JP2003004476A - Search-type signal detecting device, signal detecting method, signal detecting program, and recording medium for recording the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a signal detecting device and a method which utilize inclination and amplitude and will not beffected by noise or sampling frequencies. SOLUTION: The search-type signal detecting method includes first to fourth steps. In the first step, in the case that a signal value of a range terminal point determined by an analysis start point, an analysis range is greater than the sum of a signal value of the analysis start point and a threshold, the analysis start point is determined as a temporary start point, and the range terminal point is as a temporary peak. In the second step, a true start point is determined within a range from the temporary start point to the temporary peak. In the third step, a second temporary peak is determined to be within a range specified by a peak detecting range from the true start point. In the fourth step, a true peak is determined to be within a range from the second temporary peak to the true start pointor the temporary peak.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a signal detecting device for detecting a specific waveform portion from digital signal data, a signal detecting method, a signal detecting program, and a recording medium recording the program.

[0002]

2. Description of the Related Art The amplitude and slope of a signal can be used to detect the desired waveform portion from the signal waveform. FIG. 9 is a waveform diagram for explaining a conventional signal detection method, in which the vertical axis y and the horizontal axis t represent signal strength and time, respectively.
The downward direction of the vertical axis y and the right direction of the horizontal axis are displayed as positive.

First, the signal value at a point advanced from the designated temporary start point U by the analysis range r ₀ is the value (y) of the temporary start point U.
If it is larger than _U ) + threshold value (y _th ), the point is set as the true starting point. Next, the point W at which the signal becomes maximum is detected within the set time t ₀ from the true start point (point V). As a result, the coordinates of two points V (t _V , y _V ), W (t _W ,
y _W ), the slope of the waveform of this portion is calculated as (y _W −y _V ) / (t
_W −t _V ). That is, it is possible to detect the waveform portion having the designated inclination.

As a method which is an improvement of the above method, a point obtained by advancing data one by one from the starting point is set as a target point, and a threshold value is set as a difference between a signal value of the target point and a signal value of a point adjacent to the target point. When the value exceeds, the method of using the point of interest as the true starting point is also used.

[0005]

In the signal detection method described above, a signal having a lot of noise is affected by the noise and the detection accuracy is deteriorated. Therefore, it is necessary to preprocess the signal with a low-pass filter in advance. was there. Further, since the peak value is determined simply by taking the maximum value within the specified range, there is a problem that an erroneous peak is detected due to the influence of noise. Further, there is a problem that the signal detection accuracy is affected by the data sampling frequency.

In order to solve the above-mentioned problems, the present invention is directed to a signal detecting device, a signal detecting method, a signal detecting program, and a recording medium having the program recorded therein, which are not affected by noise or sampling frequency and which use inclination and amplitude. The purpose is to provide.

[0007]

The objects of the present invention are achieved by the following. That is, according to one aspect of the present invention, there is provided a signal detection device for detecting a partial waveform having a specified slope from a digital signal waveform, comprising a data sampling section for sampling an analog signal and converting it into digital signal data. A recording unit that records the signal data corresponding to a number, a memory that reads the signal data from the recording unit and temporarily stores the signal data, and a processing unit that performs arithmetic processing on the signal data stored in the memory. If the signal data of the range end sampling number determined by the analysis start sampling number and the analysis range is larger than the sum of the signal data of the analysis start sampling number and the threshold value, the processing unit includes the analysis start sampling number. Is the temporary start sampling number and the range end sampling number is the temporary peak Determined as sampling number, to determine the true start sampling number in the range to peak sampling number of the temporary from the temporary start sampling number, the second in the range specified by the peak detection range of the start sampling number of said vacuum
Search signal detecting apparatus for determining a true peak sampling number within the range from the second temporary peak sampling number to the true start sampling number or the temporary peak sampling number Can be provided.

In the determination of the true start sampling number, the processing unit determines from the signal data in a predetermined range with the first sampling number moved in the first direction of the time axis from the temporary start sampling number as an end point. When the calculated average value is larger than the sum of the signal data of the temporary starting sampling number and the noise value, the first sampling number is set as the true starting sampling number, and the second temporary peak sampling number is determined. In the above, the average value calculated from the signal data of the predetermined range whose end point is the second sampling number moved in the first direction from the true start sampling number is the peak noise from the signal data of the second sampling number. If it is smaller than the subtraction value obtained by subtracting the value, the second sampling number is set to the second sampling number.
Can be a temporary peak sampling number.

