Vehicle control system
The vehicle control device uses a monocular camera and distance sensor to accurately detect and prioritize warnings for moving objects, addressing inaccuracies in distance estimation and enhancing safety by reducing false alarms.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- TOYOTA JIDOSHA KK
- Filing Date
- 2023-12-05
- Publication Date
- 2026-06-23
Smart Images
Figure 0007878276000001 
Figure 0007878276000002 
Figure 0007878276000003
Abstract
Description
Technical Field
[0001] The present invention relates to a vehicle control device that issues an alarm to a driver of a host vehicle when a predetermined condition regarding the distance between the host vehicle and an object is satisfied.
Background Art
[0002] A vehicle control device that issues an alarm to a driver of a host vehicle when the distance between the host vehicle and an obstacle is less than or equal to a threshold value has been proposed (see, for example, Patent Document 1 below). This vehicle control device (hereinafter referred to as the "conventional device") includes a monocular camera and a processor. The monocular camera is directed, for example, to the rear of the host vehicle. The monocular camera provides an image obtained by photographing a rear area of the host vehicle to the processor at a predetermined frame rate. When the host vehicle is reversing, the processor calculates the distance between an object located behind the host vehicle and the host vehicle by processing a plurality of images (a plurality of images with different shooting times) acquired from the monocular camera according to a predetermined algorithm.
Prior Art Documents
Patent Documents
[0003]
Patent Document 1
Summary of the Invention
[0004] Generally, objects captured in the upper part of an image (IMG) acquired by a monocular camera are likely to be located relatively far from the camera. Conversely, objects captured in the lower part of the same image (IMG) are likely to be located relatively close to the camera. However, it is possible that an object captured in the upper part of the image (IMG) may be located relatively close to the camera and also higher than the camera. Therefore, if the distance between the vehicle and each object is estimated (calculated) based solely on the position (vertical coordinate in the image) of each object within the imaging range (angle of view) of the monocular camera mounted on the vehicle, the estimation accuracy will be low. For example, if a monocular camera captures a scene of a pedestrian approaching the vehicle while descending stairs (or a ramp), the pedestrian is likely to be captured in the upper part of the image (IMG[t0]) acquired at time t0 when the pedestrian is located near the top of the stairs. Therefore, even though the actual distance between the pedestrian and the vehicle is relatively small, the presence of the pedestrian in the upper part of image IMG[t0] may lead to the incorrect estimation that the distance between the vehicle and the pedestrian is relatively large.
[0005] One of the objectives of the present invention is to provide a vehicle control device that uses a monocular camera to detect the distance between a moving object and the vehicle and issues a warning when predetermined conditions regarding the distance are met, thereby enhancing the safety of a moving object approaching the vehicle while descending stairs or a ramp.
[0006] To solve the above problems, the vehicle control device (1) of the present invention is: An on-board sensor (20) includes a distance measuring sensor (21) that acquires the distance between the vehicle and an object in a predetermined first region in the direction of travel of the vehicle (V), and a monocular camera (22) that photographs a predetermined second region in the direction of travel of the vehicle, A processor (10) is configured to perform the following: first alarm processing, which controls the notification device (30) so that a predetermined first alarm is issued when the first condition regarding the first distance is met; and second alarm processing, which controls the notification device so that a predetermined second alarm is issued when the second condition regarding the second distance is met; based on the distance obtained by the distance measuring sensor, the distance between the vehicle and a stationary object located in the direction of travel of the vehicle is obtained as a first distance (ΔL1); and further, based on the image (IMG) obtained by the monocular camera, the distance between the vehicle and a moving object moving toward the vehicle is obtained as a second distance (ΔL2); first alarm processing, which controls the notification device so that a predetermined second alarm is issued when the second condition regarding the second distance is met. It is equipped with. The processor is configured to, when the first condition and the second condition are met, acquire the second distance based on the position (Yc) of the image of the moving object in the vertical axis direction of the imaging range of the monocular camera, and when the second distance is greater than the first distance and the difference (ΔL) exceeds a threshold (ΔLth), and when the image of the moving object moves downward and expands within the imaging range of the monocular camera, the processor increases the priority of the second alarm processing over the first alarm processing if the feature quantity (α=ΔYc / ΔYs), which is the change in the position (Yc) of the image of the moving object with respect to the change in the change in the size (Ys) of the image of the moving object (ΔYs), exceeds a predetermined value (αth).
