JP2017193207A - In-cabin conversation support device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an in-cabin conversation support device capable of transmitting voice of an interlocutor to a lister more clearly.SOLUTION: A monitor ECU 30 constitutes a sound collecting part together with a sound collecting microphone 4, and collects sound information IS inside a cabin. Also, the monitor ECU 30 as a passenger detection part detects passengers in the cabin based on photographed image IP photographed by a camera 3. Furthermore, the monitor EU 30 as a conversation determination part and an interlocutor identification part determines whether or not the collected sound information IS shows the conversation of the passengers, and in the case where the sound information IS shows the conversation of the passengers, identifies the interlocutors of the conversation. Then, in the case where the sound information IS shows the conversation of passengers, the monitor ECU 30 selectively collects the voices of the interlocutors and outputs it, by forming a selective collecting and outputting part together with the sound collecting microphone 4 and a loudspeaker 5 provided in the cabin.SELECTED DRAWING: Figure 4

Description

  The present invention relates to a vehicle interior conversation support device.

  2. Description of the Related Art Conventionally, there is a vehicle interior conversation assist device that amplifies a passenger's voice collected in a vehicle interior with a speaker. For example, the vehicle interior conversation assisting device described in Patent Document 1 includes a sound collection microphone and a speaker provided on a ceiling portion of the vehicle compartment on the front seat side and the rear seat side of the vehicle, respectively. And, by this, the sound picked up by the front seat side sound collecting microphone is output from the rear seat side speaker, and the sound picked up by the rear seat side sound collecting microphone is output from the front seat side speaker, It is the structure which assists the conversation between the passenger | crew seated in the front seat and the passenger | crew seated in a backseat.

  In this conventional example, the sound collecting microphone on the front seat side is provided on the front side of the speaker on the front seat side, and the sound collecting microphone on the rear seat side is behind the speaker on the rear seat side. On the side. The sound collecting microphone on the front seat side has a sound collecting area directed toward the vehicle front side, and the sound collecting microphone on the rear seat side has a sound collecting area directed toward the vehicle rear side.

  That is, by adopting such a configuration, it becomes difficult for the sound collecting microphone to pick up the output of the speaker. As a result, it is possible to suppress the occurrence of howling and realize a clear sound output.

Japanese Patent No. 5122414

  However, in an actual vehicle, a sound collection microphone provided in the vehicle interior picks up various sounds inside and outside the vehicle. And in the structure of the said prior art, since these noises still remain the possibility of hindering the conversation, the room for improvement is still left in this point.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide a vehicle interior conversation support device that can more clearly convey the voice of the speaker to the listener.

  A vehicle interior conversation support device that solves the above problems includes a sound collection unit that collects sound information in the vehicle interior, an occupant detection unit that detects an occupant in the vehicle interior, and an occupant in which the sound information is detected in the vehicle interior A conversation determination unit that determines whether the conversation indicates a conversation of the passenger, a speaker identification unit that identifies a speaker of the conversation when the sound information indicates the conversation of the occupant, and the speaker And a selective collection output unit that selectively collects and outputs the voice of the above.

  According to the above configuration, the voice of the speaker can be transmitted to the listener more clearly even in a situation where the sound inside and outside the vehicle hinders the conversation. And thereby, the convenience of a passenger can be improved.

  In the vehicle interior conversation support device that solves the above problem, a seating position specifying unit that specifies a seating position of the occupant in the vehicle interior, a sound source specifying unit that specifies a sound source of the sound information, and a sound source of the sound information are A sound source occupant coincidence determining unit that determines whether or not the seating position of any occupant detected in the passenger compartment coincides, and the speaker identification unit includes any one of the sound sources of the sound information It is preferable to identify the occupant as the speaker of the conversation when the seating position coincides with the occupant.

According to the above configuration, the speaker of the conversation can be specified with high accuracy. As a result, the voice of the occupant who is the speaker can be transmitted to the listener more clearly.
A vehicle interior conversation support device that solves the above problem includes a plurality of directional microphones provided in the vehicle interior in directions different from each other, and the selected collection output unit faces the direction of the speaker. Preferably, the directional microphone is selected to collect the voice of the speaker.