Further, in the determination of the true peak sampling number, the processing section uses a third sampling number moved in a direction opposite to the first direction from the second temporary peak sampling number as an end point. When the average value calculated from the signal data of the predetermined range is smaller than the signal data of the second tentative peak sampling number, the third sampling number can be regarded as the true peak sampling number. .

According to another aspect of the present invention, there is provided a signal detection method for detecting a partial waveform having a specified slope from a signal waveform, wherein a signal value at a range end point determined by an analysis start point and an analysis range. Is greater than the sum of the signal value at the analysis start point and the threshold value, the analysis start point is determined as a temporary start point and the range end point is determined as a temporary peak.
And a second step of determining a true start point in the range from the temporary start point to the temporary peak, and a second temporary value in the range specified by the peak detection range from the true start point. Search signal detection including a third step of determining a peak of the second peak and a fourth step of determining a true peak in a range from the second temporary peak to the true start point or the temporary peak. A method can be provided.

In the second step, the average value calculated from the signal values in a predetermined range with the first point moved in the first direction of the time axis from the temporary start point as an end point is the temporary start. If it is larger than the sum of the signal value of the point and the noise value, then the first
Is set as a true starting point, and in the third step, an average value calculated from signal values in a predetermined range with the second point moved from the true starting point in the first direction as an end point is When it is smaller than the subtraction value obtained by subtracting the peak noise value from the signal value of the second point, the step of setting the second point as the second temporary peak can be performed.

In the fourth step, an average value calculated from signal values in a predetermined range with a third point moved from the second tentative peak in a direction opposite to the first direction as an end point. Is smaller than the signal value of the second tentative peak, the step of setting the third point as a true peak can be performed.

According to another aspect of the present invention, a signal detecting device for detecting a partial waveform having a specified slope from a signal waveform is provided with a signal value at a range end point determined by an analysis start point and an analysis range. Is greater than the sum of the signal value at the analysis start point and the threshold value, the analysis start point is determined as a temporary start point and the range end point is determined as a temporary peak.
A second function for determining a true start point in the range from the temporary start point to the temporary peak, and a second function in a range specified by the peak detection range from the true start point.
Search-type signal that realizes a third function for determining a tentative peak of, and a fourth function for determining a true peak in the range from the second tentative peak to the true start point or the tentative peak. A detection program can be provided.

In the second function, the average value calculated from the signal values in the predetermined range with the first point moved in the first direction of the time axis from the temporary start point as an end point is the temporary start. When it is larger than the sum of the signal value and the noise value of the point, it is a function of setting the first point as a true starting point, and the third function is moving from the true starting point in the first direction. When the average value calculated from the signal values in the predetermined range with the second point as the end point is smaller than the subtraction value obtained by subtracting the peak noise value from the signal value at the second point, the second value Can be used as the second temporary peak.

Further, the fourth function is an average value calculated from signal values in a predetermined range with an end point at a third point moved in a direction opposite to the first direction from the second tentative peak. Is smaller than the signal value of the second tentative peak,
A function of making the third point a true peak can be provided.

Further, it is possible to provide a computer-readable recording medium in which the above program is recorded.

[0017]

1 is a block diagram showing the configuration of a search-type signal detection apparatus according to an embodiment of the present invention. The signal detection device 100 includes a processing unit 101 and a memory 10.
2, a recording unit 103, a video memory 104, a video output unit 105, an interface unit 106, a data sampling unit 107, a clock generation unit 108, and a bus 109.

The clock generation unit 108 includes a processing unit 101,
It generates a system clock necessary for the operation of each unit such as the memory 102, a sampling clock supplied to the data sampling unit 107 for determining the timing of data collection, and a clock supplied to the video output unit 105 for video encoding. The bus 109 is a plurality of signal lines for transmitting data, addresses and clocks.

The analog signal measured by the measuring device 110 is transmitted to the data sampling section 107, and in the data sampling section 107, it is sampled at fixed time intervals over a designated time and converted into digital data. The converted data is recorded in the recording unit 103 via the bus 109.

