JP7103069B2 - Signal processing equipment, signal processing method, and signal processing method program - Google Patents

Description

The present invention relates to a signal processing apparatus, a signal processing method, and a program of the signal processing method. In particular, it relates to an identity determination for determining whether or not the same signal is included in a plurality of signals.

For example, there is a signal processing device that processes signals such as sound received by a receiver, identifies a target object, estimates the direction of the target object, and the like. Here, in the signal processing device, an identity determination process for determining whether or not the same signal is included in the signals received by the plurality of receivers, such as being emitted from the same sound source, is performed (for example,). , Patent Document 1).

Japanese Unexamined Patent Publication No. 2003-139856

Here, the identity determination process by calculation using the cross-correlation function will be described. Assuming that the signals output by the receiver a and the receiver b are x _a [n] and x _b [n], respectively, the cross-correlation functions r _{xa and xb} between the receiver a and the receiver b [ τ] is represented by the equation (1).

When the signal x _a [n] and the signal x _b [n] contain the same signal, the cross-correlation functions r _{xa and xb} [τ] are the difference in arrival time between the signals in the receiver a and the receiver b. Has a peak in the sample according to. Then, if a threshold value γ _r that can be regarded as the same signal is set by cross-correlation, the identity determination can be performed by the equation (2). Here, max is a function that outputs the maximum value in the input sample. Further, flag _r is a flag indicating the result of the identity determination.

As shown in the equation (2), when it is determined that the same signal is included in the signal x _a [n] and the signal x _b [n], flag _r = 1. Further, when it is determined that the same signal is not included in x _b [n], flag _r = 0 is set. In this way, by using the cross-correlation function, it is possible to perform the identity determination process for determining whether or not the signals output from the receiver a and the receiver b include the same signal.

However, in the identity determination process using the cross-correlation function described above, a peak value is obtained by the integral effect as shown in the equation (1). Therefore, when the duration of the common signal is long, it is considered that the identity determination can be performed as expected. On the other hand, when a common signal having a relatively short duration is the target, the accuracy of the identity determination is significantly reduced because the determination is made with a small integral value.

Therefore, it has been desired to realize a signal processing device, a signal processing method, and a program of a signal processing method capable of performing more accurate identity determination.

The signal processing device according to the present invention is a signal processing device that performs identity determination processing for determining whether or not the same signal is included in a plurality of signals, and a plurality of types of feature quantities can be obtained from a plurality of signals. Regarding the correlation between the feature extraction unit to be extracted, the dimensionless part that makes multiple types of features dimensionless, and the dimensionless multiple types of feature quantities, the linearity between the multiple types of feature quantities is determined. A feature correlation unit that calculates the feature correlation value to be shown, a correlation determination unit that determines whether the same signal is included based on the feature correlation value, and a feature angle correlation value that represents the slope of distribution in multiple types of features. It is provided with a feature angle correlation unit for calculating the above and a feature angle correlation determination unit for determining whether or not the same signal is included based on the feature angle correlation value.

Further, the signal processing method according to the present invention is a signal processing method by a computer that performs an identity determination process for determining whether or not the same signal is included in a plurality of signals, and there are a plurality of types of signal processing methods from the plurality of signals. Regarding the correlation between the feature extraction process for extracting feature quantities, the dimensionless step for making multiple types of feature quantities dimensionless, and the correlation between the dimensionless multiple types of feature quantities, among the multiple types of feature quantities. A feature correlation value calculation process that calculates a feature correlation value indicating linearity, a correlation determination step that provisionally determines whether or not the same signal is included based on the feature correlation value, and a provisional determination that the same signal is included. When this is done, the feature angle correlation value calculation process that calculates the feature angle correlation value that represents the slope of the distribution in multiple types of feature quantities, and the feature that determines whether the same signal is included based on the feature angle correlation value. The angle correlation determination step is executed by a computer .

