WO2023013345A1 - Vehicle control device - Google Patents

Abstract

A driving assistance device (10) is provided with a camera (11), a radar device (12), and an ECU (21). The ECU (21) recognizes traffic conditions in front of a host vehicle on the basis of an image captured by the camera (11), which is for capturing the area in front of the host vehicle. The ECU (21) is provided with: a target detection unit for detecting, from an image of the camera (11), a target that pertains to the passage or stoppage of the host vehicle at an intersection in front of the host vehicle; and, an entrance determination unit for determining the entrance position of the intersection on the basis of the target detected by the target detection unit.

Description

vehicle controller Cross-reference to related applications

This application is based on Japanese Application No. 2021-126947 filed on August 2, 2021, and the contents thereof are incorporated herein.

The present disclosure relates to a vehicle control device.

A technology is known that recognizes the traffic situation in front of the own vehicle based on an image captured by a camera that captures an image of the area ahead of the own vehicle. For example, Patent Literature 1 discloses a traffic signal detection device that detects a traffic signal from at least one of a narrow-angle image captured by a narrow-angle camera and a wide-angle image captured by a wide-angle camera.

Japanese Patent No. 6447722

By the way, in recent years, driving support technology for vehicles has evolved, and there is a demand for technology that allows the vehicle to drive safely through intersections. In this case, it is considered desirable to obtain a variety of information about intersections in addition to traffic signal detection in Patent Document 1. In this regard, there is still room for improvement.

The present disclosure has been made in view of the above problems, and the main purpose thereof is to provide a vehicle control device capable of improving technology for driving support regarding traveling at intersections.

This disclosure is
A vehicle control device for recognizing traffic conditions in front of the vehicle based on an image captured by a camera that captures an image of the front of the vehicle,
a target detection unit that detects, from the image, a target related to the passage or stop of the own vehicle at an intersection in front of the own vehicle;
an entrance determination unit that determines an entrance position of the intersection based on the target detected by the target detection unit;
Prepare.

Targets detected from the images captured by the camera while the vehicle is running include targets related to the vehicle passing or stopping at the intersection in front of the vehicle. By detecting such a target from the image and determining the entrance position of the intersection based on the target, the entrance position of the intersection can be properly determined. In addition, since the entrance position of the intersection can be properly grasped, the information can be used for various types of driving support.

The above and other objects, features and advantages of the present disclosure will become more apparent from the following detailed description with reference to the accompanying drawings. The drawing is
FIG. 1 is a diagram showing a schematic configuration of a driving support device, FIG. 2 is a flowchart showing an overview of the signal determination process; FIG. 3 is a flowchart showing the entrance position determination process, FIG. 4 is a diagram for explaining the procedure for determining the stop line on the entrance side. FIG. 5 is a diagram for explaining the procedure for determining an entrance-side stop line on a road with multiple lanes on one side. FIG. 6 is a flowchart showing the process of setting the intersection recognition range; FIG. 7 is a diagram for explaining the procedure for setting the intersection recognition range. FIG. 8 is a flowchart showing signal color determination processing; FIG. 9 is a flowchart showing the signal color determination process following FIG. FIG. 10 is a diagram for explaining the procedure for determining the signal color in the intersection recognition range. FIG. 11 is a diagram for explaining the procedure for determining the signal color in the intersection recognition range. FIG. 12 is a diagram for explaining target object detection when the own vehicle enters an intersection. FIG. 13 is a diagram showing the arrangement of each traffic light at an intersection; FIG. 14 is a diagram showing score values by color for three traffic lights, FIG. 15 is a diagram showing the road in front of the intersection, FIG. 16 is a flow chart showing determination processing of arrow lights.

An embodiment in which a vehicle control device according to the present disclosure is applied to an in-vehicle driving support device 10 will be described below with reference to the drawings.

As shown in FIG. 1, the driving assistance device 10 according to this embodiment includes a camera 11, a radar device 12, and an ECU 21 as a vehicle control device. From now on, the own vehicle on which the driving support device 10 is mounted is referred to as the own vehicle VE.