According to the above configuration, the voice of the speaker can be selectively collected with high accuracy.
A vehicle interior conversation support device that solves the above-described problem includes a plurality of directional speakers provided in the vehicle interior in directions different from each other, and the selective collection output unit includes the selectively collected utterances. It is preferable to output a person's voice from the directional speaker facing the direction of an occupant other than the speaker detected in the passenger compartment.

According to the above configuration, the voice of the speaker can be efficiently transmitted to other occupants who are listeners.
The vehicle interior conversation support device that solves the above-described problem preferably includes a noise filter that attenuates noise other than voice included in the sound information.

  According to the above configuration, the voice of the speaker can be transmitted to the listener more clearly.

  According to the present invention, the voice of the speaker can be transmitted to the listener more clearly.

Explanatory drawing which shows the monitor unit (camera, sound collection microphone, and speaker) provided in the vehicle interior. Explanatory drawing of the photographed image projected by the camera (no occupant). The schematic block diagram (bottom view) of a monitor unit. The schematic block diagram of a vehicle interior conversation assistance apparatus. Explanatory drawing of the image captured by the camera (with occupants). The flowchart which shows the start procedure of vehicle interior conversation assistance control. The flowchart which shows the process sequence of vehicle interior conversation assistance control.

Hereinafter, an embodiment of a vehicle interior conversation support device will be described with reference to the drawings.
As shown in FIG. 1, the vehicle 1 of the present embodiment is provided with a camera 3 that displays an image in the passenger compartment 2. Specifically, in the vehicle 1 of the present embodiment, the camera 3 constitutes a monitor unit 10 that is integrated with the sound collecting microphone 4 and the speaker 5. And the vehicle 1 of this embodiment becomes the structure by which this monitor unit 10 is arrange | positioned in the approximate center part in the ceiling part 2a of the vehicle interior 2, the position between the front seat 7f and the rear seat 7r. Yes.

  More specifically, as shown in FIGS. 2 and 3, the monitor unit 10 according to the present embodiment has an omnidirectional image of all the seats 7 and all the vehicle windows 20 provided in the passenger compartment 2 as the camera 3. A camera 23 is provided.

  In the captured image IP shown in FIG. 2, the seat 7 shown in the lower left is the right front seat (driver's seat) 7fa, and the seat 7 shown in the lower right is the left front seat (passenger seat) 7fb. The seats 7 shown on the upper side are a right rear seat 7ra and a left rear seat 7rb, respectively. In the photographed image IP shown in FIG. 2, as the vehicle window 20, the windshields 20 fa, 20 fb, 20 ra, 20 rb located on the sides of the seats 7, the windshield 20 fc and the rear glass 20 rc of the vehicle 1 are displayed. It is reflected.

  As shown in FIG. 3, the monitor unit 10 is provided with a microphone array 24 </ b> X having a plurality of directional microphones 24 (24 a to 24 d) facing different directions as the sound collection microphone 4. . The monitor unit 10 is provided with a speaker array 25 </ b> X including a plurality of directional speakers 25 (25 a to 25 d) that are similarly directed in different directions as the speakers 5.

  Specifically, in the monitor unit 10 of the present embodiment, the microphone array 24X has a configuration in which the directional microphones 24 are arranged in a radial pattern around the position where the omnidirectional camera 23 is provided. Similarly, the speaker array 25X has a configuration in which the directional speakers 25 are arranged radially.

  As shown in FIG. 4, in the vehicle 1 of the present embodiment, the monitor unit 10 is connected to a monitor ECU 30. Further, the monitor ECU 30 of the present embodiment analyzes the captured image IP reflected by the camera 3 and the sound information IS collected by the sound collection microphone 4. The monitor ECU 30 according to the present embodiment performs image analysis processing when an authentication ECU connected via an in-vehicle network (not shown) detects the proximity of a person to the vehicle 1 by wireless communication with a portable device (electronic key). And sound information analysis processing is started. Further, the monitor ECU 30 executes sound output via the speaker 5. And in the vehicle 1 of this embodiment, the vehicle interior conversation assistance apparatus 31 using the function of this monitor unit 10 is formed by this.