Further, the program of the signal processing method according to the present invention includes a feature extraction step of extracting a plurality of types of feature quantities from a plurality of signals, a dimensionless step of making the plurality of types of feature quantities dimensionless, and a dimensionless step. The same signal is included based on the feature correlation value calculation process for calculating the feature correlation value indicating the linearity between the multiple types of feature quantities and the feature correlation value for the correlation of the multiple types of feature quantities. A correlation determination step for tentatively determining whether or not, and a feature angle correlation value calculation step for calculating a feature angle correlation value representing the slope of distribution in a plurality of types of feature quantities when it is tentatively determined that the same signal is included. A computer is made to perform a feature angle correlation determination step of determining whether or not the same signal is included based on the feature angle correlation value.

According to the present invention, the same signal is obtained when both the feature correlation value calculated by the feature correlation unit and the feature angle correlation value calculated by the feature angle correlation unit satisfy the conditions based on a plurality of types of features in the signal. Since the feature correlation unit and the feature angle correlation section are used for determination, more accurate identity determination can be performed.

It is a figure which shows the structure of the signal processing system centering on the signal processing apparatus 100 which concerns on Embodiment 1 of this invention. It is a figure which shows the relationship between the plurality of receivers 10 which concerns on Embodiment 1 of this invention, and a sound source. It is a figure explaining the feature data stored in the feature database 121 which concerns on Embodiment 1 of this invention. It is a figure which shows the structure of the feature angle correlation part 115 which concerns on Embodiment 1 of this invention. It is a figure explaining the distribution of the characteristic of the signal x which concerns on Embodiment 1 of this invention. It is a figure which shows the structure of the signal processing system centering on the signal processing apparatus 100 which concerns on Embodiment 2 of this invention. It is a figure explaining the effect which concerns on Embodiment 2 of this invention.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. Here, in the following drawings, those having the same reference numerals are the same or equivalent thereto, and are common to the whole texts of the embodiments described below. In addition, the forms of the components shown in the full text of the specification are merely examples and are not limited to these descriptions. In particular, the combination of components is not limited to the combination in each embodiment, and the components described in other embodiments can be appropriately applied to other embodiments. In addition, when it is not necessary to distinguish or specify a plurality of devices of the same type that are distinguished by subscripts, the subscripts and the like may be omitted.

Embodiment 1.
FIG. 1 is a diagram showing a configuration of a signal processing system centered on a signal processing device 100 according to a first embodiment of the present invention. Further, FIG. 2 is a diagram showing the relationship between the plurality of receivers 10 and the sound source according to the first embodiment of the present invention. In FIG. 2, each of the two sound sources emits a signal. The receiver 10 converts the received signal into an electric signal. Here, it is assumed that the receiver 10 receives a signal related to sound. In the first embodiment, it has two receivers 10a and 10b. The signal processing device 100 according to the first embodiment includes the same signal having the same specifications, such as a signal emitted from the same sound source, in the signal of the sound received by the receiver 10a and the receiver 10b, respectively. This will be described as a device for determining whether or not the signal is generated.

The signal processing device 100 according to the first embodiment includes an arithmetic processing unit 110 and a storage device 120. The arithmetic processing unit 110 according to the first embodiment is an apparatus that performs identity determination processing based on signals sent from a plurality of receivers 10. The arithmetic processing unit 110 includes a feature extraction unit 111, a dimensionless unit 112, a feature correlation unit 113, a correlation determination unit 114, a feature angle correlation unit 115, and a feature angle correlation determination unit 116. Further, the storage device 120 is a device that stores data related to the processing of the arithmetic processing unit 110. The storage device 120 has a feature database 121.

The feature extraction unit 111 (111a, 111b) of the arithmetic processing unit 110 performs a feature extraction step of extracting a plurality of feature quantities _fk extracted by processing a plurality of signals from the receiver 10. Then, the feature extraction unit 111 outputs the feature vector f, which is a column vector having the feature amount f _k as an element. Here, the feature amount f _k is a numerical value representing each feature of a plurality of types of features possessed by the signal.

