Device for detecting a driver's inability to drive

Abstract

Device (100) for detecting a driver's inability to drive, comprising: a head detection section (61) which successively detects a head section higher than the driver's neck, based on an image of a driver's seat taken by an imaging device (21) mounted in a vehicle (10); and an out-of-frame state detection section (71) which detects during the journey of the vehicle (10) that the driver is no longer able to drive when the head section detected by the head detection section (61) is outside a predetermined area (FA) in the image, characterized by the fact that the device (100) also has a deformation section that deforms the predetermined area (FA) according to an intention of a user, which includes the driver, and The deformation section deforms the predetermined area (FA) according to a user input into an HMI (80) of the device (100).

Description

TECHNICAL AREA

[0001] The present invention relates to a device for detecting a driver's inability to drive, which detects that the driver is no longer able to drive. STATE OF THE ART

[0002] While driving, a driver may become incapacitated due to a sudden illness or similar event, leading to a traffic accident. For this reason, it has been proposed to record such driver incapacity in order to prevent accidents.

[0003] An emergency evacuation device described in JP 2014-19301A detects information about a driver's line of sight, turn signals, and facial orientation from a captured image of the driver, as well as measurement information about the driver's heart rate, respiration, and brainwaves. The device described in patent literature 1 detects the driver's condition, including the degree of impaired consciousness, level of alertness, and posture, using various types of acquired information and assists in emergency evacuation if the driver's condition deteriorates.

[0004] The inventors of the present invention have found that, since the device described in JP 2014-19301A detects the driver's condition using various types of information, the processing for detecting the driver's state of incapacity to drive is complicated.

[0005] JP 2005-108 033 A discloses a device for detecting a driver's incapacity to drive, comprising: a head detection section that successively detects a head section higher than the driver's neck based on an image of a driver's seat taken by an imaging device mounted in a vehicle; and an out-of-frame state detection section that detects, while the vehicle is in motion, that the driver is no longer able to drive when the head section detected by the head detection section is outside a predetermined area in the image.

[0006] US Patent 2008 / 0080741A1 describes an occupant detection system comprising a camera, an information extraction device, a recognition processing device, an integration processing device, and a determination processing device. The camera captures 3D images of the driver of a vehicle. The information extraction device extracts information about at least one of the positions or movements of the driver's head or hand. The recognition processing device detects whether the position or movement of the driver's head or hand is in a predefined state. The integration processing device accumulates the time during which the position or movement of the driver's head or hand is not in the predefined state.The determination processing unit detects that the driver's attention is reduced when the accumulated time exceeds a preset reference value.

[0007] JP 2005-173 635 A describes a device for detecting drowsiness of a driver in a vehicle, which is equipped with a monitoring device for directly monitoring the positions of the driver's feet and a decision device which decides, on the basis of the monitoring, whether the feet are arranged in a hazard-avoiding position in relation to the pedals for vehicle operation.

[0008] JP 2011-164 825 A describes a driver state determination device comprising: a pressure sensing device for detecting the pressure distribution exerted on at least one driver's seat backrest and one driver's seat section, and a determination device for determining a driver's state of alertness based on a change in the pressure distribution detected by the pressure sensing device. Furthermore, the determination device determines the driver's state of alertness based on whether a change in the detected pressure distribution exhibits periodicity.

[0009] JP 2010-128649A describes a system for determining a driver's level of alertness, comprising: a behavior detection unit comprising a first indicator detection unit that detects the driver's eye behavior and a second indicator detection unit that detects the driver's head behavior based on an image of the driver; a criterion setting unit that sets a criterion for determining a driver's level of alertness based on the driver's head behavior while driving in a monotonous section, whereby vehicle operation is determined to be monotonous according to detected road information, vehicle information, and / or distance information; and a determination unit that determines the driver's level of alertness based on the detected head behavior of the driver and the set criterion.

[0010] DE 11 2011 105 432 T5 describes a vehicle emergency evacuation device that automatically stops a moving vehicle or maintains a stopped vehicle in the event of a driver emergency, whereby a stop maintenance process is executed to maintain the stopped state without performing an intent confirmation process for initiating an emergency evacuation process when the emergency evacuation process is required, e.g., if an abnormality or deterioration of the driver's condition occurs, or if the driver requests the emergency evacuation process under the condition that the vehicle is stopped before a traffic light, a road junction, or a level crossing. SUMMARY OF THE INVENTION

[0011] It is an object of the present invention to provide a device for detecting a driver's inability to drive, which can easily detect that the driver is no longer able to drive. This object is achieved by a device with the features of the independent claims. The dependent claims are directed to advantageous embodiments of the invention.

[0012] According to one aspect of the invention, the device for detecting a driver's incapacity to drive successively captures a section of the driver's head based on a captured image of the driver's seat. When the driver is normally operating the vehicle, the driver's head section is typically located within a predetermined area of ​​the image of the driver's seat. In contrast, if the driver loses consciousness due to a sudden illness, the driver's head section may move outside this predetermined area. Therefore, the driver's incapacity to drive can be easily detected by recognizing that the driver is no longer able to drive when the head section is located outside this predetermined area.

[0013] According to another aspect of the invention, the device for detecting a driver's incapacity to drive successively captures the driver's head section based on an image of the driver's seat and obtains a trajectory of the head section from the positions of the captured head section. If the driver is no longer able to drive due to a sudden illness, the driver's head section often shifts relative to a position it held while driving and does not return to that position. Therefore, the driver's incapacity to drive can be easily detected based on the trajectory of the driver's head section. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The above and further aspects, features and advantages of the present invention will become clear with reference to the following detailed description and the accompanying drawings. These show: Fig. 1 a block diagram representing a configuration of a device for detecting a state of incapacity to drive; Fig. 2 a view showing the interior of a vehicle equipped with the device for detecting an inability to drive; Fig. 3 a view showing a driver's seat of the vehicle equipped with the device for detecting a state of incapacity to drive; Fig. 4 a block diagram representing the functions of a control system; Fig. 5A a view representing a posture during a normal journey; Fig. 5B a view that represents an attitude when the driver has suddenly become ill and his attitude collapses (within FA); Fig. 5C a view that represents a stance in the state in which the driver has suddenly become ill and the attitude collapses (outside of FA); Fig. 6 a view that represents a posture in the state in which the driver picks up an object; Fig. 7A a view representing a posture during a normal journey; Fig. 7B a view which represents an attitude in the state in which the driver has suddenly become ill; Fig. 8 a view that represents a posture in the state in which the driver is looking to the side; Fig. 9 a view that represents a stance in the state in which the driver picks up an object; Fig. 10A a view representing a direction of travel during a normal journey; Fig. 10B a view which represents a direction of view in the state in which the driver has suddenly become ill; Fig. 11 a view which represents a change in the direction of the face in the state in which the driver is looking to the side; Fig. 12 a diagram showing the oscillation or shaking of a head section caused by an external force; Fig. 13 a diagram representing an amplitude range of head section shaking that is determined to be a state of incapacity to drive; Fig. 14A a view depicting a facial expression during a normal journey; Fig. 14B a view that represents a facial expression in the state in which the driver has suddenly become ill; Fig. 15A a view that represents a normal state; Fig. 15B a view depicting a condition in which the driver shows the white of one eye; Fig. 15C a view that represents a condition in which the driver fully shows the whites of the eyes; Fig. 16A a flowchart that represents a processing procedure for recording a state of incapacity to drive; Fig. 16B a flowchart showing the continuation of Fig. 16A represents and shows a processing procedure for recording a state of incapacity to drive; Fig. 17 a diagram representing a subroutine that represents a processing procedure for detecting a posture drop; Fig. 18 a view that represents a mode for notifying the driver regarding a posture drop level; Fig. 19 a view which represents a framework area of ​​an out-of-frame provision according to a first modification; Fig. 20 a view which represents a framework area of ​​an out-of-frame provision according to a second modification; Fig. 21 a view which represents a framework area of ​​an out-of-frame provision according to a third modification; Fig. 22 a view that represents a framework area of ​​an out-of-frame provision according to a fourth modification; and Fig. 23 a view showing the framework areas and determination times for the out-of-frame determination according to a second embodiment. DESCRIPTION OF THE EXECUTION FORMS

[0015] The following describes embodiments of the device for detecting a driver's incapacity to drive with reference to the drawings. In these embodiments, the driver's incapacity to drive includes the state in which the driver has lost consciousness due to a sudden illness and is therefore no longer able to drive, and the state in which the driver has suffered a sudden illness, such as a heart attack, is unable to move their body despite not losing consciousness, and is therefore no longer able to drive. In the following embodiments, identical or equivalent elements are designated by the same reference numerals in all drawings, and their descriptions are not repeated. (First embodiment)

[0016] A detection device 100 (device for detecting a driver's incapacity to drive or detection of a driver's incapacity to drive) according to the present embodiment is below referred to in relation to the Fig. Described in sections 1 to 3. The detection device 100 comprises a controller 50, a driver state detection device 20, a vehicle information detection device 30, a driving environment detection device 40, an HMI (human-machine interface) 80, and a storage device 52, and detects the driver's incapacity to drive state. If the driver is asked whether they are no longer able to drive and no answer is received, the detection device 100 issues a command to a vehicle controller 90 to safely stop the vehicle.

[0017] The driver condition detection device 20 includes several driver cameras 21, a seat belt sensor 22, and a seat surface sensor 23. The driver cameras 21 correspond to an imaging device, the seat belt sensor 22 corresponds to a size detection section, and the seat surface sensor 23 corresponds to a seat pressure detection section. The driver cameras 21 are, for example, each a CCD camera and serve to capture an image of the driver's seat, which is illuminated by a lighting device such as an LED in the near-infrared range. As shown in Fig. 2 and Fig. As shown in Figure 3, the driver cameras 21 are mounted on a measuring panel 14 essentially in the center of the lower end of a rearview mirror 16 and on the left and right A-pillars 17, facing the driver. The driver cameras 21 can be arranged on an instrument panel 13 (shown by a dashed line) and a steering column instead of the measuring panel 14. The driver cameras 21 can be arranged at a left or right end of the rearview mirror 16 (shown by a dashed line) instead of at the lower end of the rearview mirror 16. The four driver cameras 21 each form a driver status monitor and capture several dozen images per second of the upper half of the driver, who is seated in a driver's seat 11, from the front.