  More specifically, as shown in FIGS. 2 and 5, the monitor ECU 30 of the present embodiment analyzes the photographed image IP of the camera 3 provided in the passenger compartment 2 to analyze the occupant 40 who has boarded the vehicle 1. Presence / absence and its seating position are detected. Specifically, the monitor ECU 30 uses the well-known human authentication logic such as whether or not the captured image IP includes “a portion where it can be determined that a human face is reflected”. The occupant 40 is detected. Thus, for example, in the example shown in FIG. 5, the first occupant 41 seated in the right front seat (driver's seat) 7fa of the vehicle 1 and the second occupant 42 seated in the left rear seat 7rb are provided. It is configured to be detected.

  Further, the monitor ECU 30 of the present embodiment generates a sound source of the sound information IS collected using the sound collection microphone 4 based on the comparison of the input levels of the directional microphones 24 constituting the microphone array 24X, that is, the sound collection volume. Identify. The monitor ECU 30 determines whether or not the sound source of the sound information IS matches the seating position of any occupant 40 detected in the passenger compartment 2.

  Specifically, the monitor ECU 30 of the present embodiment identifies the sound source direction of the sound information IS with reference to the position where the monitor unit 10 including the microphone array 24X is provided, that is, the substantially central portion of the ceiling 2a. . Then, when there is a seat 7 on which the occupant 40 is seated in the sound source direction, it is determined that the seating position of the occupant 40 matches the sound source.

  Further, the monitor ECU 30 of the present embodiment determines whether or not the sound information IS collected using the sound collection microphone 4 indicates the conversation of the occupant 40. In the vehicle 1 of this embodiment, the monitor ECU 30 has a known voice recognition function. Furthermore, the monitor ECU 30 of the present embodiment is such that the sound information IS in the passenger compartment 2 indicates the conversation of the occupant 40, and the sound source of the sound information IS matches the seating position of any occupant 40. The occupant 40 is specified as the speaker 51 of the conversation. And the operation | movement of the monitor unit 10 is controlled so that the audio | voice of this speaker 51 may be selectively collected and output toward the passenger | crew 40 used as the listener 52 other than the speaker 51. FIG.

  For example, in the example shown in FIG. 5, when the first occupant 41 sitting on the right front seat 7fa of the vehicle 1 is speaking, the first of the directional microphones 24a to 24d constituting the microphone array 24X. The input level (collected sound volume) of the directional microphone 24a facing the right front seat 7fa on which the first passenger 41 is seated is maximized (see FIG. 3). 3 is a view of the monitor unit 10 provided on the ceiling 2a as viewed from below. In FIG. 3, the upper side is the front side of the vehicle and the left side is the right side of the vehicle. And monitor ECU30 of this embodiment specifies the 1st crew member 41 as speaker 51 of a conversation by this.

  Further, the monitor ECU 30 of the present embodiment collects the sound information IST collected by using the directional microphone 24a facing the direction of the first occupant 41 specified by the speaker 51, that is, the voice of the speaker 51. The sound is output from the directional speaker 25d facing the direction of the left rear seat 7rb on which the second occupant 42 is seated, that is, toward the occupant 40 who becomes the listener 52 other than the speaker 51. Furthermore, the monitor ECU 30 of the present embodiment includes a noise filter 55 that attenuates noise other than sound included in the sound information IS collected using the sound collection microphone 4. Thus, the vehicle interior conversation support device 31 of the present embodiment can clearly convey the voice of the speaker 51 to the other passengers 40 who are the listeners 52.