The non-dimensionalization unit 112 (112a, 112b) performs a non-dimensionalization step of making each feature amount f _k in the feature vector f dimensionless based on the feature data stored in the feature database 121. The feature data will be described later. Then, the dimensionless feature vector f _ND having the dimensionless feature quantity f _NDk as an element is output. Here, in FIG. 1, the feature extraction unit 111a and the feature extraction unit 111b, and the dimensionless unit 112a and the dimensionless unit 112b are installed according to the number of receivers 10, but the present invention is limited to these. is not. A feature extraction unit 111 and a dimensionless unit 112, respectively, may process a plurality of signals.

The feature correlation unit 113 performs a feature correlation value calculation step of calculating and outputting a feature correlation value _vector based on a plurality of dimensionless feature vectors f _ND sent from each dimensionless unit 112. The feature correlation value r _vector is an index for determining whether or not a plurality of signals include the same signal. The correlation determination unit 114 performs a provisional determination process based on the feature correlation value r _vector from the feature correlation unit 113, and performs a correlation determination step of outputting the provisional determination result flag _vector .

The feature angle correlation unit 115 correlates with the feature angle based on the slope of the distribution of each feature amount f _NDk of the dimensionless feature vector f _ND related to a plurality of signals tentatively determined to include the same signal in the determination result flag _vector . The feature angle correlation value calculation step of calculating the value r _angle and outputting it is performed. The feature angle correlation value r _angle is an index for determining whether or not a plurality of signals include the same signal. The feature angle correlation determination unit 116 performs a determination process based on the feature _angle correlation value rangle from the feature angle correlation unit 115, and performs a feature angle correlation determination step of outputting a determination result flag _feature .

Here, the arithmetic processing unit 110 is usually composed of an apparatus that performs control arithmetic processing such as a computer centered on a CPU (Central Processing Unit), for example. Then, the arithmetic processing unit 110 executes a pre-programmed procedure of the signal processing method performed by each unit to realize the processing of each unit. Here, for example, the storage device 120 has the data of the program. However, the present invention is not limited to this, and each part may be separately configured by a dedicated device (hardware).

As described above, the storage device 120 stores data related to the processing of the arithmetic processing unit 110. The feature database 121 of the storage device 120 is a database that stores feature data used in the processing performed by the dimensionless unit 112. The feature data is data related to the value of the feature amount _fk of each item of the signal related to the non-stationary sound of the sample obtained in advance with a plurality of features as items. Details will be described later. The storage device 120 includes a volatile storage device (not shown) such as a random access memory (RAM) capable of temporarily storing data, and a non-volatile auxiliary storage device (not shown) such as a hard disk and a flash memory capable of storing data for a long period of time. Not shown).

The processing of each part described above will be described in more detail. Here, let x _a be the signal from the receiver 10 a. Further, let x _b be the signal from the receiver 10 b. The feature extraction unit 111 outputs a feature vector f in which a plurality of feature quantities f _k obtained from the signal x _a and the signal x _b are represented as a column vector. The feature quantity f _k and the feature vector f can be expressed by the equation (3). Here, k is a number assigned to the feature. Further, K is the number of types of features used. Then, g _k (x) is a function for calculating the feature amount of the number k when the signal x output from the receiver 10 is input.

For example, as a function for calculating the feature quantity _fk related to frequency, a function for calculating the frequency at which the energy peaks, the spread of energy related to frequency, the center of gravity of energy related to frequency, etc., which can be calculated using the Fourier transform, is used. Can be adopted. Further, as a function for calculating the feature amount with respect to time, a function for calculating the duration of the signal, the time when the energy peaks, the spread of energy with respect to time, and the like can be adopted. As a function for calculating the characteristics related to the distribution of energy values, a function for calculating moments such as skewness and kurtosis can be adopted. As described above, various functions for extracting the features related to the signal can be adopted, and a plurality of feature quantities _fk obtained based on the adopted functions are used. Here, these feature quantities _fk are not easily affected by the SN ratio. Let fa be the feature vector _f obtained from the signal x _a . Further, let f _b be the feature vector f obtained from the signal x _b . Then, the identity determination using the similarity with respect to the feature vector _{fa and the feature vector f b will} be described.

