JP2016102822A - Detector for infant cry - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a detector for an infant cry, having an increased detection accuracy for the infant cry.SOLUTION: A first determination unit 16 determines that an infant cry can be contained in a collected sound signal when the collected sound signal has been determined to contain at least two alternate first and second sound signals. A second determination unit 21 determines that the collected sound signal contains an infant cry and detects the infant cry when the following determination results have been obtained, which are a determination result showing that the collected sound signal can contain the infant cry and a determination result showing that the first sound signal has a harmonic structure continuing for a predetermined period of time.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an infant cry detection device for detecting an infant cry.

  Conventionally, what was described in patent document 1 as this kind of technique is known. The document 1 describes an infant cry monitoring system that monitors whether an infant is crying at a distance from the infant. This monitoring system collects crying voices produced by infants, and when the sound pressure level of the collected crying voice exceeds a predetermined value, notifies the supervisor at a remote location that the crying voice has been detected.

  Moreover, what was described in patent document 2 is known as a technique similar to the above. The document 2 describes a cry notification device that recognizes a child's cry from the collected sound and notifies the owner of the mobile terminal when the cry is recognized. This crying notification device includes a crying sound of a child in the collected sound by comparing the collected sound with a predetermined threshold value relating to a predetermined sound volume, sound pitch, etc. stored in advance. It is determined whether or not.

  Furthermore, Patent Document 3 discloses a hearing support device that identifies various types of sounds including infants' cry and informs the user. This hearing aid apparatus identifies a sound type by a pattern matching technique that converts a collected sound into a frequency and compares the frequency characteristic with a preset frequency characteristic of a known sound.

JP 2002-367049 A JP-A-11-102493 Japanese Patent Laid-Open No. 10-282985

  In the conventional techniques described in Patent Documents 1 and 2, the cry of an infant is detected based on the magnitude of the collected sound. For this reason, in the conventional techniques described in Patent Documents 1 and 2, it is difficult to accurately detect the cry of an infant that changes depending on the surrounding environment and the condition of the infant, and the possibility of erroneous detection has increased.

  On the other hand, in the pattern matching technique described in Patent Document 3, it is difficult to accurately detect an infant's cry that changes depending on the surrounding environment and the infant's condition, simply by comparing the collected sound with the comparison pattern. It was.

  An object of the present invention is to provide an infant cry detection device with improved accuracy for detecting infant cry.

  The present invention provides a sound collection unit that collects sound, a sound pressure level detection unit that detects a sound pressure level of a sound signal collected by the sound collection unit, and a sound pressure level detected by the sound pressure level detection unit. The detected sound signal includes a first sound signal that is equal to or higher than the first sound pressure level for a first time and a second sound pressure level that is lower than the first sound signal and lower than the first time. A continuity determination unit that determines whether or not a second sound signal that is continuous for a short second time period is included, and the sound collection unit collects the first sound signal and the second sound signal. A periodicity determining unit that determines whether or not the first sound signal and the second sound signal are alternately repeated at least twice when it is determined that the sound signal is included in the sound signal; When it is determined that the first sound signal and the second sound signal are alternately repeated at least twice, the sound is collected by the sound collecting unit. It is determined that there is a possibility that an infant's cry may be included in the sound signal collected by the sound collecting unit and a first determination unit that determines that the signal may include an infant's cry A frequency converter that converts the frequency of the first sound signal, and the first sound signal that has undergone frequency conversion has a harmonic structure in which the peak of the frequency spectrum is composed of a fundamental frequency and a higher order frequency. A harmonic overtone determination unit for determining whether or not the overtone structure is continued for a preset time when the overtone structure is present; When it is determined that there is a possibility that it is included in the sound signal collected in step 1, and it is determined that the first sound signal has a harmonic structure and the harmonic structure has continued for a preset time. Infants crying in the sound signal collected by the sound collection unit Was determined to be included, to provide a second determination unit for detecting the crying infant, the infant crying detecting apparatus characterized by having a.

  The present invention also provides a sound collecting unit that collects sound, a sound pressure level detecting unit that detects a sound pressure level of a sound signal collected by the sound collecting unit, and a sound pressure level that is detected by the sound pressure level detecting unit. The sound signal whose level is detected includes a first sound signal that is equal to or higher than the first sound pressure level for a first time, and the first sound signal that is equal to or lower than a second sound pressure level lower than the first sound signal. A continuity determination unit that determines whether or not a second sound signal that is continuous for a second time shorter than the time is included, and the first sound signal and the second sound signal are the sound collection unit A periodicity determination unit that determines whether the first sound signal and the second sound signal are alternately repeated at least twice when it is determined that the sound signal is included in the sound signal collected at A frequency conversion unit that converts the frequency of the first sound signal in parallel with the operations of the continuity determination unit and the periodicity determination unit, and a frequency The converted first sound signal is determined whether or not the peak of the frequency spectrum has a harmonic structure composed of a fundamental frequency and a higher order frequency. Determining whether or not the overtone structure continues for a preset time, and performing a determination operation in parallel with the operations of the continuity determination unit and the periodicity determination unit; and the first sound signal And the second sound signal are included in the sound signal collected by the sound collecting unit, and the first sound signal and the second sound signal are alternately repeated at least twice, and the first sound signal If it is determined that the sound signal has a harmonic structure and the harmonic structure has continued for a preset time, the sound signal collected by the sound collecting unit includes an infant cry. A determination unit for determining and detecting a cry of an infant To provide a child cry detection device.

  According to the infant cry detection apparatus of the present invention, it is possible to improve the accuracy of detecting an infant cry.

It is a figure which shows the structure of the infant cry detection apparatus which concerns on 1st Embodiment of this invention. It is a figure which shows one waveform of the sound signal collected in the state where the infant is crying. It is a figure which shows an example of the model pattern of the sound pressure level transition modeled by binarizing the sound pressure level of the sound signal which has periodicity. It is the figure of the spectrogram which showed the relationship between the time and frequency in the infant's cry. FIG. 5 is a partially enlarged view of FIG. 4. It is a flowchart which shows the procedure of operation | movement of the infant cry detection apparatus which concerns on 1st Embodiment of this invention. It is a flowchart which shows the procedure of operation | movement of the infant cry detection apparatus which concerns on 1st Embodiment of this invention. It is a figure which shows the structure of the infant cry detection apparatus which concerns on 2nd Embodiment of this invention. It is a figure which shows an example of the spectrogram of the infant's cry which has a harmonic structure, and a schematic diagram which shows the mode of a frequency fluctuation. It is a figure which shows an example of the spectrogram of the notification sound which has a harmonic structure. It is a flowchart which shows the procedure of operation | movement of the infant cry detection apparatus which concerns on 2nd Embodiment of this invention. It is a figure which shows the structure of the infant cry detection apparatus which concerns on 3rd Embodiment of this invention.

  Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings.

(First embodiment)
With reference to FIG. 1, the structure of the infant cry detection apparatus 1 which concerns on 1st Embodiment of this invention is demonstrated. In FIG. 1, an infant cry detection device 1 includes a microphone 10, an ADC (analog-digital converter) 11, a first frame generation unit 12, a sound pressure level detection unit 13, a continuity determination unit 14, a periodicity determination unit 15, and A first determination unit 16 is provided. The infant cry detection apparatus 1 includes a buffer unit 17, a second frame generation unit 18, a frequency conversion unit 19, a harmonic overtone determination unit 20, and a second determination unit 21.

  The microphone 10 is provided in the vicinity of the infant as a sound collecting unit, and collects surrounding sounds. The microphone 10 converts the collected sound into an analog electrical signal. The microphone 10 gives a sound signal obtained by the conversion to the ADC 11.

  The ADC 11 converts the sound signal of the analog signal provided from the microphone 10 into a digital signal so that the infant cry detection device 1 performs processing for detecting the cry of the infant with a digital signal. The ADC 11 samples an analog sound signal at a predetermined sampling frequency and converts it into a digital signal. The ADC 11 supplies the sound signal of the digital signal obtained by the conversion to the first frame generation unit 12 and the buffer unit 17.

  The first frame generation unit 12 divides a sound signal converted into a digital signal by a predetermined number of samples corresponding to a predetermined time width, and generates a frame of the sound signal using the divided sound signal as one frame. The sound signal is processed as described later using one frame generated by the first frame generation unit 12 as a processing unit.

  The first frame generation unit 12 newly generates a frame of a sound signal based on a determination result described later given from the sound pressure level detection unit 13. The first frame generation unit 12 gives the generated sound signal frame to the sound pressure level detection unit 13.

  The sound pressure level detection unit 13 detects the sound pressure level of the sound signal included in the frame given from the first frame generation unit 12. The sound pressure level detector 13 calculates any one of the absolute value of the maximum amplitude, the absolute value of the average amplitude, the maximum mean square value of the amplitude, or the mean square value of the amplitude in the sound signal, and calculates the calculated value. Detect as sound pressure level.

  The sound pressure level detector 13 determines whether the sound signal given from the first frame generator 12 includes a sound signal with a high sound pressure level or a sound signal with a low sound pressure level. Here, a sound signal with a high sound pressure level is a sound signal that is equal to or higher than the first sound pressure level, and a sound signal with a low sound pressure level is a sound signal that is equal to or lower than the second sound pressure level.

  FIG. 2 is a diagram showing one waveform of a sound signal collected while the infant is crying. The sound signal shown in FIG. 2 represents the sound signal by the relationship between time and amplitude. Hereinafter, a sound signal represented by the relationship between time and amplitude is referred to as a time-domain sound signal.

  The sound signal shown in FIG. 2 shows the characteristics of an infant's cry. In detecting an infant's cry, considering the characteristics of the sound signal waveform shown in FIG. 2, the first sound pressure level, the second sound pressure level, and the first time and the second time described later are set. I have decided.

  In FIG. 2, a section a is a section where the cry is high (cry section), and the vowel “a” is dominant. The section b is a section where the sound pressure level is lower than the section a, and is a section (breathing section) in which breathing is performed between a high cry and a crying voice.

  The section c between the section a and the section b is a section such as the beginning of crying, the end of crying, a short cough, etc., and the section where the sound pressure level of the sound signal is unstable is excluded from the judgment of the crying voice. Interval.

  The first sound pressure level is a level set in advance to detect a state in which the crying sound is high when the infant is crying from the collected sound signal, that is, the section a shown in FIG. The first sound pressure level is set to about 80 [dB] to 90 [dB], for example.

  The second sound pressure level is a level set in advance to detect a state where the crying is temporarily stopped when the infant is crying from the collected sound signal, that is, the section b shown in FIG. The second sound pressure level is set to the same level as the environmental noise level.

  When the sound signal given from the first frame generation unit 12 includes a sound signal with a high sound pressure level or a sound signal with a low sound pressure level, the sound pressure level detection unit 13 The determination unit 14 is given.

  The continuity determination unit 14 measures the time during which a sound signal with a high sound pressure level continues based on the sound signal given from the sound pressure level detection unit 13. Based on the measurement result, the continuity determination unit 14 determines whether or not a sound signal having a high sound pressure level continues for the first time over the following frames.

  The continuity determination unit 14 measures the time during which a sound signal with a low sound pressure level continues based on the sound signal given from the sound pressure level detection unit 13. Based on the measurement result, the continuity determination unit 14 determines whether or not a sound signal having a low sound pressure level is continuous for the second time over the preceding and following frames.

  1st time is the time of the area a in the state where the cry is high shown in FIG. Therefore, the first time is a time set in advance to detect a state in which the cry is high when the infant is crying, that is, the section a shown in FIG. The first time is set to about 2 to 3 seconds, for example.

  The second time is the time of the section b where the sound pressure level is low as shown in FIG. Therefore, the second time is a time set in advance to detect a state where the crying is temporarily stopped when the infant is crying, that is, the section b shown in FIG. The second time is set to about 0.5 seconds to 1.0 seconds, for example.

  The continuity determination unit 14 gives the periodicity determination unit 15 the sound signal of the frame generated by the first frame generation unit 12. The continuity determination unit 14 provides the periodicity determination unit 15 with a determination result as to whether or not a sound signal having a high sound pressure level is continuous for the first time. Further, the continuity determination unit 14 provides the periodicity determination unit 15 with a determination result as to whether or not a sound signal having a low sound pressure level is continuous for the second time.

  The periodicity determination unit 15 determines the periodicity of the sound signal based on the determination result given from the continuity determination unit 14 and the sound signal. The periodicity determination unit 15 determines that the sound signal having a high sound pressure level is continuous for the first time, and determines that the sound signal having the second sound pressure level is continuous for the second time. In this case, the periodicity of the sound signal is determined over the previous and subsequent frames.

  The periodicity determination unit 15 includes a section a shown in FIG. 2 in which a sound signal having a high sound pressure level continues for a first time and a sound signal having a low sound pressure level in a second time in FIG. It is determined whether or not the indicated interval b is alternately repeated at least twice.

  FIG. 3 is a diagram showing an example of a model pattern of the sound pressure level transition which is modeled by binarizing the sound pressure level of the sound signal having the periodicity described above. The periodicity determination unit 15 can determine the periodicity of the sound signal as follows using, for example, the model pattern shown in FIG.

