JP2010262478A - Vehicle control system and safety confirmation determination device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a vehicle control system for preventing erroneous determination that a driver does not recognize an obstacle by stabilizing the information of the direction of the viewpoint of a driver in collating the direction of the viewpoint of the driver acquired from a driver monitor camera with the direction of the obstacle to be acquired from an exterior camera. <P>SOLUTION: The vehicle control system 10 is provided with: a driver monitor camera 34f for detecting the direction of the viewpoint of a driver in a prescribed cycle; a viewpoint cluster processing part 31b for detecting the viewpoint moving speed of the driver on angular coordinates; a viewpoint identification processing part 31c for detecting the gazing period of the viewpoint of the driver; and a safety confirmation determination part 31 for determining whether or not the direction of the obstacle detected by an exterior camera 32c is included in the range of the field of vision of the driver calculated on the basis of the viewpoint moving speed and the gazing period, and for executing vehicle control when it is determined that the direction of the obstacle is not included in the range of the field of vision of the driver. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a vehicle control system that collates a driver's viewpoint direction acquired by a driver monitor camera with an obstacle direction acquired by an outside camera and performs various vehicle controls based on the result of the comparison.

  Conventionally, the driver's viewpoint direction detected by an in-vehicle camera such as a driver monitor camera and the obstacle direction detected from the outside camera, navigation information, etc. are collated (overlaid) on the same coordinate system, and the driver is obstructed. A system is used to determine whether or not it is recognized.

  In this system, for example, a driver's viewpoint direction is determined by combining a driver's front image captured by a camera outside the vehicle such as an ultra-wide-angle CCD camera and a driver's face image captured by a driver monitor camera using a dedicated PC. It calculates where the foreground image intersects and overlays its position on the image of the camera outside the vehicle. When a dangerous obstacle is detected outside the driver's line of sight, vehicle control such as warning and braking is performed.

  By the way, there is provided a driving support device capable of providing a warning and assisting the driving of the driver only when a warning for confirmation of a pedestrian or the like is required at an intersection or the like (see, for example, Patent Document 1). In this driving support device, based on the feature information at the intersection in front of the traveling direction of the host vehicle and the detection signal of the driver's face direction and viewpoint direction acquired via the camera ECU, the driver is at the intersection. A warning is issued when it is determined that each feature is not recognized.

JP 2008-123443 A

  However, in the above conventional system, it is only determined by comparing the driver's viewpoint direction acquired from the driver monitor camera with the obstacle direction acquired by the vehicle outside camera, whether or not the driver recognizes the obstacle. There are the following problems.

  That is, (I) As shown in FIG. 2, the visual field range 201 that can be normally recognized by the driver during driving has a certain angular width from the center of the visual field. If the obstacle direction from the camera outside the vehicle does not substantially match, it is determined that the driver has not recognized the obstacle. For this reason, in the conventional system, there is a problem in that it is erroneously determined that the driver does not recognize the obstacle even when the driver actually recognizes the obstacle.

  Further, (II) as shown in the movement of the viewpoint (indicated by black dots) in FIG. 4A, the human eyeball actually performs a fine movement called tremor in the vertical direction of about 10 μm, for example, about 80 times per second. The movement of the line of sight consists of fine fine movement due to tremor and large movement due to changes in the visual target. In the conventional system described above, fine viewpoint movement due to tremor is detected faithfully as the viewpoint direction, so that the driver's line-of-sight information is not stable, and the obstacle recognition accuracy becomes unstable, resulting in erroneous determination. As a result, there is also a problem that repeated false alarms and non-sounds are repeated.

  The present invention has been made in view of the above problems, and stabilizes the information of the driver's viewpoint direction when collating the obstacle direction acquired by the outside camera with the viewpoint direction of the driver acquired by the driver monitor camera. A safety confirmation determination device for preventing the driver from erroneously determining that the obstacle is recognized but not recognizing and confirming the driver's safety based on a more accurate verification result, and the verification result An object of the present invention is to provide a vehicle control system that performs vehicle control based on the above.

