JP5410351B2 - Security device and security method - Google Patents

Description

  The present invention relates to a security device and a security method.

  2. Description of the Related Art A security device that detects an intrusion of a suspicious person in a monitoring area from an image captured by a monitoring camera, a detection signal from a sensor, and a sound collected by a microphone is known. For example, Patent Document 1 proposes a detection device that detects a movement path of an intruder from a position where a sound collected using a microphone is generated. Patent Document 1 describes that a plurality of microphones may be installed so as to surround a house, and a microphone array in which a plurality of microphones are arranged at predetermined intervals may be used. Patent Document 2 proposes a vehicle detection method in which a sound source direction is specified from the sound of a vehicle detected using a microphone array, and it is determined that the vehicle is present in a parking frame.

JP 2009-105612 A Japanese Patent Laid-Open No. 2005-115813

  However, when it is necessary to collect a wide range of sounds as in Patent Documents 1 and 2, for example, if a microphone is increased in order to detect a distant sound, a loud sound is generated near the microphone. There has been a problem that overflow may occur and the accuracy of detection processing by the collected sound may be reduced.

  This invention is made in view of the above, Comprising: It aims at providing the guard apparatus and guard method which can perform the detection process of the monitoring object using a microphone with higher precision.

  In order to solve the above-described problems and achieve the object, the present invention includes a plurality of microphones arranged in a direction substantially perpendicular to a moving surface on which a monitoring target moves, and a sound generated from the monitoring target. A microphone array that collects the sound, a direction detection unit that detects a direction of the monitoring target with respect to the microphone array based on sound collected by the microphone array, and a microphone array according to the direction of the monitoring target. A sensitivity correction unit that corrects sensitivity and an abnormality determination unit that determines occurrence of abnormality based on the direction of the monitoring target are provided.

  According to the present invention, there is an effect that detection processing of a monitoring target using a microphone can be executed with higher accuracy.

FIG. 1 is a block diagram illustrating a configuration example of a security device according to the present embodiment. FIG. 2 is a diagram illustrating a configuration example of a microphone array. FIG. 3 is a diagram for explaining the intruder detection process. FIG. 4 is a flowchart showing an example of the entire flow of intruder detection processing by the security device. FIG. 5 is a diagram for explaining the outline of the sound collection sensitivity correction process. FIG. 6 is a diagram for explaining a specific example of the sensitivity correction process. FIG. 7 is a diagram illustrating an example of a signal waveform when the sensitivity correction process is not performed. FIG. 8 is a diagram illustrating an example of a signal waveform when the sensitivity correction process is performed.

  Exemplary embodiments of a security device and a security method according to the present invention will be explained below in detail with reference to the accompanying drawings.

  FIG. 1 is a block diagram illustrating a configuration example of a security device according to the present embodiment. As shown in FIG. 1, the security device 10 includes a main body 100 and a microphone array 200. Although omitted in FIG. 1, the security device 10 is connected to a monitoring center via a network. When the monitoring center receives a notification from the security device 10 that an abnormality has been detected in the monitoring area, the monitoring center issues an instruction to the waiting guard to go to the monitoring area where the abnormality is detected, and is necessary. It is a center that reports to related organizations such as the police and fire department according to the situation.

  The microphone array 200 has a plurality of microphones and collects sound generated from a sound source including a monitoring target. The microphone array 200 is configured by arranging a plurality of microphones in a predetermined direction. As a result, the direction of the sound source can be estimated. FIG. 2 is a diagram illustrating a configuration example of the microphone array 200.

  As shown in FIG. 2, when an acoustic signal is received by a plurality of spatially arranged microphones, a time difference or an amplitude difference corresponding to the difference in arrival time of sound waves is generated between the received sound signals. The sound source direction detection unit 103 uses the microphone array 200 to estimate the direction of the sound source using the difference between these signals.

In general, by using the microphone array 200, the following functions can be realized.
(1) By arranging the microphones constituting the microphone array 200 in a vertical line, the vertical direction of the sound arrival direction is estimated.
(2) The left and right directions of the sound arrival directions are estimated by arranging the microphones constituting the microphone array 200 in a horizontal row.
(3) The sound collection sensitivity of only sound from a specific direction is increased by signal processing of the acoustic signal output from the microphone array 200.
(4) By the signal processing of the acoustic signal output from the microphone array 200, the sound collection sensitivity is reduced only for the sound from a specific direction, and the noise from that direction is reduced.