  The periodicity determination unit 15 creates and stores a sound pressure level transition pattern obtained by binarizing a sound signal having a high sound pressure level and a sound signal having a low sound pressure level with respect to the sound signal shown in FIG. Infants' crying is a physiological phenomenon, and continuity cannot always be maintained as shown in FIG. In addition, due to the influence of ambient noise, even in the breathing interval, there may be a case where a sound pressure level is temporarily detected due to mixing of sound. Therefore, in order to allow such a temporary change in the sound pressure level, the periodicity determination unit 15 may perform different situations (a low sound pressure level that temporarily appears during detection of a high sound pressure level, or vice versa). When the sound pressure level detector 13 gives a determination result that the number of times has reached a preset allowable number, the stored sound pressure level transition pattern is cleared. That is, it is considered that the audio signal being determined is not an infant cry.

  The periodicity determination unit 15 compares the stored sound pressure level transition pattern with a model pattern as shown in FIG. 3 and determines whether or not both patterns match or approximate. As a result of the determination, if both patterns match or approximate, the periodicity determination unit 15 determines that the sound signal has the above-described periodicity. On the other hand, as a result of the determination, if both patterns do not match or approximate, the periodicity determination unit 15 determines that there is no periodicity.

  The periodicity determination unit 15 gives the determination result to the first determination unit 16.

  The first determination unit 16 determines, based on the determination result given from the periodicity determination unit 15, whether there is a possibility that the collected sound signal includes an infant's cry.

  When the sound signal level is high and the sound signal that is continuous for the first time and the sound signal that is low in the sound pressure level and is continuous for the second time are generated periodically alternately. Determines that there is a possibility that the collected sound signal includes a cry of an infant. The first determination unit 16 performs a case where a sound signal having a high sound pressure level and continuing for a first time and a sound signal having a low sound pressure level and continuing for a second time are not alternately generated periodically. Determines that the collected sound signal does not include an infant's cry.

  The first determination unit 16 determines whether or not there is a possibility that an infant's cry may be included in the collected sound signal by processing the time-domain sound signal. If the first determination unit 16 determines that the collected sound signal may contain an infant cry, the first determination unit 16 gives the determination result to the second frame generation unit 18 as a transition signal. This transition signal is a control signal for shifting from the above-described processing of the sound signal in the time domain to the processing of the sound signal in the frequency domain described below.

  The buffer unit 17 temporarily holds the sound signal of the digital signal given from the ADC 11. While processing the time domain sound signal as described above, the buffer unit 17 temporarily holds the same sound signal as the sound signal subjected to this processing. The sound signal temporarily held in the buffer unit 17 is processed as a sound signal in the frequency domain as shown below.

  The buffer unit 17 gives the temporarily held sound signal to the second frame generation unit 18.

  When the transition signal is given from the first determination unit 16, the second frame generation unit 18 receives the sound given from the buffer unit 17 based on the sound signal having a high sound pressure level given from the sound pressure level detection unit 13. A sound signal having a high sound pressure level is extracted from the signal. The second frame generation unit 18 starts generating a frame for the extracted sound signal having a high sound pressure level. The second frame generation unit 18 divides the sound signal by a predetermined number of samples corresponding to a predetermined frequency width, and generates a frame of the sound signal with the divided sound signal as one frame.

  The extracted sound signal having a high sound pressure level is subjected to processing to be described later with one frame generated by the second frame generation unit 18 as a unit.

  The second frame generator 18 gives the generated frame sound signal to the frequency converter 19.

  The frequency conversion unit 19 performs frequency conversion on the sound signal in the time domain given from the second frame generation unit 18. As a frequency conversion method, for example, an orthogonal transform method such as a fast Fourier transform method, a polyphase filter having a filter bank configuration, or the like can be used. The resolution of the frequency conversion may be any resolution that can determine whether or not a harmonic structure described later is present in the frequency spectrum of the infant cry. The frequency conversion unit 19 gives the frequency spectrum of the frequency-converted sound signal to the overtone determination unit 20.

  The overtone determination unit 20 determines whether or not the sound signal has a harmonic structure based on the frequency spectrum of the sound signal given from the frequency conversion unit 19. As described above, since the main component of the sound signal is dominated by “a” of the vowel, the section in which the infant cry of the sound signal is high has a harmonic structure. The harmonic structure is a phenomenon in which the energy peak of the sound signal occurs at the fundamental frequency and a higher order frequency n times the fundamental frequency (n = 2, 3, 4,...) In the frequency spectrum of the sound signal. is there.

  FIG. 4 is a diagram showing an example of a spectrogram showing a relationship between time and frequency in an infant's cry. FIG. 4 shows the change in sound pressure level with respect to frequency (vertical axis) and time (horizontal axis) in monochrome shades, and the darker shaded portion is the portion with the higher sound pressure level.

  FIG. 5 is an enlarged view of a portion 40 surrounded by a broken line in FIG.

  In FIG. 5, the energy of the sound signal becomes stronger at the fundamental frequency f1, the higher order frequency f2 (2 × f1) twice the fundamental frequency f1, and the higher order frequency f3 (3 × f1) three times the fundamental frequency f1. ing. The energy of the sound signal is four times higher than the fundamental frequency f1 (4 × f1), five times the fundamental frequency f1 (5 × f1), and six times the fundamental frequency f1 (6 × f1). ) Is getting stronger. That is, the sound pressure level of the sound signal is high at the fundamental frequency and the higher order frequency.

  In FIGS. 4 and 5, the high-order frequency is shown only up to the high-order frequency f6 that is six times the fundamental frequency f1, but the sound pressure level is also high at a high-order frequency that is seven times or more. .

  As a result, the infant cry shown in the spectrograms in FIGS. 4 and 5 has peaks at the fundamental frequency f1 and n times higher frequency (f2 to f6), and has a harmonic structure. .

  Returning to FIG. 1, the harmonic overtone determination unit 20 determines whether or not a peak occurs at the fundamental frequency in the preset frequency band of the frequency spectrum of the sound signal. As a result of the determination, if a peak occurs at the fundamental frequency, the harmonic overtone determination unit 20 subsequently determines whether or not a peak occurs at a higher order frequency that is n times the fundamental frequency. As a result of the determination, if a peak occurs at a high-order frequency, the harmonic overtone determination unit 20 determines that the sound signal given from the frequency conversion unit 19 has a harmonic overtone structure.

  On the other hand, when no peak occurs at the fundamental frequency and its higher order frequency, the harmonic overtone determination unit 20 determines that the sound signal provided from the frequency conversion unit 19 does not have a harmonic overtone structure.