  In order to solve the above problems, a vehicle control system according to the present invention collates a driver's viewpoint direction detected using a driver monitor camera with an obstacle direction detected using an outside camera, and A vehicle control system that performs vehicle control using the result of collation, a viewpoint detection unit that detects a driver's viewpoint direction in a predetermined cycle, and a viewpoint movement speed detection unit that detects a driver's viewpoint movement speed on angular coordinates And the gaze period detecting means for detecting the gaze period of the driver's viewpoint, and the driver's visual field range calculated based on the viewpoint movement speed and the gaze period includes the obstacle direction detected by the outside camera. A safety confirmation determination means for determining whether the obstacle direction is not included in the visual field range of the driver in the safety confirmation determination means and the safety confirmation determination means Wherein the control execution and an electronic control unit capable.

  With this configuration, when the obstacle direction acquired by the outside camera and the viewpoint direction of the driver acquired by the driver monitor camera are collated, the safety confirmation determining unit detects the viewpoint moving speed detected by the viewpoint moving speed detecting unit. Then, it is possible to determine whether or not the driver recognizes an obstacle by specifying the visual field range of the driver using information on the gaze period detected by the gaze period detection unit. For this reason, in the present invention, it is possible to prevent erroneous determination that the driver recognizes an obstacle but does not recognize it, and to perform vehicle control based on a more accurate collation result.

  Further, the vehicle control system according to the present invention further records the viewpoint as a gazing point when the viewpoint movement speed detected by the viewpoint movement speed detection means is less than a threshold value, while the viewpoint movement speed is a threshold value. In the above-described case, it includes viewpoint cluster generation means for generating a cluster having a range including a plurality of consecutive points of gaze recorded previously as one field of view range, and the electronic control means includes: The vehicle control can be executed when the visual field range of the driver corresponding to the generated cluster does not include the obstacle direction detected by the outside camera.

  With this configuration, the safety confirmation determination unit can determine that the driver recognizes the same visual field range when the viewpoint is within the same cluster generated by the viewpoint cluster unit. The viewpoint direction can be stabilized.

  Further, the vehicle control system according to the present invention further detects a gaze period of a viewpoint in which the viewpoint moving speed is equal to or greater than a threshold when the cluster is generated in the viewpoint cluster generation unit, and the gaze period is equal to or less than the threshold. In this case, the record of the generated cluster is deleted. On the other hand, if the gaze period is equal to or greater than the threshold value, a new cluster including the generated cluster and the viewpoint is generated to identify the driver's visual field range. Viewpoint identification processing means for performing processing, and the electronic control means has an obstacle direction detected by the outside camera within the visual field range of the driver corresponding to the new cluster generated by the viewpoint identification processing means. When not included, vehicle control can be executed.

  With this configuration, the viewpoint identification unit uses the gaze period threshold detected by the gaze period detection unit, so that the driver recognizes the same object even when there is a viewpoint movement whose viewpoint moving speed is equal to or greater than the threshold. A new cluster can be generated, and the viewpoint direction of the driver can be stabilized based on this cluster.

  In addition, the vehicle control system according to the present invention further determines whether or not a viewpoint is detected at a predetermined cycle in the viewpoint detection unit, and when a tracking error occurs in the viewpoint detection at the predetermined cycle, 1 is determined. Determine whether the previous viewpoint data is a gazing point. If it is a gazing point, use the gazing point data. If it is not a gazing point, there is no viewpoint data at the time of tracking error. Tracking error correction means is provided.

  With this configuration, the tracking error correction means performs correction processing even when a tracking error occurs because a previous gaze point data can be used even when a tracking error occurs in the driver's viewpoint information due to the vehicle environment. be able to.

  In addition, it implement | achieves as a vehicle control method which makes the processing means which comprises the vehicle control system which concerns on this invention a step, implement | achieves those steps as a program which makes a computer perform, and records the said program on DVD, CD-ROM, etc. It goes without saying that it can be distributed via a transmission medium such as a medium or a communication network.

