing such transient signals. One method is a method of performing frequency analysis by Fourier transform (FFT). In particular, when a transient signal is analyzed, a method called short-time FFT is used because the duration time is short. The other is a method of detecting the received signal and analyzing the power.

  Japanese Patent Laid-Open No. 10-177066 discloses a technique for detecting a transient signal using power analysis (see).

Japanese Patent Laid-Open No. 10-177066

  However, when there is a large power noise outside the frequency band of the transient signal, it is difficult to detect the transient signal by power analysis.

  Accordingly, a main object of the present invention is to provide a passive sonar device, a transient signal processing method, and a signal processing program thereof that can detect a transient signal even when a high-power noise source is present outside the frequency band of the transient signal. That is.

  In order to solve the above-mentioned problem, the invention according to the passive sonar device performs PDAGC processing on the Fourier-transformed signal, outputs a normalized signal obtained by normalizing the signal, calculates the direction of the sound source from the signal, and calculates the direction. A signal processing unit that outputs as a signal, a noise removal unit that generates and outputs a noise removal signal obtained by extracting a signal having a short duration by frequency-adding at least one of a normalized signal and a direction signal, and a noise removal signal And a display unit for displaying an image based on the display unit.

  In the invention according to the transient signal processing method, the Fourier-transformed signal is subjected to PDAGC processing, and a normalized signal obtained by normalizing the signal is output, and the direction of the sound source is calculated from the signal and output as the direction signal. A signal processing procedure, a noise removal procedure for generating and outputting a noise removal signal obtained by extracting a signal having a short duration by adding the frequency of at least one of a normalized signal and a direction signal, and displaying an image based on the noise removal signal And a display procedure to be performed.

  Furthermore, the invention according to the transient signal processing program performs PDAGC processing on the signal subjected to Fourier transform, outputs a normalized signal obtained by normalizing the signal, calculates the azimuth of the sound source from the signal, and outputs the azimuth signal. A signal processing step, a noise removal step for generating and outputting a noise removal signal obtained by extracting a signal having a short duration by adding at least one of a normalized signal and an orientation signal, and displaying an image based on the noise removal signal And a display step.

  According to the present invention, it becomes possible to detect a transient signal even when a noise source having a large power is present outside the band of the transient signal, and the visibility and reliability of display are improved.

It is a block diagram of the passive sonar apparatus which concerns on embodiment of this invention. It is a flowchart which shows operation | movement of a passive sonar apparatus. It is the figure which showed the normalization signal obtained by performing PDAGC process with respect to a received signal. FIG. 6 is a distribution diagram schematically showing signals before and after PDAGC processing (FFT signal and normalized signal), where (a) is a frequency-time distribution of an FFT signal before PDAGC processing, and (b) is normalized after PDAGC processing. It is a frequency-time distribution of a signal. Distribution diagram schematically showing signals before and after PDAGC processing, (a) shows level-frequency distribution of FFT signal before PDAGC processing, and (b) shows level-frequency distribution of normalized signal after PDAGC processing. FIG. FIGS. 5A and 5B are diagrams illustrating an AGC addition process, where FIG. 5A is a level distribution diagram illustrated in FIG. 5B, and FIG. 5B is a level frequency distribution illustrating an addition value in time. FIGS. 7A and 7B are diagrams illustrating a sector addition process. FIG. 7A is a diagram illustrating an orientation signal, and FIG. 7B is a diagram illustrating an orientation frequency distribution.

  An embodiment of the present invention will be described. FIG. 1 is a block diagram of a passive sonar device 2 according to the present embodiment. The passive sonar device 2 includes a signal processing unit 4, a noise removal unit 6, and a display unit 8.

  The signal processing unit 4 includes an FFT unit 11, a normalization unit 12, and an azimuth calculation unit 13. The signal processing unit 10 receives a received signal including an omni (OMNI) signal for detecting a signal from all directions, an NS signal having directivity in the north-south direction, and an EW signal having directivity in the east-west direction. Therefore, the signal processing unit 10 generates a normalized signal based on the received signal, and also generates and outputs an orientation signal.