  When it is determined that the sound signal supplied from the frequency converter 19 has a harmonic structure, the harmonic determination unit 20 measures the time during which the harmonic structure of the sound signal is continued. The harmonic overtone determination unit 20 stores the measured duration in addition to the duration stored so far.

  The overtone determination unit 20 determines whether or not the stored duration has reached a preset time.

  Of the infant's cry, the sound signal of the portion where the sound pressure level of the main component “A”, which is a vowel, is high has a feature that it continues for a certain period of time. In consideration of such characteristics, the overtone determination unit 20 measures the duration of the overtone structure to detect the infant cry in order to distinguish the infant cry from other similar sound signals. Used.

  The overtone determination unit 20 gives the second determination unit 21 a determination result as to whether or not the stored duration has reached a preset time.

  When the overtone determination unit 20 determines that the frequency-converted sound signal does not have a harmonic structure, the overtone determination unit 20 counts the number of determinations. The harmonic overtone determination unit 20 determines whether or not the counted number of determinations has reached a preset allowable number. As a result of the determination, when the allowable number of times is reached, the overtone determination unit 20 estimates that the sound signal is very unlikely to contain an infant's cry, clears the stored duration, and determines the result. Is given to the second frame generator 18. On the other hand, if the determination result shows that the allowable number of times has not been reached, the harmonic overtone determination unit 20 holds the stored duration and gives the determination result to the second frame generation unit 18.

  This is because this process considers a situation in which the first determination unit temporarily becomes non-determined in the same manner as the exception process (a situation in which the continuity of the infant's cry is impaired due to a physiological phenomenon).

  Based on the determination result given from the harmonic overtone determination unit 20, the second determination unit 21 determines whether the collected sound signal includes an infant cry.

  When the determination result that the time during which the overtone structure of the sound signal continues has reached a preset time is given, the second determination unit 21 includes the infant cry in the collected sound signal. It is determined that the baby is crying and the infant's cry is detected. On the other hand, when the determination result that the time over which the overtone structure of the sound signal continues does not reach a preset time is given, the second determination unit 21 gives the infant's cry to the collected sound signal. Is determined not to be included.

  When detecting the cry of the infant, the second determination unit 21 notifies the outside of the infant cry detection device 1 such as a monitor.

  As a notification method, for example, the fact that an infant cry was detected is transmitted from the infant cry detection device 1 to a receiving device such as a dedicated terminal device or a general-purpose portable terminal device by communication. The receiving device notifies a monitor or the like carrying the receiving device by a visual method such as display on a display, an auditory method such as outputting sound from a speaker, or a sensory method such as mechanical vibration.

  FIG. 6 is a flowchart showing a procedure of processing the sound signal collected by the microphone 10 in the time domain.

  In FIG. 6, in step S601, the first frame generation unit 12 generates a frame of a sound signal by dividing the sound signal collected by the microphone 10 and converted into a digital signal by the ADC 11 by a predetermined number of samples. The first frame generation unit 12 provides the generated frame to the sound pressure level detection unit 13 in step S601.

  In step S602, the sound pressure level detection unit 13 detects the sound pressure level of the sound signal given from the first frame generation unit 12.

  In step S603, the sound pressure level detection unit 13 determines whether the sound signal includes a sound signal with a high sound pressure level or a sound signal with a low sound pressure level. As a result of the determination, if the sound signal includes a sound signal with a high sound pressure level or a sound signal with a low sound pressure level (YES), the process shown in step S604 is executed. On the other hand, if the result of determination is that the sound signal does not contain a sound signal with a high sound pressure level or a sound signal with a low sound pressure level (NO), the processing shown in step S609 is executed.

  In step S604, when a sound signal with a high sound pressure level is included, the continuity determination unit 14 measures a time during which a sound signal with a high sound pressure level is continuous. In step S604, if a sound signal with a low sound pressure level is included, the continuity determination unit 14 measures a time during which a sound signal with a low sound pressure level continues.

  In step S605, the continuity determination unit 14 determines whether or not a sound signal with a high sound pressure level continues for a first time and a sound signal with a low sound pressure level continues for a second time as a result of the measurement. Determine. As a result of the determination, when a sound signal with a high sound pressure level continues for the first time and a sound signal with a low sound pressure level continues for the second time (YES), the process shown in step S606 is executed. Is done. On the other hand, if the result of determination is that a sound signal with a high sound pressure level is not continuous for the first time, or a sound signal with a low sound pressure level is not continuous for the second time (NO), step S609. The process indicated by is executed. This corresponds to a mechanism for preventing erroneous detection due to the temporary state change described above.

  In step S606, the periodicity determination unit 15 creates and stores a sound pressure level transition pattern in which a sound signal having a high sound pressure level and a sound signal having a low sound pressure level are binarized. Thereafter, the process shown in step S607 is executed.

  In step S607, the periodicity determination unit 15 repeatedly generates a sound signal having a high sound pressure level for a first time and a sound signal having a low sound pressure level for a second time alternately. It is determined whether or not. As a result of the determination, when both sound signals are alternately generated repeatedly (YES), the periodicity determination unit 15 gives the determination result to the first determination unit 16 in step S607.

  On the other hand, as a result of the determination, if both sound signals are not repeatedly generated alternately, the periodicity determination unit 15 gives the determination result to the first determination unit 16 in step S607, and then the process shown in step S601. Is executed.

  In step S607, the first determination unit 16 determines whether or not there is a possibility that an infant's cry may be included in the collected sound signal based on the determination result given from the periodicity determination unit 15. judge. As a result of the determination, if the determination result that the both sound signals are alternately generated is given from the periodicity determination unit 15, the first determination unit 16 converts the collected sound signal into the collected sound signal in step S607. It is determined that there is a possibility that an infant's cry is included. If the first determination unit 16 determines in step S607 that there is a possibility that the collected sound signal includes an infant's cry (YES), then the process shown in step S608 is executed. The

  On the other hand, as a result of the determination, if the determination result that the two sound signals are not repeatedly generated alternately is given from the periodicity determination unit 15, the first determination unit 16 determines whether or not the collected sound in step S607. It is determined that there is no possibility that the signal includes an infant cry. If the first determination unit 16 determines in step S607 that there is no possibility that an infant's cry is included in the collected sound signal (NO), then the process shown in step S601 is executed. Is done.

  If the first determination unit 16 determines in step S608 that the collected sound signal may contain an infant's cry, the first determination unit 16 provides a transition signal to the second frame generation unit 18.

  In step S609, the sound pressure level detection unit 13 counts the number of times it is determined that the sound signal does not include a sound signal with a high sound pressure level and a sound signal with a low sound pressure level. In step S609, the sound pressure level detection unit 13 performs the process shown in step S610 after counting.