Next, the start procedure of the vehicle interior conversation support control executed by the monitor ECU 30 of the present embodiment and the execution procedure thereof will be described.
As shown in the flowchart of FIG. 6, when the monitor ECU 30 of the present embodiment performs image analysis on the captured image IP in the passenger compartment 2 imaged by the camera 3 (step 101), the passenger compartment 2 is based on the analysis result. It is determined whether or not there is a passenger 40 inside (step 102). When it is determined that there is an occupant 40 in the passenger compartment 2 (step 102: YES), the passenger compartment conversation support control is performed to selectively collect and output the voice of the occupant 40 as the speaker 51. It is configured to execute (step 103).

  Further, when the monitor ECU 30 of the present embodiment determines in step 102 that there is no occupant 40 in the passenger compartment 2 (step 102: NO), the vehicle 1 is in a door-locked state, that is, locked. It is determined whether or not it is in a state (step 104). If it is determined that the vehicle 1 is in the door lock state (step 104: YES), the security control of the vehicle 1 based on the image analysis is executed using the captured image IP around the vehicle 1 displayed by the camera 3. It is configured (step 105).

  More specifically, as shown in the flowchart of FIG. 7, the monitor ECU 30 of the present embodiment first detects the seating position of the occupant 40 detected by image analysis in the vehicle interior conversation support control (see FIG. 6, step 103). That is, the seat 7 on which the occupant 40 is seated is specified (step 201). Next, the monitor ECU 30 analyzes the sound information IS in the vehicle compartment 2 collected using the sound collection microphone 4 of the monitor unit 10 (step 202). Based on the analysis result, the sound source of the sound information IS is specified, specifically, the sound source direction is specified with reference to the position where the sound collection microphone 4 is provided (step 203).

  Next, the monitor ECU 30 determines whether or not the seating position of any occupant 40 identified in step 201 matches the sound source of the sound information IS identified in step 203 (step 204). Furthermore, when the sound source of the sound information IS matches the seating position of any occupant 40 (step 204: YES), the monitor ECU 30 determines whether or not the sound information IS indicates the conversation of the occupant 40. Determination is made (step 205). When the sound information IS indicates the conversation of the occupant 40 (step 205: YES), the occupant 40 that matches the sound source of the sound information IS is specified as the speaker 51 of the conversation (step 206). .

  Next, the monitor ECU 30 selectively collects the voice of the occupant 40 specified as the speaker 51 (step 207). That is, among the directional microphones 24 (24a to 24d) constituting the microphone array 24X of the monitor unit 10, the directional microphone 24 facing the seat 7 on which the speaker 51 is seated is selected and its sound information is selected. The voice of the IST, that is, the speaker 51 is collected. Further, the monitor ECU 30 of this embodiment is included in the sound information IST collected by the directional microphone 24 facing the seat 7 on which the speaker 51 is seated by using the noise filter 55 (see FIG. 4). Noise other than voice is attenuated (noise filter processing, step 208). Further, the monitor ECU 30 of the present embodiment is a seat 7 on which a passenger 40 who is a listener 52 other than the speaker 51 is seated among the directional speakers 25 (25a to 25d) constituting the speaker array 25X of the monitor unit 10. The directional speaker 25 facing the direction is selected. The selected directional speaker 25 outputs the sound information IST after the noise filter processing, that is, the voice of the speaker 51 in a state where the noise is attenuated (step 209).

As described above, according to the present embodiment, the following effects can be obtained.
(1) The monitor ECU 30 constitutes a sound collection unit together with the sound collection microphone 4 to collect sound information IS in the vehicle compartment 2. The monitor ECU 30 serving as an occupant detection unit detects the occupant 40 in the passenger compartment 2 based on the captured image IP that the camera 3 projects. Further, the monitor ECU 30 as the conversation determination unit and the speaker identification unit determines whether or not the collected sound information IS indicates the conversation of the occupant 40, and the sound information IS indicates the conversation of the occupant 40. If it is, the speaker 51 of the conversation is specified. When the sound information IS indicates the conversation of the occupant 40, the monitor ECU 30 forms a selective collection output unit together with the sound collection microphone 4 and the speaker 5 provided in the passenger compartment 2. The voice of the speaker 51 is selectively collected and output.

