JPH103549A - Device for supporting discrimination of sounding body - Google Patents

Abstract

PROBLEM TO BE SOLVED: To more accurately and quickly discriminate a sounding body by plotting a pattern corresponding to data for a harmonic sequence on an window when the data are acquired by an input device. SOLUTION: A harmonic check button 110 starts/ends harmonic checking operation. When a mouse cursor 160 exists in a picture displaying window 120, an input mode for n, X and Y is set up and a position (X, Y) becomes the positional coordinates of the cursor 160. The initial value of (n) is '2'. When the left button of a mouse 130 is clicked, '1' subtracted from 'n' when n>1. When the right button of the mouse 13 is clicked, '1' is added to 'n'. When a precedently plotted horizontal line/vertical line exists in updating any one of n, X and Y, the horizontal line/vertical line is erased, the position of the cursor 160 is set up to (F, T) in an (f-t) coordinate system to plot a horizontal straight line t=T and a straight line group f=mF/n (m=1, 2,...) is plotted on a position capable of setting up the position of the cursor 160 to an n- harmonic.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

TECHNICAL FIELD The present invention belongs to the field of pattern recognition.

[0002]

2. Description of the Related Art First, a method for identifying a sounding body will be briefly described. As the data to be identified, an image (FIG. 2) in which the frequency component (spectrum) of the observed sound is displayed in different shades of time is often used. In FIG. 2, the horizontal axis is frequency, the vertical axis is time, and the axis in the direction perpendicular to the paper is a spectrum displayed in shades. There are two types of coordinate systems for this image: an xy coordinate system that represents the number of dots from the edge of the window, and an ft coordinate system that represents the original frequency-time.

A sound produced by one sounding body is composed of a series of frequency components that are substantially constant or change continuously over time and frequency components (overtones) that are integral multiples of the frequency components. Therefore, on this image, a substantially vertical linear pattern appears at a position corresponding to the harmonic series. On the other hand, the frequency component of the sound emitted by the sounding body differs for each type of the sounding body (for example, 100H).
There is a linear pattern at the position z, the third harmonic is relatively strong, etc.). Therefore, if (1) the linear pattern is detected from the image, and (2) the linear pattern is checked and checked against the pattern of the known sounding body, the sounding body can be identified. One important operation for performing these operations is to find a series of patterns that are in a harmonic relationship with each other. If this can be achieved, it is possible to more accurately detect each linear pattern by interpolating the portions of the harmonics that are missing from each other, and it may be possible to limit the type of sounding body by confirming the presence of a specific harmonic. .

A typical conventional sounding body identification system in the field of underwater acoustic information processing is provided with an interface for assisting in finding a sequence of the above-mentioned patterns (FIG. 3). An example will be described below. Conventional systems consist of a display, a trackball, and a touch panel. The operation procedure is as follows, for example.

[0005] 1. Press the "Overtone Confirmation" button on the touch panel to start operation.

[0006] 2. Point the trackball pointer to one pattern on the image.

[0007] 3. Press the "frequency determination" button on the touch panel to determine the frequency f1.

[0008] 4. Point the trackball pointer to another pattern on the image.

[0009] 5. Press the "frequency determination" button on the touch panel to determine the frequency f2.

[0010] 6. Frequency m (f2-f1) + f1 (m
= 0, ± 1,...).

7. If the "overtone confirmation" button is pressed, the operation ends.

When a button other than the "overtone confirmation" button is pressed, go to 2.

[0013]

When a sequence of patterns having a harmonic relationship with each other is found by using the above-mentioned conventional interface, the following (a) to (e) are used.
The following problems occur.

(A) It is difficult to compare the frequency of each pattern at the same time.

[0015] For a pattern sequence whose frequency fluctuates greatly over time, comparing frequencies at different times does not make much sense.

However, in the conventional example, there is no mark for visually judging whether or not points separated in the horizontal axis (frequency) direction on the window are at the same time.

(B) Operation with one hand is difficult.

In the conventional example, since it is necessary to use both a trackball and a touch panel as input devices, it is difficult to operate quickly with one hand.

(C) It takes time and effort to find a pattern sequence for a plurality of frequencies, times, and harmonic numbers.

In the conventional example, the straight line group is not drawn unless the frequency or the like is determined on the touch panel. For this reason, for example, in order to compare two patterns in which the frequency greatly fluctuates with time, it is necessary to perform the operation of determining the frequency and the like at each time.

(D) There is a difficulty in handling “quasi-overtones”.

Although there is no overtone in the original sense, there is a series of relations in which the ratio is always maintained even if the frequency fluctuates. Here, it is tentatively called “quasi-overtone”.