After the measurement is completed, the sampling data recorded in the recording unit 103 is transferred to the memory 102,
Processing such as scale conversion of the signal strength is performed, and the two-dimensional image data representing the signal waveform is recorded in the video memory 104. The data in the video memory 104 is read by the video output unit 105, D / A converted and video-encoded, transferred to the display device 112, and displayed as an image of a signal waveform.

The input / operation device 111 is a signal detecting device 1.
00 is a device for inputting measurement parameters, signal analysis parameters, and the like via the interface 106, and for operating the signal waveform displayed on the display device 112.

FIG. 2 is a flow chart showing a detection process performed by the processing unit 101 for reading the signal data recorded in the recording unit 103 into the memory 102 and performing the same in the search type signal detection apparatus according to the present invention. is there. Figures 3-7
Represents the measured signal waveform, and the axis is set similarly to FIG.

The processing of each step in FIG. 2 will be described below with reference to FIGS. Since the signal is sampled at constant time intervals, the time axis can be represented by a sampling number corresponding to the sampling order. In the following description, the time axis t in FIGS. .

As parameters necessary for analysis,
The analysis range r ₀ , the threshold value S _th , the noise level S _n , the peak detection noise level S _pn , the peak detection range r _p , and the average value calculation range b are designated in advance from the input / operation device 111. In this case, by the signal detection processing described below,
A waveform portion whose slope is S _th / r ₀ or more is detected.

In step 201, 0 is set as the initial setting for the sampling number i at the starting point of the process and the number N assigned to each detected peak.

[0026] In step 202, compared to the signal values y (i) at the sampling number i, the signal value at i from the analysis range r ₀ only large sampling number (i + r ₀₎ y and (i + r _0), the signal value y ( i + r ₀ ) is larger than the signal value at the sampling number i + the threshold value, that is, y
If it is determined that (i + r ₀ )> y (i) + S _th , the process _proceeds to step 204. If it is judged that y (i + r ₀ )> y (i) + S _th is not satisfied, i is incremented by 1 in step 203 and the process returns to step 202. FIG. 3 shows a state where points A and B satisfy the determination condition of step 202. At this time, the point A is a “temporary start point” and the point B is a “temporary peak”.

In step 204, points AB of FIG.
In order to analyze the section of, the sampling number i of the point A is set as the initial value of the repeated counter j.

In step 205, the sampling number (j) at the end of the section j to j + b where the average value of the signal is calculated.
+ B) is larger than the sampling number (i + r ₀ ) at the point B. Here, b and r ₀ are b <r ₀
It is set to satisfy the condition of. When it is determined that the end sampling number does not exceed the sampling number of point B, the process proceeds to step 206, and when it is determined that it does not, step 206 cannot be performed, so the process proceeds to step 208. To do.

In step 206, the average value of the data in the section of sampling numbers j to j + b is calculated, and the average value, the signal value y (i) of the temporary starting point A and the noise level S _n are calculated.
When it is determined that the average value is greater than or equal to y (i) + S _n , the process proceeds to step 208. When it is determined that the average value is less than y (i) + S _n , the process proceeds to step 207, j is incremented by 1, and the process returns to step 205. FIG. 4 shows a state where the point C satisfies the condition of step 206, and the point C is a “true starting point”.
And

In step 208, step 206
The sampling number j of the point C is recorded in the recording unit 103 as the sampling number t _O (N) of the true starting point detected in 1., and the signal value y _O (N) of the true starting point C is recorded in the recording unit 103. . Here, N = 0 if point C is the first true starting point detected in the signal waveform.

In step 209, in order to detect the "true peak" by analyzing the data within the peak detection range r _p from the true start point C (data between points C to D in FIG. 5), The sampling number j = t ₀ (N) corresponding to the true starting point C is set in the counter m. here,
It is desirable that r _p is set so as to satisfy r _p > r ₀ .

In step 210, step 205
Similarly to, the end of the section for calculating the average value (sampling number = m + b) is the point D (sampling number = j +
determines whether or not beyond the r _p). If it is determined that the number has not exceeded, the process proceeds to step 212. If it is determined that the number exceeds the limit, step 212 cannot be performed. Therefore, the process proceeds to step 211, i + r ₀ is set as the true peak sampling number t _p (N), and the true peak signal value y _p ( N) the signal value y (i + r ₀ ) as the recording unit 103
, And then the process proceeds to step 219. here,
If it is the first true peak detected in the signal waveform, then N = 0.