FIG. 3 is a diagram illustrating feature data stored in the feature database 121 according to the first embodiment of the present invention. In the feature database 121, as shown in FIG. 3, the value of the feature amount _fk of the signal related to the non-stationary sound of the sample acquired in the past is organized and stored as feature data. At this time, the type of signal is clear. The dimensionless unit 112 makes the feature vector f dimensionless for each element by using the feature data stored in the feature database 121. First, for all the feature vectors f stored in the feature database 121, the mean value μ _k and the standard deviation σ _k of the feature quantity f _k for each element are calculated as expressed in the equations (4) and (5). do.

Further, the dimensionless unit 112 outputs the dimensionless feature vector f _ND represented by the equation (6) by using the average value μ _k of the feature quantity f _k and the standard deviation σ _k . In the first embodiment, the dimensionless unit 112 outputs the dimensionless feature vector f _NDa and the dimensionless feature vector f _NDb corresponding to the feature vector f _a and the feature vector f _b .

The feature correlation unit 113 outputs a feature correlation value _vector representing the correlation between the signals x of each feature based on the dimensionless feature vector f _NDa and the dimensionless feature vector f _NDb . The feature correlation value r _vector represented by the equation (7) is an index showing the linearity between a plurality of features.

As shown in the equation (8), the correlation determination unit 114 performs the determination process based on the feature correlation value _vector from the feature correlation unit 113, and outputs the determination result flag _vector . Here, the γ _vector is a threshold value that considers that the signal x of each receiver 10 includes the same signal.

The feature angle correlation unit 115 inputs the determination result flag _vector determined by the correlation determination unit 114, the dimensionless feature vector f _NDa , and the dimensionless feature vector f _NDb . When the determination result flag _vector = 0, the feature angle correlation unit 115 outputs only the determination result flag _vector because the same sound is not included in the signal.

On the other hand, when the determination result flag _vector = 1, the feature angle correlation unit 115 outputs the determination result flag _vector and the feature angle correlation value r _angle . Here, the feature _angle correlation value rangle is an index showing the degree of similarity between feature vectors.

The feature angle correlation value r _angle is an index related to the inclination when each feature amount is distributed, as shown in FIG. 5 (a) described later. The correlation determination unit 114 redetermines the identity in order to determine the signals that are regarded as the same even though they are signals of different sounds as different signals.

When obtaining the feature _angle correlation value rangle, it is necessary to calculate the slope of the distribution of each feature amount f _k . As a means for calculating the slope of the distribution, methods such as principal component analysis and least squares method are used. Here, a method of calculating the slope by performing principal component analysis using eigenvalue decomposition will be described.

FIG. 4 is a diagram showing the configuration of the feature angle correlation unit 115 according to the first embodiment of the present invention. The feature angle correlation unit 115 further includes a correlation matrix calculation unit 115A, an eigenvalue decomposition unit 115B, and a feature angle correlation value calculation unit 115C. The correlation matrix calculation unit 115A calculates the correlation matrix Σ represented by the equation (9).

As shown in the equation (10), the eigenvalue decomposition unit 115B decomposes the correlation matrix Σ into the eigenvectors φ (φ ₁ , φ ₂ ) and the eigenvalues λ (λ ₁ , λ ₂ ).

The feature angle correlation value calculation unit 115C calculates the feature _angle correlation value rangle represented by the equation ( ₁₁ ) based on the eigenvector φ1.

As shown in the equation (12), the feature angle correlation determination unit 116 performs the determination process based on the feature _angle correlation value rangle from the feature angle correlation unit 115, and outputs the determination result flag _feature . Here, γ _angle is a threshold value at which the same signal is considered to be included in the signal x from each receiver 10. Further, the determination result flag _feature is a flag indicating the result of the identity determination. When it is determined that the signal x _a and the signal x _b contain signals of the same sound, the determination result flag _feature = 1, and when it is determined that the signals of the same sound are not included, the determination result flag. It is set to _fairure = 0.