[0007] The processor of the vehicle control device according to the present invention acquires a second distance based on the position of the moving object in the image within the imaging range of the monocular camera. However, as described above, the accuracy of the second distance acquired based solely on the position of the moving object in the image may be low. For example, even if the second distance (distance between the vehicle and the moving object) is estimated to be relatively large compared to the first distance (distance between the vehicle and a stationary object), the actual difference between the first and second distances may be minute.
[0008] In this scenario, when a moving object is photographed from the front with a monocular camera in a scene where the object is moving toward the vehicle, the image of the moving object (for example, the lower edge of the image of the moving object) moves downward and expands within the imaging range of the monocular camera. Furthermore, in a scene where the moving object is moving toward the vehicle while descending stairs or a ramp, the greater the vertical distance traveled, the greater the feature quantity (the change in the position of the image of the moving object relative to the change in the size of the image of the moving object). Therefore, in this invention, if the feature quantity exceeds a predetermined value, the accuracy of the second distance (the distance estimated based solely on the position of the image of the moving object) is considered low, and the processor increases the priority of the second alarm processing. This prioritizes alerting the moving object and enhances the safety of the moving object.
[0009] In a vehicle control device according to one aspect of the present invention, The predetermined value is the amount of change in the position of the image of the moving object relative to the amount of change in the size of the image of the moving object within the imaging range when the moving object moves toward the vehicle side parallel to the optical axis (ax) of the monocular camera.
[0010] According to this, the priority of the second warning process is increased when a moving object is descending in a direction inclined with respect to the optical axis of the monocular camera and moving toward the vehicle.
[0011] In another aspect of the present invention, a vehicle control device, The aforementioned stationary object is a staircase. The moving object is a pedestrian descending the stairs.
[0012] According to this, when a pedestrian is descending stairs and moving towards the vehicle, the priority of the second warning process (priority for alerting the pedestrian) is increased.
[0013] In another aspect of the present invention, a vehicle control device, The first condition is met when the first distance is less than or equal to the first threshold, The second condition is met when the second distance is less than or equal to the second threshold, The processor temporarily changes the first threshold and the second threshold so that only the second condition is met when the feature quantity exceeds the predetermined value.
[0014] According to this, if the feature quantity exceeds a predetermined value, the first and second thresholds are temporarily changed so that only the second condition is met. For example, the first threshold is set to an extremely small value, and the second threshold is set to an extremely large value. This increases the priority of the second alarm processing.
[0015] In another aspect of the present invention, a vehicle control device, The aforementioned stationary object is a barrier that restricts the movement of the moving object toward the vehicle itself. The processor executes the second alarm process when the shielding object is absent and the second condition is met.
[0016] According to this, the processor recognizes only obstacles (e.g., walls, fences) as stationary objects, and even if other stationary objects exist, it treats them as if they do not exist. In other words, the processor obtains the first distance, which is the distance between the obstacle and the vehicle, but does not obtain the distance between the vehicle and other stationary objects. If there are no obstacles and a moving object is present, the processor executes a second alarm process if the second alarm condition related to the second distance, which is the distance between the moving object and the vehicle, is met. This enhances the safety of the moving object. [Brief explanation of the drawing]
[0017] [Figure 1] Figure 1 is a block diagram of a vehicle control device according to one embodiment of the present invention. [Figure 2] Figure 2 is a side view showing a scene in which a pedestrian is descending stairs and moving toward the vehicle, and an example of an image obtained by photographing the same scene. [Figure 3] Figure 3 is a side view showing a scene in which a pedestrian is moving toward the vehicle parallel to the optical axis of the camera, and an example of an image obtained by capturing the scene with the camera. [Figure 4] Figure 4 is a flowchart of a program executed by the CPU to implement the warning function.
Embodiments for Carrying Out the Invention
[0018] (Schematic) As shown in FIG. 1, a vehicle control device 1 according to an embodiment of the present invention is applied to a vehicle V (hereinafter referred to as “the host vehicle”) having an automatic driving function. When the distance between the host vehicle and an object located around it becomes equal to or less than a threshold value in a state where the automatic driving function is disabled, the vehicle control device 1 has a warning function that issues a warning to the driver of the host vehicle to draw attention to the object.
[0019] (Specific Configuration) As shown in FIG. 1, the vehicle control device 1 includes an ECU 10, an in-vehicle sensor 20, and a notification device 30.