[0018] The seat belt sensor 22 is a sensor for detecting the tension of a seat belt 12. In particular, the seat belt sensor 22 is a coding device that detects a motor rotation angle during the unfolding and retraction of the seat belt 12. The seat surface sensor 23 is a sensor that detects a pressure distribution of a seat section 11a of the driver's seat 11.

[0019] The vehicle information recognition device 30 includes a vehicle speed sensor 31, a steering sensor 32, an accelerator sensor 33, and a brake sensor 34. The vehicle speed sensor 31 is a sensor that detects the speed of a vehicle 10. The steering sensor 32 is a sensor that detects the steering angle of a steering wheel 15. The accelerator sensor 33 is a sensor that detects the accelerator input, i.e., the amount of pressure applied to the accelerator pedal. The brake sensor 34 is a sensor that detects the amount of pressure applied to the brake pedal.

[0020] The driving environment detection device 40 includes front and rear cameras 41, front and rear sensors 42, a vehicle navigation device 43, and a G-sensor 44. The front and rear cameras 41 each consist of a camera that captures an image of the front of the vehicle 10, which contains a white line on a road, or a camera that captures an image of the rear and the side diagonally behind the vehicle 10. The front and rear sensors 42 each consist of a sensor such as an ultrasonic sensor, a laser radar sensor, or a millimeter radar sensor and are used to detect an object located in front of or behind the vehicle 10 in order to determine the distance between the vehicle 10 and the object located in front of or behind the vehicle 10.The relative speed in relation to the front vehicle or the rear vehicle can be calculated on the basis of the distance between vehicle 10 and the front vehicle or the rear vehicle obtained from the front / rear sensor 42.

[0021] The vehicle navigation device 43 calculates the current position of the vehicle 10 using a GPS signal received from a GPS receiver and information obtained from various sensors, including the G-sensor, and calculates a path from the current position to a destination. The G-sensor 44 is a sensor installed, for example, on a seat 11, and detects the three-dimensional acceleration forward / backward, left / right, and up / down relative to the vehicle 10. The G-sensor 44 can be a sensor located in the vehicle navigation device 43 or can be a sensor located in an advanced vehicle operating system (AVOS) if the vehicle 10 has the AVOS. That is, the G-sensor 44 can be used if it is already installed for an application.

[0022] The controller 50 is a microcomputer containing a CPU, ROM, RAM, and I / O. The controller 50 receives various types of information from the driver state detection device 20, the vehicle information detection device 30, the driving environment detection device 40, the storage device 52, and the HMI 80. The controller 50 is connected to these devices via wired communication, such as CAN, or wireless communication, such as LAN and Bluetooth (registered trademark). The CPU of the controller 50 executes various programs stored in the ROM to perform the functions of an image analysis section 60, a learning section 51, and a state detection section 70 to detect the driver's state of incapacity. Each section will be described in more detail later.

[0023] The HMI 80 (corresponding to a posture notification section and a question section) includes a display 81, a speaker 82, and a cancellation switch 83. The display 81 corresponds to a display of the vehicle navigation device 43 or a display integrated into the measuring panel 14. The display 81 can be a touch-sensitive display equipped with a liquid crystal surface or an organic EL surface (EL: electroluminescent). The display 81 provides notification of a posture instability level based on the driver's posture as captured by the image. More precisely, the display 81 indicates the driver's posture in five levels. At posture instability level 5, which is the highest level, the driver is no longer able to maintain a driving posture due to a sudden illness and is therefore no longer able to drive.The driver can observe the posture status displayed on indicator 81 to check their own driving posture. If the posture level approaches level 5, the driver can correct their driving posture before it is determined that they are no longer fit to drive.

[0024] The loudspeaker 82 is a vehicle-integrated loudspeaker that is used in conjunction with the vehicle navigation device 43, audio equipment, and similar devices. If a driver impairment is detected, the loudspeaker 82 asks the driver, via an audible output, whether they are no longer able to drive. The display 81 can show a screen asking the driver whether they are no longer able to drive. The loudspeaker 82 can also provide an audible notification regarding the driver's level of impairment.

[0025] The override switch 83 is a switch that stops the detection of the incapacitation status. A single actuation of the override switch 83 stops the detection of the incapacitation status during a journey. If the override switch 83 is actuated while driving, the detection of the incapacitation status is stopped either for the duration of the override switch 83's operation or for a specific period (a few seconds) from the start of the actuation. Thus, if the driver picks up an object, the override switch 83 can be actuated in advance. In this way, even if the driver's composure suddenly collapses, a false detection indicating that the driver is no longer able to drive can be prevented.

[0026] Various functions of the 50 controller are described below with reference to Fig. 4 described. The image analysis section 60 contains a head detection section 61, a trajectory acquisition section 62, a cramp detection section 63, a tilt detection section 64, a face direction detection section 65 and a white eye detection section 66.

[0027] The head detection section 61 successively detects the head section that is higher than the driver's neck, based on the image of the driver's seat captured by each driver camera 21. More precisely, each time a respective driver camera 21 captures an image of the driver's seat, the head detection section 61 extracts an edge from the image of the driver's seat, indicating the outline of the driver's head section, and detects an area surrounded by the detected edge as the head section.

[0028] The trajectory acquisition section 62 obtains a trajectory of the driver's head section from positions of the driver's head section that are successively acquired by the head acquisition section 61. For example, the trajectory acquisition section 62 uses the center of the driver's head section, which is acquired in each image, as the position of the head section and connects the positions of the head section in the images to acquire the trajectory of the head section.

[0029] The cramp detection section 63 detects a cramp in the driver, that is, an involuntary contraction of muscles in the torso that is lower than the driver's head and neck. More precisely, cramp detection section 63 extracts edges from the outlines of the driver's head and torso in each image. If the edges extracted in successive images vibrate regularly (periodically), cramp detection section 63 registers that the driver is experiencing a cramp.

[0030] The tilt detection section 64 detects a tilt θ of the head section relative to the driver's torso based on the image of the driver's seat. More precisely, the tilt detection section 64 detects areas enclosed by the edges representing the outlines of the head section and torso as the head section and torso, as well as the center axes of the head section and torso. The tilt detection section 64 sets the tilt of the center axis of the head section relative to the center axis of the torso as the tilt θ of the head section. The direction of the torso is determined by comparing a predetermined torso orientation pattern with a detected torso direction, and the center axis of the torso is detected from the torso thus oriented.Features such as the eyes, nose, and mouth of a face contained within the head section are extracted, and the central axis of the head section is determined based on the three-dimensional arrangement of these features. When the head section is tilted forward, the distance between the facial features and the front of the vehicle decreases. Conversely, when the head section is tilted backward, the distance between the facial features and the front of the vehicle increases. The longitudinal spacing of the facial features can be used to determine the central axis of the head section.

[0031] Alternatively, the tilt detection section 64 detects the seat belt 12 of the driver's seat in the image of the driver's seat and detects the tilt θ of the head section relative to the torso from a relative position between the seat belt 12 and the head section. Since the driver's torso is constrained by the seat belt 12, the position of the seat belt 12 can be estimated from the position of the torso.

[0032] The face direction detection section 65 detects the driver's face direction in relation to the front of the vehicle 10 based on the image of the driver's seat. The face direction detection section 65 detects the inclination of the face in relation to the vertical surface opposite the front surface of the vehicle 10 as the face direction.

[0033] The white-eye detection section 66 contains a facial expression detection section 67 and a white-eye degree calculation section 68 and detects the state in which the driver displays the whites of their eyes (the driver rolls their eyes). Here, the state in which the driver displays the whites of their eyes includes a state in which the driver fully displays the whites of their eyes, as described in Fig. 15C is shown, and a condition in which a black eye area is smaller than a predetermined dimension, as shown in Fig. Figure 15B illustrates this. That is, the state in which the driver shows the whites of his eyes refers to a state in which the black of the eye is out of balance and smaller than a predetermined area.

[0034] Facial Expression Detection Section 67 captures the driver's eye outline and black eye area from the driver's seat image. Here, the driver's eye outline is a boundary line between the eyelid and the eye. The black eye area is an inner region of the eye outline, an area that has a lower brightness than the white of the eye, such as an area of ​​black as well as blue, brown, gray, or similar. Facial Expression Detection Section 67 captures the opening of the driver's mouth from the captured edge that defines the mouth outline.

[0035] The white eye degree calculation section 68 calculates the white eye degree of the driver's eye based on the eye outline and black eye area captured by the facial expression detection section 67.

[0036] In particular, section 68 on the white eye degree calculation calculates the white eye degree from a ratio of a longitudinal length Lb of the black eye area to a longitudinal length Lw + Lb of an area surrounded by the eye outline ( Fig. 15A to 15C). If the length Lb is smaller relative to the length Lw + Lb, the white eye degree is greater. Alternatively, white eye degree calculation section 68 calculates the white eye degree from the distance Lb between the top of the eye outline and the bottom of the black eye area. If the distance Lb is smaller, the white eye degree is greater. Alternatively, white eye degree calculation section 68 calculates the white eye degree based on the ratio of an area of ​​the black eye area to an area of ​​the white eye area, obtained by subtracting the area of ​​the black eye area from an area of ​​the total eye enclosed by the eye outline. If the area of ​​the black eye area is smaller relative to the area of ​​the white eye area, the white eye degree is greater.

[0037] Alternatively, section 68 of the white eye degree calculation calculates the white eye degree based on the flattening of the black eye area. In the white eye condition, the black eye area is oriented upwards, the flattening of the black eye area is obviously large, and the greater the flattening of the black eye area, the greater the white eye degree. Alternatively, section 68 of the white eye degree calculation calculates the white eye degree from a distance Lc between a midline of the longitudinal center of the area enclosed by the eye outline and the lower edge of the black eye area.

