APPARATUS AND METHOD FOR DETECTING SOUND

• US 20100290632A1
• Filed: 11/15/2007
• Published: 11/18/2010
• Est. Priority Date: 11/20/2006
• Status: Active Grant

• Current Assignee: Panasonic Corporation (Panasonic Holdings Corporation)
• Original Assignee: Panasonic Corporation (Panasonic Holdings Corporation)
• Inventors: Lin, Che-Ming
• Granted Patent: US 8,098,832 B2
• Time in Patent Office: Days
• US Class Current: 381/56
• CPC Class Codes: G01S 3/8006 Multi-channel systems specially adapted for direction-finding, i.e. having a single aerial system capable of giving simultaneous indications of the directions of different signals
• CPC Class Codes: H04R 1/1083 Reduction of ambient noise active noise reduction per se G10K11/175; protective devices for the ear, e.g. providing acoustic protection A61F11/06
• CPC Class Codes: H04R 5/033 Headphones for stereophonic communication details thereof, e.g. relating to batteries, cables or control elements H04R1/10

Abstract

An apparatus for detecting sound includes a plurality of microphones, a sound inspecting unit, a direction estimating unit, a background noise removing unit, and an alerting unit. The microphones are used to collect sounds around a user. The sound inspecting unit is used to calculate the feature values of a background noise within a preset time interval, and to determine if a latest collected sound satisfies a preset condition. When the preset condition is satisfied, the direction estimating unit is used to estimate the occurrence direction of the latest collected sound, and to determine if the occurrence direction is within a preset range behind the user. When the preset range is satisfied, the background noise removing unit is used to remove the background noise in the latest collected sound so as to obtain a detected sound. The alerting unit is used to inform the user of the detected sound via an alert message. A method for detecting sound is also disclosed.

Claims

1. An apparatus for detecting sound adapted to be worn on both ears of a user so as to detect sounds from behind the user, wherein the sounds include a background noise, said apparatus comprising:

a plurality of microphones for collecting sounds around the user;

a sound inspecting unit for calculating the feature values of the background noise within a preset time interval and for determining if a latest collected sound satisfies a preset condition;

a direction estimating unit for estimating the occurrence direction of the latest collected sound and for determining if the occurrence direction is within a preset range behind the user, when the preset condition is satisfied;

a background noise removing unit for removing the background noise in the latest collected sound so as to obtain a detected sound, when the occurrence direction is within the preset range; and

an alerting unit for informing the user of the detected sound via an alert message.

2. The apparatus for detecting sound as claimed in claim 1, wherein said sound inspecting unit calculates an average value and a standard deviation of the energy spectrum of the background noise within the preset time interval as the feature values, and the preset condition be that the energy of the latest collected sound is larger than the sum of the average value and the standard deviation.

3. The apparatus for detecting sound as claimed in claim 2, wherein said background noise removing unit includes a direction control module and a spectrum subtraction module, said direction control module being used to remove the background noise which is not within the preset range from the latest collected sound that satisfies the preset condition, said spectrum subtraction module being used to remove the background noise which is within the preset range from the latest collected sound that satisfies the preset condition.

4. The apparatus for detecting sound as claimed in claim 3, wherein said direction control module removes the background noise which is not within the preset range from the latest collected sound that satisfies the preset condition by adjusting a weight vector of the direction control of a spatial filter.

5. The apparatus for detecting sound as claimed in claim 3, wherein said spectrum subtraction module first calculates the energy spectrum in the preset range from the latest collected sound that satisfies the preset condition, and then subtracts the average value of the energy spectrum of the background noise from the calculated energy spectrum so as to remove the background noise which is within the preset range from the latest collected sound that satisfies the preset condition, thereby obtaining the detected sound.

6. The apparatus for detecting sound as claimed in claim 1, wherein said background noise removing unit includes a direction control module and a spectrum subtraction module, said direction control module being used to remove the background noise which is not within the preset range from the latest collected sound that satisfies the preset condition, said spectrum subtraction module being used to remove the background noise which is within the preset range from the latest collected sound that satisfies the preset condition.

7. The apparatus for detecting sound as claimed in claim 6, wherein said direction control module removes the background noise which is not within the preset range from the latest collected sound that satisfies the preset condition by adjusting a weight vector of the direction control of a spatial filter.

