JP2000171331A - Signal detecting method and device therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a signal detecting methcod and its device capable of detecting the presence of an effective signal with high reliability and extremely good working efficiency even if an interference sound is generating, or a sound signal and a silence signal are included. SOLUTION: An energy integrating characteristic calculating circuit including a sensor 1, an energy integrating circuit 2, an integrating characteristic calculating circuit 3, a judgment circuit 4, and an energy integrating characteristic calculation circuit including a time marker 5 and an analysis scheduled data memory 6 is provided. The detection data continued in time is finely divided into short monitoring unit time areas, and the level of the detected signal is monitored every monitoring unit time area. When silence level is present even in one monitoring unit time, 'normality' is judged, and the monitoring work is immediately ended. When the detected level fluctuation of signal is large, inclusion of crosstalk is judged, this data is removed as out the subject, and the procedure is transferred to the monitoring of the following monitoring unit time area data to analyze only the data continued for a required time, having a small level fluctuating width and difficult to fudge the silence signal by highly precise frequency analyzing method.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a signal detection method and apparatus for detecting the presence or absence of a signal in the presence of interfering sound or background noise and the presence or absence of interfering sound in a signal. It is about.

[0002]

2. Description of the Related Art Conventionally, in order to specify the approximate location of a leak point in a water supply pipe, a leak sound detection device having a sensor installed on the ground surface close to the water pipe and a water leak installed on a water meter and a fire hydrant have been known. There is a sound search knowledge-based device. In the device for detecting water leak sound, a high-precision frequency analysis method is applied to the output signal of the sensor, and a spectrum emphasis process, an integration process, a frequency fluctuation numerical process, etc. obtained by the frequency analysis (hereinafter referred to as a “high-precision frequency (Referred to as “analysis processing”) to detect water leak noise (for example,
No. 18451).

[0003]

However, in the above-described identification method using the conventional apparatus for detecting water leak sound, identification processing is performed by high-precision frequency analysis processing regardless of the presence or absence of interfering sound and the presence or absence of a signal. Since the result was forcibly obtained, the reliability of the data was difficult, and there was a possibility that an erroneous determination was made. Thus, the work efficiency and the technology were not satisfactory.

The present invention has been made to solve the above problems, and has a wide range of dark noise signal groups (hereinafter referred to as "signals") composed of continuous signals or intermittent but continuous signals for a certain period of time. The background noise is excluded from a group of signals mixed with a high-level, irregularly and spontaneously generated signal (hereinafter referred to as a "disturbance sound") to detect the presence or absence of a signal. be able to.

[0005] In other words, the present invention can be applied to a case where a disturbing sound is generated or a signal and a non-signal are mixed.
It is an object of the present invention to provide a signal detection method and device capable of detecting the presence or absence of an effective signal with high reliability and extremely high work efficiency.

[0006]

According to the present invention, in order to achieve the above object, there is provided [1] a signal detection method, comprising: a sensor, an energy integration circuit, an integration characteristic calculation circuit, a judgment circuit, a time marker, An energy integration characteristic calculation circuit including an analysis scheduled data memory is provided, and time-continuous detection data is subdivided into short monitoring unit time regions, and the level of a detection signal is monitored for each monitoring unit time region. "Normal" if no signal level exists even in the unit time domain
And immediately terminate the monitoring work.If the level fluctuation of the detected signal is large, it is determined that interfering sounds are mixed, and this data is excluded as a detection target and removed to monitor the next monitoring unit time domain data. The processing is continued, the required time is continuous, the level fluctuation width is small, and only the data for which it is difficult to determine no signal is analyzed by the high-precision frequency analysis method.