[0038] When the driver is able to drive, learning section 51 learns the tilt θ of the head section, which is detected by tilt detection section 64. When the driver is able to drive, learning section 51 learns the face direction, which is detected by face direction detection section 65. When the driver is able to drive, learning section 51 also learns the amplitude of head section shaking, which is detected by head detection section 61. That is, learning section 51 learns a driving posture habit of the driver. If several drivers use vehicle 10, learning section 51 learns a driving posture habit of each driver.

[0039] The condition monitoring section 70 contains an out-of-frame condition monitoring section 71, a posture intrusion condition monitoring section 72, a direction intrusion condition monitoring section 73, a shaking condition monitoring section 74, and a white eye condition calculation section 75.

[0040] While the vehicle 10 is in motion, the out-of-frame detection section 71 determines the out-of-frame state. In the event that the head section is out of frame, the out-of-frame detection section 71 detects that the driver is no longer able to drive. More precisely, the out-of-frame detection section 71 detects that the driver is no longer able to drive when the driver's head section, as detected by the head detection section 61, is outside of a defined area FA of the image. Here, area FA is a predetermined area in the image captured by the driver cameras 21. During normal driving, the driver's head section is never outside of area FA. Area FA can encompass the entire captured image.

[0041] While the driver drives the vehicle 10 normally, as it is in Fig. As shown in Figure 6, even when the driver picks up an object, the driver's head section is usually within the FA area of ​​the image. In contrast, if the driver suddenly becomes ill and loses consciousness, as shown in the Fig. As shown in sections 5A to 5C, the driver's head section moves outside the FA area. Therefore, if the driver's head section moves outside the FA area of ​​the image, the out-of-frame detection section 71 registers that the driver is no longer able to drive.

[0042] If, at this point, the out-of-frame state detection section 71 takes into account the trajectory along which the head section moved before it entered the FA area, as obtained by the trajectory acquisition section 62, it is possible to improve the accuracy of detecting the driver's incapacity. The case in which the driver's head section cannot be detected due to image blurring in the FA area can be distinguished, based on the head section's trajectory, from the case in which the driver's head section cannot be detected due to the driver's movement within the FA area, thus improving the accuracy of detecting the driver's incapacity.

[0043] If the driver temporarily shifts their head section or if the head section cannot be detected due to image blur, the head section is frequently re-detected near a last position of the trajectory. Thus, if the head detection section 61 cannot detect the head section, it searches near the last position of the trajectory obtained by the trajectory acquisition section 62. In this way, even if the driver's head section cannot be detected, it can be efficiently re-detected using the head section's trajectory.

[0044] The posture instability detection section 72 determines whether the driver's posture is instability while the vehicle 10 is in motion. If it is determined that the driver's posture is instability, the posture instability detection section 72 detects that the driver is no longer able to drive. More precisely, if the inclination θ of the head section, detected by the inclination detection section 64, is greater than a threshold Th1 (a relative inclination threshold), the posture instability detection section 72 detects that the driver is no longer able to drive.

[0045] Since the driver's torso is generally restricted by the driver's seat 11 and the seat belt 12, the torso hardly moves even if the driver loses consciousness. Since the driver's head is generally not restricted, the driver must consciously control their head. Therefore, if the driver suddenly becomes ill and loses consciousness, the driver can no longer control their head, and as described in Fig. 7A and Fig. As shown in Figure 7b, the head section is often strongly tilted in some direction towards the torso.

[0046] When the driver looks to the side while driving, they generally turn their neck. As it says in Fig. As shown in Figure 8, the inclination of the head section relative to the torso is therefore small. When the driver picks up an object located away from the driver's seat, the driver generally consciously tilts their torso, and thus, as shown in Figure 8, the inclination of the head section relative to the torso is small. Fig. As shown in Figure 9, the inclination θ of the head section relative to the torso is small. Therefore, if the inclination θ of the head section is greater than the threshold Th1, the posture intrusion state detection section 72 detects that the driver is no longer able to drive. If, at this point, the driver turns their face away from the front of the vehicle 10, the posture intrusion state detection section 72 detects that the driver is no longer able to drive, thus preventing a false detection of the inability to drive condition.

[0047] The Direction Breakdown State Detection Section 73 determines whether the driver's direction of face is breaking down while the vehicle 10 is in motion, and if it is determined that the direction of face is breaking down, the Direction Breakdown State Detection Section 73 detects that the driver is no longer able to drive. More precisely, the Direction Breakdown State Detection Section 73 detects that the driver is no longer able to drive if the direction of face, with respect to the front of the vehicle 10 as detected by the Direction of Face Detection Section 65, has been greater than a threshold Th2 (a direction of face threshold) for a time T2 (a Direction Breakdown Determination Time) or longer.

[0048] If the driver suddenly becomes ill, the driver is generally unable to maintain the direction of travel, and as is the case in Fig. 10A and Fig. As shown in 10B, the direction of view remains locked in relation to the front of the vehicle 10. In contrast, if the driver looks to the side while driving, the driver generally often returns to the direction of view immediately afterward. Thus, in the above case, the direction lock state detection section 73 detects that the driver is no longer able to drive.

[0049] Alternatively, the direction intrusion state detection section 73 detects that the driver is no longer able to drive if the direction of face relative to the front of the vehicle 10, detected by the face direction detection section 65, is greater than the threshold Th2 and the driver removes or has removed their hands from the steering wheel 15. Whether the driver removes or has removed their hands from the steering wheel 15 can be detected from an image or using a pressure sensor installed on the steering wheel 15.

[0050] If the driver suddenly becomes ill, the driver is generally unable to maintain their direction of travel, and, as is stated in Fig. 10A and Fig. As shown in Figure 10B, the driver's facing direction in relation to the front of the vehicle 10 changes, and the driver takes their hands off the steering wheel 15. If, on the other hand, the driver looks to the side while driving, as shown in Figure 10B, the driver's face changes direction in relation to the front of the vehicle 10, and the driver takes their hands off the steering wheel 15. Fig. As shown in Figure 11, the driver generally changes their facing direction while holding the steering wheel 15. If the driver looks to the side, they can turn their neck to change their facing direction without tilting their head section. Thus, in the above case, the directional intrusion state detection section 73 detects that the driver is no longer able to drive.

[0051] If the face direction detected by the face direction detection section 65 is greater than the threshold Th2 and the accelerator opening is greater than a predetermined opening, the direction intrusion state detection section 73 detects that the driver is no longer able to drive.

[0052] If the driver looks to the side while driving, they generally do not press the accelerator pedal as much. Accordingly, if the direction of gaze relative to the front of the vehicle is greater than the threshold Th2 and the accelerator opening is greater than the predetermined opening, the direction of gaze break is likely due to a sudden illness rather than looking to the side. Thus, in the above case, the direction break condition detection section 73 detects that the driver is no longer able to drive.

[0053] If the direction of face detected by the face direction detection section 65 is greater than the threshold Th2 and the accelerator and brake pedals have not been operated or actuated for a time T3 (operating time) or longer, the direction intrusion state detection section 73 detects that the driver is no longer able to drive.

[0054] If the driver suddenly becomes ill, their orientation generally changes with respect to the front of the vehicle 10, and the accelerator and brake pedals are not operated for a period longer than time T3. Conversely, if the driver looks to the side while driving, they are generally more likely to change their orientation and operate the accelerator and brake pedals within time T3. Thus, in the above case, the orientation break state detection section 73 detects that the driver is no longer able to drive.

[0055] The shake state detection section 74 determines the shake state of the driver's head section due to an external force while the vehicle 10 is in motion. If the head section shakes abnormally, the shake state detection section 74 detects that the driver is no longer able to drive. More precisely, the shake state detection section 74 detects that the driver is no longer able to drive if, during a period of time T5 (shake determination time) following the application of an external force to the vehicle 10, the amplitude of any shaking of the head section detected by the head detection section 61 is less than an amplitude Am1 (first amplitude) or greater than an amplitude Am2 (second amplitude). Amplitude Am2 is greater than amplitude Am1.

[0056] As it is in Fig. As shown in Figure 12, when an external force is applied to the vehicle 10, a vibration is transmitted to the driver after a predetermined time. Normally, in a case where the driver has not lost consciousness, the driver's head section oscillates, shakes, or wobbles when an external force (more precisely, a vertical external force) is applied to the vehicle 10, with an amplitude in the range of amplitude Am1 to amplitude Am2, as shown in Figure 12. Fig. Figure 13 illustrates this. Conversely, if the driver suddenly becomes ill and stiff, the amplitude of the head section's shaking is smaller than usual. If the driver suddenly becomes ill and then suddenly relaxes, the amplitude of the head section's shaking is larger than usual. Thus, in such a case, the shaking state detection section 74 detects the driver's incapacitation.

[0057] Time T5 is the time it takes for the external force to become irrelevant to the driver after it has been applied to vehicle 10. Amplitude Am1 and amplitude Am2 are time functions, and Fig. Figure 13 provides an example. When determining the shaking state, a minimum value of the amplitude Am1 and a maximum value of the amplitude Am2 are simply set as threshold values ​​during a period in which the time T5 elapses after the application of an external force.

[0058] The white-eye condition calculation section 75 determines the whiteness of the eye while the vehicle 10 is in motion, and when the white-eye detection section 66 detects the white-eye condition, the white-eye condition calculation section 75 detects that the driver is no longer fit to drive. More precisely, if the white-eye degree calculated by the white-eye degree calculation section 68 is greater than a threshold Th3 (white-eye threshold), the white-eye condition calculation section 75 detects that the driver is no longer fit to drive.

[0059] If the driver is able to drive, they normally never show the whites of their eyes. However, if the driver suddenly becomes ill, as is often the case, this can happen. Fig. 14A and Fig. As shown in Figure 14B, the driver must show the whites of their eyes. When the white-eye state is detected, the white-eye state calculation section 75 detects that the driver is no longer able to drive.