  In the vehicle control system according to the present invention, when the obstacle direction acquired by the outside camera and the driver's viewpoint direction acquired by the driver monitor camera are collated, information on the driver's viewpoint direction acquired by the driver monitor camera is stabilized. Thus, it is possible to prevent the driver from erroneously determining that the obstacle is recognized but not recognized, and perform various vehicle controls based on more accurate matching results.

Overall view of vehicle control system according to embodiment The figure which shows an example of the visual field range of the driver during general driving Functional block diagram of a vehicle control system according to an embodiment Explanatory drawing of the viewpoint stabilization process in the safety confirmation determination part which concerns on embodiment The flowchart which shows the operation | movement procedure of the whole vehicle control system which concerns on embodiment. Flow chart showing the operation procedure at the time of tracking error correction shown in step S11 The flowchart which shows the operation | movement procedure at the time of the viewpoint cluster process shown to step S12 The flowchart which shows the operation | movement procedure at the time of the viewpoint identification process shown to step S13. The figure explaining another form of the cluster in viewpoint identification processing

  Hereinafter, embodiments of a vehicle control system and a safety confirmation determination device according to the present invention will be described with reference to the drawings.

(Embodiment)
FIG. 1 shows an overall view of a vehicle control system 10 according to the present embodiment.
The vehicle control system 10 acquires the viewpoint direction of the driver 100 using the driver monitor camera 101, and acquires the relative direction between the obstacle 103 in the predetermined area 104 and the vehicle from the front image using the outside camera 102.

  As shown in FIG. 1B, the vehicle control system 10 determines whether or not the driver is in a safety confirmation state based on whether or not the obstacle 103 exists in the visual field range 105 having a predetermined angular width from the viewpoint center 106 of the driver 100. When the obstacle 103 is not recognized within the driver's visual field range 105, various vehicle controls such as warning and automatic braking are performed.

  FIG. 2 is a diagram illustrating an example of a driver's visual field range 201 during general driving. The visual field range 201 in which the driver can accurately recognize a front obstacle during driving is, for example, approximately ± 5 to 20 from the center of the visual field. It becomes an elliptical shape having an angular width of degrees (however, it changes depending on the driver's consciousness state).

  FIG. 3 shows a functional block diagram of the vehicle control system 10 according to the present embodiment.

  In the vehicle control system 10, a safety confirmation determination unit 31 (corresponding to a safety confirmation determination device in claims), an environment information acquisition unit 32, an alarm notification unit 33, and a driving information acquisition unit 34 are connected to each other via a bus. Has been.

  The environmental information acquisition unit 32 is a processing unit that acquires various environmental information around the vehicle and transmits the environmental information to the safety confirmation determination unit 31, and includes a GPS information acquisition unit 32a that acquires GPS (Global Positioning System) information, and a road map. A road database 32b in which information is recorded, an outside camera 32c such as a CCD (charged coupled device) camera provided near the upper part of the windshield of the vehicle body, and a distance measurement function for measuring the distance to obstacles ahead and to the side And a radar 32d such as a millimeter wave radar having a speed measuring function for measuring a speed with respect to an obstacle.

  The alarm notification unit 33 is a display 33a that is a display unit such as a liquid crystal panel that displays a predetermined warning to the driver when receiving a signal transmitted from the safety confirmation determination unit 31 when vehicle control is required. And a speaker 33b for emitting a warning sound.

  The driving information acquisition unit 34 is a processing unit that acquires driving information of the driver and transmits the driving information to the safety confirmation determination unit 31, and includes a vehicle speed sensor 34a that acquires a vehicle speed signal, an accelerator sensor 34b that acquires an accelerator pedal signal, such as a magnetic rotor. Brake sensor 34c for detecting a change in magnetism due to rotation of the steering wheel, rudder angle sensor 34d for sensing the steering angle of the front wheels when the driver turns the steering wheel, turn signal sensor 34e for obtaining a left / right signal signal, and the driver's viewpoint A driver monitor camera 34f for detecting the direction is provided.