FIG. 5 is a diagram illustrating the distribution of the characteristics of the signal x according to the first embodiment of the present invention. Next, the effect of the signal processing device 100 according to the first embodiment will be described. When the receiver _10a and the receiver 10b receive the same signal, the signal xa and the signal are extracted from _xb , and the dimensionless feature vector f _NDa and the dimensionless feature vector f _NDb are made dimensionless. , The feature quantities f _{NDa, k} and the feature quantities f _{NDb, k} of the same index k have similar values. Therefore, as shown in FIG. 5A, each feature is distributed in the vicinity of a straight line having a slope of 1.

The feature correlation value r _vector calculated by the feature correlation unit 113 is an index showing the linearity of the feature distribution. When the same signal is included in the signal received by each receiver 10, the feature correlation value _reflector becomes a large value. On the other hand, when the signals received by each receiver 10 do not include the same signal and are different signals, the feature correlation value _vector is a small value. Therefore, the correlation determination unit 114 can determine the feature correlation value r _vector based on the set threshold value γ _vector , thereby performing the identity determination.

Here, as shown in FIG. 5B, even when the straight line derived by the distribution by the feature has a slope different from 1, the feature correlation value r _vector becomes a large value when the distribution is linear. Therefore, even though they are different signals, they are determined to be the same signal.

On the other hand, the feature angle correlation value _rangle calculated by the feature angle correlation unit 115 is an index indicating the angle between the slope of the feature distribution and the straight line of the slope 1. When the same signal is included in the signal received by each receiver 10, the angle between the slope of the feature distribution and the straight line of slope 1 becomes small (the slope of the feature distribution becomes close to 1). , The feature angle correlation value r _angle is a large value. On the other hand, when the signals received by each receiver 10 do not include the same signal and are different signals, the feature _angle correlation value range is a small value. Therefore, the feature angle correlation determination unit 116 determines the feature _angle correlation value rangle based on the set threshold value γ _angle , thereby preventing erroneous determination of different signals as the same in the feature determination. can.

Here, even when the signals received by each receiver 10 are different, there is a case where the feature _angle correlation value rangle becomes a large value. For example, if the two signals are uncorrelated, the features will be randomly distributed, as shown in FIG. 5 (c). Therefore, the slopes are also random, and even though the signals are different, they may be determined to be the same signal.

Therefore, in the signal processing device 100 of the first embodiment, the feature correlation unit 113 and the feature angle correlation unit 115 calculate both the feature correlation value _vector and the feature angle correlation value r _angle . Then, when the correlation determination unit 114 and the feature angle correlation determination unit 116 determine that there is a linear relationship between the dimensionless feature vectors f _ND related to the plurality of signals and the slope of the correlation characteristic is close to 1. It is determined that the signals are the same. Therefore, the identity determination of the signal can be performed more accurately.

Embodiment 2.
The signal processing device 100 of the first embodiment uses each feature amount f _NDk of the dimensionless feature vector f _ND processed by the dimensionless unit 112 for the identity determination. Here, the effectiveness of the feature amount f _NDk of each feature differs depending on the type of non-stationary sound. If there is little variation in the feature amount f _NDk , it is difficult to distinguish them, so it is not effective for determination. For example, when performing identity determination, for a signal x having energy in a wide frequency band, a feature amount that quantifies a feature related to energy spread is effective. Further, for the signal x having a short duration, a feature amount obtained by quantifying the characteristics related to the duration is effective.

Therefore, the signal processing device 100 of the second embodiment calculates in advance the effectiveness of each feature amount from the variation of each feature amount stored in the feature database 121. A weight based on the effectiveness is given to each feature amount, and the identity determination is performed.

FIG. 6 is a diagram showing a configuration of a signal processing system centered on the signal processing device 100 according to the second embodiment of the present invention. In FIG. 6, those having the same reference numerals as those in FIG. 1 perform the same operations as those described in the first embodiment. In FIG. 6, the weighting unit 117 (117a, 117b) performs a process of assigning a weight to each element of the dimensionless feature vector f _ND in which the signal received by each receiver 10 is made dimensionless. Further, the feature database 121 in the second embodiment stores data related to the weight for correcting the variation related to each feature amount _fk in association with the feature data.

