Visual recognition determination device and visual recognition determination device control method
The visibility determination device improves the accuracy of determining whether a driver sees objects by adjusting the line-of-sight cone area based on face orientation, vehicle speed, and turning, addressing the issue of decreased accuracy in existing systems.
Patent Information
- Application Number
- JP2023077075
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-05-09
- Publication Date
- 2024-11-21
AI Technical Summary
The accuracy of estimating the line of sight vector based on images captured by a driver monitor camera decreases due to variations in the driver's face orientation, leading to incorrect determination of objects being seen or not seen by the driver.
A visibility determination device that estimates the driver's line of sight vector using a driver monitor camera, determines the driver's eye point, and adjusts the line-of-sight cone area based on face orientation, vehicle speed, turning state, and object movement to improve accuracy.
Enhances the accuracy of determining whether the driver is visually recognizing objects by adjusting the line-of-sight cone area according to face orientation, vehicle speed, turning, and object movement, thereby improving visual recognition judgment.
Smart Images

Figure 2024161969000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to a visibility determination device and a control method for the visibility determination device. [Background technology]
[0002] Japanese Patent Application Publication No. 2014-120113 is known as a technical document related to a visibility determination device. This publication describes a device that detects the driver's line of sight, line of sight circle, and field of view from an image captured by a driver monitor camera, detects a feature ahead in the vehicle's traveling direction from an image captured by a front camera, and identifies the driver's visibility of the feature based on each detection result. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] JP 2014-120113 A Summary of the Invention [Problem to be solved by the invention]
[0004] However, the estimation accuracy of the gaze vector based on the image captured by the driver monitor camera decreases depending on the direction of the driver's face. Due to the decrease in the estimation accuracy of the gaze vector, there is a risk of erroneously determining that the driver is not looking at an object that the driver is actually looking at. [Means for solving the problem]
[0005] One aspect of the present invention is a visibility determination device that determines whether or not a driver is viewing an object detected by an external sensor of the vehicle, and includes a gaze cone area estimation unit that estimates the driver's gaze vector using the vehicle's driver monitor camera and estimates a gaze cone area that expands in a cone shape based on the driver's gaze vector the farther away from the driver's eyepoint, and a visibility determination unit that determines whether or not the driver is viewing the object from the overlap of the gaze cone area and the object, and the gaze cone area estimation unit estimates the gaze cone area depending on the direction of the driver's face.
[0006] According to one embodiment of the visibility assessment device of the present invention, since the estimation accuracy of the gaze vector based on the image captured by the driver monitor camera decreases depending on the driver's facial direction, the accuracy of the driver's visibility assessment of the target object can be improved by estimating the gaze cone area according to the facial direction.
[0007] In a visibility assessment device according to one aspect of the present invention, the gaze cone area estimation unit may increase the expansion rate of the gaze cone area depending on the distance from the eye point when the driver's face is directed away from the front of the vehicle, compared to when the driver's face is directed forward. According to this visibility assessment device, when the driver's face is directed away from the front of the vehicle, the estimation accuracy of the gaze vector based on the image captured by the driver monitor camera decreases, so the accuracy of the driver's visibility assessment of the target object can be improved by increasing the magnification rate of the gaze cone area.
[0008] In one embodiment of the visibility assessment device of the present invention, the gaze cone area estimation unit may estimate a gaze cone area whose cross section perpendicular to the gaze vector is a vertically elongated ellipse when the driver's face is facing forward toward the vehicle, and may estimate a gaze cone area whose cross section perpendicular to the gaze vector is a horizontally elongated ellipse when the driver's face is facing laterally away from the vehicle's front. According to this visibility assessment device, when the driver's face is facing the front of the vehicle, the estimation accuracy of the line of sight vector in the pitch direction of the vehicle is low, so a line of sight cone area with a vertically elongated elliptical cross section is estimated. Also, when the driver's face is facing away from the front of the vehicle to the side, the estimation accuracy of the line of sight vector in the yaw direction of the vehicle is low, so a line of sight cone area with a horizontally elongated elliptical cross section is estimated. As a result, this visibility assessment device can improve the accuracy of the driver's visibility assessment of the object.
[0009] In a visibility determination device according to one embodiment of the present invention, the gaze cone area estimation unit may reduce the expansion rate of the gaze cone area with respect to the distance from the eye point when the vehicle speed is equal to or greater than a vehicle speed determination threshold, compared to when the vehicle speed is less than the vehicle speed determination threshold. According to this visibility determination device, since the effective viewing angle of the driver narrows when the vehicle speed is high, the expansion rate of the line of sight cone area is reduced when the vehicle speed is equal to or higher than the vehicle speed determination threshold, thereby improving the accuracy of the visibility determination device for the driver's visibility of the object.
[0010] In a visibility assessment device according to one aspect of the present invention, the gaze cone area estimation unit may expand a cross-section perpendicular to the gaze vector of the gaze cone area in the turning direction of the vehicle when the vehicle is turning, compared to when the vehicle is not turning. According to this visibility assessment device, since it can be assumed that when the vehicle is turning, the driver will be consciously looking at the area ahead, by expanding the cross-section perpendicular to the line of sight vector of the line of sight cone area in the direction of turning the vehicle, the accuracy of the driver's visibility assessment of the target object can be improved.
[0011] In a visibility assessment device according to one embodiment of the present invention, the gaze cone area estimation unit may estimate the gaze cone area so that a cross section perpendicular to the gaze vector expands in the direction of movement of the object as seen by the driver when the movement speed of the object as seen by the driver is equal to or greater than a movement speed threshold. According to this visibility assessment device, when the moving speed of an object seen by the driver is fast, the driver will attempt to view the object while taking its movement into account. Therefore, by estimating the line of sight cone area so that the cross section perpendicular to the line of sight vector expands in the direction of the object's movement, the accuracy of the driver's visibility assessment of the object can be improved.
[0012] In a visibility assessment device according to one embodiment of the present invention, the gaze cone area estimation unit may increase the expansion rate of the gaze cone area after determining that the driver is viewing an object, until determining that the driver is not viewing the object, compared to before the driver viewed the object. According to this visibility assessment device, since it is believed that there is a high probability that the driver will still be viewing an object even if he or she slightly shifts their line of sight from an object that they have once viewed, the accuracy of the driver's visibility assessment of an object can be improved by increasing the expansion rate of the line of sight cone area.