[0020] The ECU 10 includes a microcomputer including a CPU 10a, a ROM 10b (rewritable non-volatile memory), a RAM 10c, a timer 10d, etc. The CPU realizes various functions by executing programs (instructions) stored in the ROM. The ECU 10 is connected to other ECUs via a CAN (Controller Area Network).
[0021] The in-vehicle sensor 20 includes a sonar 21 as a sensor (distance measuring sensor) for measuring the distance between the host vehicle and an object located around it. The in-vehicle sensor 20 also includes a camera 22 for photographing the surrounding area of the host vehicle.
[0022] The sonar 21 intermittently emits ultrasonic waves to the right rear diagonal and left rear diagonal of the host vehicle, and receives ultrasonic waves (reflected waves) reflected by a three-dimensional object. The sonar 21 calculates the distance between the host vehicle and the three-dimensional object based on the time from when the ultrasonic wave is transmitted until the reflected wave is received, and provides the calculation result to the ECU 10.
[0023] Camera 22 includes an imaging device and an image analysis device. The imaging device is, for example, a monocular camera with a built-in CCD. The imaging device is installed at the rear of the vehicle and directed towards the rear of the vehicle. The imaging device captures the area behind the vehicle at a predetermined frame rate and acquires image data. The image analysis device analyzes the image data acquired from the imaging device and recognizes (identifies) objects present around the vehicle from the image IMG. The image analysis device can recognize, for example, a pedestrian P. That is, the image analysis device distinguishes between the area occupied by the pedestrian P and other areas in the acquired image IMG. Then, as shown in Figures 2 and 3, the image analysis device acquires the vertical coordinate Yc, which is the position of the pedestrian P's feet (feature point) in the vertical axis (Y axis) direction, and the vertical size Ys, which is the size (number of pixels) in the vertical axis direction, and provides the ECU 10 with information representing the vertical coordinate Yc and vertical size Ys. Furthermore, the image analysis device recognizes walls, fences, and other objects reflected in the image (IMG) and provides the recognition results to the ECU10.
[0024] Furthermore, the on-board sensor 20 includes a vehicle speed sensor 23 for acquiring the vehicle speed. The vehicle speed sensor 23 includes a rotation speed measurement circuit and a vehicle speed calculation device. The rotation speed measurement circuit includes a pulse generation circuit that outputs a pulse (electrical signal) each time the vehicle's wheels rotate by a predetermined angle, and a counter circuit that counts the number of such pulses. The vehicle speed calculation device acquires the output value (number of pulses) of the counter circuit at predetermined intervals (each unit of time elapsed) and resets the count value to "0". In this way, the vehicle speed calculation device acquires the number of wheel rotations N per unit time. The vehicle speed calculation device acquires the vehicle speed vs (absolute value) of the vehicle by multiplying the rotation speed N by a coefficient k. The vehicle speed calculation device then provides the ECU 10 with information representing the acquired vehicle speed vs.
[0025] Furthermore, the on-board sensor 20 includes a shift position sensor 24 for acquiring the current position of the vehicle's shift lever (forward position, reverse position, etc.). The shift position sensor 24 provides the ECU 10 with information representing the acquired current shift position.
[0026] The notification device 30 includes an image display device and an audio device. The image display device displays an image based on an image display command obtained from the ECU 10. The audio device plays sound based on an audio playback command obtained from the ECU 10.
[0027] (Alarm function) The ECU 10 sequentially acquires information representing the shift position from the shift position sensor 24. If the current shift position is the reverse position, the ECU 10 sequentially acquires the distance between the vehicle and a three-dimensional object from the sonar 21. The ECU 10 also sequentially acquires the vehicle speed vs from the vehicle speed sensor 23. Based on this information, the ECU 10 acquires the distance ΔL1 between the vehicle and a stationary object located behind the vehicle. If the distance ΔL1 is less than or equal to the threshold ΔL1th, the ECU 10 determines that the first alarm condition has been met.
[0028] Furthermore, if the current shift position is the reverse position, the ECU 10 sequentially acquires the vertical coordinate Yc (the position of the pedestrian P's feet in the image IMG) from the camera 22. Based on the vertical coordinate Yc, the ECU 10 estimates the distance ΔL2 between the vehicle and the pedestrian P (the distance in the direction parallel to the optical axis ax of the camera 22). Specifically, the ROM 10b stores a map MP (database) that defines the relationship between the vertical coordinate Yc and the distance ΔL2, and the ECU 10 obtains the distance ΔL2 (current value) by referring to the map MP. If the distance ΔL2 is decreasing (i.e., the pedestrian P is moving towards the vehicle) and the distance ΔL2 is less than or equal to the threshold ΔL2th, the ECU 10 determines that the second alarm condition has been met.