An example is a sound produced by a sounding body having two rotating shafts connected by gears.

The sound produced by this sounding body includes two overtone sequences corresponding to the two rotation axes, and the frequency ratio of the fundamental sound is always kept equal to the gear ratio.

For example, it is assumed that there are two quasi-overtone sequences having fundamental sounds of 70 Hz and 50 Hz, respectively.

At this time, the correct overtone sequence is 70, 14
0,... And 50, 100,. However, when 50 Hz and 70 Hz are selected as candidates for the harmonic series according to the conventional example, vertical straight line groups are displayed at positions corresponding to 10, 30, 50, 70... One of these is
.. Do not include correct overtones such as 00, 200..., And conversely include incorrect overtones such as 10, 30, 90.

In the conventional example, a series of straight lines to be displayed is determined and displayed based on two frequencies and their difference. However, it is difficult to naturally expand the series when there are three or more semi-overtone sequences. .

(E) There is no means for automatically calculating the frequency ratio of the fundamental sound of the semi-overtone series.

In the example of (d), the frequency ratio of the fundamental sound is 7:
It is not immediately apparent that it is 5.

An object of the present invention is to solve these problems.

[0031]

Means for Solving the Problems The symbols used in the claims or figures will be used without exception in this section and below.

The means for solving the above-mentioned problems consist of the following points.

(A) Draw a horizontal line at a specified position on the window.

The point near the horizontal line can be easily visually checked to determine the same time, so that the problem (a) can be solved.

(B) As an input interface, a pointing device having switches capable of user input is used.

In this section, a mouse will be described as an example of such a pointing device, and a mouse button will be described as switches that can be input by a user. If the mouse button is simply used instead of the touch panel, the operation can be easily performed even with one hand, and the problem (b) can be solved.

(C) Designation of coordinate values X and Y is automatically performed according to the position of the mouse cursor. Positive integer (overtone number) n
The initial value is increased or decreased by a mouse button operation after the initial value (refer to claims 2 and 4).
Further, for n, X, and Y designated in this manner, the straight line group described in claims 1 and 2 is displayed again.

Thus, various frequencies, times, times, etc. can be obtained only by moving the mouse or operating the mouse button without explicitly determining the frequency and harmonic number as in the conventional example.
(#) Can be performed for the overtone number, and the problem (c) can be solved.

(C ') A specific overtone pattern is stored in advance.

Thus, even when the pattern has a complicated shape, it is possible to confirm the harmonic relationship at a plurality of times at a time.

(D) The frequency and harmonic number are input parameters.

The straight line group to be displayed is simply the union of the straight line groups to be displayed when each overtone sequence is specified alone.

Thus, it is easily understood that the disadvantage described in the problem (d), which occurs when two frequencies are used as input parameters as in the conventional example, can be solved.

(E) An integer ratio close to the given frequency ratio is displayed.

In order to obtain an integer ratio close to a given real number, for example, an algorithm that performs the Euclidean algorithm (an algorithm for obtaining the greatest common divisor of two integers) in the opposite direction may be used.

Thus, the problem (e) can be solved.

[0047]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS First, a first embodiment of the present invention will be described.

FIG. 1 is a block diagram of a sounding body identification support apparatus (100) according to the present embodiment. This is composed of an image display window, a harmonic confirmation button, and a pointing device (mouse in the following example), and operates as follows.

The overtone confirmation button is a toggle button. Clicking this button starts and ends the overtone confirmation operation.

When the mouse comes out of the image display window, the horizontal and vertical lines are deleted.

When the mouse cursor is in the image display window, the input mode is n, X, Y.

When the mouse cursor is in the image display window, (X, Y) is the coordinates of the mouse cursor position. (FIG. 4) ○ The initial value of n is 2.

○ When the left mouse button is clicked,
If n> 1, subtract 1 from n. When the right mouse button is clicked, 1 is added to n. (FIG. 5) When any one of n, X, and Y is updated, the previously drawn horizontal line and vertical line, if any, are deleted, and the position of the mouse cursor ((F, T in the ft coordinate system) )), A horizontal straight line t = T is drawn, and a straight line group f = mF /
n (m = 1, 2,...) is drawn (FIGS. 4, 5).

Since it is not necessary to see the same time at frequencies other than the frequency at which harmonics are to be checked, the display of the horizontal straight line t = T may be limited to only the vicinity of each vertical straight line f = mF / N.