[0013] Another aspect of the present invention is a control method for a visibility determination device that determines whether or not a driver is viewing an object detected by an external sensor of the vehicle, in which the driver's gaze vector is estimated using the vehicle's driver monitor camera, and a gaze cone area that expands in a cone shape based on the driver's gaze vector the further away from the driver's eyepoint is, the gaze cone area is estimated according to the direction of the driver's face, and whether or not the driver is viewing the object is determined from the overlap between the gaze cone area and the object.
[0014] According to a control method for a visibility assessment device of another aspect of the present invention, since the estimation accuracy of the gaze vector based on the image captured by the driver monitor camera decreases depending on the direction of the driver's face, the accuracy of the driver's visibility assessment of the target object can be improved by estimating the size of the gaze cone area according to the direction of the driver's face. Effect of the Invention
[0015] According to each aspect of the present invention, it is possible to improve the accuracy of determining whether a driver is able to see an object. [Brief description of the drawings]
[0016] [Figure 1] 1 is a block diagram showing a visibility determination device according to an embodiment; [Diagram 2] FIG. 2 is a plan view for explaining a line-of-sight vector of a driver. [Diagram 3] 1A is a diagram for explaining a situation in which the driver is visually recognizing an object, and FIG. 1B is a diagram for explaining a situation in which the driver is not visually recognizing an object. [Figure 4] 1A is a diagram showing an example of a cross section of a gaze cone region, and FIG. 1B is a diagram showing an example of a cross section of a gaze cone region when the magnification is increased. [Diagram 5] 1A is a diagram showing an example of a vertically elongated cross section of a gaze cone region, and FIG. 1B is a diagram showing an example of a horizontally elongated cross section of a gaze cone region. [Figure 6] 13 is a flowchart illustrating an example of a visual recognition determination process. [Figure 7] 10A and 10B are flowcharts illustrating an example of a line-of-sight cone region estimation process according to a vehicle speed and a turning of a vehicle, respectively; [Figure 8] 10A is a flowchart showing an example of a gaze cone region estimation process according to a moving speed of an object, and FIG. 10B is a flowchart showing an example of a gaze cone region estimation process according to a visibility state of a driver. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0017] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
[0018] Fig. 1 is a block diagram showing a visibility determination device 100 according to an embodiment. The visibility determination device 100 shown in Fig. 1 is a device mounted on a vehicle such as a passenger car or a freight vehicle, and determines whether or not a driver of the vehicle is visually recognizing an object outside the vehicle. Even if a vehicle is equipped with an advanced driving assistance function, the driver has an obligation to monitor the surroundings, and is required to determine whether the driver is properly visually recognizing an object.
[0019] [Configuration of the visibility determination device] The configuration of the visual identification determination device 100 according to this embodiment will be described below. As shown in FIG. 1, the visual identification determination device 100 includes a visual identification determination ECU [Electronic Control Unit] 10 that manages the device in an integrated manner. The visual identification determination ECU 10 is an electronic control unit having a CPU [Central Processing Unit] and a storage unit such as a ROM [Read Only Memory] or a RAM [Random Access Memory]. In the visual identification determination ECU 10, various functions are realized, for example, by the CPU executing a program stored in the storage unit. The visual identification determination ECU 10 may be composed of multiple electronic units.
[0020] The visual confirmation determination ECU 10 is connected to a driver monitor camera 1 , an external sensor 2 , an internal sensor 3 , and an HMI [Human Machine Interface] 4 .
[0021] The driver monitor camera 1 is a camera for capturing an image of a driver of a vehicle. The driver monitor camera 1 can be a digital camera having an image sensor such as a CCD [Charge Coupled Device] or a CIS [CMOS Image Sensor]. The driver monitor camera 1 is provided, for example, on a steering column cover in front of the driver's seat, and captures an image of the driver's head at a predetermined frame rate. The driver monitor camera 1 may be embedded in an instrument panel in front of the driver's seat, or may be provided on the upper part of the instrument panel. The driver monitor camera 1 may be composed of multiple cameras. The driver monitor camera 1 transmits the driver captured image to the visibility determination ECU 10.
[0022] The external sensor 2 is a detection device that detects the situation around the vehicle. The external sensor 2 includes at least one of a camera and a radar sensor. The camera is an imaging device that captures the situation outside the vehicle. The camera is provided, for example, on the back side of the windshield of the vehicle and captures an image in front of the vehicle. The camera may include multiple imaging devices that capture images in multiple directions around the vehicle. The camera transmits imaging information related to the situation outside the vehicle to the visibility determination ECU 10.
[0023] A radar sensor is a detection device that detects objects around a vehicle using radio waves (e.g., millimeter waves) or light. Radar sensors include, for example, millimeter wave radar or LIDAR (Light Detection and Ranging). The radar sensor may include multiple detection devices to detect objects in multiple directions around the vehicle. The radar sensor transmits information about the detected objects to the visibility determination ECU 10.
[0024] The internal sensor 3 is a detection device that detects the running state of the vehicle. The internal sensor 3 includes a vehicle speed sensor and a yaw rate sensor. The vehicle speed sensor is a detector that detects the speed of the vehicle. The vehicle speed sensor transmits the detected vehicle speed information (wheel speed information) to the visual confirmation determination ECU 10.
[0025] The yaw rate sensor is a detector that detects the yaw rate (rotational angular velocity) around the vertical axis of the center of gravity of the vehicle. For example, a gyro sensor can be used as the yaw rate sensor. The yaw rate sensor transmits the detected yaw rate information of the vehicle to the visual recognition determination ECU 10.
[0026] The HMI 4 is an interface for inputting and outputting information between the visibility determination ECU 10 and the driver. The HMI 4 includes, for example, a display, a speaker, and the like provided in the vehicle cabin. The HMI 4 outputs images on the display and sounds from the speaker in response to a control signal from the visibility determination ECU 10. The display may be a MID [Multi Information Display], a navigation system display, or a HUD [Head Up Display]. The HMI 4 may include various indicators.
[0027] Next, a description will be given of the functional configuration of the visual identification determination ECU 10. The visual identification determination ECU 10 has an object recognition unit 11, a line of sight cone region estimation unit 12, and a visual identification determination unit 13. Note that some of the functions of the visual identification determination ECU 10 described below may be executed by a server capable of communicating with the vehicle.