[2] In the signal detection device, a sensor, an energy integration circuit connected to the sensor,
An analysis schedule data memory connected to the sensor, an integration characteristic calculation circuit connected to the energy integration circuit, a determination circuit connected to the integration characteristic calculation circuit, and the analysis schedule data connected to the determination circuit Memory and
A frequency analysis circuit connected to the analysis-scheduled data memory, a display connected to the frequency analysis circuit, the determination circuit, and the analysis-scheduled data memory, the energy integration circuit, the integration characteristic calculation circuit, and the determination circuit; And a time marker that is connected to the time marker.

[0008]

Embodiments of the present invention will be described below in detail. Here, the signal is not specified as long as it can be captured by converting it into an electric signal such as a sound wave or a vibration wave. Further, in the following description, as an application example, a device for detecting water leak sound will be described. However, the present invention is not limited to only the method for detecting water leak sound, and generally, the above-described dark noise and The present invention can be applied to a case where the presence or absence of a specific signal is detected from a signal group mixed with an interfering sound.

[0009] In the case of the water leak sound detection method, the dark noise is generated when the used water flows through the water supply / distribution pipe, the noise of the used water, the sewage sound, the compressor operation sound of the vending machine on the beach, and the distance from the water supply / distribution pipe to be measured. It is a hybrid signal such as a vibration sound and other general noises that are continuously generated by the operation of a machine at a non-existent point. Interference noise includes running noise of vehicles traveling on the road surface near the sensor installed at the measurement target point, pile driving noise of construction work, and human footsteps.This level is transient and is usually dark noise. Larger than.

As described above, in the field, there is a background noise and a disturbing sound in addition to the water leakage sound which is an object of the water leakage sound detection knowledge. Background noise is classified into a signal having a certain level or higher (hereinafter, referred to as “present signal”) and a signal having a certain level or lower (hereinafter, referred to as “no signal”). The existing signal is a water leak sound at a certain level or more, a dark noise at a certain level or more, or a signal group in which these are mixed. The non-signal is a water leak sound having a certain level or less, a dark noise having a certain level or less, or a signal group in which these are mixed.

The signal detection method according to the present invention focuses on the temporal signal level fluctuation state and the mixing of transient and large-level interference sound data, and determines a predetermined short unit time (0. (About 5 seconds) (hereinafter referred to as "monitoring unit time")
The signal level is monitored for each area. First, it is determined whether the output waveform has no signal or signal. If no signal is detected even in one monitoring unit time, "Normal" immediately
("No water leak" by water leak sound detection knowledge) is determined. In this case, it is not necessary to perform the high-precision frequency analysis processing, and it is not necessary to wait for a predetermined time. If it cannot be determined to be “normal”, then it is determined whether there is no loud interfering sound at a high level and the level fluctuation range is continuously below a predetermined value. Determine.

If the fluctuation range is large and it is determined that the data is unsuitable for detection, the data is discarded, the monitoring of the data in the monitoring unit time is ended, and the next monitoring unit time area is automatically or manually operated immediately. Review the data and repeat this automatically or manually. Then, high-precision frequency analysis processing is performed only on the data in the area determined to be suitable for detection (signal data in which the ratio between the lowest level and the highest level falls within a predetermined range), and the data content is specified and identified.

The above-mentioned signal level for each area is expressed not by integration of the waveform amplitude at every moment, but by the integral of energy at each moment within each monitoring unit time area. With such a configuration, in the case of data containing low-level background noise (original data considered to be no signal), or data having a certain level or more but containing interfering sounds and having a large fluctuation range, There is no need to perform high-precision frequency analysis processing, and erroneous determination can be reduced. Also,
When there is no signal, it is immediately determined to be "normal", so that the analysis processing time can be significantly reduced. As a result, it is possible to provide an excellent device with a quick response and high reliability.

FIG. 2 is a signal detection flowchart showing an embodiment of the present invention. Hereinafter, the signal detection method according to the present invention will be described with reference to FIG. (1) First, the signal level is monitored for each monitoring unit time area. That is, it is monitored whether or not the sensor output waveform is a signal without a signal (step S1). (2) Next, it is checked whether or not there is no signal even in one monitoring unit time area (step S2). As a result, if there is no signal even in one monitoring unit time area, the detection is terminated.