[0060] The memory device 52 (a memory section) stores a threshold value and a determination value used by a state detection section. Additionally, the memory device 52 stores the tilt θ of the head section, the face direction, and the amplitude of head section shaking, which are learned by the learning section 58. The memory device 52 records personal information, including the driver's medical history and age. If there are multiple drivers, personal information is recorded for each driver. The memory device 52 records the driver's posture not determined to be an impairment state and the driver's posture determined to be an impairment state. The driver's posture not determined to be an impairment state includes a normal driving posture and commonly assumed postures.The driver's posture, which is determined to be a state of unfitness to drive, includes a posture the driver assumes when chronically ill or suffering from a seizure. The driver takes pictures of their posture in the driver's seat in advance using the driver cameras 21 and records the posture in the storage device 52.

[0061] The following describes a processing procedure for recording the driver's state of unfitness to drive with reference to the flowcharts of the Fig. 16A and Fig. 16B described. This processing procedure is executed by controller 50.

[0062] It is determined whether the vehicle speed of vehicle 10 is equal to or less than V (S10). V can be 0 km / h (Stop) or sufficiently low to indicate a stop (for example, 1 km / h). If the vehicle speed is equal to or less than V (S10: Yes), the determination in S10 is repeated until the vehicle speed is greater than V. If the vehicle speed is greater than V (S10: No), it is determined that the vehicle is moving, and processing to detect the driver's incapacity state begins in order to determine the out-of-frame state.

[0063] Here, in the S10 determination, it can be determined whether the driver is driving. If the driver is driving, the determination in S10 can be repeated, and if the vehicle is moving but the driver is not, the processing to detect the driver's incapacity can be initiated. For example, it is determined whether the vehicle speed is equal to or greater than V2 (e.g., 50 km / h), whether the steering angle detected by the steering sensor 32 is equal to or greater than a predetermined angle, or whether the steering angular velocity is equal to or greater than a predetermined angular velocity. V2 is a value indicating that the driver is pressing the accelerator pedal, and the predetermined angle and predetermined angular velocity are values ​​indicating that the driver is turning the steering wheel.If at least one of the three conditions is met, it is determined that the driver is driving, and the processing to detect the driver's inability to drive is not started.

[0064] When determining the out-of-frame state, the driver's head and torso are first captured in the image of the driver's seat (S11). At this point, facial features contained in the driver's head are captured to authenticate the driver. The driver can be authenticated in advance by communication with a mobile device such as a smartphone or by communication with a vehicle key 10 in which personal information is registered.

[0065] The next step is to determine whether the head section position can be detected (S12). If the head section position is not detected (S12: No), head section trajectory information is obtained from head section positions recorded in the processing described below in S19, and it is determined whether the trajectory information indicates an out-of-frame state (S13). That is, it is determined whether the reason for not detecting the head section position is that the head section lies outside an image area or that the image is blurry.

[0066] If the head section trajectory information does not indicate an out-of-frame state (S13: No), it is determined that the head section cannot be captured due to image blurring, and the procedure returns to processing in S10. Conversely, if the head section trajectory information indicates an out-of-frame state (S13: Yes), it is determined that the head section is outside the imaging area. In this case, to confirm that the head section is outside the imaging area, information from the seat belt sensor 22 and the seat surface sensor 23 is used to support this (S14).

[0067] The system then determines whether the out-of-frame condition lasted for a time T0 (out-of-frame determination time) or longer (S15). More precisely, it determines whether, during time T0 or longer, the position of the head section was outside the area FA, and furthermore, whether the pull magnitude of the seat belt 12 is greater by a first pull magnitude or more than the magnitude detected when the seat belt 12 was fastened, and furthermore, whether a high-pressure section in the pressure distribution of the seat section 11a is located at one end of the seat section 11a. If these three conditions are met, it is determined that the driver was out of frame for time T0 or longer. The following condition can be specified here: The pull magnitude per magnitude detection time, detected by the seat belt sensor 22, that is, the pull velocity of the seat belt 12, is greater than a second pull magnitude.

[0068] Information from the seat belt sensor 22 and the seat surface sensor 23 does not need to be used. That is, the processing in S14 does not need to be carried out, and the processing in S15 can only determine whether the position of the head section was continuously outside the area FA during time T0 or longer.

[0069] Here, the time T0 is set based on personal information recorded in memory device 52. For example, if the person is older, the time T0 is shorter. A shorter time T0 is set for a person with a specific medical history compared to a person without one. Additionally, the time T0 is modified according to the driver's condition and the driving environment. In a condition where the driver shows signs of impairment, in a condition where the likelihood of the driver becoming incapacitated is high, or in a driving environment where a collision is likely to occur if the driver becomes incapacitated, the time T0 is reduced to facilitate the detection of the driver's impairment.

[0070] In particular, if the position of the head section, recorded in processing S19, vibrates with an amplitude greater than a predetermined amplitude—that is, if the head section shakes or wobbles erratically—the probability of a state of incapacity is high, and thus the time T0 is reduced. If the head section's movement speed is faster, the postural collapse due to sudden illness occurs more quickly than when picking up an object. Thus, in the acquired trajectory information of the head section, if the head section's movement speed is higher, the time T0 is reduced. In the case of a postural collapse due to sudden illness, the head section's movement speed often increases as it approaches the end of the FA range.Therefore, if the movement speed of the head segment increases as its recorded position approaches the end of the FA range, the collapse of posture is more likely due to a sudden illness than to picking up an object. Consequently, if the movement speed of the head segment increases as its recorded position approaches the end of the FA range, the time T0 is reduced. If a seizure is detected, the probability of the incapacitation state is high, and thus the time T0 is reduced.

[0071] To avoid a collision, a suitable vehicle control system must start earlier when the vehicle speed is higher. Thus, the time T0 is reduced when the vehicle speed of vehicle 10 is higher. To avoid a collision, a suitable vehicle control system must start earlier than the TTC (time to collision), which is obtained by dividing the distance between a vehicle ahead and the vehicle itself by their relative speeds. Thus, the time T0 is reduced when the TTC is shorter. If a driver assistance control system such as ACC (adaptive cruise control) and LKA (lane keeping assist) is implemented in vehicle 10, the driver's attitude may drop for a long time. Thus, the time T0 is increased.Furthermore, the time T0 can be reduced in the week or in a time frame in which a sudden illness such as a heart attack statistically occurs frequently.

[0072] If the out-of-frame condition has lasted for time T0 or longer (S15: No), the procedure proceeds to processing in S21. If the out-of-frame condition has lasted for time T0 or longer (S15: Yes), it is determined that the driver is no longer able to drive, and the driver is questioned about their incapacity to drive. More specifically, the driver is informed of the detection of the incapacity to drive by means of an audible signal from speaker 82, a display on indicator 81, a flashing indicator (not shown), or similar means, and it is determined whether the driver responds within a predetermined time (S16).

[0073] If, within the predetermined time, any contact by the driver with a touch-sensitive display, any speech by the driver, any operation of the steering wheel 15, the brake, etc. of the vehicle 10, or any operation of a specific switch is detected, it is determined that the driver responds (S16: No). If none of these are detected, it is determined that the driver does not respond (S16: Yes).

[0074] If it is determined that the driver is responding, the recognition that the driver is able to drive is indicated by an audible output from the loudspeaker 82 or by a display on the screen 81 (S17). Conversely, if it is determined that the driver is not responding, the vehicle control unit 90 is instructed to brake and steer the vehicle appropriately to bring it to a stop. To notify surrounding vehicles of this, the vehicle control unit 90 is instructed to switch on a headlight and sound the horn (S18). This fact is also communicated to other occupants of the vehicle 10.

[0075] If processing in S12 determines that the head section position can be captured (S12: Yes), the head section and torso positions are recorded (S19). Trajectory information about the head section can be obtained from head section positions recorded in each image.

[0076] The procedure then determines whether the head section's position lies outside the preset area FA (S20). If the head section's position is within the mapping area but outside FA (S20: Yes), the procedure proceeds to S15 to determine the out-of-frame state.

[0077] Subsequently, based on the relative position of the head section to the torso, it is determined whether the driver's posture is the posture identified as the incapacitated driving state, which is registered in memory device 52 (S21). If the driver's posture is the posture identified as the incapacitated driving state (S21: Yes), it is determined that the driver is no longer able to drive, and the procedure proceeds to processing in S16.

[0078] If the driver's posture is not one identified as an incapacitating condition (S21: No), it is determined whether the driver's posture is the position not identified as an incapacitating condition and registered in memory device 52 (S22). If the driver's posture is the position not identified as an incapacitating condition (S22: Yes), the procedure returns to processing in S10. If the driver's posture differs from the position not identified as an incapacitating condition (S22: No), a posture breakdown is determined.

[0079] It is determined whether a postural drop is detected (S23). More precisely, the postural drop is determined according to the processing of the subroutine of the Fig. 17. First, the tilt of the head section and its direction of tilt are calculated (S231). Then, the tilt of the torso and its direction of tilt are calculated (S232). Next, the angle formed by the calculated torso tilt with the calculated head section tilt, that is, the tilt θ of the head section relative to the torso, is calculated (S233). The calculated tilt θ of the head section is learned if the driver's incapacity to drive is not detected. If the calculated tilt θ of the head section is then greater than the threshold Th1 (relative tilt threshold), a postural fall is detected (S234). If the postural fall is not detected (S23: No), the procedure for determining a face direction fall proceeds in S25.

[0080] If a posture collapse is detected (S23: Yes) it is determined whether during time T1 or longer the inclination θ of the head section was continuously greater than the threshold Th1 and the face was not continuously directed towards the front of the vehicle 10 (S24).

[0081] Here, the following simple condition can be defined: During time T1 or longer, the tilt θ of the head section is continuously greater than the threshold Th1, or the face is not continuously oriented towards the front of the vehicle 10. Alternatively, the following condition can be defined: During time T1 or longer, the calculated tilt direction of the head section and the tilt direction of the torso do not vary continuously; that is, the position of the head section and the position of the torso lie continuously within a region UA ​​(non-motion determination region). Within region UA, the head section and the torso are assumed to be stationary. Another alternative condition can be defined: The calculated tilt θ of the head section is greater than a learned tilt of the head section by a determination value D1 (a tilt determination value) or more.Alternatively, the following condition can be set: The steering wheel 15 is not operated during time T3 (an operating determination time) or longer.