  According to the above configuration, the voice of the speaker 51 can be transmitted to the listener 52 more clearly even in a situation where the sound inside and outside the vehicle 1 hinders the conversation. And thereby, the convenience of passenger 40 can be improved. In addition, security control of the vehicle 1 can be executed by performing image analysis on the captured image IP projected by the camera 3.

  (2) The monitor ECU 30 serving as the seating position specifying unit specifies the seating position (seat 7 to be seated) of the occupant 40 in the passenger compartment 2 based on the captured image IP captured by the camera 3. Further, the monitor ECU 30 as the sound source identification unit identifies the sound source of the sound information IS by analyzing the sound information IS collected using the sound collection microphone 4 in the passenger compartment 2. Further, the monitor ECU 30 as the sound source occupant coincidence determination unit determines whether or not the sound source of the specified sound information IS matches the seating position of any occupant 40 detected in the passenger compartment 2. When the sound source of the sound information IS matches the seating position of any occupant 40, the monitor ECU 30 as the utterer specifying unit specifies the occupant 40 as the speaker 51 of the conversation.

  According to the said structure, the speaker 51 of the conversation can be pinpointed accurately. As a result, the voice of the occupant 40 as the speaker 51 can be transmitted to the listener 52 more clearly.

  (3) In the vehicle compartment 2, a microphone array 24X having a plurality of directional microphones 24 (24a to 24d) facing different directions is provided. Then, the monitor ECU 30 as the selection collection output unit selects the directional microphone 24 facing the direction of the speaker 51 and collects the voice of the speaker 51. By adopting such a configuration, the voice of the speaker 51 can be selectively collected with high accuracy.

  (4) The monitor ECU 30 as the sound source specifying unit and the sound source occupant match determination unit specifies the sound source direction of the sound information IS based on the position where the microphone array 24X is provided based on the input level of each directional microphone 24. When the seating position of the occupant 40 is in the sound source direction of the sound information IS, it is determined that the seating position of the occupant 40 matches the sound source of the sound information IS. Thereby, it is possible to accurately determine whether the sound source of the sound information IS collected using the sound collection microphone 4 matches the seating position of any occupant 40 detected in the passenger compartment 2.

  (5) In the passenger compartment 2, a speaker array 25X having a plurality of directional speakers 25 (25a to 25d) facing different directions is provided. Then, the monitor ECU 30 as the selective collection output unit outputs the selectively collected voice of the speaker 51 from the directional speaker 25 facing the occupant 40 other than the speaker 51 serving as the listener 52. Thereby, the voice of the speaker 51 can be efficiently transmitted to the other occupant 40 who becomes the listener 52.

  (6) The monitor ECU 30 serving as the speaker sound information output unit includes a noise filter 55 that attenuates noise other than speech included in the sound information IS collected using the sound collection microphone 4. Thereby, the voice of the speaker 51 can be transmitted to the listener 52 more clearly.

In addition, you may change the said embodiment as follows.
In the above embodiment, the camera 3 constitutes the monitor unit 10 integrated with the sound collecting microphone 4 and the speaker 5. And it decided to arrange | position to the approximate center part in the ceiling part 2a of the compartment 2. As shown in FIG. However, the present invention is not limited to this, and the camera 3 does not necessarily have to be provided integrally with the sound collecting microphone 4 and the speaker 5. You may change arbitrarily, such as a pillar. In addition, a configuration using a wide-angle camera such as the omnidirectional camera 23 is not necessarily required, and the number of cameras may be arbitrarily changed. Further, the arrangement and configuration of the sound collecting microphone 4 and the speakers 5 and the number thereof may be arbitrarily changed.