In step 212, the average value of the data in the section of sampling numbers m to m + b is calculated, and the noise level at peak detection S _pn is calculated from this average value and the signal value y (m).
Is compared with the value obtained by subtracting, and the average value is y (m) -S
If it is less than or equal to _{pn, the process proceeds} to step 214. If the average value is larger, the process proceeds to step 213 and m is set to 1
Only, and returns to step 210. FIG. 5 shows a state in which the point E satisfies the determination condition of step 212, and the point E is the “second temporary peak”.

In step 214, the counter n is set to m.
Set. At this point, m is the sampling number at point E.

At step 215, since the analysis is performed in the direction in which the sampling number is subtracted from the second tentative peak (point E) (leftward in FIG. 6), the left end of the section nb to n for calculating the average value is calculated. (Sampling number = n−b)
, Is located to the left of the true starting point C. Here, the sampling number j of the point C is j = t ₀ (N)
Is. When it is determined that n−b is t ₀ (N) or more,
That is, when it is determined that the left end for calculating the average value is not located to the left of the point C, the process proceeds to step 216.
When it is determined that n−b is smaller than t ₀ (N), that is, when the left end for calculating the average value is located to the left of the point C, the process of step 216 may be performed. Therefore, the process proceeds to step 217, and n is recorded as the true peak sampling number t _p (N) and the signal value y (n) is recorded as the true peak signal value y _p (N) in the recording unit 103. Then, the process proceeds to step 219.

In step 216, the average value of the data in the section of the sampling numbers nb to n is calculated, and this average value is compared with the signal value y (m) of the second tentative peak.
If it is determined that the average value is larger, step 21
8, the n is decreased by 1, and the process returns to step 215. When it is determined that the average value is less than or equal to y (m), the process proceeds to step 217, where n is set as the true peak sampling number t _p (N), and the true peak signal value y _p (N) is set. Then, the signal value y (n) is recorded in the recording unit 103, and then the process proceeds to step 219. FIG. 6 shows a state in which the point F satisfies the determination condition of step 216, and the point F is a true peak.

In step 219, it is confirmed whether or not the difference between the signal value of the true peak (point F) and the signal value of the true start point (point C) is equal to or more than the set threshold value S _th .
When it is determined that the condition of step 219 is not satisfied, the process proceeds to step 203, i is incremented by 1, and steps after step 202 are repeated. When it is determined that the condition of step 219 is satisfied, it means that the signal portion having the slope equal to or _larger than the desired slope (S _th / r ₀ ) can be detected.

Finally, in step 220, when it is determined that the data to be detected does not have any remaining, the signal detection processing is ended, and when it is determined that the data to be detected has the remaining, the section already processed. In order to restart the detection process from the sampling number exceeding ₀ , the counter i is set to the sampling number t ₀ (N) of the true peak (point F) in step 221, and the process proceeds to step 203. As a result, the steps after step 202 are repeated until there is no data to be detected.

As described above, it is possible to detect, from the measured signal waveform, a portion having a slope equal to or more than the designated slope as the true start point and the true peak, and to correspond to the number N assigned to each peak. Then, the sampling number and the signal value of the true start point and the true peak can be recorded in the recording unit 103.

In the above description, in order to perform the analysis in step 216 in the range of points E to C, the case of determining in step 215 whether the left end of the analysis range is located to the left of point C has been described. However, the process of step 216 is point E
You may limit to the range of -B. In that case, the determination condition of step 215 is n−b <i + r ₀ .

In the above description, the detection process is performed by moving the point of interest in the direction in which the sampling number increases. However, starting from the maximum sampling number,
Similar detection processing is possible even if the point of interest is moved in the direction in which the sampling number decreases.