[0082] Here, time T1 is set similarly to time T0 based on personal information recorded in memory device 52 and is modified according to vehicle speed, TTC, and the presence or absence of a driver assistance control execution. The threshold T1 decreases if a seizure is detected. Furthermore, time T1 can be reduced, similarly to time T0, during a week or timeframe in which a sudden illness, such as a heart attack, occurs statistically frequently.

[0083] If, during time T1 or longer, the inclination θ of the head section was greater than the threshold Th1 and the face was not directed towards the front of the vehicle 10 (S24: Yes), it is determined that the driver is no longer able to drive, and the procedure proceeds to question processing in S16. If, during time T1 or longer, the inclination θ of the head section was not continuously greater than the threshold Th1, or if the face was continuously directed towards the front of the vehicle (S24: No), a face direction dip is determined next.

[0084] The system determines whether a face direction instability is detected (S25). More specifically, the driver's face direction is detected in relation to the front of the vehicle. If the detected face direction is greater than the threshold Th2 (a face direction threshold), a face direction instability is detected. If the face direction instability is not detected (S25: No), the procedure proceeds to a shake state determination in S28. The detected face direction is learned if the driver's incapacity to drive state is not detected.

[0085] If the face direction dip is detected (S25: Yes), it is determined whether the face direction was continuously greater than the threshold Th2 during time T2 (face direction dip determination time) or longer (S26).

[0086] Here, time T2 is set similarly to time T0 based on personal information recorded in memory device 52 and is modified according to vehicle speed, TTC, and the presence or absence of a driver assistance control execution. Furthermore, time T2 can be reduced, similarly to time T0, during a week or timeframe in which a sudden illness, such as a heart attack, occurs statistically frequently. If a seizure is detected, the Th2 threshold is reduced.

[0087] If the face direction was continuously greater than the threshold Th2 during time T2 or longer (S26: Yes), it is determined that the driver is no longer able to drive, and the procedure proceeds to question processing in S16.

[0088] If the face direction was not continuously greater than the threshold Th2 during time T2 or longer (S26: No), it is determined whether the driver removed their hands from the steering wheel 15 during time T3 or longer, whether the accelerator opening was greater than a predetermined opening, or whether the accelerator and brake pedals were not pressed (S27). If at least one of these three conditions is met in the processing in S27 (S27: Yes), it is determined that the driver is no longer able to drive, and the procedure proceeds to the question processing in S16. If none of the three conditions is met in the processing in S27 (S27: No), the shaking state is determined.

[0089] Here, in the determination processing in S26, the condition that at least one of the three conditions in the processing in S27 is met can be set. In the determination processing in S26 and S27, the following condition can be set: The detected facial direction is greater than the learned facial direction by the determination value D1 (inclination determination value) or more. Since the driver generally does not enter a state of incapacity to drive if the driver's hands are higher than the driver's neck, the condition that the driver's hands are positioned lower than the driver's neck can be set.

[0090] In the following, the shaking state determination is performed to determine whether an abnormal discontinuity of the head section is detected when an external force is applied (S28). More precisely, it is determined whether, until the time T5 (a shaking determination time) has elapsed after an external force has been applied to the vehicle 10, the amplitude of the oscillation or shaking of the head section is smaller than the amplitude Am1 (first amplitude) or larger than the amplitude Am2 (second amplitude).

[0091] The following condition can be set here: The head section vibrates with an abnormal amplitude during the time T5 elapses after an external force has been applied to the vehicle 10, and after the time T5 has elapsed, the position of the head section lies within the range UA. That is, the following condition can be set: The head section vibrates in accordance with an external force, and after the effect of the external force has subsided, the position of the head section does not change. Alternatively, the following condition can be set: If the driver's incapacity is not detected, an amplitude of head section shaking is learned, and the detected amplitude of head section shaking is greater than the learned amplitude of head section shaking by the determination value D2 (amplitude determination value) or more.A condition can be set that the steering wheel 15 has not been continuously operated during time T3 or longer.

[0092] If abnormal discontinuity of the head segment is detected upon application of an external force (S28: Yes), it is determined that the driver is no longer able to drive, and the procedure proceeds to question processing in S16. If abnormal discontinuity of the head segment is not detected upon application of an external force (S28: No), the white eye condition is determined.

[0093] It is determined whether the white-eye condition is being recorded (S29). More precisely, if the calculated white-eye level is greater than the threshold Th3 (white-eye threshold), it is determined that the white-eye condition is being recorded. Here, the white-eye level of each of the driver's two eyes is calculated, and if both white-eye levels of both eyes are greater than the threshold Th3, it is determined that the white-eye condition is being recorded. However, if only one eye is being recorded, or if the white-eye condition is being recorded in a simplified manner, the recording of the white-eye condition can also be determined based on the white-eye level of only one eye.

[0094] If it is determined that the white eye condition is not recorded (S29: No), the driver's inability to drive was not recorded in the out-of-frame determination, the posture intrusion determination, the face direction intrusion determination, the shaking condition determination and the white eye condition determination, and thus the procedure proceeds to processing in S10.

[0095] If it is determined that the white-eye condition is being recorded (S29: Yes), it is determined whether the white-eye level was greater than the threshold Th3 during time T4 (white-eye determination time) or longer (S30). A condition can be set here that the steering wheel 15 was not continuously operated during time T3 or longer.

[0096] Here, time T4 is set similarly to time T0, based on personal information recorded in memory device 52, and is modified according to vehicle speed and TTC. If a seizure is detected, the threshold Th3 is reduced. Similar to time T0, time T4 can be reduced during the week or within a timeframe in which a sudden illness, such as a heart attack, occurs statistically frequently.

[0097] If the white eye level was not continuously greater than the threshold Th3 during time T4 or longer (S30: No), the procedure proceeds to processing in S10. If the white eye level was continuously greater than the threshold Th3 during time T4 or longer (S30: Yes), the driver's unfitness to drive state is recorded, and the procedure proceeds to question processing in S16. This processing ends here.

[0098] As it is in Fig. As shown in Figure 18, the driver's postural incapacity level is displayed on the screen 81 based on the detected head section tilt θ relative to the torso, the detected facial direction, and the detected head section position. If the detected head section tilt θ is greater, the postural incapacity level is increased. If the detected facial direction is greater, the postural incapacity level is increased. If the detected head section position deviates from a standard position during driving, the postural incapacity level is increased. The standard position during driving is the head position at vehicle start 10 or a mid-range head section position at a time when the driver's incapacity is not detected.

[0099] According to the present embodiment, the following effects can be achieved.

[0100] By recognizing that the driver is no longer able to drive when the driver's head segment moves outside the FA area, the driver's incapacity can be detected quickly. Furthermore, by considering the trajectory achieved until the head segment moves outside the FA area, the accuracy of detecting the driver's incapacity can be improved.

[0101] When the driver picks up an object located away from the driver's seat, the head section normally returns to area FA, even if the head section is temporarily outside of area FA. Therefore, by setting the condition that the driver's head section has been outside of area FA for time T0 or longer, a false detection of the driver's incapacity to drive can be prevented.

[0102] If the FA area occupies the entire image, the capture processing can be further simplified, as the driver's incapacity to drive is captured even if the driver's head section is not present in the image.

[0103] If the driver's posture collapses due to a sudden illness, it is estimated that the tension of the seat belt 12 at that time is greater than the tension at the time the seat belt was fitted by one or more tension increments. Even if the driver's head section cannot be detected, if the tension of the seat belt 12 is greater than the tension at the time of fitting by one or more tension increments, it is determined that the driver's head section has moved outside the imaging area. Thus, by setting the condition that the tension of the seat belt 12 is greater than the tension at the time of fitting by one or more tension increments, the driver's incapacity to drive can be detected with high accuracy.

[0104] The driver's posture deteriorates even more rapidly when the driver suddenly becomes ill than when the driver picks up an object. Therefore, the seatbelt 12 is deployed more quickly when the driver suddenly becomes ill than when the driver picks up an object. Thus, by setting a condition that the pull magnitude of the seatbelt 12 is greater than a second pull magnitude per measurement time, an inaccurate detection of the driver's incapacity to drive can be prevented.

[0105] If the driver's posture collapses due to a sudden illness, it is estimated that a high-pressure section of the pressure distribution of seat section 11a of the driver's seat is inclined towards one end of seat section 11a. Even if the driver's head section cannot be detected, if the high-pressure section of the pressure distribution of seat section 11a of the driver's seat is located at the end of seat section 11a, it is determined that the driver's head section lies outside the imaging area. Thus, by setting the condition that the high-pressure section of the pressure distribution of seat section 11a of the driver's seat is located at the end of seat section 11a, the driver's incapacity to drive can be detected with high accuracy.

[0106] By reducing the time T0 when the head section vibrates with an amplitude greater than a predetermined amplitude, that is, when the head section oscillates or shakes discontinuously, the time required to determine the driver's incapacity to drive can be reduced.

[0107] By reducing the time T0 when the movement speed of the head section is higher, the time required to determine the driver's incapacity to drive can be reduced.

[0108] By reducing the time required to determine the inability to drive state as the head section's movement speed increases when the head section approaches the end of the FA area, the time required to determine the driver's inability to drive state can be reduced.

[0109] If the inclination θ of the head section relative to the torso is greater than the threshold Th1, the loss of posture is more likely due to a sudden illness rather than picking up an object. Therefore, by detecting that the driver is no longer able to drive when the inclination θ of the head section relative to the torso is greater than the threshold Th1, the driver's state of incapacity to drive can be detected with high accuracy.

[0110] If the driver consciously tilts their head sharply towards their torso while driving, it is assumed that the driver continues to face the front of the vehicle 10 for safety reasons. Therefore, by setting the condition that the face is not facing the front of the vehicle 10, a false detection of the driver's incapacity to drive can be prevented.

[0111] If the driver deliberately tilts the head section sharply towards the torso while driving, it is assumed that the driver will quickly return the head section to its initial position for safety reasons. Therefore, by setting the condition that the head section was tilted sharply relative to the torso for a period of time T1 or longer, a false detection of the driver's incapacity to drive can be prevented.