  Based on the information acquired from the environment information acquisition unit 32 and the driving information acquisition unit 34, the safety confirmation determination unit 31 develops the driver's viewpoint direction and the obstacle direction imaged by the outside camera 32c on the same coordinate system. To determine whether to perform vehicle control.

  In addition, the safety confirmation determination unit 31 includes a tracking error correction unit 31a, a viewpoint cluster processing unit 31b, and a viewpoint identification processing unit 31c, and performs stabilization processing for the viewpoint direction of the driver acquired from the driver monitor camera 34f.

  The tracking error correction unit 31a determines whether or not the driver's line-of-sight information can be accurately acquired at a predetermined interval (for example, 100 ms) by the driver monitor camera 34f. For example, when the sunlight is reflected on the driver's glasses when the weather is good, the driver's viewpoint direction cannot be detected by the driver monitor camera 34f depending on the reflection angle. Accordingly, in such a case, the tracking error correction unit 31a performs a process of detecting the occurrence of the tracking error and correcting the driver's line-of-sight data.

  Note that the near-infrared camera is used to detect the driver's viewpoint using the driver monitor camera 34f, for example, in order to easily extract the pupil of the driver. In the near-infrared image, the pupil shows a feature having the lowest luminance value compared to the surroundings. Therefore, more accurate pupil center coordinates are obtained from the image of the pupil region using the acquired driver image information, and the positions of the pupil center coordinates in the image and the center coordinates of the image of the point light source reflected on the surface of the eyeball It can be detected by calculating the viewpoint direction in the global coordinate system from the relationship.

  The viewpoint cluster processing unit 31b detects the movement speed between the detected viewpoints detected in a predetermined cycle of the driver, detects whether the driver recognizes the same object based on the viewpoint movement speed, A process of taking in a visual field area determined to recognize the same object as a “cluster” is performed.

  The viewpoint identification processing unit 31c detects a gaze period, which is a gaze period in which the driver's viewpoint has not moved to another viewpoint position. For example, even when the viewpoint moving speed moves quickly at a predetermined value or more, If the gaze period of the viewpoint after movement is greater than or equal to a predetermined value, the driver's viewpoint identification process is performed, such as recognizing the same object.

  The safety confirmation determination unit 31 calculates an obstacle direction in the traveling direction based on image information acquired from the outside camera 32c. The viewpoint of the driver whose viewpoint direction is stabilized by using the tracking error correction unit 31a, the viewpoint cluster processing unit 31b, and the viewpoint identification processing unit 31c from the image information acquired from the driver monitor camera 34f of the driving information acquisition unit 34. Get direction. Then, the safety confirmation determination unit 31 determines whether there is an obstacle within the visual field range of the driver based on the obstacle direction and the viewpoint direction.

  When the safety confirmation determination unit 31 determines that the driver has not confirmed the safety against the obstacle, for example, when the obstacle is outside the visual field range of the driver and is closer than a predetermined distance, the safety confirmation determination unit 31 warns using the display 33a. A control signal for issuing a warning sound to the driver using the display or the speaker 33 b is transmitted to the alarm notification unit 33.

  FIG. 4 is an explanatory diagram of viewpoint stabilization processing in the safety confirmation determination unit 31 according to the present embodiment. Note that the black dots shown in this figure are the coordinates of the viewpoint detected at a predetermined interval.

  FIG. 4A shows the cluster 401, the cluster 402, the cluster 403, and the cluster 404 generated by the viewpoint cluster processing unit 31b. In this figure, clusters 401 to 404, which are one field of view range, are generated by movement by “saccades 405 to 407”, which are movements with a large viewpoint movement speed. In the present invention, it is determined that the driver recognizes the same visual field range in the same cluster, and the viewpoint direction of the driver can be stabilized.

  FIG. 4B shows the relationship between the movement of the driver's viewpoint and the cluster 410. When the viewpoint is in the same cluster 410, the safety confirmation determining unit 31 recognizes that the driver recognizes the same visual field range. It becomes possible to judge.