(3) Next, it is checked whether or not there is no high-level interfering sound in a certain monitoring unit area, and whether the level fluctuation is continuously below a predetermined value (step S).
3). (4) Next, in step S3, if NO,
The monitoring in the monitoring unit area is completed, and the monitoring in the next monitoring unit area is examined (step S4). Repeat this.

(5) Next, in step S3, YE
In the case of S, high-precision frequency analysis processing is performed (step S5). FIG. 1 is a block diagram of an energy integration characteristic calculation circuit showing an embodiment of the present invention. In this figure, 1 is a sensor, 2 is an energy integration circuit, 3 is an integration characteristic calculation circuit, 4 is a judgment circuit, 5 is a time marker, 6 is an analysis scheduled data memory, 7 is a frequency analysis circuit, 8 is a display, ₁
The output waveform, S ₂ the time marker signal, S ₃ is reset pulse, S ₄ are energy integration course characteristic output, S ₅ are energy level per unit time, S ₆ the lowest output level throughout the entire area, S ₇ the main analysis result pulse, S ₈ is normality determination pulse, S ₉ is incapable of being analyzed result pulse, S ₁₀ is a fragmentary analytical data, S ₁₁ the analysis start instruction pulse, S ₁₂ is the determination data.

For example, the "output waveform S ₁ " of the sensor 1
Is the energy integration circuit 2 and the data memory 6 to be analyzed
Is connected to the input side, the energy "energy integration course characteristic output S _4" of the integrating circuit 2, integral characteristic is connected to the input side of the calculation circuit 3, the integrating characteristic calculation circuit which is the output of the 3 'each unit time energy The energy level S ₅ or the “minimum output level S ₆ throughout the entire area” is the “input side for the energy level per unit time” of the determination circuit 4 or the “input side for the minimum output level throughout the entire area”, respectively. Is connected.

Further, the "analysis determination pulse S ₇ " output from the determination circuit 4 is supplied to the "analysis determination pulse input side" of the analysis-scheduled data memory 6 by the "normal determination pulse S7" output from the determination circuit 4. "S ₈ " is the "normal judgment pulse input side" of the time marker 5, the analysis scheduled data memory 6, and the display 8.
The “analysis-impossible determination pulse S ₉ ” output from the determination circuit 4 is connected to the analysis-scheduled data memory 6 and the “analysis-impossible determination pulse input side” of the display 8. A certain “time marker signal S ₂ ” is connected to the integration characteristic calculation circuit 3 and the determination circuit 4.

The “reset pulse S ₃ ” is connected to the energy integrating circuit 2, and the “analysis required data S ₁₀ ” or the “analysis start instruction pulse S ₁₁ ” which is the output of the analysis scheduled data memory 6 is “Input side for analysis required data” or “Input side for analysis start instruction pulse” of the frequency analysis circuit 7
The "judgment data S ₁₂ " output from the frequency analysis circuit 7 is connected to the "judgment data input side" of the display 8.

In FIG. 1, the signal format in the signal processing may be either an analog type or a digital type, and is not particularly limited to either type. Next, the operation of the above-described energy integration characteristic calculation circuit will be described. Here, a description will be given of a method of detecting sound leakage sound. FIG. 3 is a diagram showing an example of a sensor output waveform, and FIG.
FIG. 5 shows a time marker signal and a reset pulse, and FIG.
FIG. 5A is a diagram showing a relationship between an energy integration progress characteristic and an energy amount level characteristic per unit time. FIG. 5A shows an energy integration level per unit time, and FIG. 5B shows an energy amount per unit time. Each level characteristic is shown.
FIG. 6 is a diagram illustrating an energy amount per unit time, which is a target of determination as to whether or not the high-precision frequency analysis processing is appropriate.