Next, the weight given by the weight giving unit 117 will be described. Regarding the weight, the weight vector w is calculated in advance from the feature data in the feature database 121 based on the equation (13). Here, α is an arbitrary coefficient. Further, when obtaining the mean value μ _k and the standard deviation σ _k of the feature amount f _k stored as the feature data, the signal is obtained by using only the signal of the same type as the signal received by the receiver 10. It is possible to calculate the effectiveness of features according to the type of.

The weighting unit 117 performs a process of assigning the weight vector w stored as data in the feature database 121 to the dimensionless feature vector f _ND by the equation (14). Here, ○ means the element product (Hadamard product).

FIG. 7 is a diagram illustrating the effect according to the second embodiment of the present invention. When all the features are treated equally and the identity judgment is performed even though the variation of a certain feature amount f _k is large, a valid judgment result may not be obtained. In FIG. 7A, although the same signal is received by each receiver 10, a valid determination result cannot be obtained due to the influence of handling the feature amount _fk which is easily affected by the SN ratio. An example is shown.

Therefore, in the signal processing device 100 of the second embodiment, the weighting unit 117 assigns weights to the dimensionless feature vector f _ND according to the effectiveness of each feature amount f _k , whereby FIG. 7 ( As shown in b), the distribution of the feature quantity f _k can be modified. Therefore, the accuracy of the identity determination can be improved.

Embodiment 3.
Although not particularly shown in the first and second embodiments described above, the signal processing device 100 receives a signal received by each receiver 10 of the receiver array having a plurality of receivers 10. May perform a process of determining the identity.

Further, in the first embodiment and the second embodiment described above, the identification processing is performed on the signals received by the plurality of receivers 10 arranged at different positions, but the present invention is not limited to this. do not have. The signal processing device 100 may perform a process of determining the identity of signals received at different times in a receiver array having a single receiver 10 or a plurality of receivers 10.

In the above-described first embodiment, as the characteristics related to the frequency, the frequency at which the energy peaks, the spread of the energy related to the frequency, the center of gravity of the energy related to the frequency, and the like are given as examples. In addition, as characteristics related to time, the duration of the signal, the time when the energy peaks, the spread of energy with respect to time, etc. are given as examples. Then, as a feature regarding the distribution of energy values, moments such as skewness and kurtosis are given as examples.

In addition, the feature amount f _k can be adopted for various features. As for the frequency, a feature amount _fk characterized by frequency fluctuation, the lowest frequency, the skewness of the energy distribution related to the frequency, the kurtosis of the energy distribution related to the frequency, and the like can be adopted. Regarding time, a feature quantity _fk characterized by the center of gravity of energy with respect to time, the strain of energy distribution with respect to time, the sharpness of energy distribution with respect to time, the rate of increase of energy with respect to time, the rate of decrease of energy with respect to time, and the like. Can be adopted. In addition to these feature quantities _fk , features unique to the signal received by the receiver 10 can be adopted.

In the above-described first embodiment, the method of using the principal component analysis and the least squares method when the feature angle correlation unit 115 calculates the feature _angle correlation value range has been described, but the distribution of feature quantities _fk other than these has been described. A method of calculating the slope can be used.

Further, in the first embodiment described above, after performing the identity determination based on the linearity using the feature correlation value _vector by the correlation determination unit 114, the feature angle correlation value rangle by the feature _angle correlation determination unit 116 is used. The identity was determined. However, the order of determination may be reversed.