  FIG. 4C is an explanatory diagram of “saccade”, which is a quick movement of the viewpoint defined in the present embodiment. When the start point 420 is set at the center of the cluster 421, the viewpoints 2a, 2b, and 2c in the cluster 421 are labeled as “gaze points”, while the viewpoints 2d and 2e outside the cluster 421 correspond to the quick movement of the viewpoint. Labeled as "saccade".

  FIG. 5A is a flowchart showing an overall operation procedure of the vehicle control system according to the present embodiment.

  First, the safety confirmation determination unit 31 performs the stabilization process of the driver's viewpoint direction acquired from the driver monitor camera 34f (step S501), and the driver's viewpoint direction and the camera outside the vehicle after the stabilization process is performed. A process of acquiring the obstacle direction ahead and overlaying on the same coordinate system is performed (step S502).

  Next, the safety confirmation determination unit 31 determines whether or not there is an obstacle in the visual field range of the driver (step S503). If there is a danger that the obstacle cannot be recognized (YES in step S503), An alarm is given using the speaker 33b (step S504).

  FIG. 5B is a flowchart showing an operation procedure in the stabilization process of the viewpoint direction of the driver shown in step S501.

First, the safety confirmation determination unit 31 reads the current viewpoint data of the driver accumulated in the real-time buffer (step S10).
Next, the tracking error correction unit 31a performs a correction process by detecting a tracking error from the viewpoint of the driver acquired at a predetermined interval (step S11). This detailed operation procedure will be described with reference to FIG.

Then, the viewpoint cluster processing unit 31b performs a process of generating a cluster using a plurality of viewpoint information (step S12). This detailed operation procedure will be described with reference to FIG.
Finally, the viewpoint identification processing unit 31c determines whether or not the driver recognizes the same object, and performs a driver's viewpoint identification process (step S13). This detailed operation procedure will be described with reference to FIG.

  FIG. 6 is a flowchart showing an operation procedure at the time of tracking error correction shown in step S11.

  First, the tracking error correction unit 31a determines whether or not “missing data” from which viewpoint data cannot be acquired is detected from the viewpoint data of the driver of the real-time buffer read at a predetermined interval (step S601).

  Here, the mixing data is a state in which, for example, data in the viewpoint direction of the driver that should be detected every 100 ms cannot be detected. For example, when the driver's line-of-sight data can be acquired, truck_status = 0, and in the period when the driver's line-of-sight cannot be acquired, it is defined as truck_status = 1. Therefore, the tracking error correction unit 31a generates a tracking error due to the numerical value of the truck_status It can be detected whether or not.

  Next, in the case of a regular process in which no mixing data is detected (NO in step S601), the tracking error correction unit 31a proceeds to the viewpoint cluster process shown in step S12.

  On the other hand, when the mixing data is detected (YES in step S601), the tracking error correction unit 31a determines whether or not the viewpoint acquired immediately before is a gazing point (step S602).

  If the previous viewpoint is a gazing point (YES in step S602), the tracking error correction unit 31a uses the previous gazing point as the current viewpoint (step S604). On the other hand, if the previous point is not the gazing point (NO in step S602), the tracking error correction unit 31a ignores the tracking error (step S603), returns to step S10, and again the driver in the real-time buffer. Reads the viewpoint data.

  As described above, when the driver's viewpoint is not detected due to the vehicle environment, the tracking error correction unit 31a can perform correction processing even if a tracking error occurs using the data of the previous gazing point.

  FIG. 7 is a flowchart showing an operation procedure during the viewpoint cluster processing shown in step S12.

  First, the viewpoint cluster processing unit 31b detects a viewpoint movement speed that is a movement speed between different viewpoints (step S701), and determines whether or not the detected viewpoint movement speed is less than a threshold value (step S702). Here, for the detection of the viewpoint movement speed of the viewpoint cluster processing unit 31b, for example, when the sampling frequency is 60 Hz, the threshold of the viewpoint movement speed is 2 deg (= 120 deg / sec), 2.5 deg ( = 150 deg / sec), 3 deg (= 180 deg / sec), etc. Further, the sampling frequency of 60 Hz and the viewpoint moving speed threshold are examples, and the content of the present invention is not limited to this.