[0028] The object recognition unit 11 recognizes objects outside the vehicle based on the detection results of the external sensor 2. An object is an object that is visually recognized by the driver of the vehicle. The object includes moving objects such as other vehicles, bicycles, personal mobility, and pedestrians. The object may include stationary objects such as traffic lights, road signs, road guide boards, and road markings. The object may be limited to those located within a distance visible to the driver. The object may be limited to those located in front of the vehicle.
[0029] The object recognition unit 11 recognizes the relative position of the object with respect to the vehicle and the relative speed of the object with respect to the vehicle. The object recognition unit 11 may recognize the object by sensor fusion that combines the captured image of the camera in the external sensor 2 and the detection result of the radar sensor. Since well-known technology can be used for recognizing the object, detailed explanation will be omitted.
[0030] The object recognition unit 11 may recognize the position of an object as seen by the driver, based on the detection result of the external sensor 2. The position of an object as seen by the driver means the position of the object as seen from the driver's eye point.
[0031] The driver's eye point is a position where the eyes of a driver seated in the driver's seat are assumed to be present in the vehicle cabin. The driver's eye point may be expressed as coordinates in a three-dimensional space. The driver's eye point is set, for example, based on the vehicle body. The driver's eye point may be a fixed position or an adjustable position.
[0032] The driver's eye point may be adjusted based on the position of the driver's eyes recognized from the driver's captured image captured by the driver monitor camera 1. The driver's eye point may be adjusted according to the position of the driver's seat (e.g., the front-to-rear position of the seat cushion). The driver's eye point may be adjusted according to the inclination of the seat back of the driver's seat. The driver's eye point may be adjusted based on personal information such as the driver's height and sitting height shared with the visibility determination device 100. The driver's eye point can be adjusted using other well-known techniques.
[0033] The object recognition unit 11 determines the position of the object as seen by the driver, for example, based on the recognition result of the object by sensor fusion. The object recognition unit 11 may determine the position of the object as seen by the driver by performing viewpoint conversion processing on the image captured by the camera of the external sensor 2 so that it becomes an image as seen from the driver's eye point. The object recognition unit 11 calculates the moving speed of the object as seen by the driver based on the change over time in the position of the object as seen by the driver. The object recognition unit 11 also recognizes the size of the object as seen by the driver.
[0034] The object recognition unit 11 may recognize the type of object. The type of object is a four-wheeled vehicle, a two-wheeled vehicle, a pedestrian, a traffic light, a road sign, or the like. The object recognition unit 11 recognizes the type of object by, for example, pattern matching. The object recognition unit 11 may recognize the type of object from the captured image of the camera of the external sensor 2 by matching using image patterns prepared in advance for each type.
[0035] Note that various types of recognition of objects may be performed using a machine learning model. The machine learning model is, for example, a neural network such as a convolutional neural network (CNN). The neural network may include multiple layers including multiple convolution layers and pooling layers. A deep learning network using deep learning may be used as the neural network.
[0036] The gaze cone area estimation unit 12 estimates the gaze vector of the driver based on the image of the driver captured by the driver monitor camera 1 of the vehicle, and estimates the gaze cone area based on the gaze vector of the driver. The gaze cone area estimation unit 12 also recognizes the face direction of the driver.
[0037] The driver's gaze vector is a vector in three-dimensional space that corresponds to the driver's gaze, extending in a straight line starting from the driver's eye point. For example, a gaze vector can be estimated using the face direction and the center position of the pupil measured by a corneal reflex method. A machine learning model may be used to estimate the gaze vector. Since the estimation of the gaze vector is a well-known technique, a detailed description will be omitted. Note that the gaze vector does not necessarily need to include a length, and can be used in the sense of the gaze direction.
[0038] Here, Fig. 2 is a plan view for explaining the line of sight vector of the driver. Fig. 2 shows a vehicle M, a road R on which the vehicle M is traveling, a pedestrian 50 (target object), and a line of sight vector EB. The pedestrian 50 is about to cross the crosswalk in front of the vehicle M. The traveling direction of the pedestrian 50 is indicated by an arrow P. As shown in Fig. 2, the line of sight vector EB corresponds to the line of sight of the driver of the vehicle M toward the pedestrian 50, who is the target object.
[0039] FIG. 3(a) is a diagram showing an example of a line of sight cone area. FIG. 3(a) shows a line of sight cone area EC that starts from the driver's eye point EP and spreads from a line of sight vector EB. FIG. 3(a) also shows a bounding box 60 used when determining the driver's visibility of a pedestrian 50, which is an object. The bounding box 60 is set in a three-dimensional space, and is set as, for example, a rectangular box that encloses the pedestrian 50 with a minimum volume. Note that setting of the bounding box 60 is not essential. Note that FIG. 3(b) is a diagram for explaining a situation in which the driver does not view an object. The situation in FIG. 3(b) will be described later.
[0040] As shown in FIG. 3(a), the gaze cone area estimation unit 12 estimates the gaze cone area EC based on the gaze vector EB extending from the driver's eye point EP. The gaze cone area EC is a cone-shaped area that is expanded from the gaze vector EB of the driver. Note that the gaze cone area EC does not need to be strictly conical. The cross-sectional shape of the gaze cone area EC is not limited to a circular or elliptical shape, and may be a shape close to a circle or ellipse overall.
[0041] The gaze cone region estimation unit 12 may estimate a gaze cone region of a different size for each object. When multiple objects exist at the same time, the gaze cone region estimation unit 12 can estimate multiple gaze cone regions used to determine visibility of each object with respect to the driver's gaze vector.
[0042] The gaze cone area EC may be estimated taking into consideration the driver's effective viewing angle. The effective viewing angle is located in the center of a person's viewing field, and is an area where objects can be recognized with high resolution. The effective viewing angle (the viewing angle of the effective viewing field) varies from person to person, but may be, for example, 20°. Outside the effective viewing field, there is a wide range of peripheral vision. The peripheral vision is an area where objects can be recognized with low resolution. The gaze cone area estimation unit 12 may use the effective viewing angle as the apex angle of the gaze cone area EC. Note that the gaze cone area estimation unit 12 may estimate a value unrelated to the effective viewing angle as the apex angle of the gaze cone area EC.
[0043] The gaze cone area estimation unit 12 can estimate the gaze cone area EC in consideration of the estimation accuracy of the gaze vector EB. The estimation accuracy of the gaze vector EB is significantly affected by the direction of the driver's face.