When there are two patterns that are harmonics, if the mouse cursor is positioned at one point on one pattern and the left and right buttons of the mouse are clicked appropriately and n is adjusted, the other pattern and a certain vertical pattern are obtained. Straight line f = mF / n
Can be found on the horizontal straight line t = T.
At this time, the former pattern is n harmonics, and the latter pattern is m
Proof that it is of an overtone is obtained. Here, if the mouse position (F, T) is further moved along the former pattern, the intersection of f = mF / n and t = T moves along the latter pattern. In this way, it can be confirmed that the two patterns are those of harmonics.

Next, a second embodiment will be described. The configuration is such that an integer ratio display window is added as the output device (610) to the first example. In the integer ratio display window, the real number F ′ / F converted to an integer ratio is displayed at any time according to the movement of the cursor (FIG. 6).

An example of an algorithm for converting a real number x into an integer ratio b / a is as follows.

1. X (0) = x X (n + 1) = 1 / (X (n)-[X (n)]) 2. For N such that X (N) is an integer, A (N) = 1, B (N) = X (N). A (n-1) = B (n) B (n-1) = B (n) * [X (n-1)] + A (n) a = A (0), b = B (0) where [] is a symbol representing an integer part. In this way, the frequency ratio of a series of overtones or “quasi-overtones” can be known as an integer ratio. For example, if the gear ratio of the gear included in the sounding body is known, this greatly contributes to identification.

As a third example, a fourth embodiment will be described.

The display device according to claim 5 comprises an image display window, a pointing device (a mouse in the following example), and a storage device (710). FIG. 7 is an operation image diagram. First, in the initial state, nothing is stored in the storage device (FIG. 7, left). Next, by dragging the pattern while holding down the mouse button on the window, the locus of the mouse is stored in the storage device, and when the mouse button is released, the storage is completed (FIG. 7). Finally, when the mouse cursor is moved, a pattern (740) obtained by expanding the stored pattern (730) in the horizontal direction is displayed at the position of the pointing device. Here, the center of enlargement and the magnification are defined in claim 5.
As described in the above.

As described above, by storing the patterns in advance, it is possible to confirm the overtone at many times between the patterns having a complicated shape at once.

As a fourth example, an embodiment according to claim 5 will be described.

The configuration is such that an integer ratio display window is added as an output device to the third example. In the integer ratio display window, the real number F ′ / F converted to an integer ratio is displayed at any time according to the cursor movement. However, F and F 'are as described in claim 4. An example of an algorithm for converting a real number x into an integer ratio b / a is the same as the second example.

As a fifth example, a sixth embodiment will be described.

FIG. 8 is a diagram showing the configuration. Here, the frequency ratio output device outputs the frequency ratio in the form of a real number and an integer ratio, and also outputs the frequency itself. The sounding body database holds data of various sounding bodies using the name of a sounding body, a typical frequency of a basic sound, and, when there are a plurality of basic sounds, a frequency ratio (real number, integer ratio) of the basic sound as one unit. And

The database matching device receives the frequency ratio a: b and the frequency f from the frequency ratio output device, compares them with the sounding body database, and outputs the name of the sounding body that is “matched”.

The determination of “conformity” is made as follows, for example.
The frequency ratio entry of the data in the database is a ':
When b ′ and its corresponding frequencies are fa and fb, a is a ′, b is b ′, and f / a and fa
Is within 1 ± 5% ○ If any one of the above conditions where a is replaced by b and fa is replaced by fb is satisfied, then it is not suitable.

It is conceivable that f is not a fundamental sound but an appropriate overtone of the basic sound. Therefore, it is preferable to use the prime factors for the conditions a, b, etc. (= If they are multiples, they need not be equal).
Generally, it is considered that the frequency of a rotating shaft to which an external load is applied by a gear or the like is likely to fluctuate due to a change in load. Then f /
It is preferable that the condition of the ratio of a to fa has a certain range. In this example, the case where there are three or more fundamental sounds (the case of a sounding body in which three or more rotating shafts are connected by a plurality of gears) is not considered.

When a sound produced by the sounding body is observed, the frequency itself fluctuates due to the Doppler effect, uneven rotation of the sounding body itself, etc., and thus is not always stable. However, the frequency ratio between overtones and the frequency ratio of two-axis sounds connected by gears (both are integer ratios) are kept constant irrespective of frequency fluctuations. Therefore, in the case where the sounding body to be identified includes gears and has large rotation unevenness, the object can be identified more accurately by using the frequency ratio than by using the frequency itself.

[0070]

The present invention provides an operator interface for supporting the confirmation of overtones, which is one of the important steps in sounding body identification. This allows the operator of the sounding body identification system to check the overtones accurately and quickly,
As a result, the sounding body can be identified more accurately and quickly.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration example of a sounding body identification support device according to claim 1 of the present invention.