[0112] By setting the condition that the positions of the head section and torso remained unchanged during time T1 or longer, while the head section was strongly inclined relative to the torso, a false detection of the driver's incapacity to drive can be prevented.

[0113] By setting the condition that the detected inclination θ of the head section is greater than the learned inclination of the head section by the determined value D1 or more, a false detection of the driver's inability to drive can be prevented even if the driver has the habit of tilting the head section in relation to the torso.

[0114] If the driver's gaze direction relative to the front of the vehicle 10 was greater than the threshold Th2 for time T2 or longer, the collapse of the gaze direction is more likely due to a sudden illness rather than looking to the side. Accordingly, by detecting such a case in which the driver is no longer able to drive, the driver's state of incapacity can be detected with high accuracy.

[0115] If the driver's gaze direction relative to the front of the vehicle 10 is greater than the threshold Th2 and the driver takes their hands off the steering wheel 15, the gaze direction is more likely to collapse due to a sudden illness rather than looking to the side. Therefore, by detecting such a case in which the driver is no longer able to drive, the driver's state of incapacity can be detected with high accuracy.

[0116] If the direction of gaze relative to the front of the vehicle is greater than the threshold Th2 and the accelerator opening is larger than the predetermined opening, the change in gaze direction is more likely due to a sudden illness rather than looking to the side. Accordingly, by detecting such a situation in which the driver is no longer able to drive, the driver's state of incapacity can be detected with high accuracy.

[0117] If the driver's gaze direction relative to the front of the vehicle is greater than the threshold Th2 and the accelerator and brake pedals have not been operated for a period longer than time T3, the gaze direction is more likely to collapse due to a sudden illness than due to looking to the side. Accordingly, by detecting such a situation in which the driver is no longer able to drive, the driver's state of incapacity can be detected with high accuracy.

[0118] If the driver is capable of driving, the steering wheel is generally operated within the time T3. Therefore, by setting the condition that the steering wheel has not been operated for a period longer than T3, a false detection of the driver's incapacity to drive can be prevented.

[0119] If the driver is capable of driving, the driver never continuously depresses the accelerator pedal for a period longer than time T3. Thus, by setting the condition that the accelerator opening was greater than the predetermined opening for a period longer than time T3, an incorrect determination of the driver's incapacity to drive can be prevented.

[0120] If the driver intentionally changes their facing direction relative to the front of the vehicle 10 while driving, they will likely return their face to the front. Therefore, by setting the condition that the facing direction towards the front of the vehicle 10 was greater than the threshold Th2 for time T2 or longer, a false detection of the driver's incapacity to drive can be prevented.

[0121] If the driver suddenly becomes ill, they will generally not be classified as unfit to drive if their hands are higher than their neck. Therefore, by setting the condition that the driver's hands are lower than their neck, a false detection of the driver's unfitness to drive can be prevented.

[0122] During the period T5 elapsed after an external force has been applied to the vehicle, the driver's inability to drive can be easily detected by sensing that the driver is no longer able to drive if the amplitude of the head section shaking is less than the amplitude Am1 or greater than the amplitude Am2.

[0123] If the driver has lost consciousness, the head section generally vibrates in response to the external force. Once the effect of the external force subsides, the head section becomes stationary. Conversely, if the driver normally vibrates the head section, it vibrates regardless of the effect of the external force. Thus, by setting the condition that the head section vibrates in response to the external force and that its position does not change after the effect of the external force subsides, a false detection of the driver's incapacitation can be prevented.

[0124] By setting the condition that the detected amplitude of head shaking is greater than the learned amplitude of head shaking by a value exceeding a determination value D2, even if the driver has a habit of swinging or shaking their head, a false detection of the driver's inability to drive can be prevented.

[0125] By recording the white-eye condition, the driver's inability to drive can be determined with high accuracy by detecting that the driver is no longer able to drive.

[0126] The driver's eye outline and black eye area are captured based on the image. The degree of white-eye impairment is then calculated based on this captured outline and black eye area. If the degree of white-eye impairment exceeds the threshold Th3, the system detects that the driver is unfit to drive. Thus, the driver's white-eye condition can be detected with high accuracy, and the driver's unfitness to drive can be detected with high accuracy.

[0127] Even if the driver rolls their eye upwards to temporarily increase the white-eye level, the level of white-eye impairment remains low in a state where they are capable of driving. Therefore, by setting the condition that the white-eye level has been greater than the threshold Th3 for a period of time T4 or longer, a false detection of the driver's unfitness to drive can be prevented.

[0128] By setting the condition that the degree of white vision in each eye is greater than the threshold Th3, even if the white vision is erroneously detected because the driver is wearing a blindfold or one eye is an artificial eye, the driver's unfitness to drive will not be detected if the driver does not show the white of the other eye. This prevents an incorrect detection of the driver's unfitness to drive.

[0129] Since the ratio of the longitudinal length Lb of the black eye area to the longitudinal length Lw + Lb of the eye correlates with the proportion of the white eye area to the total area of ​​the eye, the degree of white eye can be calculated from the ratio of the longitudinal length Lb of the black eye area to the longitudinal length Lw + Lb of the eye.

[0130] If the distance Lb between the upper edge of the eye outline and the lower edge of the black eye area is smaller, the degree of whiteness in the eye increases. Therefore, the degree of whiteness in the eye can be calculated from the distance Lb between the upper edge of the eye outline and the lower edge of the black eye area.

[0131] The degree of whiteness in the eyes can be calculated from the ratio of the area of ​​the white eye region to the area of ​​the black eye region.

[0132] In the white-eye state, the black eye area is oriented upwards, thus making the flattening of the black eye area appear greater. Therefore, the degree of white eye can be calculated from the flattening of the black eye area.

[0133] If the distance Lc between the longitudinal midline of the entire eye and the lower edge of the black eye area is greater, the black eye area becomes larger, and the degree of whiteness increases. Therefore, the degree of whiteness can be calculated from the distance Lc between the midline and the lower edge of the black eye area.

[0134] In general, the probability of a driver becoming incapacitated is higher if the driver experiences a seizure than if the driver does not. Therefore, if a seizure is detected, the driver's incapacitation can be detected even more quickly by lowering the Th1 and Th2 thresholds.

[0135] By reducing or shortening the times T0, T1, T2 and T4 when the vehicle speed is higher, the time required to determine the driver's incapacity to drive can be reduced when the vehicle speed is higher, thus allowing appropriate vehicle control to be initiated earlier.

[0136] By reducing the times T0, T1, T2 and T4 as the TTC decreases, the time required to determine the driver's incapacity to drive can be reduced as the TTC decreases, thus enabling appropriate vehicle control to be initiated earlier.

[0137] By setting the times T0, T1, T2 and T4 based on personal information including medical history and the driver's age, the time required to determine the state of unfitness to drive can be adjusted according to the characteristics of each driver.

[0138] In the case where a driver assistance control system is implemented in the vehicle, incorrect detection of the driving incapacity state can be prevented by extending the times T0, T1 and T2.

[0139] By reporting the degree of the driver's posture to the system, the driver can recognize their own posture. Therefore, even if the driver's posture deteriorates, they can correct it, preventing the impairment from being detected. This can prevent an incorrect assessment of the impairment.

[0140] By asking the driver whether they are no longer able to drive when the driver's incapacity to drive is detected, vehicle control measures to safely stop the vehicle can be avoided if the driver's incapacity to drive is incorrectly detected. (First modification of the first embodiment)

[0141] When determining the out-of-frame state by the out-of-frame state detection section 71, the predetermined area FA can be elliptical, as shown in Fig. Figure 19 shows the ellipse. It has a major axis along the width of the vehicle 10 and a minor axis along the height of the vehicle 10.

[0142] When multiple movements of the driver's head segment are considered during driving, the driver's head segment moves right and left as the vehicle turns right and left, and the driver's head segment moves within an elliptical area that has a principal axis in the left-right direction. Thus, because the FA area is elliptical, even if the driver moves their head segment during normal operation, a false detection of the driving impairment state can be prevented. (Second modification of the first embodiment)

[0143] The detection device 100 can include a deformation section that deforms the area FA according to the user's intent (which includes the driver) or driver information. The deformation section can be configured as a function of the state detection section 70. Deformation involves enlarging, reducing, and changing the shape of the area FA. The deformation section deforms the area FA according to user input to the HMI 80. The HMI 80 can include a zoom switch and a zoom switch, and the display 81 can be a touch-sensitive display.

[0144] As it is in Fig. As shown in Figure 20, if the driver has a habit of shifting their driving posture from a standard position, or if it is difficult to detect the inability to drive, the driver may operate a magnification switch or similar device to enlarge the FA area into an FAa area.

[0145] In the event that the driver is an elderly person or someone with a chronic illness, the driver or the driver's family can operate a reduction switch or similar device to reduce the FA area to an FAb area for even faster detection of the incapacity to drive. In the event that the driver is a bus or taxi driver, a bus or taxi administrator can operate a reduction switch or similar device to reduce the FA area to an FAb area for even faster detection of the incapacity to drive. Fig. 20 areas FA, FAa and FAb are elliptical, but can also be rectangular.

[0146] Regardless of the user's intent, the deformation section can enlarge or reduce the FA area according to the driver information. The driver information includes the driver's age, medical history, and / or driving posture habits and is stored in memory device 52. Driving posture habits can be learned from the driver's head section by learning section 51 or can be registered by the driver pre-mapping their posture.

[0147] The deformation section, for example, increases the FA range to FAa if the driver has a habitual driving posture, and decreases the FA range to FAb if the driver is elderly or has a medical history. The deformation section may increase or decrease the FA range according to vehicle speed, TTC, and the presence or absence of driver assistance controls, or it may decrease the FA range during a week or timeframe in which a sudden illness, such as a heart attack, statistically occurs more frequently.

[0148] The deformation section modifies the FA area according to the driver's intent and / or driver information, and can modify the area according to both the user's intent and the driver information. The deformation section can increase or decrease the FA area. The deformation section can deform the FA area according to the user's intent or the driver information.