  Specifically, a configuration in which the microphone array 24X is not used for the sound collection microphone 4 or a configuration in which the speaker array 25X is not used for the speaker 5 may be employed. For example, the configuration may include a plurality of directional microphones 24 facing in different directions, and the directional microphones 24 arranged in the vehicle compartment 2 without forming the microphone array 24X. Even when such a configuration is adopted, the sound source of the sound information IS can be specified based on the input level of each directional microphone 24. And while having the several directional speaker 25 which faced the mutually different direction, the structure by which these each directional speaker 25 is arrange | positioned in the compartment 2 in the state which does not form the speaker array 25X may be sufficient.

  Further, the configuration may not necessarily include the directional microphone 24 and the directional speaker 25. For example, in the configuration in which the sound collecting microphones 4 are provided at a plurality of locations in the passenger compartment 2, a sound collecting microphone 4 that is arranged near the occupant 40 specified by the speaker 51 is selected. Then, the voice of the speaker 51 may be collected. Moreover, it is good also as a structure which selects the speaker 5 arrange | positioned in the vicinity of the passenger | crew 40 used as the listener 52, and outputs the audio | voice of the speaker 51. FIG. In the stage of outputting the voice of the speaker 51 collected selectively, the voice output does not necessarily have to be directed to the passenger 40 other than the speaker 51 who becomes the listener 52.

  In the above embodiment, when the authentication ECU detects the proximity of a person to the vehicle 1 by wireless communication with a portable device (electronic key), image analysis processing and sound information analysis processing used for in-vehicle conversation support control are started. It was decided to be done. However, the present invention is not limited to this. For example, the start timing of the image analysis processing and the sound information analysis processing is arbitrarily set including the start trigger of the camera 3 and the sound collection microphone 4 when the vehicle 1 is turned on (IG on). May be.

  In the above embodiment, after determining whether the sound information IS collected using the sound collection microphone 4 matches the seating position of the occupant 40, it is determined whether or not the sound information IS indicates a conversation. (Refer to FIG. 7, steps 204 and 205). However, the present invention is not limited to this, and after the conversation determination, it is possible to determine whether the sound source matches the seating position of the occupant 40. As a result, it is possible to specify the speaker 51 of the conversation with higher accuracy by screening the “conversation” in advance.

  In the above embodiment, the presence / absence of the occupant 40 who has boarded the vehicle 1 and the seating position thereof are detected by analyzing the captured image IP of the camera 3 provided in the passenger compartment 2. However, the present invention is not limited to this, and a configuration may be adopted in which occupant detection is performed using a seating sensor provided on the vehicle seat. The seating sensor may be of any type, such as a pressure sensitive type using a membrane switch or a strain gauge, a capacitance type, etc. As a result, it is possible to accurately detect the presence / absence of the occupant 40 in the passenger compartment 2 and the seating position thereof with a simple configuration without requiring high calculation processing capability.

  In the above embodiment, when the sound source of the sound information IS matches the seating position of any occupant 40, the occupant 40 is specified as the speaker 51 of the conversation. May be arbitrarily changed. For example, the movement of the mouth of the passenger 40 in the passenger compartment 2 is detected by analyzing a captured image IP of the camera 3. Thus, it may be determined whether or not the occupant 40 is a conversation speaker 51.

  In the above embodiment, the voice of the speaker 51 that is selectively collected is output toward the other passenger 40 that becomes the listener 52 detected in the passenger compartment 2. However, the present invention is not limited to this, and human detection around the vehicle 1 is performed through the vehicle window 20 by analyzing the captured image IP of the camera 3. When it is determined that the listener 52 of the conversation is a person detected outside the vehicle (for example, when there is only one passenger 40 detected in the passenger compartment 2), the listener 52 is directed toward the listener 52 outside the vehicle. It is good also as a structure which outputs the audio | voice of the speaker 51. FIG.

  For example, in the configuration in which the speaker array 25X similar to that in the above embodiment is provided in the passenger compartment 2, among the directional microphones 24 that form the speaker array 25X, human detection based on the captured image IP of the camera 3 is performed. The voice of the speaker 51 may be output by selecting the one facing the direction of the vehicle window 20 that is determined to be a person outside the vehicle. In this case, the speaker 5 provided outside the vehicle compartment 2 may be used to output sound to the listener 52 outside the vehicle.