  Next, when the detected viewpoint moving speed is equal to or higher than the threshold (NO in step S702), the viewpoint cluster processing unit 31b labels the viewpoint as “saccade” meaning quick movement (step S703). The process proceeds to the viewpoint identification process shown in step S13.

  On the other hand, when the speed is less than the threshold value (YES in step S702), the viewpoint cluster processing unit 31b records the data in the real-time buffer by labeling as “gazing point” because the eyes are not moving quickly. (Step S704), the process returns to step S10 again, and the process of reading the viewpoint data of the driver in the real-time buffer is performed.

  FIG. 8 is a flowchart showing an operation procedure during the viewpoint identification process shown in step S13.

  First, the viewpoint identification processing unit 31c performs a process of taking a plurality of continuous gazing points before the viewpoint labeled as saccade as the same cluster (step S801).

  Next, the viewpoint identification processing unit 31c detects the gaze period of the viewpoint labeled as saccade, and determines whether or not the gaze period is less than the threshold (step S802). Here, the threshold is 85 to 100 ms when the sampling frequency is 60 Hz, for example. That is, when the gaze period is 100 ms, it is considered that the driver recognizes the same object when the viewpoint is gazing for 0.1 second or longer.

  Then, when the gaze period is less than the threshold value (YES in step S802), the viewpoint identification processing unit 31c deletes the previous cluster information on the assumption that another object has been recognized (step S803).

  On the other hand, if the gaze period is equal to or greater than the threshold value (NO in step S802), it is captured as the same cluster, and the center point of the cluster and the gaze period are calculated (step S804). That is, according to the present invention, by using the threshold of the gaze period, even when there is a viewpoint movement labeled as saccade, the viewpoint can be taken into the same cluster assuming that the driver recognizes the same object.

  When the viewpoint identification process is completed in the viewpoint identification processing unit 31c, the process returns to step S10 again, and the process of reading the viewpoint data of the driver from the real time buffer is performed.

  FIG. 9 is a diagram for explaining another form of cluster in the viewpoint identification processing. In this figure, the viewpoint identification processing unit 31c creates a new cluster 900 with the center point 902 because the viewpoint 901 labeled saccade has a long gaze period. Then, the viewpoint identification processing unit 31c can define that the movement of the viewpoint within the cluster 900 recognizes the same object.

  As described above, in the vehicle control system 10 according to the present invention, the viewpoint cluster processing unit 31b determines a range of continuous gazing points as a cluster based on the viewpoint movement speed, and the safety confirmation determination unit 31 In the collation process using the obstacle direction acquired by the camera 32c and the viewpoint direction of the driver acquired by the driver monitor camera 34f, if the obstacle direction matches in the cluster, it is determined that the driver recognizes the obstacle. It becomes possible to do.

  In addition, the safety confirmation determination unit 31 can use a cluster to handle fine movements as the same viewpoint and have the same viewpoint direction, stabilize the line-of-sight information, and also stabilize the collation determination. Therefore, in the vehicle control system 10 according to the present invention, the accuracy of the collation recognition / determination using the obstacle direction acquired by the outside camera 32c and the viewpoint direction of the driver acquired by the driver monitor camera 34f is improved and more accurately. Vehicle control can be performed.

  The vehicle control system and the safety confirmation device according to the present invention can be used for a moving body such as a vehicle.