[0149] By deforming the FA area in this way according to the user's intent or the driver information, the incapacity to drive state can be detected accordingly. For example, if the probability of an incapacity to drive is high, the incapacity to drive can be detected quickly. If the driver has a habit of shifting their driving posture from the standard position, deforming the FA area according to this habit can prevent an incorrect detection of the incapacity to drive state. (Third modification of the first embodiment)

[0150] The detection device 100 includes a transmission section that transfers the FA area to an image according to the user's intent (which includes the driver) or driver information. The transmission section can be configured as a function of the state detection section 70. The transmission section moves the FA area according to an input to the HMI 80 by the user, for example, a drag on the touch-sensitive display. That is, this modification differs from the second modification in that the FA area is moved according to the user's intent or driver information, rather than being deformed.

[0151] As it is in Fig. As shown in Figure 21, if the driver has a habit of shifting their driving position from the standard position, moving area FA to area FAc according to this habit can prevent an incorrect detection of the driving impairment state. Fig. In 21, the area FA is rectangular, but can also be elliptical. The second modification can be combined with the third modification so that the area FA is both moved and deformed. (Fourth modification of the first embodiment)

[0152] The driver cameras 21, for example, include a stereo camera. As it says in Fig. As shown in Figure 22, the area FA is a three-dimensional area extending in the width and front-to-back directions of the vehicle 10. The stereo cameras are installed, for example, at the right and left ends of the rearview mirror 16. The shape of the area FA is not limited to a cube but can also be an oval sphere. If the area FA is an oval sphere, a principal axis of the oval sphere extends along the width of the vehicle 10. The second or third modification can be used together with the fourth modification.

[0153] The stereo camera can determine the position of the driver's head in three dimensions. This means that the position of the head in the lateral direction of the vehicle (10) as well as the front-to-back position can be determined. Therefore, by adjusting the FA area in a three-dimensional manner, even if the driver is no longer able to drive and falls to the side, or is no longer able to drive and falls forward, the driver's incapacity can be detected through an out-of-frame state determination. (Second embodiment)

[0154] The following describes the difference between a detection device 100 according to the second embodiment and the detection device 100 according to the first embodiment. The detection device 100 according to the second embodiment differs from the detection device 100 according to the first embodiment in the way in which the out-of-frame state is determined by the out-of-frame state detection section 71.

[0155] In this embodiment, the out-of-frame state detection section 71 sets three areas as the area FA, which is a predetermined area for out-of-frame state detection: an area FA1 (a first predetermined area), an area FA2 (a second predetermined area), and an area FA3. Area FA1 is the smallest area. Area FA2 contains an area larger than area FA1 and includes area FA1 and the area outside area FA1 (end of the image). Area FA3 contains an area larger than area FA2 and includes area FA2 and the area outside area FA2. In this embodiment, area FA1, area FA2, and area FA3 share a common center. In this embodiment, three areas are set as the predetermined area for out-of-frame state detection, but at least two areas can also be set.

[0156] The out-of-frame state detection section 71 sets an out-of-frame determination time for each predetermined area. That is, the out-of-frame state detection section 71 sets the times T01, T02, and T03 as the out-of-frame determination times for frame FA1, frame FA2, and frame FA3, respectively.

[0157] The out-of-frame detection section 71 detects that the driver is no longer able to drive if the driver's head section has gone outside at least one predetermined area for the out-of-frame determination time corresponding to the predetermined area or for a longer period.

[0158] This means that in this embodiment there are three conditions: (1) the driver's head section is outside the FA1 area for time T01 or longer; (2) the driver's head section is outside the FA2 area for time T02 or longer; and (3) the driver's head section is outside the FA3 area for time T03 or longer. If at least one of the three conditions (1) to (3) is met, the out-of-frame condition detection section 71 detects that the driver is no longer able to drive.

[0159] Thus, in the flowcharts of the Fig. 16A and Fig. 16B determines whether at least one of the three conditions (1) to (3) is met, instead of determining in processing S15 whether the head section was continuously outside the area FA for time T0 or longer. If at least one condition is met, the procedure proceeds to processing in S14, and if none of the three conditions are met, the procedure proceeds to processing in S21.

[0160] When the driver picks up an object located away from the driver's seat, a dwell time is generally reduced for safety reasons when the driver is away from the normal driving position. In this case, the out-of-frame detection section 71 adjusts the out-of-frame determination time to a shorter value if the predetermined range is larger. That is, the out-of-frame detection section 71 adjusts the out-of-frame determination time such that T01 > T02 > T03.

[0161] For example, if time T01 is set to three seconds and time T02 is set to one second, even if three seconds have not elapsed after the driver's head section has moved outside of area FA1, the driver's inability to drive will be detected at the time one second has elapsed after the driver's head section has moved outside of area FA2.

[0162] Similar to time T0, times T01, T02, and T03 are set based on personal information recorded in memory device 52 and modified according to vehicle speed, TTC, and the presence or absence of the driving assistance control. Also similar to time T0, times T01, T02, and T03 can be reduced during a week or timeframe in which a sudden illness, such as a heart attack, occurs statistically frequently. Furthermore, times T01, T02, and T03 can be reduced, similarly to time T0, according to the amplitude of the head section position, the speed of head section movement, and the detection of a seizure.

[0163] In Fig.In the first embodiment, areas FA1, FA2, and FA3 are elliptical. However, FA1, FA2, and FA3 can also be rectangular. Furthermore, the second or third modification of the first embodiment can be used to deform or move each of FA1, FA2, and FA3. The fourth modification of the first embodiment can be used to adjust FA1, FA2, and FA3 three-dimensionally.

[0164] In the second embodiment described above, several predetermined areas of varying sizes are set with the same center point, and the out-of-frame detection time is set for each predetermined area. That is, the out-of-frame detection time is set according to the degree of deviation from the normal driving position. Thus, the driver's incapacity to drive can be detected according to the degree of deviation from the normal driving position.

[0165] Since the out-of-frame determination time is set to a shorter duration when the degree of deviation from the normal driving position is greater, incorrect detection of the driving incapacity can be prevented, and the driver's driving incapacity can be detected early. (Other embodiments)

[0166] The driver camera 21 can be one of the four cameras mounted in a vehicle interior. At least one driver camera 21 must or can be provided.

[0167] The out-of-frame state detection section 71 can detect that the driver is no longer able to drive, based on the trajectory obtained by the trajectory acquisition section 62. If the driver suddenly becomes ill and is no longer able to drive, the driver's head section often moves from a position while driving and does not return to that position; therefore, the driver's inability to drive state can be detected based on the head section trajectory.

[0168] The Directional Intrusion State Detection section 73 can also detect that the driver is no longer able to drive if the driver's face is directed further downward than a Th2d threshold (downward threshold) or further upward than a Th2u threshold (upward threshold). If the driver suddenly becomes ill and loses consciousness, the driver's face will generally be directed further downward than the Th2d threshold or further upward than the Th2u threshold. Thus, if the driver's face direction is strongly downward or upward, it will be determined that the driver is no longer able to drive.

[0169] The shaking state detection section 64 can detect that the driver is no longer able to drive if, during the movement of the vehicle 10, when an external force is applied to the vehicle 10, the head section detected by the head detection section 61 is or was tilted in the direction of the external force for a time T6 (return determination time) or longer. If the driver has not lost consciousness, even when an external force (especially an external force in the right-left or forward-backward direction) is applied to the vehicle 10, the driver's head section is normally tilted in the direction of the external force, but the head returns to its original position within the time T6.In contrast, if the driver suddenly becomes ill and loses consciousness, the driver's head section exhibits low resistance to the external force and remains tilted in the direction of the external force for a time T6 or longer. Thus, the shaking state detection section 74 can detect the driver's incapacitation state in the above case.

[0170] If the mouth opening (particularly the longitudinal opening) detected by facial expression detection section 67 is larger than a defined opening size, the white-eye state calculation section 75 can also detect that the driver is no longer fit to drive. If the driver suddenly becomes ill and the whites of their eyes are showing, the driver tends to open their mouth. Thus, the driver's unfitness to drive can be detected if the driver's mouth opening is larger than a defined opening size.

[0171] Even if all the off-frame assessment, posture intrusion assessment, facial orientation intrusion assessment, shaking state assessment, and white-eye state assessment are performed, to achieve the highest accuracy in determining the driver's unfitness to drive, at least one of the above assessments can be performed. Any number of the assessments can be performed. In this case, the off-frame assessment, posture intrusion assessment, facial orientation intrusion assessment, shaking state assessment, and white-eye state assessment can be performed in that order.

[0172] For example, in the case where the determination of instability is combined with the determination of the shaking state, the determination of the shaking state can be carried out with high accuracy if the driver's inability to drive is not recorded in the determination of instability.

[0173] In the case where the determination of face direction intrusion is combined with the determination of the shaking state, if the driver's inability to drive cannot be detected during the determination of face direction intrusion, the driver's inability to drive can be detected with high accuracy by performing the shaking state determination.

[0174] In the case where the shaking state determination is combined with the white-eye state determination, if the driver's incapacity to drive is not detected by the shaking state determination, the driver's incapacity to drive can be detected with high accuracy by performing the white-eye state determination.

[0175] Learning section 51 can learn the driver's posture when it is incorrectly detected that the driver is no longer able to drive. That is, the posture the driver adopts when the incapacity to drive is detected, but the driver responds, can be learned. Then, the learned posture can be the posture that is not correctly identified as incapacitated to drive.

[0176] Statistical values ​​of the threshold and the target value can each be stored in the storage device 52 and used as initial values. The statistical values ​​of the threshold and the target value are, in each case, statistical values ​​of the threshold and the target value for the driver of the respective vehicles. The vehicle 10 can transmit each threshold and each target value set for the driver to an information center, and the information center can collect statistics on the values.

[0177] If the driver can be asked about their unfitness to drive and responds, it is recognized that the driver is capable of driving for a certain period of time after the response. The processing to detect the driver's unfitness to drive can be performed at a time interval set by the driver (for example, once every hour).