FIG. 2 is a diagram illustrating an image used as identification data of a sounding body.

FIG. 3 is a diagram showing a configuration example of a conventional sounding body identification support device.

FIG. 4 is a diagram showing a relationship between a movement of a pointing device and a change in display according to claim 2 of the present invention.

FIG. 5 is a diagram showing a relationship between operation of switches and a change in display according to a second embodiment of the present invention.

FIG. 6 is a diagram showing a processing screen image of the sounding body identification support device according to claim 3 of the present invention when the sounding body has a plurality of overtone sequences.

FIG. 7 is a diagram showing an operation image of the display device according to claims 4 and 5 of the present invention.

FIG. 8 is a diagram showing a configuration of a semi-automatic sounding body identification device according to claim 6 of the present invention.

[Explanation of symbols]

100: sounding body identification support device 110: overtone confirmation start / end button 120: image display window as data to be identified 130: input device
140 ... horizontal straight line l (1), 150 ... vertical straight line group l (m, 1), 160 ... cursor of pointing device, 200 ... image image of data to be identified, 310 ...
Touch panel, 320: window for displaying an image as data to be identified, 330: trackball, 350: vertical straight line group indicating overtone sequence candidates, 360: trackball cursor, 610: output window for displaying the frequency ratio as an integer ratio, 700 ... The display device according to claim 5, 7
10: storage device, 720: pattern to be stored,
730: stored pattern, 740: generated / displayed pattern, 810: database matching device, 82
0 ... Pronunciation database.

 ──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Hiroshi Takahashi 216 Totsuka-cho, Totsuka-ku, Yokohama-shi, Kanagawa Prefecture Inside the Information and Communications Division, Hitachi, Ltd. (72) Inventor Kazuhiko Hino 216 Totsuka-cho, Totsuka-ku, Yokohama-shi, Kanagawa, Japan In-house Information and Communications Division

Claims (6)

[Claims] A display window and an input device;
The window has two orthogonal coordinate systems (tentatively, an xy coordinate system and an ft coordinate system), and the input device displays coordinate values (X, Y) in the xy coordinate system and a positive integer n. When given,
The line t = T and the line group f = mF / n (m = 1, 2,...) Where (X, Y) is expressed in the ft coordinate system as (F, T).
And a display method (100) for drawing on a window in a color designated in advance. 2. The pointing device according to claim 1, further comprising a pointing device having switches that can be input by a user as an input device, and updating the value of a positive integer n when a specific operation of the switches is performed. The xy coordinate value (X, Y) is updated with the new cursor position coordinate as needed along with the cursor movement, and the ft of (X, Y) when any one of n, X, and Y is updated as described above. Assuming that the expression in the coordinate system is (F, T), a straight line t = T, a straight line group f = mF / n
A display method and a device for drawing (m = 1, 2,...). 3. The frequency ratio output device according to claim 2,
10), when another specific operation of switches is performed, a positive integer n, xy coordinate values X, Y are fixed, and n,
After the fixing of X and Y is completed, the cursor of the pointing device is located at (F ′, T ′) in the ft coordinate system, and the expression (X, Y) in the ft coordinate system is (F, T). ), A display method and apparatus for drawing a straight line t = T ′, converting the ratio F ′ / F of the f-coordinate into an integer ratio, and outputting the result to an output device. 4. A trace comprising a window, a pointing device, and a storage device (710). The trace is stored in the storage device by tracing a pattern (720) on the window with the pointing device. In the case where the cursor of the pointing device is located at (F ′, T ′) and the stored pattern includes (F, T ′) in the representation in the t coordinate system, a trajectory (730) that draws and stores a straight line t = T ′ and A display method and apparatus (700) for drawing a pattern (740) obtained by multiplying the f-coordinate by F ′ / F. 5. The frequency ratio output device according to claim 4,
10) A display system and apparatus for converting the ratio F / F ′ of the cursor position and the f coordinate of the stored pattern into an integer ratio and outputting the integer ratio to an output device as needed. 6. The database matching device according to claim 3, wherein the database matching device and the pronunciation body database.
The frequency ratio output device and the database matching device, the database matching device and the pronunciation body database are respectively connected, the frequency ratio output device outputs the frequency ratio, the pronunciation body database, the pronunciation body ID, the frequency of its overtone sequence,
When there are a plurality of overtone sequences, data of a set of these frequency ratios is stored, and the database matching device searches for a pronunciation ID matching the frequency ratio and outputs the same.