DESCRIPTION OF SYMBOLS 10 Vehicle control system 31 Safety confirmation determination part 31a Tracking error correction part 31b View point cluster process part 31c View point identification process part 32 Environmental information acquisition part 32a GPS information acquisition part 32b Road database 32c, 102 Outside camera 32d Radar 33 Alarm notification part 33a Display 33b Speaker 34 Driving information acquisition unit 34a Vehicle speed sensor 34b Acceleration sensor 34c Brake sensor 34d Rudder angle sensor 34e Turn signal sensor 34f, 101 Driver monitor camera 100 Driver 103 Obstacle

Claims (6)

A vehicle control system that collates a viewpoint direction of a driver detected using a driver monitor camera and an obstacle direction detected using a camera outside the vehicle, and performs vehicle control using a result of the comparison,
Viewpoint detection means for detecting the driver's viewpoint direction in a predetermined cycle;
Viewpoint movement speed detection means for detecting the viewpoint movement speed of the driver on the angle coordinates;
Gaze period detecting means for detecting the gaze period of the driver's viewpoint;
Safety confirmation determination means for determining whether or not an obstacle direction detected by the outside camera is included in the visual field range of the driver calculated based on the viewpoint moving speed and the gaze period;
A vehicle control system comprising: the safety confirmation determination means; and an electronic control means capable of executing vehicle control when it is determined that the obstacle direction is not included in the visual field range of the driver. The vehicle control system further includes:
When the viewpoint movement speed detected by the viewpoint movement speed detection means is less than a threshold value, the viewpoint is recorded as a gazing point, whereas when the viewpoint movement speed is equal to or greater than the threshold value, the previously recorded continuous A viewpoint cluster generation means for generating a cluster having a range including a plurality of gazing points as one visual field range;
The electronic control means can execute vehicle control when the driver's visual field range corresponding to the cluster generated by the viewpoint cluster generation means does not include an obstacle direction detected by the outside camera. The vehicle control system according to claim 1, wherein: The vehicle control system further includes:
When a cluster is generated by the viewpoint cluster generation means, a gaze period of a viewpoint whose viewpoint moving speed is equal to or greater than a threshold is detected, and when the gaze period is equal to or less than a threshold, the generated cluster record is deleted On the other hand, when the gaze period is equal to or greater than a threshold value, it includes viewpoint identification processing means for generating a new cluster including the generated cluster and the viewpoint and performing identification processing of the visual field range of the driver,
The electronic control unit can execute vehicle control when an obstacle direction detected by the outside camera is not included in the visual field range of the driver corresponding to the new cluster generated by the viewpoint identification processing unit. The vehicle control system according to claim 2, wherein: The vehicle control system further includes:
It is determined whether or not a viewpoint is detected at a predetermined cycle in the viewpoint detection unit, and if a tracking error has occurred in the viewpoint detection at a predetermined cycle, whether or not the viewpoint data acquired immediately before is a gazing point And a tracking error correction unit that uses the data of the gazing point in the case of a gazing point, and that there is no viewpoint data when a tracking error occurs when the gazing point is not. The vehicle control system according to any one of 1 to 3. A safety confirmation determination device that determines a driver's safety by collating a driver's viewpoint direction detected using a driver monitor camera and an obstacle direction detected using an outside camera on the same coordinate system,
Viewpoint detection means for detecting the driver's viewpoint direction in a predetermined cycle;
Viewpoint movement speed detection means for detecting the viewpoint movement speed of the driver on the angle coordinates;
Gaze period detecting means for detecting the gaze period of the driver's viewpoint;
Safety confirmation determination means for determining whether or not an obstacle direction detected by the outside camera is included in the visual field range of the driver calculated based on the viewpoint movement speed and the gaze period. Safety confirmation judgment device. A vehicle control method for collating a driver's viewpoint direction detected using a driver monitor camera and an obstacle direction detected using an outside camera, and performing vehicle control using the result of the verification,
A viewpoint detection step of detecting the viewpoint direction of the driver at a predetermined cycle;
A viewpoint movement speed detection step for detecting the viewpoint movement speed of the driver on the angular coordinates;
A gaze period detecting step for detecting a gaze period of the driver's viewpoint;
A safety confirmation determination step for determining whether or not an obstacle direction detected by the outside-vehicle camera is included in the visual field range of the driver calculated based on the viewpoint movement speed and the gaze period;
An electronic control step capable of executing vehicle control when it is determined in the safety confirmation determination step that the obstacle direction is not included in the visual field range of the driver.

Images