[0044] The gaze cone area estimation unit 12 estimates the gaze cone area EC according to the face direction of the driver. The gaze cone area estimation unit 12 recognizes the face direction of the driver based on the image of the driver captured by the driver monitor camera 1. Since the driver's face direction can be recognized by using a well-known technique, a detailed description will be omitted.
[0045] Specifically, when the driver's face is directed away from the front of the vehicle M, the gaze cone area estimation unit 12 increases the expansion rate of the gaze cone area EC according to the distance from the eye point EP compared to when the driver's face is directed toward the front of the vehicle M.
[0046] The case where the driver's face is facing forward of the vehicle M is, for example, the case where the driver's face is within a set angle range with respect to an axis parallel to the longitudinal axis of the vehicle M passing through the eye point EP. The set angle range corresponds to, for example, a range in which the driver monitor camera 1 can estimate the line of sight vector with high accuracy. The set angle range may be changed according to the specifications of the driver monitor camera 1, or may be changed according to the positional relationship between the driver monitor camera 1 and the eye point EP. The case where the driver's face is deviated from the forward direction of the vehicle M is the case where the driver's face is outside the set angle range. Note that the direction in which the driver's face is deviated is not limited to the horizontal direction, and may include the vertical direction.
[0047] Here, FIG. 4(a) is a diagram showing an example of a cross section of the gaze cone area EC. FIG. 4(b) is a diagram showing an example of a cross section of the gaze cone area EC when the magnification is increased. The cross section of the gaze cone area EC in FIG. 4(b) is a larger circle than the cross section in FIG. 4(a). When the driver's face is facing forward of the vehicle M, the gaze cone area estimation unit 12 estimates the gaze cone area EC having a size shown in FIG. 4(a). When the driver's face is deviated from the forward direction of the vehicle M, the gaze cone area estimation unit 12 estimates the gaze cone area EC having a size larger than that of FIG. 4(a) as shown in FIG. 4(b).
[0048] Since the gaze cone area estimation unit 12 is likely to have a lower accuracy in estimating the gaze vector using the image captured by the driver monitor camera 1 the more the driver's face deviates, the accuracy of the driver's visibility judgment of the object can be improved by increasing the magnification rate of the gaze cone area EC when the driver's face deviates from the forward direction of the vehicle M.
[0049] Furthermore, the gaze cone area estimation unit 12 may estimate a gaze cone area EC whose cross section perpendicular to the gaze vector EB is a vertically long ellipse when the driver's face is facing forward of the vehicle M, and may estimate a gaze cone area whose cross section is horizontally long ellipse when the driver's face is facing laterally away from the forward direction of the vehicle M. "Vertical" means that the vertical (up-down) length is longer than the horizontal length of the cross section perpendicular to the gaze vector EB. Similarly, "horizontal" means that the horizontal length is longer than the vertical length of the cross section perpendicular to the gaze vector EB.
[0050] Here, Fig. 5(a) is a diagram showing an example of a vertically long cross section of the gaze cone area EC. Fig. 5(b) is a diagram showing an example of a horizontally long cross section of the gaze cone area EC. When the driver's face is facing forward of the vehicle M, the gaze cone area estimation unit 12 estimates the gaze cone area EC whose cross section is a vertically long ellipse as shown in Fig. 5(a). When the driver's face is deviated to the side from the forward direction of the vehicle M, the gaze cone area estimation unit 12 estimates the gaze cone area whose cross section is a horizontally long ellipse as shown in Fig. 5(b).
[0051] The gaze cone area estimation unit 12 can avoid erroneously determining that the driver does not see the object by making the cross section of the gaze cone area EC a vertically long ellipse, since the estimation accuracy of the gaze vector in the pitch direction of the vehicle M is likely to be low when the driver's face is facing forward of the vehicle M. Similarly, the gaze cone area estimation unit 12 can improve the accuracy of the driver's visibility determination of the object by making the cross section of the gaze cone area EC a horizontally long ellipse, since the estimation accuracy of the gaze vector in the yaw direction of the vehicle M is likely to be low when the driver's face is deviated to the lateral direction of the vehicle M.
[0052] The estimation accuracy of the gaze vector EB varies depending on the driver's face direction as described above, as well as individual differences such as the lighting conditions that illuminate the driver's face, glasses, makeup, and facial expression. For example, the estimation accuracy of the gaze vector EB decreases when strong light from outside hits the driver's face and whiteout occurs around the driver's eyes. The estimation accuracy of the gaze vector EB also decreases when only the area around the driver's eyes is in shadow due to the way the light from outside hits the driver's face.
[0053] The estimation accuracy of the gaze vector EB may be lower when the driver wears glasses than when the driver does not wear glasses. The estimation accuracy of the gaze vector EB also varies depending on the shape of the rim of the glasses. The estimation accuracy of the gaze vector EB may be lowered by makeup such as eyeliner or false eyelashes when the driver wears makeup around the eyes. The estimation accuracy of the gaze vector EB may be lowered when the driver's eyelids tend to droop due to old age or other individual differences. When there is a factor that reduces the estimation accuracy of the gaze vector EB as described above, the gaze cone area estimation unit 12 may increase the enlargement rate of the gaze cone area EC compared to when there is no factor that reduces the estimation accuracy.
[0054] Furthermore, the gaze cone area estimation unit 12 may estimate the gaze cone area EC according to the vehicle speed of the vehicle M. It is known that the driver's effective viewing angle changes depending on the vehicle speed. Specifically, when the vehicle speed of the vehicle M is equal to or higher than the vehicle speed determination threshold, the gaze cone area estimation unit 12 may reduce the expansion rate of the gaze cone area EC according to the distance from the eye point EP compared to when the vehicle speed is less than the vehicle speed determination threshold. The vehicle speed determination threshold is a threshold value set in advance. The vehicle speed determination threshold may be 40 km / h, 50 km / h, or 60 km / h. The value of the vehicle speed determination threshold is not particularly limited.
[0055] Since the driver's effective field of view angle narrows when the vehicle speed is high, the gaze cone area estimation unit 12 can improve the accuracy of the driver's visibility judgment of the target object by reducing the expansion rate of the gaze cone area EC when the vehicle speed is equal to or greater than the vehicle speed judgment threshold.