[0029] In the map MP, the distance ΔL2a associated with the vertical coordinate Yca is greater than the distance ΔL2b associated with the vertical coordinate Ycb (<Yca) below the vertical coordinate Yca. That is, it is estimated that the distance between the pedestrian P and the host vehicle when the pedestrian P is reflected in the upper part of the image IMG is greater than the distance between the pedestrian P and the host vehicle when the pedestrian P is reflected in the lower part of the same image IMG.
[0030] When the first warning condition is satisfied and the second warning condition is not satisfied, the ECU 10 controls the notification device 30 so that a predetermined first warning for alerting the driver to a stationary object behind the host vehicle is issued (first warning process). Specifically, the ECU 10 causes the notification device 30 to display a predetermined first image and reproduce a predetermined first voice.
[0031] When the first warning condition is not satisfied and the second warning condition is satisfied, the ECU 10 controls the notification device 30 so that a predetermined second warning for alerting the driver to the pedestrian P behind the host vehicle is issued (second warning process). Specifically, the ECU 10 causes the notification device 30 to display a predetermined second image and reproduce a predetermined second voice.
[0032] When the first warning condition is satisfied and the second warning condition is satisfied, the ECU 10 controls the notification device 30 so that either one of the first warning and the second warning is issued as described below. That is, the ECU 10 controls the notification device 30 so that the first warning is issued when it is determined that attention should be given priority to alerting the driver to a stationary object. On the other hand, the ECU 10 controls the notification device 30 so that the second warning is issued when it is determined that attention should be given priority to alerting the driver to the pedestrian P.
[0033] Specifically, the ECU 10 determines whether or not the following condition X regarding the magnitude relationship between the distance ΔL1 and the distance ΔL2 is satisfied. (Condition X) The difference ΔL (=ΔL2 - ΔL1) between distance ΔL1 and distance ΔL2 exceeds the threshold ΔLth.
[0034] Condition X is not met if "pedestrian P is estimated to be closer to the vehicle than to a stationary object" or if "pedestrian P is estimated to be farther away from the vehicle than to a stationary object, and pedestrian P and the stationary object are estimated to be relatively close to each other." In this case, it is preferable to prioritize issuing a warning about potential contact between the vehicle and pedestrian P. Therefore, if condition X is not met, the ECU 10 controls the notification device 30 to issue a second warning.
[0035] On the other hand, condition X is met when it is estimated that pedestrian P is located considerably farther away from the vehicle than stationary objects. Here, the distance ΔL2 is a value obtained (estimated) based solely on the vertical coordinate Yc of the image of pedestrian P captured in the image IMG (the coordinate of pedestrian P's feet within the imaging range of the monocular camera). As mentioned above, the accuracy of the distance ΔL2 obtained by this method is low. Therefore, even if condition X is met (it is estimated that pedestrian P is located considerably farther away from stationary objects), in reality, the stationary object and pedestrian P may be in close proximity. Such false detections occur, for example, when pedestrian P is located at a relatively high position compared to the vehicle (camera 22), as shown in Figure 2.
[0036] Figure 2 shows a scene in which pedestrian P is descending the stairs STP and approaching the vehicle. In this example, as shown in Figure 2(A), pedestrian P is visible at the top of the image IMG[t0] acquired at time t0 when pedestrian P is located at the top step of the stairs STP. Therefore, the distance ΔL2 (the distance acquired based on the vertical coordinate Yc) is considerably larger than the distance ΔL1 (the distance between the vehicle and the stairs STP acquired by the sonar 21), and condition X is satisfied (ΔL2 - ΔL2 > ΔLth). In other words, in this example, it is estimated (falsely detected) that pedestrian P is located considerably farther away from the stairs STP.
[0037] The ECU 10 determines whether such a false detection has occurred as follows. Specifically, if condition X is met, the ECU 10 sequentially acquires the vertical coordinate Yc and vertical size Ys from the camera 22. Then, based on the change in the vertical coordinate Yc (change in distance ΔL2 (estimated value)) in relation to the change in vertical size Ys, the ECU 10 determines whether the pedestrian P is moving toward the vehicle while descending the stairs STP or ramp. The determination process will be explained in detail below with reference to Figures 2 and 3. Note that in the example shown in Figures 2 and 3, the vehicle is temporarily stopped.