  In the present embodiment, an example will be described in which microphones constituting the microphone array 200 are arranged in a vertical row (substantially vertical direction) with respect to a moving surface (ground) on which an intruder or the like moves in a monitoring area. Note that the microphones may be further arranged in a horizontal row with respect to the moving surface so that the horizontal direction of the sound source can also be estimated.

  Returning to FIG. 1, the main unit 100 includes a database 121, a storage unit 122, an AD (analog / digital) conversion unit 101, a suspicious sound determination unit 102, a sound source direction detection unit 103, a footstep determination unit 104, an abnormality A determination unit 105, an alarm output processing unit 106, and a correction amount calculation unit 107 are provided.

  The database 121 stores a footstep feature amount referred to when the footstep determination unit 104 determines that it is a footstep. In the following, a case is described in which a person is a monitoring target and it is determined that the collected sound is a footstep sound of the person, but the monitoring target is not limited to a person, and the sound to be determined is It is not limited to footsteps. If the monitoring target is a specific sound that is predetermined as a sound when moving on a predetermined moving surface such as the ground, any sound can be targeted. In this case, the database 121 may store the feature amount of the specific sound.

  As the feature amount of sound, any feature amount conventionally used, such as frequency feature of sound obtained by frequency analysis of an acoustic signal, can be used.

  The storage unit 122 stores various types of information such as sound source direction information, suspicious sound detection time information, monitoring area information, and abnormality determination value information. The sound source direction information is information indicating the direction of the sound source detected by the sound source direction detecting unit 103. The suspicious sound detection time information represents the time when the suspicious sound determination unit 102 detects the suspicious sound. The monitoring area information is information that identifies the range of the monitoring area, the position of the monitoring area with respect to the microphone array 200, and the like. The abnormality determination value information is information that the abnormality determination unit 105 refers to when determining whether an abnormality has occurred. The monitoring area information and the abnormality determination value information are given as initial settings.

  The AD conversion unit 101 converts an acoustic signal that is an analog signal collected by the microphone array 200 into a digital signal. In addition, the AD conversion unit 101 includes a sound collection sensitivity correction unit 101a that adjusts sound collection sensitivity. The sound collection sensitivity correction unit 101 a adjusts the sound collection sensitivity of the microphone array 200 according to the correction amount calculated by the correction amount calculation unit 107.

  The suspicious sound determination unit 102 extracts a suspicious sound such as a loud sound as pre-processing of the sound source direction estimation processing by the sound source direction detection unit 103 and the footstep determination processing by the footstep determination unit 104. For example, the suspicious sound determination unit 102 calculates the sound pressure level of the sound from the converted digital signal, and determines that the sound is suspicious when the calculated value is greater than a predetermined threshold.

  The sound source direction detection unit 103 calculates the sound source direction of the suspicious sound extracted by the suspicious sound determination unit 102. For example, the sound source direction detection unit 103 calculates an angle α that represents the direction of the suspicious sound relative to the vertically downward direction.

  The footstep determination unit 104 determines whether the suspicious sound extracted by the suspicious sound determination unit 102 is a footstep. For example, the footstep determination unit 104 analyzes the suspicious sound and calculates a feature amount of the suspicious sound. The calculated feature amount matches or is similar to the footstep feature amount stored in the database 121, and the sound source direction of the suspicious sound However, it is determined that the suspicious sound is the footstep when the footstep is within the range of a predetermined direction in which the footstep is assumed to be detected.

  The abnormality determination unit 105 determines the occurrence of abnormality with reference to the direction of the sound source. For example, the abnormality determination unit 105 performs addition and subtraction of abnormality determination points using the footstep direction information (sound source direction information) and the temporal change of the sound source direction information. The abnormality determination point is a value for determining whether or not an abnormality has occurred. The abnormality determination unit 105 determines that an abnormality has occurred when the abnormality determination point exceeds a predetermined threshold.

  The alarm output processing unit 106 outputs an alarm when the abnormality determination unit 105 determines that an abnormality has occurred. For example, the alarm output processing unit 106 outputs an alarm from a speaker or the like (not shown). Further, the alarm output processing unit 106 notifies the monitoring center of abnormality detection via the network.

  The correction amount calculation unit 107 calculates the correction amount of the sound collection sensitivity of the microphone array 200 using the direction information of the footsteps and the time change of the direction information. The calculated correction amount is transmitted to the AD conversion unit 101.