8. The apparatus for detecting sound as claimed in claim 1, further comprising a sound classifying unit that includes an alert message sound database, said sound classifying unit receiving the detected sound sent by said background noise removing unit, identifying from said alert message sound database an alert message corresponding to the detected sound, and transmitting the alert message to said alerting unit.

9. The apparatus for detecting sound as claimed in claim 8, wherein said sound classifying unit first calculates a Mel-Frequency Cepstral Coefficient of the detected sound, and calculates the Mahalanobis distances between the detected sound and various sound samples in said alert message sound database according to the Mel-Frequency Cepstral Coefficient, an alert message corresponding to a sound sample with the smallest Mahalanobis distance thus calculated serving as the alert message corresponding to the detected sound.

10. The apparatus for detecting sound as claimed in claim 1, wherein said direction estimating unit uses a Direction of Arrival technique to estimate the occurrence direction of the latest collected sound that satisfies the preset condition.

11. The apparatus for detecting sound as claimed in claim 10, wherein the Direction of Arrival technique is the Multiple Signal Classification method.

12. A method for detecting sound, comprising the following steps:

(a) collecting sounds around a user, wherein the sounds include a background noise;

(b) calculating the feature values of the background noise within a preset time interval, and determining if a latest collected sound satisfies a preset condition;

(c) when the preset condition is satisfied, estimating the occurrence direction of the latest collected sound, and determining if the occurrence direction is within a preset range behind the user;

(d) when the occurrence direction is within the preset range, removing the background noise in the latest collected sound so as to obtain a detected sound; and

(e) informing the user of the detected sound via an alert message.

13. The method for detecting sound as claimed in claim 12, wherein step (b) includes calculating an average value and a standard deviation of the energy spectrum of the background noise within the preset time interval as the feature values, and the preset condition be that the energy of the latest collected sound is larger than the sum of the average value and the standard deviation.

14. The method for detecting sound as claimed in claim 13, wherein step (d) includes removing the background noise which is not within the preset range from the latest collected sound that satisfies the preset condition, followed by removing the background noise which is within the preset range from the latest collected sound that satisfies the preset condition.

15. The method for detecting sound as claimed in claim 14, wherein, in step (d), the removal of the background noise which is not within the preset range from the latest collected sound that satisfies the preset condition is conducted by adjusting a weight vector of the direction control of a spatial filter.

16. The method for detecting sound as claimed in claim 14, wherein step (d) includes calculating the energy spectrum in the preset range from the latest collected sound that satisfies the preset condition, followed by subtracting the average value of the energy spectrum of the background noise from the calculated energy spectrum so as to remove the background noise which is within the preset range from the latest collected sound that satisfies the preset condition, thereby obtaining the detected sound.

17. The method for detecting sound as claimed in claim 12, wherein step (d) includes removing the background noise which is not within the preset range from the latest collected sound that satisfies the preset condition, followed by removing the background noise which is within the preset range from the latest collected sound that satisfies the preset condition.

18. The method for detecting sound as claimed in claim 17, wherein, in step (d), the removal of the background noise which is not within the preset range from the latest collected sound that satisfies the preset condition is conducted by adjusting a weight vector of the direction control of a spatial filter.

19. The method for detecting sound as claimed in claim 12, further comprising, between steps (d) and (e), a step (f) of performing sound classification upon obtaining the detected sound in step (d) to identify an alert message corresponding to the detected sound, and transmitting the alert message to step (e).

20. The method for detecting sound as claimed in claim 19, wherein step (f) includes:

calculating a Mel-Frequency Cepstral Coefficient of the detected sound, and calculating the Mahalanobis distances between the detected sound and various sound samples according to the Mel-Frequency Cepstral Coefficient, an alert message corresponding to a sound sample with the smallest Mahalanobis distance thus calculated serving as the alert message corresponding to the detected sound.

21. The method for detecting sound as claimed in claim 12, wherein a Direction of Arrival technique is used in step (c) to estimate the occurrence direction of the latest collected sound that satisfies the preset condition.

22. The method for detecting sound as claimed in claim 21, wherein the Direction of Arrival technique is the Multiple Signal Classification method.