[0056] The gaze cone region estimation unit 12 may set the effective viewing angle to a fixed value common to all drivers, or may set a value for each driver individually if the driver is personally authenticated. The value of the individual effective viewing angle of the driver may be set based on the driver's age, height, sex, etc., for example. The gaze cone region estimation unit 12 may perform a driver inspection regarding the effective viewing angle by projecting a virtual object on the windshield at the time of driver registration and having the driver move his / her line of sight within a range that the driver considers to be visible. The gaze cone region estimation unit 12 may set the value of the individual effective viewing angle of the driver based on the result of the driver inspection regarding the effective viewing angle.
[0057] Furthermore, the gaze cone region estimation unit 12 may estimate the gaze cone region EC depending on the traveling state of the vehicle M. Specifically, when the vehicle M is turning, the gaze cone region estimation unit 12 may estimate the gaze cone region such that a cross section perpendicular to the line of sight vector EB expands in the turning direction (rightward or leftward) of the vehicle M compared to when the vehicle M is not turning.
[0058] The line of sight cone region estimation unit 12 determines whether or not the vehicle M is turning based on, for example, yaw rate information detected by a yaw rate sensor of the internal sensor 3. When the yaw rate of the vehicle M is equal to or greater than a turning threshold, the line of sight cone region estimation unit 12 determines that the vehicle M is turning. The line of sight cone region estimation unit 12 may determine whether or not the vehicle M is turning based on the steering angle of the vehicle M. When the steering angle of the vehicle M is equal to or greater than a turning steering angle threshold, the line of sight cone region estimation unit 12 determines that the vehicle M is turning. The turning threshold and the turning steering angle threshold are thresholds of preset values.
[0059] When the vehicle M is turning, the gaze cone region estimation unit 12 estimates the gaze cone region EC so that the cross section expands in the turning direction of the vehicle M. Explaining with reference to Fig. 5(b), when the turning direction coincides with the arrow P shown in Fig. 5(b), for example, the gaze cone region estimation unit 12 estimates the gaze cone region EC whose cross section expands in the turning direction as shown in Fig. 5(b). At this time, the vertical width of the gaze cone region EC may be the same as when the vehicle M is not turning, or may be shorter than when the vehicle M is not turning.
[0060] Since it can be assumed that when the vehicle is turning, the driver looks consciously at the area ahead, the gaze cone area estimation unit 12 can improve the accuracy of the driver's visibility judgment of the target object by expanding the cross section perpendicular to the gaze vector of the gaze cone area in the vehicle turning direction.
[0061] Furthermore, when the moving speed of the object as seen by the driver is equal to or greater than the moving speed threshold, the gaze cone area estimation unit 12 may estimate the gaze cone area EC so that the cross section expands in the moving direction of the object as seen by the driver. The moving speed of the object as seen by the driver is recognized by the above-mentioned object recognition unit 11. The moving speed threshold is a threshold value set in advance.
[0062] Explaining with reference to Fig. 5(b), when the moving speed of the pedestrian 50 is equal to or higher than the moving speed threshold, the gaze cone region estimation unit 12 estimates a gaze cone region EC whose cross section expands in the traveling direction P of the pedestrian 50 as shown in Fig. 5(b). Since it is considered that the driver tries to view an object moving at a fast speed while taking into consideration its movement, the gaze cone region estimation unit 12 estimates the gaze cone region such that the cross section expands in the moving direction of the object, thereby making it possible to improve the accuracy of the driver's visibility judgment of the object.
[0063] In addition, the gaze cone region estimation unit 12 may increase the expansion rate of the gaze cone region EC after it is determined that the driver is viewing the object until it is determined that the driver is not viewing the object, compared to before the driver viewed the object. The viewing determination of the object is performed by the viewing determination unit 13, which will be described later.
[0064] Explaining with reference to Figures 4(a) and 4(b), for example, after it is determined that the driver shown in Figure 4(a) is viewing the pedestrian 50, the gaze cone region estimation unit 12 increases the expansion rate of the gaze cone region EC as shown in Figure 4(b) compared to Figure 4(a) before the driver views the object, during the period until it is determined that the driver does not view the pedestrian 50. Since it is considered that there is a high possibility that the driver continues to view the object even if the driver slightly shifts his / her gaze from the object once viewed, the gaze cone region estimation unit 12 can improve the accuracy of the driver's visibility determination of the object by increasing the expansion rate of the gaze cone region.
[0065] Furthermore, the gaze cone area estimation unit 12 may estimate the gaze cone area EC in consideration of the recognition accuracy of the object (the recognition accuracy of the object's position and the object's size). The recognition accuracy of the object varies depending on the detection error of the external sensor 2, the relative speed between the vehicle M and the object, etc. A well-known technique can be used to calculate the detection error of the external sensor 2, so a detailed description will be omitted.
[0066] When the detection error of the external sensor 2 is equal to or greater than the error threshold, the gaze cone area estimation unit 12 may increase the magnification rate of the gaze cone area EC since there is a possibility that the recognition accuracy of the target object is reduced compared to when the detection error is less than the error threshold.
[0067] When the relative speed between the vehicle M and the object is equal to or greater than the relative speed threshold, the gaze cone area estimation unit 12 may increase the magnification rate of the gaze cone area EC because the recognition accuracy of the object may be lower than when the relative speed between the vehicle M and the object is less than the relative speed threshold. The error threshold and the relative speed threshold are thresholds of preset values.
[0068] When a size misrecognition occurs in which the size of the object is mistakenly recognized as being larger than it actually is, the gaze cone area estimation unit 12 may reduce the magnification rate of the gaze cone area EC compared to when no size misrecognition occurs. The occurrence of a size misrecognition can be determined from the detection error of the external sensor 2 or information on the size estimation error included in the fusion information. The occurrence of a size misrecognition can be determined by a well-known technique.
[0069] The gaze cone region estimation unit 12 may also estimate the gaze cone region EC according to the position of the object relative to the vehicle M. Specifically, the gaze cone region estimation unit 12 may estimate a gaze cone region EC whose cross section is longer horizontally as the position of the object deviates laterally from the forward direction of the vehicle M. The gaze cone region estimation unit 12 may also estimate the gaze cone region EC according to the type of object. Specifically, the gaze cone region estimation unit 12 may reduce the magnification rate of the gaze cone region EC when the type of object is a two-wheeled vehicle, since the object is more likely to be overlooked than when the type of object is a four-wheeled vehicle. The reduction in the magnification rate of the gaze cone region EC can be applied to objects that are easy to overlook, such as pedestrians, in addition to two-wheeled vehicles.