The analysis range r ₀ , the signal threshold S _th , the noise level S _n , the peak detection noise level S _pn , the peak detection range r _p , and the average value calculation range b, which are the parameters to be set first, are the detection target. It is desirable to set a numerical value that takes into account the characteristics of the signal. For example, the average value calculation range b,
It is desirable that the analysis range r ₀ and the peak detection range r _p are set so as to satisfy the relationship of b <r ₀ <r _p . Also,
The noise level S _n and the noise level at the time of peak detection S _pn are
Although it is possible to specify a separate value, S _n such that S _n> S _pn is the case of the signal waveforms for neuronal activity, S
It is desirable to set _pn .

In the above, the case where the values of the signal data y (i) are all positive has been described. However, when the value of y (i) becomes negative or includes positive and negative values, all the data are written. If you add an appropriate value to make all data positive or negative and then take an absolute value, convert all data to a positive value and then process it, analyze it in the same way as above. Is possible.

[0044]

The search-type signal detection apparatus according to the present invention can reduce the influence of noise in the signal detection processing by using the average value of the signals calculated within the set average value calculation range. . Therefore, it is not necessary to process the signal with a low-pass filter in advance to reduce noise.

Since the processing is relatively simple, extremely high speed processing is possible. Furthermore, the calculation of the average value is
Since it is performed only in the vicinity of the temporary start point and the temporary peak after detecting the temporary start point and the temporary peak, it is more efficient than the case where the entire data is processed by the low-pass filter. In particular, this is effective in the analysis of a signal in which a significant signal is observed for a short period and a period equal to or less than the set detection threshold is long.

Further, by performing a backward search for analyzing from the second tentative peak in the direction opposite to the direction of the time axis, a correct peak value is detected in the case of a trapezoidal waveform as shown in FIG. It is possible. In FIG. 8, a point P1 is a peak that is erroneously detected under the influence of noise when the conventional detection method is used. Point P2
Are peaks that were erroneously detected when the backward search was not performed. On the other hand, P3 is a true peak detected by the signal detecting device according to the present invention,
This makes it possible to detect the desired tilted portion with high accuracy.

The search-type signal detection device according to the present invention has an effect of being able to detect all signals having a steeper slope than the slope specified by the user.

Since no specific waveform model is assumed, it can be applied to all signals characterized by the slope and the amplitude by setting the detection parameters appropriately. For example, if the average value calculation range is set appropriately for signals observed from the living body such as extracellular potentials and synaptic currents, accurate signal detection is possible for data collected at various sampling frequencies. .

[Brief description of drawings]

FIG. 1 is a block diagram showing an internal configuration of a search-type signal detection device according to an embodiment of the present invention.

FIG. 2 is a flowchart showing processing performed by a processing unit of the search-type signal detection device according to the embodiment of the present invention.

FIG. 3 is a waveform diagram showing a state in which a temporary start point (point A) and a temporary peak (point B) are detected.

FIG. 4 is a waveform diagram showing a state in which a true start point (point C) is detected.

FIG. 5 is a waveform diagram showing a state in which a second temporary peak (point E) is detected.

FIG. 6 is a waveform diagram showing a state in which a true peak (point F) is detected.

FIG. 7 is a waveform diagram showing the finally detected true start point (point C) and true peak (point F).

FIG. 8 is a waveform diagram showing peaks detected by different detection methods for the same signal waveform.

FIG. 9 is a waveform diagram showing a conventional signal detection method.

[Explanation of symbols]

100 signal detector 101 processing unit 102 memory 103 storage 104 video memory 105 Video output section 106 interface 107 data sampling unit 108 clock generator 109 bus 110 Measuring device 111 Input / operation device 112 display device

Claims (12)