  In the above embodiment, the voice of the speaker 51 is output in a state where the noise is attenuated using the noise filter 55, but such noise attenuation processing is not necessarily performed.

Next, technical ideas that can be grasped from the above embodiments will be described together with effects.
(A) a microphone array provided in the vehicle interior having a plurality of directional microphones facing different directions, wherein the sound source specifying unit is provided by the microphone array based on an input level of each directional microphone; The sound source direction of the sound information with reference to the determined position is determined, and the sound source occupant match determination unit determines that the seat position of the occupant is the sound source when the occupant's seating position is in the sound source direction of the sound information. A vehicle interior conversation support device, characterized in that it is determined that the two match. Thereby, it is possible to accurately determine whether the sound source of the sound information collected using the sound collection microphone matches the seating position of any occupant detected in the passenger compartment.

  (B) The vehicle interior conversation support device, wherein the vehicle occupant detection unit detects a vehicle occupant on the basis of an image captured by a camera provided in the vehicle interior. Thereby, the captured image can be shared with vehicle security control.

  DESCRIPTION OF SYMBOLS 1 ... Vehicle, 2 ... Car compartment, 2a ... Ceiling part, 3 ... Camera (occupant detection part, speaker identification part, seating position identification part, and sound source occupant coincidence determination part), 4 ... Sound collection microphone (sound collection part, Speaker identification unit, sound source occupant match determination unit, and selection collection output unit), 5 ... Speaker (selection collection output unit), 7 ... Seat, 7f ... Front seat, 7r ... Rear seat, 7fa ... Right front seat, 7rb ... left rear seat, 10 ... monitor unit, 20 ... car window, 23 ... omnidirectional camera, 24 (24a-24d) ... directional microphone, 24X ... microphone array, 25 (25a-25d) ... directional speaker, 25X ... speaker Array, 30 ... monitor ECU (sound collecting unit, occupant detection unit, conversation determination unit, speaker identification unit, seating position identification unit, sound source identification unit, sound source occupant coincidence determination unit, and selective collection output unit), 31 ... vehicle interior Conversation support device, 40 ... Crew, 1 ... first occupant, 42 ... second passenger, 51 ... speaker, 52 ... listener, 55 ... noise filter, IP ... captured image, IS, IST ... sound information.

Claims (5)

A sound collection unit that collects sound information in the passenger compartment;
An occupant detection unit for detecting an occupant in the vehicle interior;
A conversation determination unit that determines whether or not the sound information indicates a conversation of an occupant detected in the vehicle interior;
A speaker identifying unit that identifies a speaker of the conversation when the sound information indicates the conversation of the occupant;
A selective collection output unit for selectively collecting and outputting the voice of the speaker;
A vehicle interior conversation support device. The in-vehicle conversation support device according to claim 1,
A seating position specifying unit for specifying the seating position of the occupant in the passenger compartment;
A sound source identifying unit that identifies a sound source of the sound information;
A sound source occupant match determination unit that determines whether the sound source of the sound information matches the seating position of any occupant detected in the vehicle interior;
The in-vehicle conversation support device, wherein the speaker specifying unit specifies the occupant as the speaker of the conversation when a sound source of the sound information matches a seating position of any of the occupants. In the vehicle interior conversation support device according to claim 1 or 2,
A plurality of directional microphones provided in the vehicle interior in directions different from each other,
The in-vehicle conversation support apparatus, wherein the selection collection output unit collects the voice of the speaker by selecting the directional microphone facing the speaker. In the vehicle interior conversation support device according to any one of claims 1 to 3,
A plurality of directional speakers provided in the vehicle interior in directions different from each other,
The selective collection output unit outputs the selectively collected voice of the speaker from the directional speaker facing a passenger other than the speaker detected in the vehicle interior;
A vehicle interior conversation support device. In the vehicle interior conversation support device according to any one of claims 1 to 4,
Comprising a noise filter for attenuating noise other than speech contained in the sound information;
A vehicle interior conversation support device.