  Next, an overview of intruder detection processing by the security device 10 according to the present embodiment will be described with reference to FIG. FIG. 3 is a diagram for explaining the intruder detection process.

  As shown in FIG. 3, in the present embodiment, the up-down direction of the collected sound is estimated by using microphone arrays 200 arranged in a vertical row. FIG. 3 shows an example in which a vertical microphone array 200 is installed in the vicinity of a window of a building that is a target and the outdoors is wary. As shown in FIG. 3, when the installation position of the microphone array 200 and the position of the boundary of the monitoring area are determined, the monitoring boundary angle αth indicating the direction from the microphone array 200 to this boundary is determined.

  When the angle α representing the direction estimation result of the detected suspicious sound is larger than αth (α> αth), the abnormality determining unit 105 detects the detected suspicious sound outside the monitoring region boundary, Judged as noise.

  When the angle α representing the direction estimation result of the detected suspicious sound is smaller than αth (α <αth), the footstep determination unit 104 determines whether or not it is a human footstep from the frequency characteristics of the sound. . At this time, sound generation means for amplifying footsteps such as crime prevention gravel may be laid on the ground to make it easier to recognize the footsteps.

When the direction of the footstep detected at a certain time t1 is α (t1) and the direction of the footstep detected at time t2 after time t1 is α (t2), the abnormality determination unit 105 The movement is judged as follows.
α (t2)> α (t1): it is determined that the footsteps are far away α (t2) <α (t1): it is determined that the footsteps are approaching

  The abnormality determination unit 105 gives an abnormality determination point P according to the angle when a footstep is detected in the monitoring area. For example, the abnormality determination unit 105 gives a high abnormality determination point P when the angle is sufficiently small with respect to αth (when sufficiently close to the target). The abnormality determination point P is added when the footstep approaches, and is subtracted when the footstep moves away. When the footsteps continue to approach and the abnormality determination point P exceeds a predetermined set value Pth, the abnormality determination unit 105 determines that an abnormality has occurred due to an intruder. Then, the alarm output processing unit 106 performs measures such as alarm output and notification to the monitoring center. Depending on the abnormality determination point P, appropriate measures such as threatening may be performed.

  Next, intruder detection processing by the security device 10 of the present embodiment will be described. FIG. 4 is a flowchart showing an example of the overall flow of the intruder detection process by the security device 10.

  When sound collection is started by the microphone array 200 (step S101), the sound source direction detection unit 103 calculates an angle α representing the direction of the collected sound (step S102). The abnormality determination unit 105 determines whether or not the angle α is smaller than the monitoring boundary angle αth (step S103). If not smaller (step S103: No), the process returns to step S101 and repeats the process.

  When the angle α is smaller than the monitoring boundary angle αth (step S103: Yes), the footstep determination unit 104 determines whether or not the collected sound is a footstep (step S104). As described above, the suspicious sound determination unit 102 determines whether or not the collected sound is a suspicious sound, and the footstep determination unit 104 determines whether or not the suspicious sound is a footstep when it is a suspicious sound. To do. The processing by the suspicious sound determination unit 102 may be omitted.

  If the collected sound is not a footstep (No at Step S104), the process returns to Step S101 and the process is repeated. When the collected sound is a footstep (step S104: Yes), the abnormality determination unit 105 determines whether or not the detection of the footstep is the first time (step S105). In the first case (step S105: Yes), the abnormality determination unit 105 sets an initial value of the abnormality determination point P (step S106). The initial value of the abnormality determination point P is set to a value that increases when the angle α is small, for example.

  If it is not the first time (step S105: No), or after setting the initial value of the abnormality determination point P, a sound collection sensitivity correction process for correcting the sound collection sensitivity of the microphone array 200 is executed (step S107). Details of the sound collection sensitivity correction processing will be described later. Note that the execution timing of the sound collection sensitivity correction processing is not limited to the timing shown in FIG. 4, and any timing may be used as long as the angle α representing the direction of the sound source is obtained.

  Next, the abnormality determination unit 105 determines a time change of the angle α. For example, the abnormality determination unit 105 determines whether or not the angle α (t2) at the current time t2 is smaller than the angle α (t1) at the time t1 when the previous footstep was collected (step S108).