[Claims] 1. A signal detection device for detecting a partial waveform having a specified slope from a digital signal waveform, comprising: a data sampling section for sampling an analog signal and converting it into digital signal data; The processing unit includes a recording unit that records signal data, a memory that reads the signal data from the recording unit and temporarily stores the signal data, and a processing unit that performs arithmetic processing on the signal data stored in the memory. If the signal data of the range end sampling number determined by the analysis start sampling number and the analysis range is larger than the sum of the signal data of the analysis start sampling number and the threshold value, the analysis start sampling number is set to a temporary start sampling number. , The range end sampling number is a temporary peak sampling number No., the true start sampling number is determined in the range from the provisional start sampling number to the provisional peak sampling number, and the true start sampling number is determined to be the second in the range specified by the peak detection range. A search, characterized in that a temporary peak sampling number is determined, and a true peak sampling number is determined in a range from the second temporary peak sampling number to the true start sampling number or the temporary peak sampling number. Type signal detector. 2. The processing unit, in the determination of the true start sampling number, a signal in a predetermined range with a first sampling number moved in the first direction of the time axis from the temporary start sampling number as an end point. If the average value calculated from the data is larger than the sum of the signal data of the temporary start sampling number and the noise value, the first sampling number is set as the true starting sampling number, and the second temporary peak sampling number is set. In the determination of, the average value calculated from the signal data of the predetermined range having the second sampling number moved in the first direction from the true start sampling number as an end point is calculated from the signal data of the second sampling number. When it is smaller than the subtraction value obtained by subtracting the peak noise value, the second sampling number is set to the second temporary peak. Exploratory signal detection apparatus according to claim 1, characterized in that the sampling number. 3. The processing unit determines, in the determination of the true peak sampling number, a third sampling number moved in a direction opposite to the first direction from the second tentative peak sampling number as an end point. The average value calculated from the signal data of the predetermined range
3. The third sampling number is set as a true peak sampling number when the signal data is smaller than the signal data of the second tentative peak sampling number.
The search-type signal detection device described in 1. 4. A signal detection method for detecting a partial waveform having a specified slope from a signal waveform, wherein a signal value at a range end point determined by an analysis start point and an analysis range is a signal value at the analysis start point. And a threshold value greater than the sum of the thresholds, a first step of determining the analysis start point as a temporary start point and the range end point as a temporary peak, and a range from the temporary start point to the temporary peak. A second step of determining a true starting point; a third step of determining a second tentative peak within the range specified by the peak detection range from the true starting point; and the second tentative peak To the true start point or the tentative peak, and a fourth step of determining a true peak. 5. The first step of moving in the first direction of the time axis from the temporary starting point in the second step.
If the average value calculated from the signal value in a predetermined range with the point as the end point is larger than the sum of the signal value and the noise value at the temporary start point, the first point is set as the true start point, and In the step of 3, an average value calculated from signal values in a predetermined range with the second point moved in the first direction from the true start point as an end point is
The second point is defined as a second tentative peak when it is smaller than a subtraction value obtained by subtracting a peak noise value from the signal value of the second point. Search type signal detection method. 6. The fourth step is an average value calculated from signal values in a predetermined range having an end point at a third point moved in a direction opposite to the first direction from the second tentative peak. Is smaller than the signal value of the second tentative peak, the third point is set as a true peak, and the search-type signal detection method according to claim 5. 7. A signal detection device for detecting a partial waveform having a specified slope from a signal waveform, wherein a signal value at a range end point determined by an analysis start point and an analysis range is a signal value at the analysis start point and a threshold value. A first function of determining the analysis start point as a provisional start point and the range end point as a provisional peak, and a true start in the range from the provisional start point to the provisional peak. A second function of determining a point, a third function of determining a second provisional peak in a range specified by a peak detection range from the true start point, a true start of the second provisional peak A search-type signal detection program, which realizes a fourth function of determining a true peak in a range up to a point or the temporary peak. 8. The first function of moving the second function in a first direction of a time axis from the temporary starting point.
When the average value calculated from the signal values in the predetermined range with the point of as the end point is larger than the sum of the signal value of the temporary start point and the noise value, the function of setting the first point as the true start point is used. And the third function is that an average value calculated from signal values in a predetermined range with a second point moved in the first direction from the true start point as an end point is,
The function of setting the second point as a second temporary peak when the difference is smaller than the subtraction value obtained by subtracting the peak noise value from the signal value of the second point. 7
The search-type signal detection program described in 1. 9. The fourth function is an average value calculated from signal values in a predetermined range with an end point at a third point moved in a direction opposite to the first direction from the second tentative peak. 9. The search-type signal detection program according to claim 8, characterized in that, when is smaller than the signal value of the second tentative peak, the third point is a true peak. 10. A computer-readable recording medium in which the program according to claim 7 is recorded. 11. A computer-readable recording medium in which the program according to claim 8 is recorded. 12. A computer-readable recording medium in which the program according to claim 9 is recorded.