  In step S610, the sound pressure level detection unit 13 determines whether or not the counted number has reached a preset allowable number. As a result of the determination, if the allowable number of times has been reached (YES), the periodicity determination unit 15 estimates that the sound signal is very unlikely to contain an infant's cry, and stores it in step S611. Clear the sound pressure level transition pattern. Thereafter, the process shown in step S601 is executed to perform the above-described process on the sound signal of the next frame.

  On the other hand, if the allowable number of times has not been reached as a result of the determination (NO), the process shown in step S601 is executed to perform the above-described process on the sound signal of the next frame.

  If the sound signal is processed in the time domain according to the procedure shown in the flowchart of FIG. 6 and it is determined that the collected sound signal may contain an infant's cry, the frequency is continued. The sound signal collected in the region is processed.

  FIG. 7 is a flowchart showing a procedure for processing the sound signal collected by the microphone 10 in the frequency domain.

  In FIG. 7, the buffer unit 17 temporarily holds the sound signal to be processed in the time domain, which is the digital signal sound signal supplied from the ADC 11 in step S <b> 701. In step S <b> 701, the buffer unit 17 gives the temporarily held sound signal to the second frame generation unit 18.

  The second frame generation unit 18 determines whether or not a transition signal is given from the first determination unit 16 in step S702. If the result of determination is that a transition signal has been given (YES), the second frame generation unit 18 proceeds to step S703. On the other hand, when the transition signal is not given (NO), the process shown in step S701 is executed.

  In step S703, the second frame generation unit 18 divides a sound signal having a high sound pressure level among the sound signals given from the buffer unit 17 by a predetermined number of samples corresponding to a predetermined frequency width. A frame of a sound signal is generated with the signal as one frame. In step S703, the second frame generation unit 18 provides the frequency conversion unit 19 with a frame of a sound signal having a high sound pressure level.

  In step S704, the frequency conversion unit 19 performs frequency conversion on the sound signal given from the second frame generation unit 18 to generate a frequency spectrum. In step S704, the frequency conversion unit 19 provides the harmonic overtone determination unit 20 with the frequency spectrum of the frequency-converted sound signal.

  In step S705, the overtone determination unit 20 determines whether the sound signal has a harmonic structure based on the frequency spectrum of the sound signal given from the frequency conversion unit 19. As a result of the determination, if the sound signal has a harmonic structure (YES), the harmonic determination unit 20 measures the time during which the harmonic structure continues in step S706 and stores the measured duration so far. Memorize in addition to the duration that has been. On the other hand, if the result of determination is that the sound signal does not have a harmonic structure (NO), the processing shown in step S710 is executed.

  In step S707, the harmonic overtone determination unit 20 determines whether the stored duration has reached a preset time. As a result of the determination, if it has been reached (YES), the harmonic overtone determination unit 20 gives the determination result to the second determination unit 21 in step S707. On the other hand, if the result of determination is not reached (NO), the harmonic overtone determination unit 20 gives the determination result to the second determination unit 21 in step S707, and the process shown in step S703 is executed.

  In step S707, the second determination unit 21 determines whether or not the collected sound signal includes an infant's cry. If the determination result that the stored duration has reached the preset time in step S707 is given from the harmonic overtone determination unit 20 (YES), the second determination unit 21 collects the sound. It is determined that the infant's cry is included in the sound signal. Thereafter, the process shown in step S708 is executed.

  On the other hand, when the second determination unit 21 receives a determination result from the harmonic overtone determination unit 20 in step S707 that the stored duration has not reached the preset time (NO), It is determined that the cry of the infant is not included in the sound signal that is sounded. Thereafter, the process shown in step S703 is executed to perform the above-described process on the sound signal of the next frame.

  If the second determination unit 21 determines in step S708 that the collected sound signal includes an infant's cry, the second determination unit 21 notifies the monitor or the like to that effect.

  If the harmonic overtone determination unit 20 determines in step S709 that the frequency-converted sound signal does not have a harmonic overtone structure, the overtone determination unit 20 counts the number of determinations. Thereafter, the process shown in step S710 is executed.

  In step S710, the overtone determination unit 20 determines whether or not the counted number of determinations has reached a preset allowable number. As a result of the determination, if the allowable number of times has been reached (YES), the harmonic overtone determination unit 20 estimates that the possibility that the sound signal includes an infant's cry is extremely low, and stores it in step S711. Clear duration. Thereafter, the process shown in step S703 is executed to perform the above-described process on the sound signal of the next frame.

  On the other hand, if it is determined that the allowable number of times has not been reached (NO), the processing shown in step S703 is executed.

  By performing the above-described series of processing on the sound signal in the frequency domain in units of frames, it is determined whether or not the collected sound signal includes an infant cry.

  As described above, the infant cry detection device 1 according to the first embodiment can obtain the following effects by adopting the following configuration.

  The infant cry detection device 1 includes a microphone 10 that collects sound as a sound collection unit, and a sound pressure level detection unit 13 that detects a sound pressure level of a sound signal collected by the microphone 10.

  The infant cry detection device 1 includes a continuity determination unit 14. The continuity determination unit 14 continues to the collected sound signal for the first time that is equal to or higher than the first sound pressure level for the first time and for the second time that is equal to or lower than the second sound pressure level. It is determined whether or not the second sound signal is included. The second sound pressure level is lower than the first sound pressure level, and the second time is shorter than the first time.

  The infant cry detection device 1 includes a periodicity determination unit 15. When it is determined that the first sound signal and the second sound signal are included in the collected sound signal, the periodicity determination unit 15 performs the first sound signal and the second sound signal, Determines whether or not alternately repeat at least twice.

  The infant cry detection device 1 includes a first determination unit 16. If the first determination unit 16 determines that the first sound signal and the second sound signal are alternately repeated at least twice, the collected sound signal may include an infant cry. Judge that there is sex.

  The infant cry detection device 1 includes a frequency conversion unit 19. The frequency conversion unit 19 converts the frequency of the first sound signal when it is determined that there is a possibility that the collected sound signal includes an infant cry.

  The infant cry detection device 1 includes a harmonic overtone determination unit 20. The harmonic overtone determination unit 20 determines whether or not the frequency-converted first sound signal has a harmonic overtone structure in which the peak of the frequency spectrum is composed of a fundamental frequency and a higher order frequency. When the harmonic overtone determination unit 20 has a harmonic overtone structure, the overtone determination unit 20 determines whether or not the overtone structure continues for a preset time.

  The infant cry detection device 1 includes a second determination unit 21. When the following determination result is obtained, the second determination unit 21 determines that the infant's cry is included in the collected sound signal, and detects the infant's cry. The determination result includes a determination result that may include the cry of the infant in the collected sound signal, the first sound signal has a harmonic structure, and the harmonic structure is continued for a preset time. It is a judgment result.