  When the angle α (t2) is not smaller than the angle α (t1) (step S108: No), the abnormality determining unit 105 further determines whether the angle α (t2) is larger than the angle α (t1) (step S108). S109). When the angle α (t2) is larger than the angle α (t1) (step S109: Yes), the abnormality determination unit 105 subtracts the abnormality determination point P (step S110). If the angle α (t2) is not larger than the angle α (t1) (step S109: No), the angle α has not changed, so the abnormality determination unit 105 does not subtract the abnormality determination point P and returns to step S101. repeat.

  When the angle α (t2) is smaller than the angle α (t1) (step S108: Yes), the abnormality determination unit 105 adds the abnormality determination point P (step S111).

  Next, the abnormality determination unit 105 determines whether or not the abnormality determination point P is greater than the set value Pth (step S112). When the abnormality determination point P is not larger than the set value Pth (step S112: No), the process returns to step S101 and is repeated.

  When the abnormality determination point P is larger than the set value Pth (step S112: Yes), the alarm output processing unit 106 outputs an alarm from a speaker or the like (step S113) and reports via the network (step S114). To do.

  Next, the sound collection sensitivity correction process in step S107 will be described.

  In general, the energy of sound waves (sound pressure level) attenuates in inverse proportion to the square of the distance from the sound source. When footsteps are detected using the sound pressure level, if the distance to the sound source is long, the sound pressure level may be attenuated, resulting in detection failure. At this time, it is possible to collect a sound far away by increasing the sound collection sensitivity of the microphone. However, if a nearby large sound is collected with the sound collecting sensitivity that can collect a distant sound, an overflow may occur such that the amplitude of the acoustic signal of the collected sound exceeds a predetermined upper limit value. When an overflow occurs, the acoustic signal cannot be accurately captured, and there is a possibility that the frequency analysis or the like may be disturbed.

  Therefore, in the present embodiment, the sound collection sensitivity is corrected using the sound source direction (angle) obtained by using the microphone array 200.

FIG. 5 is a diagram for explaining the outline of the sound collection sensitivity correction process. The definitions of symbols used in FIG. 5 are described below.
L [m]: Installation height of microphone array 200 from the ground t1, t2: Time when footsteps are detected (t2> t1)
d [m]: Distance from microphone array 200 to footstep generation position (ground) α [°]: Angle in footstep sound source direction S [dB]: Sound pressure level of footstep

(Derivation of formula)
From FIG. 5, the distance d to the sound source is calculated by the following equation (1).
d = L / cos α (1)

Since the footstep sound as the target sound is considered as a point sound source, the distance attenuation formula is as shown in the following formula (2). S (t) represents the sound pressure level of footsteps at time t.
S (t1) -S (t2) =-20log (d (t1) / d (t2)) (2)

Substituting equation (1) into equation (2) yields the following equation (3) as a sensitivity correction equation for calculating the sensitivity correction amount C.
C = S (t1) −S (t2)
= -20log (cosα (t2) / cosα (t1)) (3)

  The correction amount calculation unit 107 calculates the correction amount C of the sensitivity of the microphone array 200 according to the angle α by the above equation (3). Then, the sound collection sensitivity correction unit 101a corrects the sensitivity of the microphone array 200 using the calculated correction amount C.

  As a result, the sensitivity of the microphone array 200 can be appropriately corrected even when the microphone array 200 installed in a building window or the like collects sounds from a relatively wide monitoring area. That is, it is possible to maintain the optimum sound collection by adjusting the sound collection sensitivity so that the corrected sound pressure level becomes equal to that before the correction. In addition, it is possible to suppress the occurrence of overflow due to excessively high sound collection sensitivity. That is, it is possible to execute the detection process of the monitoring target using the microphone with higher accuracy.

  Note that the initial value of the sound collection sensitivity is set in advance to a sensitivity at which footsteps generated at the farthest position in the monitoring area can be collected.

  Next, a specific example of the sensitivity correction process will be described with reference to FIGS. FIG. 6 is a diagram for explaining a specific example of the sensitivity correction process.

  Hereinafter, a case where the estimated angle α of the footstep direction changes as 60 °, 45 °, 30 °, and 15 ° will be described as an example. Footstep generation positions P1 to P4 in FIG. 6 represent positions of sound sources (footsteps) corresponding to angles α of 60 °, 45 °, 30 °, and 15 °, respectively. Hereinafter, the angle α and the sound pressure level of the footsteps at the footstep generation position P (P1 to P4) are respectively expressed as α (P) and S (P).