[0038] As shown in Figure 2, in the process of pedestrian P moving from time t0, when they are located on the top step of the stairs STP, to time t1, when they have moved to the middle step (Figure 2(B)), pedestrian P moves a distance Δd in a direction parallel to the optical axis ax of camera 22. The vertical coordinate Yc2A of pedestrian P's feet in image IMG[t0], acquired at time t0, is lower than the vertical coordinate Yc2B of pedestrian P's feet in image IMG[t1], acquired at time t1. Also, the vertical size Ys2B of pedestrian P in image IMG[t1] is larger than the vertical size Ys2A of pedestrian P in image IMG[t0]. In other words, in the process of pedestrian P descending one step of the stairs STP, the vertical coordinate Yc moves downward and the vertical size Ys expands within the imaging range of camera 22.
[0039] On the other hand, Figure 3 shows a scene in which pedestrian P moves toward the vehicle parallel to the optical axis ax of camera 22. In this example, the distance pedestrian P moves toward the vehicle during the transition from time t0 to time t1 is the same as the distance Δd pedestrian P moved in the scene shown in Figure 2. The vertical coordinate Yc3B in image IMG[t1] is located below the vertical coordinate Yc3A in image IMG[t0]. Also, the vertical size Ys3B of pedestrian P in image IMG[t1] is larger than the vertical size Ys3A of pedestrian P in image IMG[t0]. In other words, as pedestrian P moves forward, the vertical coordinate Yc moves downward and the vertical size Ys expands within the imaging range of camera 22.
[0040] Here, in the scene where pedestrian P descends the stairs STP (Figure 2), the change in vertical size Ys ΔYs2 (=ΔYs2B-ΔYs2A) when pedestrian P moves (forward) by a distance Δd is the same as the change in vertical size Ys ΔYs3 (=ΔYs3A-ΔYs3B) in the scene where pedestrian P moves (forward) by a distance Δd parallel to the optical axis of camera 22 (Figure 3). In contrast, in the scene where pedestrian P descends the stairs STP (Figure 2), the change in vertical coordinate Yc ΔYc2 (=Yc2A-Yc2B) when pedestrian P moves (forward) by a distance Δd is greater than the change in vertical coordinate Yc ΔYc3 (=Yc3A-Yc3B) in the scene where pedestrian P moves (forward) by a distance Δd parallel to the optical axis ax of camera 22 (Figure 3). Thus, the "change in vertical size Ys" is positively correlated with the "distance pedestrian P has moved in the optical axis direction." In addition, not only is there a positive correlation between the "change in vertical coordinate Yc" and the "distance pedestrian P has moved in the optical axis direction," but there is also a positive correlation between the "change in vertical coordinate Yc" and the "distance pedestrian P has moved in the vertical direction." According to this finding, when the change in vertical coordinate Yc ΔYc (hereinafter referred to as "feature quantity α") is relatively large relative to the change in vertical size Ys ΔYs when the image of pedestrian P is enlarged in the image IMG (within the imaging range of camera 22), it can be estimated that pedestrian P is descending from a position higher than the vehicle (camera 22) and approaching the vehicle.
[0041] Therefore, the ECU 10 sequentially acquires the vertical coordinate Yc and vertical size Ys, and based on this time-series data, acquires the change in the vertical coordinate Yc in relation to the change in the vertical size Ys as a feature α. Then, the ECU 10 determines whether or not the feature α exceeds the threshold αth. Here, when the pedestrian P moves toward the vehicle side parallel to the optical axis direction of the camera 22 (Figure 3), the change in the vertical coordinate Yc ΔYc3 in relation to the change in vertical size Ys ΔYs3 is measured in advance, and this measurement result is stored in the ROM 10b as the threshold αth.
[0042] If the feature quantity α exceeds the threshold αth, it is highly likely that the pedestrian P is descending from a relatively high position and moving towards the vehicle, and the distance ΔL2 estimated based solely on the vertical coordinate Yc is likely to be inaccurate. In other words, the actual distance between the vehicle and pedestrian P may be smaller than the distance ΔL2 (estimated value). Therefore, in this case (α > αth), the ECU 10 selects pedestrian P as the object to be warned and controls the notification device 30 to issue a second alarm. That is, the ECU 10 prioritizes the second alarm over the first alarm. On the other hand, if the feature quantity α is less than or equal to the threshold αth, the ECU 10 selects a stationary object as the object to be warned and controls the notification device 30 to issue a first alarm.