  The noise removal unit 6 includes an addition unit 14, a subtraction unit 15, and a signal detection unit 16. The noise removal unit 6 receives the normalization signal and the azimuth signal, and performs AGC addition processing and sector addition processing on these signals to generate a noise removal signal from which noise has been removed. An extracted signal obtained by extracting the signal is generated and output.

  The detailed configuration and operation of each element will be described below with reference to the flowchart shown in FIG. 2 and FIGS.

  Step S1: The FFT unit 11 performs an FFT process on the received signal. The FFT process is a process of deriving a frequency spectrum by cutting out a signal in a predetermined time window and performing fast Fourier transform on the received signal for each cut out time.

  Step S2: The normalizing unit 12 normalizes the FFT-processed signal and outputs it as a normalized signal. A known method can be applied as the normalization method. For example, in order to suppress the fluctuation of the frequency spectrum derived by the FFT unit 11 on the time axis, a post-detection automatic gain control (PDAGC) that integrates and normalizes the input signal with a predetermined time constant. : Post Detection Automatic Gain Control) is applicable.

  That is, this PDAGC process performs local normalization by, for example, obtaining an exponential moving average (integration) in the time direction set in advance for each frequency and dividing the amplitude level of the frequency by the average value. Then, such normalization is sequentially performed for all frequencies. At this time, when the frequency is low, the time constant is set to a long time, and when the frequency is high, the time constant is set to a short time, whereby noise can be appropriately removed. At this time, since no filter is used in the PDAGC process, noise can be removed with a simple configuration.

  FIG. 3 is a diagram illustrating a normalized signal obtained by performing PDAGC processing on a received signal. 4 is a distribution diagram schematically showing signals before and after PDAGC processing (FFT signal and normalized signal). FIG. 4A is a frequency-time distribution of the FFT signal before PDAGC processing. (B) is a frequency-time distribution of the normalized signal after PDAGC processing. FIG. 5 is also a distribution diagram schematically showing signals before and after the PDAGC process as in FIG. 5A shows the level-frequency distribution of the FFT signal before PDAGC processing, and FIG. 5B shows the level-frequency distribution of the normalized signal after PDAGC processing. In particular, as clearly shown in FIG. 4, by performing PDAGC processing on the FFT signal, a signal with a long duration is suppressed or removed (that is, a transient signal remains), and a normalized signal is obtained. .

  Step S3: The direction calculation unit 13 calculates the direction indicated by the frequency spectrum by Fourier transform of the OMNI signal based on the NS signal and the EW signal. This calculation result is output to the noise removal unit 6 as an orientation signal.

  Step S4: The adding unit 14 adds the frequency of the normalized signal or the azimuth signal and outputs a frequency signal. At this time, whether the frequency addition process (hereinafter referred to as AGC addition process) is performed on the normalized signal or the frequency addition process (hereinafter referred to as sector addition process) on the azimuth signal is determined by the user's selection instruction It is possible. However, the present embodiment is not limited to such a configuration, and both processes may be automatically performed as a default so that the user can select while viewing the display screen on the display unit 8.

(AGC addition processing)
When the user selects the AGC addition process, the addition unit 14 adds a normalized signal by the PDAGC process as shown in FIG. FIG. 6 is a diagram for explaining this. FIG. 6A is a level distribution diagram shown in FIG. 5B, and FIG. 6B is an added value (value obtained by AGC addition processing). The level frequency distribution is shown by time.