[0070] The visibility determination unit 13 determines whether or not the driver is viewing the object based on the object recognized by the object recognition unit 11 and the gaze cone area EC of the driver estimated by the gaze cone area estimation unit 12. The visibility determination unit 13 determines whether or not the driver is viewing the object based on, for example, the overlap between the gaze cone area EC and the object in a coordinate system of a three-dimensional space based on the vehicle M or the driver's eye point EP.
[0071] For example, the visibility determination unit 13 determines that the driver is viewing the object when the overlap rate between the line of sight cone area EC and the object is equal to or greater than the visibility determination threshold (see FIG. 3(a)). The visibility determination unit 13 determines that the driver is not viewing the object when the overlap rate between the line of sight cone area EC and the object is less than the visibility determination threshold (see FIG. 3(b)). The visibility determination threshold is a threshold value that is set in advance. The visibility determination threshold may be 30%, 40%, or 50%. The value of the visibility determination threshold is not particularly limited.
[0072] The overlap ratio is the ratio of the object to the line of sight cone region EC. The overlap ratio may be calculated using the object itself or a bounding box 60 as shown in Figs. 3 and 4.
[0073] The visibility determination unit 13 may determine whether the driver is viewing the object based on the overlap between the line of sight cone area EC and the object in the plane coordinate system. The visibility determination unit 13 may project the object and the line of sight cone area EC onto a captured image that has been viewpoint converted so that the driver's eye point EP is the reference point, and calculate the ratio of the object to the range of the line of sight cone area EC on the captured image as the overlap ratio. The visibility determination unit 13 may calculate the overlap ratio between the object and the line of sight cone area EC using other well-known techniques. The visibility determination unit 13 determines that the driver is viewing the object when the overlap ratio between the line of sight cone area EC and the object is equal to or greater than a visibility determination threshold.
[0074] The visibility determination unit 13 may use a machine learning model to determine whether the driver is viewing an object from the overlap between the line of sight cone area EC and the object in a three-dimensional coordinate system or a planar coordinate system.
[0075] The visibility determination unit 13 may determine that the driver is viewing the object when the overlap rate between the line of sight cone area EC and the object continues to be equal to or greater than the visibility determination threshold for a certain period of time. The visibility determination unit 13 may change the certain period of time to a longer period during nighttime compared to daytime. The certain period of time may be set to a different period of time depending on the type of object.
[0076] The visibility determination unit 13 may call the driver's attention or provide information based on the visibility determination result. The visibility determination unit 13 executes the call to the driver or the provision of information by transmitting a control signal to the HMI 4. The visibility determination unit 13 calls the driver's attention to the presence of an object that the driver does not visually recognize, for example, by outputting sound from a speaker of the HMI 4 or displaying an image on a display. The visibility determination unit 13 may provide information on an object that the driver visually recognizes, by outputting sound from a speaker or displaying an image on a display. In addition, the visibility determination unit 13 may transmit the driver's visibility determination result of the object to a driving assistance system or an automatic driving system of the vehicle M.
[0077] [Method of controlling the visibility determination device] Next, a control method of the visual confirmation determination device 100 according to this embodiment will be described with reference to the drawings. Fig. 6 is a flowchart showing an example of a visual confirmation determination process. The visual confirmation determination process is executed, for example, when the ignition of the vehicle M is in an on state.
[0078] As shown in Fig. 6, the visibility determination ECU 10 of the visibility determination device 100 determines whether or not an object outside the vehicle M has been recognized by the object recognition unit 11 in S10. The object recognition unit 11 recognizes the object by, for example, sensor fusion. If it is determined that the object has been recognized (S10: YES), the visibility determination ECU 10 proceeds to S11. If it is determined that the object has not been recognized (S10: NO), the visibility determination ECU 10 ends the visibility determination process. After a certain period of time has elapsed, the visibility determination ECU 10 repeats the process from S10.
[0079] In S11, the visual confirmation determination ECU 10 recognizes the facial direction of the driver by the gaze cone region estimation unit 12. The gaze cone region estimation unit 12 recognizes the facial direction of the driver based on an image of the driver captured by the driver monitor camera 1. After that, the visual confirmation determination ECU 10 proceeds to S12.
[0080] In S12, the visual confirmation determination ECU 10 estimates the driver's line of sight vector EB by the line of sight cone region estimation unit 12. The line of sight cone region estimation unit 12 estimates the driver's line of sight vector EB extending from the driver's eye point EP based on the driver's captured image by the driver monitor camera 1. After that, the visual confirmation determination ECU 10 proceeds to S13. Note that the recognition of the driver's face direction and the estimation of the driver's line of sight vector EB may be performed simultaneously.
[0081] In S13, the visual confirmation determination ECU 10 determines, by the line of sight cone area estimation unit 12, whether or not the driver's facial direction deviates from the forward direction of the vehicle M. If the visual confirmation determination ECU 10 determines that the driver's facial direction deviates from the forward direction of the vehicle M (S13: YES), the visual confirmation determination ECU 10 proceeds to S14. If the visual confirmation determination ECU 10 does not determine that the driver's facial direction deviates from the forward direction of the vehicle M (S13: NO), the visual confirmation determination ECU 10 proceeds to S15.
[0082] In S14, the visual recognition determination ECU 10 uses the gaze cone area estimation unit 12 to estimate a gaze cone area EC that is enlarged compared to the case of S15 described later. The enlarged gaze cone area EC means that the magnification rate of the gaze cone area EC according to the distance from the eye point EP is increased. The gaze cone area estimation unit 12 sets the gaze cone area EC that spreads in a cone shape based on the gaze vector EB. The gaze cone area estimation unit 12 may estimate a gaze cone area EC with a horizontally elongated cross section as shown in FIG. 5(b). After that, the visual recognition determination ECU 10 proceeds to S16.
[0083] In S15, the visual confirmation determination ECU 10 estimates the visual cone area EC by the visual cone area estimation unit 12. Here, the visual cone area EC is not enlarged according to the face direction of the driver. The visual cone area estimation unit 12 may estimate a visual cone area EC having a vertically elongated cross section as shown in FIG. 5(a). After that, the visual confirmation determination ECU 10 proceeds to S16.