[0043] Furthermore, when the vehicle is moving in reverse, the ECU 10 calculates the feature quantity α by subtracting the correction amounts ΔYc[vs] and ΔYs[vs] corresponding to the vehicle speed vs from the change amount ΔYc and change amount ΔYs, respectively. A map representing the relationship between the vehicle speed vs and the correction amount is pre-designed and stored in the ROM 10b.
[0044] Next, referring to Figure 4, we will explain the program PR1 that the CPU 10a of the ECU 10 (hereinafter simply referred to as "CPU") executes to realize the above alarm function.
[0045] When the ignition switch is ON, the CPU sequentially acquires the current shift position from the shift position sensor 24. If the current shift position is the reverse position, the CPU starts executing program PR1 at a predetermined interval. The CPU starts executing program PR1 from step 100 and proceeds to step 101.
[0046] In step 101, the CPU determines whether the first alarm condition is met. If the CPU determines that the first alarm condition is met (ΔL1≦ΔL1th) (101:Yes), it proceeds to step 102. On the other hand, if the CPU does not determine that the first alarm condition is met (101:No), it proceeds to step 103.
[0047] In step 102, the CPU determines whether the second alarm condition is met. If the CPU determines that the second alarm condition is met (ΔL2≦ΔL2th) (102:Yes), it proceeds to step 104. On the other hand, if the CPU does not determine that the second alarm condition is met (102:No), it proceeds to step 107.
[0048] In step 103, the CPU determines whether the second alarm condition is met. If the CPU determines that the second alarm condition is met (ΔL2≦ΔL2th) (103:Yes), it proceeds to step 106. On the other hand, if the CPU does not determine that the second alarm condition is met (103:No), it proceeds to step 108, and in step 108, it terminates the execution of program PR1.
[0049] In step 104, the CPU determines whether condition X is true. If the CPU determines that condition X is true (ΔL2 - ΔL1 > ΔLth) (104: Yes), it proceeds to step 105. On the other hand, if the CPU does not determine that condition X is true (104: No), it proceeds to step 106.
[0050] In step 105, the CPU calculates the feature α and determines whether the feature α exceeds the threshold αth. If the CPU determines that the feature α exceeds the threshold αth (105: Yes), it proceeds to step 106. On the other hand, if the CPU does not determine that the feature α exceeds the threshold αth (105: No), it proceeds to step 107.
[0051] In step 106, the CPU controls the alarm device 30 so that the first alarm is issued. Next, the CPU proceeds to step 108 and terminates the execution of program PR1.
[0052] In step 107, the CPU controls the alarm device 30 so that a second alarm is issued. Next, the CPU proceeds to step 108 and terminates the execution of program PR1.
[0053] (effect) The ECU 10 of the vehicle control device 1 acquires the distance ΔL2 based on the position (vertical coordinate Yc) of the pedestrian P image within the imaging range of the camera 22. However, as mentioned above, the accuracy of the distance ΔL2 acquired by this method is low. Therefore, the distance ΔL2 (the distance between the vehicle and the pedestrian descending the stairs STP, acquired based on the vertical coordinate Yc) may be estimated to be relatively large (false detection) compared to the distance ΔL1 (the distance between the vehicle and the stairs STP acquired by the sonar 21).
[0054] In this scenario, when pedestrian P is moving toward the vehicle, and camera 22 captures the pedestrian P from the front, the image of pedestrian P (the position of their feet) moves downward and expands within the imaging range of camera 22. Furthermore, in the scenario where pedestrian P is moving toward the vehicle while descending stairs STP (or a ramp) (Figure 2), the greater the vertical movement, the larger the feature quantity α (the change in the vertical coordinate Yc relative to the change in vertical size Ys ΔYs, i.e., ΔYc). Therefore, in this embodiment, if the feature quantity α exceeds the threshold αth, the distance ΔL2 (the distance between the vehicle and pedestrian P estimated based solely on the vertical coordinate Yc) is considered inaccurate, and the ECU 10 increases the priority of the second alarm. This prioritizes alerting pedestrian P, thereby enhancing pedestrian P's safety.