The adder 14 adds levels (level frequency) over all frequencies to calculate an added value, and outputs the added value as a frequency signal. FIG. 6A shows the level distribution with respect to the frequency at a certain time, and this added value (level frequency) corresponds to the diagonal line in FIG. 6B.
(Sector addition processing)
On the other hand, when the user selects the sector addition process, the frequency of the azimuth signal is added and a frequency signal is output. FIG. 7 is a diagram illustrating the sector addition process. Hereinafter, the result obtained in the sector addition process is output as an orientation frequency, and this orientation frequency distribution is output as a frequency signal. FIG. 7A is a diagram showing the azimuth signal, and FIG. 7B is a diagram showing the azimuth frequency distribution. FIG. 7B illustrates the case where the addition azimuth range is 45 degrees. For example, the central azimuth of 90 degrees is an added value (azimuth frequency) in the range of 67.5 degrees to 112.5 degrees of azimuth.

  Hereinafter, the level frequency and the azimuth frequency are simply referred to as frequency.

  Step S5: The subtraction unit 15 calculates an average value of the level frequency or the azimuth frequency (frequency average value), and subtracts the frequency average value from the level frequency or the azimuth frequency.

  In the case of the AGC addition process, the subtraction unit 15 calculates an average value with respect to the time of the level frequency. On the other hand, in the case of the sector addition process, the subtracting unit 15 calculates an average value of the azimuth frequency with respect to the central direction. The dotted line in FIG.6 (b) and FIG.7 (b) has shown the frequency average value.

  Then, the subtraction unit 15 subtracts the frequency average value from the level frequency or the azimuth frequency. Thereby, the noise removal signal from which the omnidirectional noise was removed can be acquired. That is, the transient signal can be detected with high accuracy even when a noise source having a large power exists outside the band of the transient signal. The noise removal signal is output to the signal detection unit 16 and the display unit 8.

  Step S6: The signal detection unit 16 performs threshold processing with a preset threshold for the noise removal signal. Then, a signal having a value larger than the threshold value is extracted and output to the display unit 8 as an extracted signal. The extracted signal indicates a transient signal.

  Step S7: The display unit 8 displays an image of the noise removal signal and the extracted signal. At this time, the display unit 8 displays the image of the noise removal signal and displays the extracted signal on the image. By displaying the extracted signal as an image, the transient signal can be displayed with emphasis. Of course, the present embodiment is not limited to displaying the extracted signal as an image, and may be displayed as character information.

  Then, the user instructs the adding unit 14 to display an image based on the AGC addition process or an image based on the sector addition process. Thereby, when the image based on the AGC addition process is displayed, the image based on the sector addition process can be displayed.

  As described above, the transient signal can be detected even when a noise source with a large power exists outside the band of the transient signal.

  Further, since an image based on the AGC addition process and the sector addition process can be selected, it is possible to display an image suitable for the user's request, and the convenience is improved.

2 Passive sonar device 4 Signal processing unit 6 Noise removal unit 8 Display unit 10 Signal processing unit 12 Normalization unit 13 Direction calculation unit 14 Addition unit 15 Subtraction unit 16 Signal detection unit

Claims (12)