First, the “output waveform S ₁ ” of the sensor 1 installed for detecting and identifying a leak location is, for example, as shown in FIG.
Is shown in The “output waveform S ₁ ” is sent to the energy integration circuit 2 and the analysis scheduled data memory 6. As shown in FIG. 4, the time interval "time marker signal S _2" t _p
Is not specified here because it differs depending on the measurement target.However, it is necessary to use the interval where no interfering sound exists to detect the leakage location, so it is limited to a short time, and in many cases, It is considered that about 0.5 seconds is appropriate.

The energy integrating circuit 2 is activated by the “reset pulse S ₃ ”, calculates the momentary energy amount from the momentary amplitude waveform of the “output waveform S ₁ ”, and calculates the next “time marker signal S 3”. Until “ ₂ ”, the energy amount is integrated, and the energy amount data is held in the energy integration circuit 2. When the next “time marker signal S ₂ ” appears, the result of the final level obtained by integrating each monitoring unit time (ie, the final value of the energy integration progress characteristic output S ₄ for each monitoring unit time) is calculated. The signal is transferred to the integration characteristic calculation circuit 3 and, at the same time as the disappearance of the “time marker signal S ₂ ”,
The integrated energy amount is reset by the “reset pulse S ₃ ” of the energy integration characteristic, which is transmitted with a delay of the reset pulse delay time t _d from the “time marker signal S ₂ ”, and is reset again in the next unit time. The operation for integrating the amount of energy is started.

As shown in FIG. 4, the "reset pulse S ₃ " is transmitted after being delayed by the delay time t _d of the reset pulse so as not to overlap with the "time marker signal S ₂ " in time. Due to the “reset pulse S ₃ ”, the integral has already been added and does not affect the data sent to the integral characteristic calculation circuit 3. Every time the time marker signal S _{2 is} received from the time marker 5, the integration characteristic calculation circuit 3 works on and holds the level of the “energy integration progress characteristic output S ₄ ” at that time from the energy integration circuit 2. It is sent to the determination circuit 4 as the energy level S ₅ per unit time, and is held until the next “time marker signal S ₂ ” is received.

At the same time, the data held as "the lowest output level S ₆ throughout the entire area" is also determined by the determination circuit 4.
To provide. If the “energy level S ₅ per unit time” is lower than the “minimum output level S ₆ throughout the entire area”, the value of the “minimum output level S ₆ throughout the entire area” is also changed. The detected value is corrected to the “energy level S ₅ per unit time” and provided to the determination circuit 4 as “minimum output level S ₆ ”.

As shown in FIG. 5B, the level in the integration characteristic calculation circuit 3 changes stepwise in units of the monitoring unit time with reference to the "time marker signal S ₂ ". An energy amount level characteristic curve for each is obtained. That is, the low a _i of the energy amount level characteristic curve
Is the energy level when there is no signal or signal, and b _j having a high energy amount characteristic curve is no signal + interference sound or signal +
As can be seen from FIG. 5, the energy amount level at the time of the interference sound indicates that the characteristic curve of the “energy amount level S ₅ per unit time” is shorter than the corresponding data in the “energy integration progress characteristic output S ₄ ”. Is delayed by one monitoring unit time.

If the output level value a _{i of the} “energy level S ₅ per unit time” can be regarded as a no-signal level, the determination circuit 4 determines that there is no signal, and returns to the “normal determination pulse S ₈ Is sent to the time marker 5, the analysis-scheduled data memory 6, and the display 8 to end the monitoring operation.
As shown in FIG. 6, the propriety of the high-precision frequency analysis processing is determined based on the energy level b _j when (no signal + disturbance sound or signal + interference sound) and the minimum energy level a _min when there is no signal or signal. The ratio b _j / a _min is the allowable maximum level ratio b
If it exceeds / a _min , it is recognized that interfering sounds with a high level are mixed, and an “analysis impossible determination pulse S ₉ ” is sent to the analysis scheduled data memory 6 and the display 8, and within this monitoring unit time, The data is discarded, data for the next monitoring unit time is waited, and the process is repeated until b _j / a _min becomes equal to or less than a predetermined value.