[0055] The present invention is not limited to the embodiments described above, and various modifications can be adopted within the scope of the present invention, as described below.
[0056] <Example 1> When the first and second alarm conditions are met and the feature quantity α exceeds the threshold αth, the ECU 10 temporarily changes the thresholds ΔL1th and ΔL2th so that only the second alarm condition is met. For example, the ECU 10 sets the threshold ΔL1th to an extremely small value and the threshold ΔL2th to an extremely large value. Then, for example, when the ECU 10 can no longer recognize the pedestrian P (when the pedestrian P moves outside the imaging range of the camera 22), it returns the thresholds ΔL1th and ΔL2th to their original values (standard values).
[0057] <Modification 2> If the ECU 10 detects a stationary object other than an obstruction and a pedestrian P in the direction of travel of its own vehicle (rearward in the example of Figure 2), and there are no stationary objects such as low walls or fences (hereinafter referred to as "obstructions") that restrict the approach of a pedestrian P to the vehicle, the ECU 10 may prioritize the second alarm regardless of the distance ΔL1 and distance ΔL2. This enhances the safety of the pedestrian P. On the other hand, if an obstruction exists in the direction of travel of the vehicle, the ECU 10 determines whether or not a pedestrian P is present on the vehicle side relative to the obstruction using the same procedure as in the above embodiment. Thus, in this modified example, the stationary object in the above embodiment is limited to an obstruction. The ECU 10 can determine the presence or absence of an obstruction based on information acquired from the camera 22. Furthermore, based on information acquired from the sonar 21, the ECU 10 may recognize as an obstruction if it detects stationary objects behind its own vehicle that are continuous in the lateral direction or stationary objects arranged at equal intervals, whose lateral length (width) is greater than the width of its own vehicle and whose height is substantially constant.
[0058] <Other> Regardless of whether the first and second alarm conditions are met, the ECU 10 may calculate a feature quantity α and control the notification device 30 so that the second alarm is issued preferentially if the feature quantity α exceeds a threshold αth. [Explanation of Symbols]
[0059] 1...Vehicle control unit, 10...ECU, 20...On-board sensor, 30...Notification device
Claims
1. An on-board sensor including a distance measuring sensor that acquires the distance between the vehicle and an object in a predetermined first region in the direction of the vehicle's movement, and a monocular camera that photographs a predetermined second region in the direction of the vehicle's movement, A processor configured to perform the following: first alarm processing, which controls the alarm device to issue a predetermined first alarm when the first condition regarding the first distance is met; and second alarm processing, which controls the alarm device to issue a predetermined second alarm when the second condition regarding the second distance is met; based on the distance acquired by the distance measuring sensor, acquire the distance between the vehicle and a stationary object located in the direction of travel of the vehicle as the first distance; further, based on the image acquired by the monocular camera, acquire the distance between the vehicle and a moving object moving toward the vehicle as the second distance; A vehicle control device equipped with, The processor is configured to acquire the second distance based on the position of the moving object's image in the vertical axis direction of the imaging range of the monocular camera when the first condition and the second condition are met, and to increase the priority of the second alarm processing over the first alarm processing when the second distance is greater than the first distance and the difference exceeds a threshold, and the feature amount which is the change in the position of the moving object's image relative to the change in the size of the moving object's image when the image of the moving object moves downward and expands within the imaging range of the monocular camera exceeds a predetermined value.
2. In the vehicle control device according to claim 1, The predetermined value is the amount of change in the position of the image of the moving object within the imaging range, with respect to the amount of change in the size of the image of the moving object when the moving object moves toward the vehicle side parallel to the optical axis of the monocular camera, in a vehicle control device.
3. In the vehicle control device according to claim 1 or claim 2, The aforementioned stationary object is a staircase. The moving object is a pedestrian descending the stairs, and the vehicle control device is also a vehicle control device.
4. In the vehicle control device according to claim 1, The first condition is met when the first distance is less than or equal to the first threshold, The second condition is met when the second distance is less than or equal to the second threshold, A vehicle control device configured such that, when the feature quantity exceeds a predetermined value, the processor temporarily changes the first threshold and the second threshold so that only the second condition is met.
5. In the vehicle control device according to claim 1, The aforementioned stationary object is a barrier that restricts the movement of the moving object toward the vehicle itself. The processor executes the second alarm process when the obstruction does not exist and the second condition is met. A vehicle control device configured as follows.