[0084] In S16, the visibility determination ECU 10 determines whether the overlap ratio between the line of sight cone region EC and the object is equal to or greater than the visibility determination threshold by the visibility determination unit 13. If the visibility determination ECU 10 determines that the overlap ratio between the line of sight cone region EC and the object is equal to or greater than the visibility determination threshold (S16: YES), the visibility determination ECU 10 proceeds to S17. If the visibility determination ECU 10 does not determine that the overlap ratio between the line of sight cone region EC and the object is equal to or greater than the visibility determination threshold (S16: NO), the visibility determination ECU 10 proceeds to S18.
[0085] In S17, the visual confirmation determination ECU 10 determines that the driver is visually recognizing the object by the visual confirmation determination unit 13. The visual confirmation determination unit 13 transmits the visual confirmation determination result of the driver to the driving assistance system or the automatic driving system of the vehicle M. Thereafter, the visual confirmation determination ECU 10 ends the visual confirmation determination process, and repeats the process from S10 after a certain period of time has elapsed.
[0086] In S18, the visual confirmation determination ECU 10 determines that the driver does not visually confirm the object by the visual confirmation determination unit 13. The visual confirmation determination unit 13 transmits the driver's visual confirmation determination result to the driving assistance system or the automatic driving system of the vehicle M. Thereafter, the visual confirmation determination ECU 10 ends the visual confirmation determination process, and repeats the process from S10 after a certain period of time has elapsed.
[0087] The visibility determination ECU 10 may estimate the gaze cone area EC according to various conditions other than the driver's face direction. Fig. 7(a) is a flowchart showing an example of a gaze cone area estimation process according to a vehicle speed. The gaze cone area estimation process can be executed between S12 and S13 in the flowchart of Fig. 6, for example. The gaze cone area estimation process may be executed between S14 or S15 and S16. The same applies to various gaze cone area estimation processes described later.
[0088] 7(a), in S20, the visual confirmation determination ECU 10 determines whether or not the vehicle speed of the vehicle M is equal to or greater than the vehicle speed determination threshold using the line of sight cone area estimation unit 12. If the visual confirmation determination ECU 10 determines that the vehicle speed is equal to or greater than the vehicle speed determination threshold (S20: YES), the visual confirmation determination ECU 10 proceeds to S21. If the visual confirmation determination ECU 10 does not determine that the vehicle speed is equal to or greater than the vehicle speed determination threshold (S20: NO), the visual confirmation determination ECU 10 proceeds to S22.
[0089] In S21, the visual confirmation determination ECU 10 estimates the visual cone area EC by the visual cone area estimation unit 12. In this case, the visual confirmation determination ECU 10 does not change the visual cone area EC according to the vehicle speed. After that, the visual confirmation determination ECU 10 ends the visual cone area estimation process.
[0090] In S22, the visual confirmation determination ECU 10 estimates a gaze cone area EC that is reduced compared to the case of S21 by the gaze cone area estimation unit 12. The reduced gaze cone area EC means that the magnification rate of the gaze cone area EC according to the distance from the eye point EP is reduced. After that, the visual confirmation determination ECU 10 ends the gaze cone area estimation process.
[0091] Fig. 7(b) is a flowchart showing an example of a line-of-sight cone region estimation process according to a turning of a vehicle. As shown in Fig. 7(b), the visual identification determination ECU 10 determines whether or not the vehicle M is turning by the line-of-sight cone region estimation unit 12 in S30. The line-of-sight cone region estimation unit 12 determines whether or not the vehicle M is turning based on the yaw rate of the vehicle M or the steering angle of the vehicle M. If it is determined that the vehicle M is turning (S30: YES), the visual identification determination ECU 10 proceeds to S31. If it is not determined that the vehicle M is turning (S30: NO), the visual identification determination ECU 10 proceeds to S32.
[0092] In S31, the visibility determination ECU 10, by the gaze cone area estimation unit 12, estimates a gaze cone area EC expanded in the turning direction of the vehicle M compared to the case of S32 described later. The gaze cone area EC expanded in the turning direction of the vehicle M means a gaze cone area EC whose cross section perpendicular to the line of sight vector EB is expanded in the turning direction of the vehicle M. Thereafter, the visibility determination ECU 10 ends the gaze cone area estimation process.
[0093] In S32, the visibility determination ECU 10 estimates the gaze cone area EC by the gaze cone area estimation unit 12. In this case, the gaze cone area EC is not changed in accordance with the turning of the vehicle M. Thereafter, the visibility determination ECU 10 ends the gaze cone area estimation process.
[0094] Fig. 8(a) is a flowchart showing an example of a gaze cone region estimation process according to the moving speed of an object. As shown in Fig. 8(a), the visual identification determination ECU 10 determines whether or not the moving speed of the object seen by the driver is equal to or greater than the moving speed threshold by the gaze cone region estimation unit 12 in S40. If the visual identification determination ECU 10 determines that the moving speed of the object seen by the driver is equal to or greater than the moving speed threshold (S40: YES), the visual identification determination ECU 10 proceeds to S41. If the visual identification determination ECU 10 does not determine that the moving speed of the object seen by the driver is equal to or greater than the moving speed threshold (S40: NO), the visual identification determination ECU 10 proceeds to S42.
[0095] In S41, the visual identification determination ECU 10 estimates a visual cone area EC expanded in the moving direction of the object by the visual cone area estimation unit 12 compared to S42 described later. The visual cone area EC expanded in the moving direction of the object means a visual cone area EC whose cross section perpendicular to the visual line vector EB is expanded in the moving direction of the object. After that, the visual identification determination ECU 10 ends the visual cone area estimation process.
[0096] In S42, the visual recognition determination ECU 10 estimates the visual cone area EC by the visual cone area estimation unit 12. In this case, the visual recognition determination ECU 10 does not change the visual cone area EC according to the moving speed of the object. After that, the visual recognition determination ECU 10 ends the visual cone area estimation process.
[0097] Fig. 8(b) is a flowchart showing an example of a gaze cone region estimation process according to the driver's viewing situation. As shown in Fig. 8(b), the viewing determination ECU 10 determines, in S50, by the gaze cone region estimation unit 12, whether or not the object of the viewing determination process is an object determined to be viewed by the driver. If the viewing determination ECU 10 determines that the object is an object determined to be viewed by the driver (S50: YES), the process proceeds to S51. If the viewing determination ECU 10 does not determine that the object is an object determined to be viewed by the driver (S50: NO), the process proceeds to S52.