(1) When the angle α (P1) = 60 ° is moved from α (P2) = 45 ° When the angle value is substituted into the above equation (3), the correction amount C can be obtained as follows.
C = S (P1) -S (P2)
= -20 log {cos α (p2) / cos α (p1)}
= -20log (2 / √2)
= −3.01 [dB]

  Therefore, the sound collection sensitivity correction unit 101a corrects the sound collection sensitivity of the microphone array 200 by −3.01 dB.

(2) When the angle α (P2) = 45 ° is moved to α (P3) = 30 ° When the angle value is substituted into the above equation (3), the correction amount C can be obtained as follows.
C = S (P2) -S (P3)
= −20 log {cos α (p3) / cos α (p2)}
= -20log (√6 / 2)
= -1.76 [dB]

  Therefore, the sound collection sensitivity correction unit 101a corrects the sound collection sensitivity of the microphone array 200 by −1.76 dB.

  FIG. 7 is a diagram illustrating an example of a signal waveform when such sensitivity correction processing is not performed. FIG. 8 is a diagram illustrating an example of a signal waveform when the sensitivity correction process is performed. 7 and 8 show signal waveforms when the footstep generation position changes from P1 to P4 as shown in FIG.

  As shown in FIG. 7, when the sensitivity correction process is not performed, an overflow in which the amplitude of the signal waveform exceeds the upper limit occurs as the angle α decreases as it approaches the target (P3 and P4). On the other hand, as shown in FIG. 8, when the sensitivity correction process is performed, the sensitivity is adjusted according to the angle α. Therefore, if the initial sensitivity value is set appropriately, the occurrence of overflow of the signal waveform is avoided. It becomes possible.

DESCRIPTION OF SYMBOLS 10 Security apparatus 100 Main body part 101a Sound collection sensitivity correction part 101 AD conversion part 102 Suspicious sound determination part 103 Sound source direction detection part 104 Foot sound determination part 105 Abnormality determination part 106 Alarm output process part 107 Correction amount calculation part 121 Database 122 Storage part 200 Microphone array

Claims (6)

A microphone array having a plurality of microphones arranged in a direction substantially perpendicular to a moving surface on which the monitoring target moves, and collecting sounds generated from the monitoring target;
A direction detection unit that detects a direction of the monitoring target with respect to the microphone array, based on sound collected by the microphone array;
A sensitivity correction unit for correcting the sensitivity of the microphone array according to the direction of the monitoring target;
An abnormality determination unit that determines the occurrence of an abnormality based on the direction of the monitoring target;
A security device comprising: The sensitivity correction unit reduces the sensitivity of the microphone array as the direction of the monitoring target is closer to a substantially vertical direction;
The security device according to claim 1. The direction detection unit detects an angle α (t) indicating the direction of the monitoring target with respect to a vertical downward direction at time t,
Correction amount C of sensitivity of the microphone array at time t2 (t2> t1) with respect to sensitivity of the microphone array at time t1 is calculated by C = −20log (cosα (t2) / cosα (t1)). A calculation unit;
The sensitivity correction unit corrects the sensitivity of the microphone array by the calculated correction amount C;
The security device according to claim 1. The direction detection unit detects an angle α (t) indicating the direction of the monitoring target with respect to a vertical downward direction at time t,
The abnormality determination unit compares the angle α (t) with a predetermined threshold and determines that an abnormality has occurred when the angle α (t) is smaller than the threshold;
The security device according to claim 1. The feature amount of the sound collected by the microphone array is calculated, and the calculated feature amount is compared with the feature amount of a predetermined specific sound representing the sound that the monitoring target moves on the moving surface. A determination unit that determines that the sound collected by the microphone array is the specific sound when the characteristic amount matches the characteristic amount of the specific sound;
The abnormality determining unit compares the angle α (t) with the threshold when the sound collected by the microphone array by the determining unit is determined to be the specific sound, and compares the angle α (t determining that an abnormality has occurred when t) is less than the threshold;
The security device according to claim 4. A security method executed by a security device having a plurality of microphones arranged in a substantially vertical direction with respect to a moving surface on which a monitoring target moves and having a microphone array for collecting sounds generated from the monitoring target. And
A direction detecting step for detecting a direction of the monitoring target with respect to the microphone array, based on sound collected by the microphone array;
A sensitivity correction step of correcting the sensitivity of the microphone array according to the direction of the monitoring target;
An abnormality determination step for determining occurrence of an abnormality based on the direction of the monitoring target;
Security method characterized by including.

Images