"Energy level S ₅ per unit time"
Output level value a _i exceeds the no-signal level, and b _j / a
If _min is equal to or below the allowable maximum level ratio b / a _min, it sends a "principal analytical determination pulse S _7" to analyze schedule data memory 6, that is continuously "necessary analysis result pulse S _7" is sent Expect and wait for data within the following monitoring unit time. In time marker 5, sent to monitor the output waveform S ₁ and subdivision for each monitoring unit time, the "time marker signal S _2" to the integral characteristic calculation circuit 3 and the decision circuit 4 in the monitoring unit time t _p interval Then, the “reset pulse S ₃ ” is sent to the energy integration circuit 2 to perform the monitoring process for each monitoring unit. However, when the time marker 5 receives the “normality determination pulse S ₈ ” from the determination circuit 4, the time marker 5 stops sending the “time marker signal S ₂ ” and the “reset pulse S ₃ ” and ends the system monitoring operation. . In this case, the display 8 shows a normal display.

[0028] In the analysis schedule data memory 6, waits for the instruction from the determination circuit 4 while inputting the "output waveform S _1" at every monitoring unit time, when the "analysis allowed determining pulse S _9" is input, to date discard analyze schedule data entered "output waveform S _1", again, expected for each monitoring unit time while inputting the "output waveform S _1" is input continuously "necessary analysis result pulse S _7" I do.

The "analysis required determination pulse S ₇ " is counted by the "analysis required instruction pulse counter" of the analysis scheduled data memory 6, and the "analysis required determination pulse S ₇ " is continuously input to the required number. Once you reach, "the analysis start instruction pulse S _11"
Is generated and input to the frequency analysis circuit 7. Since the required number n is a condition that data of the time t _R required for the frequency analysis is obtained, the following equation is obtained in consideration of the time delay t _d (= + α) due to the data processing.

N ≧ (t _R + t _p + α) / t _p (1) For each leak sound detection knowledge, t _p = 0.5 sec, t _R = 1
It is sufficient to consider about n ≧ 22 as 0 seconds. When the "normality determination pulse S ₈ " is input, the analysis schedule data memory 6
Clears the secured data and stops the operation. "Normal" is displayed on the display 8, and the system also ends the monitoring operation.

When the analysis start instruction pulse S ₁₁ is input from the analysis scheduled data memory 6, the frequency analysis circuit 7 fetches a required amount of “analysis required data S ₁₀ ” from the analysis scheduled data memory 6. Perform high-precision frequency analysis processing. This high-precision frequency analysis processing is disclosed in detail in Japanese Patent Application Laid-Open No. 7-318451, entitled "Method of identifying water leakage noise by high-precision frequency analysis", which has already been proposed by the present inventors.

As a result of such high-precision frequency analysis processing, "judgment data (one of normal / remeasurement / leakage)"
S ₁₂ ”is supplied to the display 8, a display corresponding to the judgment data is made, and the system monitoring operation is terminated. According to the present invention, an energy integration circuit, an integration characteristic calculation circuit, a determination circuit, a time marker, and an analysis schedule data memory are provided, subdivided into extremely short monitoring unit time regions, and monitored signal levels are monitored. First, it is determined whether there is no signal or signal from the energy integration characteristic of
If no signal level data is detected even in one monitoring unit time domain, it is immediately determined to be no signal ("normal"), and the monitoring operation is terminated.

In the case of a signal, the presence / absence of an interfering sound is determined based on the fluctuation range of the level. If the “ratio of the detected signal level to the lowest level detected so far” is larger than a predetermined level ratio, In order to determine that interfering sounds are mixed and to eliminate erroneous recognition caused by the interfering sounds, this data is discarded, and the data is moved to the data of the next monitoring unit time area again to perform monitoring work.