An exponential moving average in the time direction set in advance for each frequency of the Fourier-transformed signal is obtained, and a normalized signal obtained by dividing the amplitude level of the frequency by the average value is output. A signal processing unit that calculates an azimuth and outputs it as an azimuth signal;
The orientation signal by the frequency adder as well as frequency adding the normalized signal, a noise elimination unit for generating and outputting a noise removing signal obtained by extracting the short signal duration,
A display unit for displaying an image based on the noise removal signal,
The noise removal unit includes:
An addition unit for outputting a frequency signal the direction signal and the frequency sum with frequency adding the normalized signal,
A subtractor that calculates an average value of the frequency signal and subtracts the average value from the frequency signal and outputs it as a noise removal signal;
A signal detection unit that extracts the noise removal signal that is larger than a preset threshold and outputs the signal as an extraction signal;
A passive sonar device comprising: The passive sonar device according to claim 1,
The adding unit derives a level frequency obtained by adding the levels of the normalized signal at a predetermined time interval when calculating the frequency signal from the normalized signal, and calculates a level frequency distribution over a predetermined time. A passive sonar device characterized by output. The passive sonar device according to claim 1,
The adding unit derives the azimuth frequency obtained by adding the azimuth signals included in the predetermined azimuth range when calculating the frequency signal from the azimuth signal, and calculates and outputs the azimuth frequency distribution over the predetermined azimuth range. A passive sonar device characterized by that. The passive sonar device according to any one of claims 1 to 3,
The display unit displays a noise removal signal from the subtraction unit as an image, and displays the extracted signal from the signal detection unit as an image or character information. An exponential moving average in the time direction set in advance for each frequency of the Fourier-transformed signal is obtained, and a normalized signal obtained by dividing the amplitude level of the frequency by the average value is output. A signal processing procedure for calculating an azimuth and outputting it as an azimuth signal;
The orientation signal by the frequency adder as well as frequency adding the normalized signal, and noise removal procedure for generating and outputting a noise removing signal obtained by extracting the short signal duration,
A display procedure for displaying an image based on the noise removal signal,
The denoising procedure includes:
An addition procedure for adding the frequency of the normalized signal and adding the frequency of the azimuth signal to output as a frequency signal;
A subtraction procedure for calculating an average value of the frequency signal, subtracting the average value from the frequency signal and outputting as a noise removal signal;
And a signal detection procedure for extracting the noise removal signal larger than a preset threshold value and outputting the extracted signal as an extraction signal. The transient signal processing method according to claim 5,
When the frequency signal is calculated from the normalized signal, the addition procedure derives a level frequency obtained by adding the levels of the normalized signal at a predetermined time interval, and calculates a level frequency distribution over a predetermined time. A transient signal processing method comprising: outputting a transient signal. The transient signal processing method according to claim 5,
In the addition procedure, when calculating the frequency signal from the azimuth signal, the azimuth frequency obtained by adding the azimuth signals included in the predetermined azimuth range is derived, and the azimuth frequency distribution over the predetermined azimuth range is calculated and output. And a transient signal processing method. The transient signal processing method according to any one of claims 5 to 7,
In the display procedure, the noise removal signal from the subtraction procedure is displayed as an image, and the extracted signal from the signal detection procedure is displayed as an image or displayed as character information. An exponential moving average in the time direction set in advance for each frequency of the Fourier-transformed signal is obtained, and a normalized signal obtained by dividing the amplitude level of the frequency by the average value is output. A signal processing step of calculating an azimuth and outputting it as an azimuth signal;
The orientation signal by the frequency adder as well as frequency adding the normalized signal, and noise removal step for generating and outputting a noise removing signal obtained by extracting the short signal duration,
Displaying an image based on the noise removal signal, and
The noise removal step includes:
An addition step of adding the frequency of the normalized signal and adding the frequency of the azimuth signal to output as a frequency signal;
A subtraction step of calculating an average value of the frequency signal, subtracting the average value from the frequency signal and outputting as a noise removal signal;
And a signal detection step of extracting the noise removal signal larger than a preset threshold value and outputting the extracted signal as an extraction signal. The transient signal processing program according to claim 9,
In the addition step, when calculating the frequency signal from the normalized signal, a level frequency obtained by adding the level of the normalized signal at a predetermined time interval is derived, and a level frequency distribution over a predetermined time is calculated. A transient signal processing program for outputting. The transient signal processing program according to claim 9,
In the adding step, when calculating the frequency signal from the azimuth signal, the azimuth frequency obtained by adding the azimuth signals included in the predetermined azimuth range is derived, and the azimuth frequency distribution over the predetermined azimuth range is calculated and output. The transient signal processing program characterized by the above-mentioned. A transient signal processing program according to any one of claims 9 to 11,
In the transient signal processing program, the display step displays an image of the noise removal signal from the subtraction step, and displays the extracted signal from the signal detection step as an image or character information.

Images