  By adopting the above-described configuration, the infant cry detection apparatus 1 can detect an infant cry by distinguishing it from various noises in the background, particularly notification sounds similar in characteristics to the infant cry. As a result, the frequency of erroneous detection can be reduced, and the accuracy of detecting an infant's cry can be increased.

  If the infant cry detection device 1 does not determine that there is a possibility that the collected sound includes an infant cry, the frequency converter 19 does not perform frequency conversion, and the harmonic determination unit 20 and the second The determination unit 21 does not perform a determination operation.

  By adopting the above technical features, the infant cry detection apparatus 1 does not operate the frequency conversion unit 19, the harmonic overtone determination unit 20, and the second determination unit 21, so that the load on the entire device can be reduced. Moreover, the infant cry detection apparatus 1 can reduce power consumption by adopting the above technical features.

(Second Embodiment)
With reference to FIG. 8, the structure of the infant cry detection apparatus 2 which concerns on 2nd Embodiment of this invention is demonstrated.

  In FIG. 8, the infant cry detection device 2 is different from the configuration of the infant cry detection device 1 of the first embodiment in that a frequency variation determination unit 80 is added, and a new second determination unit 81 is substituted for the second determination unit 21. Is provided.

  In FIG. 8, the same components as those in FIG. 1 are denoted by the same reference numerals, and the description thereof is omitted.

  Here, differences from the first embodiment described above will be mainly described.

  In FIG. 8, the harmonic overtone determination unit 20 gives the frequency fluctuation determination unit 80 the determination result as to whether or not the stored duration has reached a preset time.

  When the determination result that the duration of the harmonic structure has reached a preset time is given from the harmonic determination unit 20, the frequency variation determination unit 80 is a fundamental frequency that becomes the peak of the frequency spectrum of the frequency-converted sound signal. And determining whether or not the higher-order frequency is fluctuating.

  FIG. 9A is a diagram showing an example of a spectrogram of an infant cry having a harmonic structure. FIG. 9A shows changes in the sound pressure level with respect to frequency (vertical axis) and time (horizontal axis) in monochrome shades, where dark shades are portions where the sound pressure level is high. In FIG. 9 (a), the sound pressure level of the infant's cry is periodically increased (parts with darker shades) at times t1, t2, and t3.

  In such a change in the sound pressure level, the fundamental frequency f1 and the higher-order frequencies f2 to f6 of the respective frequency components when the sound pressure level is large (the dark and shaded portion) are shown in the schematic diagram of FIG. So that it changes over time. That is, the fundamental frequency f1 and the higher-order frequencies f2 to f6 gradually increase from the time when the sound pressure level of the cry increases, and then return to the original frequency before decreasing and increasing the frequency. ing. As described above, the crying voice of the infant child has a characteristic that the fundamental frequency and the higher order frequency fluctuate as described above in the harmonic structure.

  In contrast to such a feature, for example, a notification sound having a harmonic structure such as a horn of a vehicle intermittently sounded becomes a spectrogram as shown in FIG. FIG. 10 shows the change in the sound pressure level of the notification sound with respect to frequency (vertical axis) and time (horizontal axis) in monochrome shades, and the dark shaded portion is the portion where the sound pressure level is large.

  As can be seen from FIG. 10, the frequency when the sound pressure level is large is flat regardless of the change of time, and the frequency fluctuation as shown in FIG.

  Therefore, the frequency variation determination unit 80 distinguishes artificial sounds such as notification sounds and infant crying by detecting the above-described frequency variation.

  The frequency variation determination unit 80 gives the determination result to the second determination unit 81.

  The second determination unit 81 determines based on the determination result given from the frequency variation determination unit 80 whether or not the collected sound signal includes an infant cry.

  When the second determination unit 81 is given a determination result that the fundamental frequency and the higher order frequency are fluctuating due to the harmonic structure of the sound signal, the collected sound signal includes the infant cry. It is determined that the child is crying. On the other hand, when the second determination unit 81 is given a determination result that the fundamental frequency and the higher-order frequency are not fluctuating due to the harmonic structure of the sound signal, the collected sound signal includes the cry of the infant. Judge that it is not.

  When detecting the cry of the infant, the second determination unit 81 notifies the outside of the infant cry detection device 2 such as a monitor to that effect as in the first embodiment.

  FIG. 11 is a flowchart showing a procedure for processing a sound signal collected by the microphone 10 in the frequency domain in the second embodiment.

  In FIG. 11, the same processes as those in FIG. 7 are denoted by the same reference numerals. That is, in FIG. 11, the processes shown in steps S701 to S706 and S708 to S711 are the same as the processes shown in steps S701 to S706 and S708 to S711 in FIG.

  Here, differences from the processing shown in the flowchart of FIG. 7 will be mainly described.

  In the process shown in the flowchart of FIG. 11, the process shown in step S <b> 1101 is added between the process shown in step S <b> 705 and the process shown in step S <b> 706 by adding the frequency variation determination unit 80. In the process shown in the flowchart of FIG. 11, the process shown in step S1102 is executed instead of the process shown in step S708.

  In step S1101, the frequency variation determination unit 80 determines whether or not the fundamental frequency and the higher order frequency that are the peak of the frequency spectrum of the frequency-converted sound signal are varied. As a result of the determination, when the fundamental frequency and the higher order frequency are fluctuating (YES), the processing shown in step S706 is executed. On the other hand, as a result of the determination, if the fundamental frequency and the higher order frequency are not changed (NO), the process shown in step S709 is executed.

  If the determination result that the fundamental frequency and the higher order frequency in the harmonic structure are fluctuating is given from the frequency fluctuation determining unit 80 in step S1102 (YES), the second determining unit 81 collects the sound. It is determined that the infant's cry is included in the sound signal. Thereafter, the process shown in step S708 is executed.

  On the other hand, if the determination result that the fundamental frequency and the higher order frequency in the harmonic structure are not fluctuated is given from the frequency variation determination unit 80 in step S1102, the second determination unit 81 (NO). It is determined that the cry of the infant is not included in the sound signal that is sounded. Thereafter, the process shown in step S703 is executed to perform the above-described process on the sound signal of the next frame.

  As described above, the infant cry detection device 2 of the second embodiment can obtain the following effects by adopting the following configuration.

  The infant cry detection apparatus 2 includes a frequency variation determination unit 80 in addition to the configuration of the infant cry detection apparatus 1 of the first embodiment. The frequency fluctuation determination unit 80 generates a frequency fluctuation that returns to the original frequency after the peak fundamental frequency and the higher order frequency in the frequency spectrum of the frequency-converted first sound signal fluctuate to a high frequency with time. It is determined whether or not.