By this processing, it is possible to detect the presence / absence of a specific signal with high reliability and extremely high work efficiency even in the case where an interfering sound is generated or a mixed signal or non-signal is present. Can be. It should be noted that the present invention is not limited to the above embodiment, and various modifications are possible based on the spirit of the present invention, and these are not excluded from the scope of the present invention.

[0035]

As described above in detail, according to the present invention, even when an interfering sound is generated or a signal or a non-signal is mixed, the reliability is high, and It is possible to detect the presence or absence of an effective signal with extremely high work efficiency.

[Brief description of the drawings]

FIG. 1 is a block diagram of an energy integration characteristic calculation circuit showing an embodiment of the present invention.

FIG. 2 is a signal detection flowchart showing an embodiment of the present invention.

FIG. 3 is a diagram showing an example of a sensor output waveform.

FIG. 4 is a diagram showing a time marker signal and a reset pulse.

FIG. 5 is a diagram showing a relationship between an energy integration progress characteristic and an energy amount level characteristic per unit time.

FIG. 6 is a diagram showing an energy amount per unit time, which is a target of a determination as to whether or not a high-precision frequency analysis process is appropriate;

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Sensor 2 Energy integration circuit 3 Integration characteristic calculation circuit 4 Judgment circuit 5 Time marker 6 Analysis planned data memory 7 Frequency analysis circuit 8 Display S ₁ Output waveform S ₂ Time marker signal S ₃ Reset pulse S ₄ Energy integration progress characteristic output S Energy level per ₅ unit time S ₆ Minimum output level throughout the whole area S ₇ Analysis required pulse S ₈ Normal determination pulse S ₉ Analysis impossible determination pulse S ₁₀ Analysis required data S ₁₁ Analysis start instruction pulse S ₁₂ Determination data

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Eki Obata, Inventor Eiki 1-7-12 Toranomon, Minato-ku, Tokyo Oki Electric Industry Co., Ltd. (72) Inventor Hideo Kimura 1-7-112 Toranomon, Minato-ku, Tokyo Inside Electric Industry Co., Ltd. (72) Inventor Tsuneo Miyamoto 1-12-6 Michiba, Yoshikawa-shi, Saitama (72) Inventor Tohru 7-30-4 Horikiri, Katsushika-ku, Tokyo (72) Inventor Shunji Tanamura Suginami, Tokyo 3-8-5 Izumi Ward F-term (reference) 2G067 BB11 CC02 DD13 EE04 EE05

Claims (2)

[Claims] 1. A sensor, an energy integration circuit, an integration characteristic calculation circuit, a determination circuit, and an energy integration characteristic calculation circuit including a time marker and an analysis scheduled data memory are provided. The detected data is subdivided into short monitoring unit time regions, and the level of the detection signal is monitored for each monitoring unit time region. (C) If there is no signal level even in one monitoring unit time region, “normal” (D) When the level fluctuation of the detected signal is large, it is determined that interfering sounds are mixed, and the data is removed as a non-detection target, and the next monitoring unit time domain data And (e) analyzing only data that is continuous for a required time, has a small level fluctuation width, and is difficult to determine the absence of a signal by a high-precision frequency analysis method. 2. A sensor, (b) an energy integration circuit connected to the sensor, (c) an analysis data memory connected to the sensor, and (d) an energy integration circuit connected to the energy integration circuit. (E) an integration characteristic calculating circuit;
A determination circuit connected to the integration characteristic calculation circuit; (f) the analysis-scheduled data memory connected to the determination circuit;
(G) a frequency analysis circuit connected to the analysis-scheduled data memory; (h) a display connected to the frequency analysis circuit, the determination circuit, and the analysis-scheduled data memory; and (i) the energy integration circuit. A signal detection device comprising a time marker connected to the integration characteristic calculation circuit and a determination circuit.