10, 10a, 10b Receiver 100 Signal processing device 110 Arithmetic processing device 111, 111a, 111b Feature extraction unit 112, 112a, 112b Dimensionless unit 113 Feature correlation unit 114 Correlation judgment unit 115 Feature angle correlation unit 115A Correlation matrix calculation Part 115B Eigenvalue decomposition part 115C Feature angle correlation value calculation part 116 Feature angle correlation judgment part 117, 117a, 117b Weighting part 120 Storage device 121 Feature database

Claims (10)

A signal processing device that performs identity determination processing that determines whether or not the same signal is included in a plurality of signals.
A feature extraction unit that extracts a plurality of types of features from the plurality of signals,
A dimensionless unit that makes the plurality of types of the feature quantities dimensionless, and
With respect to the dimensionless correlation of the plurality of types of the feature quantities, a feature correlation unit for calculating a feature correlation value indicating linearity between the plurality of types of feature quantities, and a feature correlation unit.
A correlation determination unit that determines whether or not the same signal is included based on the feature correlation value,
A feature angle correlation unit for calculating a feature angle correlation value representing the slope of distribution in the plurality of types of the feature amounts, and a feature angle correlation unit.
A signal processing device including a feature angle correlation determination unit that determines whether or not the same signal is included based on the feature angle correlation value. The first aspect of claim 1, wherein the feature extraction unit extracts one or more of the feature amounts related to the time-related feature amount, the frequency-related feature amount, and the moment of the energy value distribution from the plurality of signals. Signal processing equipment. It is provided with a feature database that stores data related to the feature amount obtained by analyzing a sample signal obtained in advance as feature data.
The dimensionless unit is
Based on the feature data stored in the feature database, the average and standard deviation of each feature amount are calculated, the difference between the value of the feature amount and the average is divided by the standard deviation, and the plurality of types of the feature The signal processing apparatus according to claim 1 or 2, wherein the amounts are made non-dimensional, respectively. The signal processing apparatus according to any one of claims 1 to 3, wherein the feature angle correlation unit calculates the feature angle correlation value by using principal component analysis. The signal processing device according to any one of claims 1 to 3, wherein the feature angle correlation unit calculates the feature angle correlation value by using the least squares method. The signal processing apparatus according to any one of claims 1 to 5, further comprising a weighting unit that assigns weights to each of the plurality of types of dimensionless feature quantities based on the distribution of the feature quantities. .. It is provided with a feature database that stores data related to the feature amount obtained by analyzing a sample signal obtained in advance as feature data.
The signal processing device according to claim 6, wherein the weight is calculated for each feature amount from the average and standard deviation of the feature amount based on the stored feature data, and is associated with the feature data. The feature angle correlation unit of claims 1 to 7 calculates the feature angle correlation value with respect to the plurality of types of the feature quantities determined by the correlation determination unit to include the same signal. The signal processing apparatus according to any one of the following items. It is a signal processing method by a computer that performs an identity determination process for determining whether or not the same signal is included in a plurality of signals.
A feature extraction process that extracts multiple types of features from multiple signals,
A dimensionless step of making each of the plurality of types of the feature quantities dimensionless,
With respect to the dimensionless correlation of the plurality of types of the feature quantities, a feature correlation value calculation step of calculating a feature correlation value indicating linearity between the plurality of types of feature quantities, and a feature correlation value calculation step.
A correlation determination step for tentatively determining whether or not the same signal is included based on the feature correlation value, and
When it is tentatively determined that the same signal is included, the feature angle correlation value calculation step of calculating the feature angle correlation value representing the slope of the distribution in the plurality of types of the feature amounts, and the feature angle correlation value calculation step.
With the feature angle correlation determination step of determining whether or not the same signal is included based on the feature angle correlation value.
Is a signal processing method executed by the computer . A feature extraction process that extracts multiple types of features from multiple signals,
A dimensionless step of making each of the plurality of types of the feature quantities dimensionless,
With respect to the dimensionless correlation of the plurality of types of the feature quantities, a feature correlation value calculation step of calculating a feature correlation value indicating linearity between the plurality of types of feature quantities, and a feature correlation value calculation step.
A correlation determination step for tentatively determining whether or not the same signal is included based on the feature correlation value, and
When it is tentatively determined that the same signal is included, the feature angle correlation value calculation step of calculating the feature angle correlation value representing the slope of the distribution in the plurality of types of the feature amounts, and the feature angle correlation value calculation step.
A program of a signal processing method for causing a computer to perform a feature angle correlation determination step of determining whether or not the same signal is included based on the feature angle correlation value.

Images