[0178] The external force exerted on the vehicle 10 can be detected by a component other than the G-sensor 44, for example the seat surface sensor 23.

[0179] According to a first aspect, a device for detecting a driver's incapacity to drive includes a head detection section which successively detects a head section higher than the driver's neck based on an image of a driver's seat taken by an imaging device mounted in a vehicle, and an out-of-frame state detection section which detects, during the vehicle's movement, when the head section detected by the head detection section is outside a predetermined area in the image, that the driver is no longer able to drive.

[0180] The device for detecting a driver's incapacity to drive sequentially captures a section of the driver's head based on a captured image of the driver's seat. When the driver is operating the vehicle normally, the driver's head section is usually within the predetermined area of ​​the image of the driver's seat. However, if the driver loses consciousness due to a sudden illness, the driver's head section may move outside the predetermined area. Therefore, the driver's incapacity to drive can be easily detected by recognizing that the driver is no longer able to drive when the head section moves outside the predetermined area.

[0181] According to a second aspect, a device for detecting a driver's incapacity to drive includes a head detection section which successively detects a head section higher than the driver's neck based on an image of a driver's seat taken by an imaging device mounted in a vehicle, a trajectory acquisition section which obtains a trajectory of the head section from positions of the head section detected by the head detection section, and a state detection section which detects, based on the trajectory obtained by the trajectory acquisition section, that the driver is no longer able to drive.

[0182] The device for detecting a driver's incapacity to drive sequentially captures the driver's head segment based on an image of the driver's seat and obtains a trajectory of the head segment from the positions captured. If the driver is unable to drive due to a sudden illness, the driver's head segment often shifts relative to its position while driving and does not return to that position. Therefore, a driver's incapacity to drive can be detected simply and quickly based on the trajectory of the driver's head segment.

[0183] Note that in the present application, a flowchart or the processing of the flowchart contains several steps (also referred to as sections), each designated, for example, as S10. Furthermore, each step can be subdivided into several substeps, and several steps can be combined into a single step.

Claims

[1] Device (100) for detecting a driver's inability to drive, comprising: a head detection section (61) which successively detects a head section higher than the driver's neck, based on an image of a driver's seat taken by an imaging device (21) mounted in a vehicle (10); and an out-of-frame state detection section (71) which detects during the journey of the vehicle (10) that the driver is no longer able to drive when the head section detected by the head detection section (61) is outside a predetermined area (FA) in the image, characterized by , that the device (100) also has a deformation section that deforms the predetermined area (FA) according to an intention of a user, which includes the driver, and The deformation section deforms the predetermined area (FA) according to a user input into an HMI (80) of the device (100). [2] Device (100) for detecting a driver's inability to drive, comprising: a head detection section (61) which successively detects a head section higher than the driver's neck, based on an image of a driver's seat taken by an imaging device (21) mounted in a vehicle (10); and an out-of-frame state detection section (71) which detects during the journey of the vehicle (10) that the driver is no longer able to drive when the head section detected by the head detection section (61) is outside a predetermined area (FA) in the image, characterized by , that the device (100) also has a transmission section that moves the predetermined area (FA) according to an intention of a user, which includes the driver, and the transmission section moves the predetermined area (FA) according to a user input into an HMI (80) of the device (100). [3] Device (100) for detecting a driver's inability to drive according to claim 1 or 2, wherein the driver information includes the driver's age, medical history and / or driving posture. [4] Device (100) for detecting a state of incapacity to drive of a driver according to any one of claims 1 to 3, wherein a shape of the predetermined area (FA) is an ellipse having a principal axis extending along a width of the vehicle (10). [5] Device (100) for detecting a driver's inability to drive, comprising: a head detection section (61) which successively detects a head section higher than the driver's neck, based on an image of a driver's seat taken by an imaging device (21) mounted in a vehicle (10); and an out-of-frame state detection section (71) which detects during the journey of the vehicle (10) that the driver is no longer able to drive when the head section detected by the head detection section (61) is outside a predetermined area (FA) in the image, characterized by , that The out-of-frame state detection section (71) detects that the driver is no longer able to drive if the head section was outside the predetermined area (FA) during an out-of-frame determination time (T0); at least one first predetermined area (FA1) and a second predetermined area (FA2) containing the first predetermined area (FA1) and an area outside the first predetermined area (FA1) are set as predetermined areas; a first out-of-frame determination time (T01) is set as the out-of-frame determination time (T0) for the first predetermined area (FA1), and a second out-of-frame determination time (T02) is set as the out-of-frame determination time (T0) for the second predetermined area (FA2); The out-of-frame state detection section (71) detects that the driver is no longer able to drive if the driver's head section was outside at least one of the predetermined areas during the out-of-frame determination time (T0) corresponding to the at least one predetermined area; and the first out-of-frame period (T01) is longer than the second out-of-frame period (T02). [6] Device (100) for detecting a state of incapacity to drive of a driver according to any one of claims 1 to 5, wherein the imaging device (21) is provided by a stereo camera; and The predetermined area (FA) is provided by a three-dimensionally defined area. [7] Device (100) for detecting a driver's inability to drive according to any one of claims 1 to 6, which further comprises: a trajectory acquisition section (62) that acquires a trajectory of the head section from positions of the head section that is acquired by the head detection section (61), wherein During the vehicle's journey (10), the out-of-frame state detection section (71) detects, based on the fact that the head section detected by the head detection section (61) is outside a predetermined area (FA) of the image, and based on the trajectory acquired by the trajectory acquisition section (62) until the head section has gone outside the predetermined area (FA), that the driver is no longer able to drive. [8] Device (100) for detecting a driver's inability to drive according to any one of claims 1 to 7, wherein the predetermined area (FA) is provided by an entire area of ​​the image. [9] Device (100) for detecting a driver's inability to drive according to claim 7, wherein if the head detection section (61) does not detect the head section, the head detection section (61) searches in a neighborhood of a last position of the trajectory obtained by the trajectory acquisition section (62). [10] Device (100) for detecting a driver's inability to drive according to claim 5, which further comprises: a size detection section (22) that detects a pull size of a seat belt of the driver's seat, wherein The out-of-frame condition detection section (71) detects that the driver is no longer able to drive when the pull size detected by the size detection section (22) is greater by a first pull size than a pull size that is present at the time of the seat belt being fitted. [11] Device (100) for detecting a driver's inability to drive according to claim 10, wherein the out-of-frame state detection section (71) detects that the driver is no longer able to drive when the pull size per size detection time is greater than a second pull size, wherein the pull size is detected by the size detection section (22). [12] Device (100) for detecting a driver's inability to drive according to one of claims 5, 10 and 11, which further comprises: a seat pressure sensing section (23) which detects a pressure distribution of a seat section (11a) of the driver's seat, wherein The out-of-frame condition detection section (71) detects that the driver is no longer able to drive when a high-pressure section in the pressure distribution detected by the seat pressure detection section (23) is biased towards one end of the seat section (11a). [13] Device (100) for detecting a driver's inability to drive according to one of claims 5 and 10 to 12, wherein if a position of the head section, which is successively detected by the head detection section (61), vibrates with an amplitude greater than a predetermined amplitude, the out-of-frame time (T0) is reduced. [14] Device (100) for detecting a driver's inability to drive according to one of claims 5 and 10 to 13, wherein if the speed of movement of the head section being successively detected by the head detection section (61) increases, the out-of-frame time (T0) is reduced compared to a set out-of-frame time (T0). [15] Device (100) for detecting a driver's inability to drive, comprising: a head detection section (61) which successively detects a head section higher than the driver's neck, based on an image of a driver's seat taken by an imaging device (21) mounted in a vehicle (10); and an out-of-frame state detection section (71) which detects during the journey of the vehicle (10) that the driver is no longer able to drive when the head section detected by the head detection section (61) is outside a predetermined area (FA) in the image, characterized by , that The out-of-frame state detection section (71) detects that the driver is no longer able to drive if the head section was outside the predetermined area (FA) during an out-of-frame determination time (T0); and when the movement speed of the head section increases, when a position of the head section, which is successively detected by the head detection section (61), approaches an end of the predetermined area (FA), the out-of-frame time (T0) is reduced compared to a set out-of-frame time (T0). [16] Device (100) for detecting a driver's inability to drive according to one of claims 5 and 10 to 15, which further comprises: a cramp detection section (63) that detects a cramp of the driver, wherein when the seizure detection section (63) detects the seizure, the out-of-frame determination time (T0) is reduced. [17] Device (100) for detecting a driver's inability to drive according to one of claims 5 and 10 to 16, wherein if the vehicle speed of the vehicle (10) is higher, the out-of-frame determination time (T0) is reduced. [18] Device (100) for detecting a driver's incapacity to drive according to any one of claims 5 and 10 to 17, wherein if the time to a collision obtained by sharing a distance to a vehicle ahead by a relative speed with respect to the vehicle ahead is smaller, the out-of-frame time (T0) is reduced. [19] Device (100) for detecting a driver's inability to drive, comprising: a head detection section (61) which successively detects a head section higher than the driver's neck, based on an image of a driver's seat taken by an imaging device (21) mounted in a vehicle (10); and an out-of-frame state detection section (71) which detects during the journey of the vehicle (10) that the driver is no longer able to drive when the head section detected by the head detection section (61) is outside a predetermined area (FA) in the image, characterized by , that The out-of-frame state detection section (71) detects that the driver is no longer able to drive if the head section was outside the predetermined area (FA) during an out-of-frame determination time (T0); and the device (100) also includes a memory section (52) that records personal information, including the driver's medical history and age; and The out-of-frame period (T0) is set on the basis of the personal information recorded in the storage section (52). [20] Device (100) for detecting a driver's inability to drive according to one of claims 5 and 10 to 19, wherein when a driving assistance control is carried out in the vehicle (10), the out-of-frame determination time (T0) is increased. [21] Device (100) for detecting a state of incapacity to drive of a driver according to any one of claims 1 to 20, wherein a state of incapacity to drive is provided by a condition in which the driver suddenly becomes ill.

Images