  In addition to the determination result of the second determination unit 21, the second determination unit 81 of the infant cry detection apparatus 2 collects sound when it is determined that the first sound signal has a frequency variation. It is determined that the infant's cry is included in the signal, and the infant's cry is detected.

  By adopting the above-described configuration, the infant cry detection apparatus 2 detects the infant cry by further reducing the frequency of false detection in addition to the effects obtained by the infant cry detection apparatus 1 of the first embodiment. Accuracy can be increased.

  If the infant cry detection device 2 does not determine that there is a possibility that the collected sounds include an infant cry, the frequency converter 19 does not perform frequency conversion, the harmonic determination unit 20, and frequency fluctuations. The determination unit 80 and the second determination unit 81 do not perform a determination operation.

  By adopting the above technical features, the infant cry detection device 2 does not operate the frequency conversion unit 19, the harmonic overtone determination unit 20, the frequency variation determination unit 80, and the second determination unit 81, thereby reducing the load on the entire device. can do. Moreover, the infant cry detection apparatus 2 can reduce power consumption by adopting the above technical features.

(Third embodiment)
With reference to FIG. 12, the structure of the infant cry detection apparatus which concerns on 3rd Embodiment of this invention is demonstrated.

  In the infant cry detection devices 1 and 2 of the first embodiment and the second embodiment, after processing the collected sound signal in the time domain, it is determined whether or not to process in the frequency domain based on the processing result. Yes. On the other hand, the infant cry detection device 3 according to the third embodiment is characterized in that the processing of the collected sound signal in the time domain and the processing in the frequency domain are performed in parallel.

  In FIG. 12, the same components as those in FIG. 8 are denoted by the same reference numerals, and the description thereof is omitted. Here, differences from the second embodiment described above will be mainly described.

  12, the infant cry detection device 3 of the third embodiment includes a first frame generation unit 12, a sound pressure level detection unit 13, a continuity determination unit 14, a periodicity determination unit 15, and a frequency fluctuation determination unit illustrated in FIG. 80 has the same configuration.

  In the infant cry detection apparatus 3, the buffer unit 17 and the second frame generation unit 18 shown in FIG. 8 are deleted. The infant cry detection device 3 includes a new frequency conversion unit 121 instead of the frequency conversion unit 19 shown in FIG. In addition, the infant cry detection device 3 includes a new harmonic determination unit 122 instead of the harmonic determination unit 20 illustrated in FIG. 8, and includes a new determination unit 124 instead of the second determination unit 81 illustrated in FIG. 8. . In addition, the infant cry detection device 3 is newly provided with a holding unit 123.

  The infant cry detection device 3 processes the sound signal in the time domain by the first frame generation unit 12, the sound pressure level detection unit 13, the continuity determination unit 14, and the periodicity determination unit 15, as in the second embodiment. This process is performed by executing the process shown in the flowchart of FIG.

  The frequency converter 121 receives the sound signal from the first frame generator 12, and the sound signal given from the first frame generator 12 by the sound pressure level detector 13 includes a sound signal having a high sound pressure level. If it is determined that the sound signal has a high sound pressure level, the frequency of the sound signal is converted.

  As a frequency conversion method, for example, an orthogonal transform method such as a fast Fourier transform method, a polyphase filter having a filter bank configuration, or the like can be used. The resolution of frequency conversion may be any resolution that can recognize a harmonic structure described later in the frequency structure (spectral structure) of an infant's cry. The frequency conversion unit 121 gives the overtone determination unit 122 the frequency-converted sound signal.

  The harmonic overtone determination unit 122 has the same function as that of the overtone determination unit 20 shown in FIG.

  The holding unit 123 receives the determination result of the periodicity determination unit 15 and the determination result of the frequency fluctuation determination unit 80 and holds the determination result of the periodicity determination unit 15 and the determination result of the frequency fluctuation determination unit 80. The holding unit 123 gives the held determination result to the determination unit 124.

  Based on the determination result given from the holding unit 123, the determination unit 124 determines whether or not the collected sound signal includes an infant cry.

  The determination unit 124 determines that the periodicity determination unit 15 has periodicity in the sound signal, the harmonic overtone determination unit 122 detects the characteristics of the harmonic structure in the sound signal, and the frequency variation determination unit 80 detects the fundamental frequency and higher order. When it is determined that the frequency is fluctuating, it is determined that the collected sound signal includes an infant's cry. The determination unit 124 determines that the periodicity determination unit 15 determines that the sound signal is not periodic, or the harmonic overtone determination unit 122 determines that the sound signal does not have a harmonic structure characteristic, or the frequency variation determination unit 80 determines that the sound signal is fundamental. When it is determined that the frequency and the higher order frequency are not changed, it is determined that the collected sound signal does not include the infant's cry.

  In addition, although the infant cry detection device 3 of the third embodiment includes the frequency variation determination unit 80, the frequency variation determination unit 80 is deleted as in the infant cry detection device 1 of the first embodiment, and the overtone determination unit. An infant's cry may be detected using the determination result 122.

  As described above, the infant cry detection device 3 according to the third embodiment can obtain the following effects by adopting the following configuration.

  The infant cry detection device 3 includes a microphone 10 that collects sound as a sound collection unit, and a sound pressure level detection unit 13 that detects a sound pressure level of a sound signal collected by the microphone 10.

  The infant cry detection device 3 includes a continuity determination unit 14. The continuity determination unit 14 continues to the collected sound signal for the first time that is equal to or higher than the first sound pressure level for the first time and for the second time that is equal to or lower than the second sound pressure level. It is determined whether or not the second sound signal is included.

  The infant cry detection device 3 includes a periodicity determination unit 15. When it is determined that the first sound signal and the second sound signal are included in the collected sound signal, the periodicity determination unit 15 performs the first sound signal and the second sound signal, Determines whether or not alternately repeat at least twice.

  The infant cry detection device 3 includes a frequency conversion unit 121. The frequency conversion unit 121 converts the frequency of the first sound signal in parallel with the operations of the continuity determination unit 14 and the periodicity determination unit 15.

  The infant cry detection device 3 includes a harmonic overtone determination unit 122. The harmonic overtone determination unit 122 determines whether the frequency-converted first sound signal has a harmonic overtone structure in which the peak of the frequency spectrum is composed of a fundamental frequency and a higher order frequency. The overtone determination unit 122 determines whether or not the overtone structure continues for a preset time when the first sound signal has a overtone structure. The harmonic overtone determination unit 122 performs the determination operation in parallel with the operations of the continuity determination unit 14 and the periodicity determination unit 15.