[0098] In S51, the visual recognition determination ECU 10 estimates a visual cone region EC that is enlarged compared to the case of S52 described later, by the visual recognition cone region estimation unit 12. Thereafter, the visual recognition determination ECU 10 ends the visual cone region estimation process.
[0099] In S52, the visual recognition determination ECU 10 estimates the visual cone area EC by the visual cone area estimation unit 12. In this case, the visual recognition determination ECU 10 does not change the visual cone area EC depending on whether or not the target object is visually recognized. After that, the visual recognition determination ECU 10 ends the visual cone area estimation process.
[0100] According to the visibility assessment device 100 (and the control method for the visibility assessment device 100) of this embodiment described above, the estimation accuracy of the gaze vector EB based on the image captured by the driver monitor camera 1 decreases depending on the direction of the driver's face, so the accuracy of the driver's visibility assessment of the target object can be improved by estimating the gaze cone area EC according to the direction of the driver's face.
[0101] In addition, according to the visibility assessment device 100, when the driver's face is facing forward toward the vehicle M, the estimation accuracy of the gaze vector in the pitch direction of the vehicle M is low, so a gaze cone area EC with a cross-section that is vertically elongated elliptical is estimated, and when the driver's face is deviated to the side from the forward direction of the vehicle M, the estimation accuracy of the gaze vector in the yaw direction of the vehicle M is low, so a gaze cone area EC with a cross-section that is horizontally elongated elliptical is estimated, thereby improving the accuracy of the driver's visibility assessment of the target object.
[0102] In addition, according to the visibility assessment device 100, since the driver's effective field of view angle narrows when the vehicle speed is high, the accuracy of the driver's visibility assessment of the target object can be improved by reducing the expansion rate of the line of sight cone area when the vehicle speed is equal to or greater than the vehicle speed assessment threshold.
[0103] Furthermore, according to the visibility assessment device 100, since it can be assumed that when the vehicle M is turning, the driver will be consciously looking at the area ahead, by expanding the cross section perpendicular to the line of sight vector of the line of sight cone area EC in the turning direction of the vehicle M, the accuracy of the driver's visibility assessment of the target object can be improved.
[0104] In addition, according to the visibility assessment device 100, when the moving speed of an object as seen by the driver is fast, the driver will attempt to view the object while taking its movement into consideration. Therefore, by estimating the line of sight cone area EC so that the cross section perpendicular to the line of sight vector expands in the direction of the object's movement, the accuracy of the driver's visibility assessment of the object can be improved.
[0105] In addition, according to the visibility assessment device 100, since it is considered that there is a high possibility that the driver will still be able to see an object even if he or she slightly shifts his or her line of sight from an object that he or she has once seen, the accuracy of the driver's visibility assessment of an object can be improved by increasing the expansion rate of the line of sight cone area EC.
[0106] Although the embodiment of the present invention has been described above, the present invention is not limited to the above-mentioned embodiment. The present invention can be embodied in various forms including the above-mentioned embodiment and various deductions and improvements based on the knowledge of those skilled in the art. [Explanation of symbols]
[0107] 1...driver monitor camera, 2...external sensor, 10...visibility determination ECU, 11...object recognition unit, 12...gaze cone area estimation unit, 13...visibility determination unit, 50...pedestrian (object), 100...visibility determination device, EB...gaze vector, EC...gaze cone area, EP...eye point, M...vehicle.
Claims
1. A visibility determination device that determines whether a driver is viewing an object detected by an external sensor of a vehicle, a gaze cone area estimation unit that estimates a gaze vector of the driver using a driver monitor camera of the vehicle and estimates a gaze cone area that expands in a cone shape as it moves away from the eye point of the driver based on the gaze vector of the driver; a visibility determination unit that determines whether the driver is viewing the object based on an overlap between the line of sight cone area and the object; Equipped with The gaze cone area estimation unit estimates the gaze cone area according to a facial direction of the driver.
2. 2. The visibility determination device according to claim 1, wherein the gaze cone area estimation unit increases an expansion rate of the gaze cone area depending on a distance from the eye point when the driver's face direction is deviated from a forward direction of the vehicle compared to when the driver's face direction is facing the forward direction.
3. 3. The visibility determination device according to claim 1, wherein the gaze cone area estimation unit estimates the gaze cone area having a cross section perpendicular to the gaze vector that is a vertically elongated ellipse when the driver's face is facing forward of the vehicle, and estimates the gaze cone area having a cross section perpendicular to the gaze vector that is a horizontally elongated ellipse when the driver's face is facing laterally away from the forward direction of the vehicle.
4. 3. The visibility determination device according to claim 1, wherein the gaze cone region estimation unit reduces an expansion rate of the gaze cone region depending on the distance from the eye point when the vehicle speed of the vehicle is equal to or greater than a vehicle speed determination threshold, compared to when the vehicle speed is less than the vehicle speed determination threshold.
5. 3. The visibility determination device according to claim 1, wherein the gaze cone area estimation unit estimates the gaze cone area such that, when the vehicle is turning, a cross section perpendicular to the gaze vector expands in a turning direction of the vehicle compared to when the vehicle is not turning.
6. 3. The visibility determination device according to claim 1, wherein when the moving speed of the object as seen by the driver is equal to or greater than a moving speed threshold, the gaze cone area estimation unit estimates the gaze cone area so that a cross section perpendicular to the gaze vector expands in a moving direction of the object as seen by the driver.
7. 3. The visibility determination device according to claim 1, wherein the gaze cone area estimation unit increases an expansion rate of the gaze cone area after determining that the driver is viewing the object, until determining that the driver is not viewing the object, compared to before determining that the driver is viewing the object.
8. A control method for a visibility determination device that determines whether a driver is viewing an object detected by an external sensor of a vehicle, comprising: A line-of-sight vector of the driver is estimated by a driver monitor camera of the vehicle, and a line-of-sight cone region that expands in a cone shape as it moves away from the eye point of the driver based on the line-of-sight vector of the driver is estimated; Estimating the line of sight cone area according to a face direction of the driver; A control method for a visibility determination device, the control method determining whether or not the driver is viewing the object based on an overlap between the line of sight cone area and the object.
Citation Information
Patent Citations
Driving support device, method, and program
JP2015127937A
Driving sensibility adjustment system of vehicle
JP2016037187A
Sight line learning system and sight line learning program
JP2017168024A
Recognition device, recognition method, and program
JP2020201670A
JP120113A