  The infant cry detection device 3 includes a determination unit 124. When the following determination result is obtained, the determination unit 124 determines that the collected sound signal includes the infant's cry, and detects the infant's cry. The determination result includes a determination result in which the first sound signal and the second sound signal are alternately repeated at least twice, and the first sound signal has a harmonic structure and the harmonic structure continues for a preset time. It is a judgment result.

  By adopting the above-described configuration, the infant cry detection apparatus 3 can detect an infant cry by distinguishing it from various noises in the background, in particular, a notification sound similar in characteristics to the infant cry. As a result, the frequency of erroneous detection can be reduced, and the accuracy of detecting an infant's cry can be increased.

  The infant cry detection device 3 performs the operations of the continuity determination unit 14 and the periodicity determination unit 15 in parallel with the operations of the frequency conversion unit 121 and the harmonic overtone determination unit 122.

  The infant cry detection device 3 employs the above technical features to detect the infant cry more quickly than the infant cry detection device 1 of the first embodiment and the infant cry detection device 2 of the second embodiment. be able to.

  The infant cry detection device 3 includes a frequency variation determination unit 80. The frequency fluctuation determination unit 80 generates a frequency fluctuation that returns to the original frequency after the peak fundamental frequency and the higher order frequency in the frequency spectrum of the frequency-converted first sound signal fluctuate to a high frequency with time. It is determined whether or not.

  If the determination unit 124 of the infant cry detection apparatus 3 determines that the first sound signal has a frequency variation in addition to the above determination result, the collected sound signal includes the infant cry. It is determined that the baby is crying and the infant's cry is detected.

  By providing the frequency variation determination unit 80, the infant cry detection device 3 can further reduce the frequency of erroneous detection and increase the accuracy of detecting an infant cry.

10 Microphone (sound collector)
DESCRIPTION OF SYMBOLS 13 Sound pressure level detection part 14 Continuity determination part 15 Periodicity determination part 16 1st determination part 19,121 Frequency conversion part 20,122 Overtone determination part 21,81 2nd determination part 80 Frequency fluctuation determination part 124 Determination part

Claims (6)

A sound collection section for collecting sounds;
A sound pressure level detection unit for detecting a sound pressure level of a sound signal collected by the sound collection unit;
The sound signal for which the sound pressure level is detected by the sound pressure level detection unit includes a first sound signal that is equal to or higher than the first sound pressure level for a first time, and a second sound signal that is lower than the first sound signal. A continuity determination unit that determines whether or not a second sound signal that is lower than a sound pressure level and shorter than the first time for a second time is included;
When it is determined that the first sound signal and the second sound signal are included in the sound signal collected by the sound collection unit, the first sound signal and the second sound signal A periodicity determining unit for determining whether or not the signal repeats alternately at least twice;
When it is determined that the first sound signal and the second sound signal are alternately repeated at least twice, the sound signal collected by the sound collecting unit may include a crying sound of an infant A first determination unit that determines that there is
A frequency conversion unit that converts the frequency of the first sound signal when it is determined that the sound signal collected by the sound collection unit may contain a cry of an infant;
When the frequency-converted first sound signal has a harmonic structure by determining whether or not the peak of the frequency spectrum has a harmonic structure composed of a fundamental frequency and a higher-order frequency. A harmonic overtone determination unit that determines whether or not the overtone structure continues for a preset time;
It is determined that there is a possibility that an infant's cry is included in the sound signal collected by the sound collection unit, and the first sound signal has a harmonic structure, and the harmonic structure is continued for a preset time. A second determination unit that determines that the infant cry is included in the sound signal collected by the sound collection unit and detects the infant cry;
An infant cry detecting device characterized by comprising: If it is not determined that there is a possibility that the collected sound contains an infant cry, the frequency converter does not perform frequency conversion, and the harmonic overtone determination unit and the second determination unit perform the determination operation. The infant cry detection device according to claim 1, wherein the detection is not performed. It is determined whether or not there is a frequency fluctuation that returns to the original frequency after the peak fundamental frequency and the higher order frequency in the frequency spectrum of the first sound signal that has been frequency-converted fluctuate to a high frequency with time. It has a frequency variation determination unit,
The second determination unit determines that an cry of an infant is included in the sound signal collected by the sound collection unit when it is determined that a frequency variation occurs in the first sound signal. The infant cry detection apparatus according to claim 1, wherein an infant cry is detected. If it is not determined that there is a possibility that the collected sound includes an infant's cry, the frequency conversion unit does not perform frequency conversion, and the harmonic overtone determination unit, the frequency fluctuation determination unit, and the second The infant cry detection apparatus according to claim 3, wherein the determination unit does not perform a determination operation. A sound collection section for collecting sounds;
A sound pressure level detection unit for detecting a sound pressure level of a sound signal collected by the sound collection unit;
The sound signal whose sound pressure level is detected by the sound pressure level detection unit includes a continuous first sound signal that is equal to or higher than the first sound pressure level for a first time, and lower than the first sound signal. A continuity determination unit that determines whether or not a second sound signal that is continuous for a second time that is less than or equal to a sound pressure level of 2 and shorter than the first time is included;
When it is determined that the first sound signal and the second sound signal are included in the sound signal collected by the sound collection unit, the first sound signal and the second sound signal A periodicity determining unit for determining whether or not the signal repeats alternately at least twice;
A frequency conversion unit that converts the frequency of the first sound signal in parallel with the operations of the continuity determination unit and the periodicity determination unit;
When the frequency-converted first sound signal has a harmonic structure, it is determined whether or not the peak of the frequency spectrum has a harmonic structure composed of a fundamental frequency and a higher order frequency. Determining whether the overtone structure continues for a preset time, and performing a determination operation in parallel with the operations of the continuity determination unit and the periodicity determination unit; and
The first sound signal and the second sound signal are included in the sound signal collected by the sound collecting unit, and the first sound signal and the second sound signal are alternately alternately at least twice. When it is determined that the first sound signal has a harmonic structure and the harmonic structure continues for a preset time, the infant's crying sound is added to the sound signal collected by the sound collecting unit. And a determination unit for detecting an infant's cry,
An infant cry detecting device characterized by comprising: It is determined whether or not there is a frequency fluctuation that returns to the original frequency after the peak fundamental frequency and the higher order frequency in the frequency spectrum of the first sound signal that has been frequency-converted fluctuate to a high frequency with time. It has a frequency variation determination unit,
When the determination unit determines that a frequency variation has occurred in the first sound signal, the determination unit determines that the sound signal collected by the sound collection unit includes an infant cry, 6. The infant cry detection apparatus according to claim 5, wherein an infant cry is detected.