The camera 11 is a monocular camera. The cameras 11 are attached, for example, to the front end, the rear end, and both side surfaces of the vehicle VE, and capture images of the surroundings of the vehicle. The camera 11 transmits the image information of the captured image to the ECU 21 . Camera 11 may be a compound eye camera.

The radar device 12 is a ranging device that transmits high-frequency signals in the millimeter wave band. The radar devices 12 are mounted, for example, on the front end, rear end, and both side surfaces of the vehicle VE, and measure distances to objects around the vehicle. Specifically, a search wave is transmitted at predetermined intervals, and reflected waves are received by a plurality of antennas. A plurality of detection points on the object are detected from the transmission time of the search wave and the reception time of the reflected wave, thereby measuring the distance to the object. In addition, the azimuth of the object is calculated from the phase difference of the reflected waves received by the multiple antennas. If the distance to the object and the azimuth of the object can be calculated, the relative position of the object with respect to the own vehicle VE can be specified.

Also, the radar device 12 calculates the relative velocity of the object from the frequency of the reflected wave reflected by the object, which is changed by the Doppler effect. As a result, the object existing around the vehicle is detected as a stationary object or a moving object. Specifically, when the sum of the relative velocity of the object and the vehicle speed, which is the vehicle speed of the vehicle VE, is zero, the object is detected as a stationary object, and the relative velocity of the object and the vehicle speed are detected as zero. If the sum is non-zero, the object is detected as moving. The radar device 12 transmits detection information of stationary objects and moving objects around the own vehicle to the ECU 21 .

The ECU 21 is a control device equipped with a well-known microcomputer consisting of CPU, ROM, RAM, flash memory, and the like. The ECU 21 acquires various signals and performs various controls based on the acquired information.

The ECU 21 detects objects around the vehicle based on the image captured by the camera 11 . Specifically, the ECU 21 calculates the relative position of each object captured by the camera 11 with respect to the own vehicle VE. Objects to be detected include moving objects such as other vehicles, pedestrians, and bicycles around the own vehicle, and fixed objects such as traffic lights, curbs, and guardrails provided around the road. Based on the image captured by the camera 11, the ECU 21 also recognizes lane marking information regarding lane markings on the road on which the vehicle VE is traveling. Specifically, the presence or absence of a white line or the like as a marking line is recognized, and the form (specifically, the type, position, width, length, curvature, etc. of the marking line) is recognized. In addition to lane markings, stop lines, pedestrian crossings, road arrows, and the like can be detected as road markings on the road.

The ECU 21 performs various controls based on detection information of objects around the vehicle transmitted from the radar device 12 and detection information of objects around the vehicle detected from the image captured by the camera 11 . For example, based on the detection information, the ECU 21 causes the display 31 provided in the instrument panel or the like in the vehicle to display information indicating the attention to the traffic light detected from the captured image of the camera 11, or An alarm device 32, such as a speaker or a buzzer installed in the room, is activated.

A yaw rate sensor 13, a steering angle sensor 14, and a vehicle speed sensor 15 are also connected to the ECU 21. The yaw rate sensor 13 is provided, for example, at a central position of the vehicle VE, and outputs to the ECU 21 a signal indicating a yaw rate corresponding to the rate of change of the steering amount of the vehicle VE. The steering angle sensor 14 is attached to, for example, a steering rod of the vehicle, and outputs a steering angle signal to the ECU 21 according to the amount of change in the steering angle of the steering wheel caused by the driver's operation. The vehicle speed sensor 15 is attached, for example, to a wheel portion of the own vehicle VE, detects the rotation direction of the wheel, and outputs a vehicle speed signal corresponding to the wheel speed to the ECU 21 .

By the way, in recent years, driving support technology for vehicles has evolved, and there is a demand for technology that allows the own vehicle VE to drive safely through intersections. Based on this idea, the ECU 21, which is the vehicle control device of the present embodiment, detects a target related to the passage or stop of the vehicle VE at the intersection in front of the vehicle from the image captured by the camera 11, and detects the target. , the entrance position of the intersection is determined. Further, a predetermined range on the far side from the entrance position is defined as an intersection recognition range CR that is recognized as being within the intersection, and the display color of the traffic light that the host vehicle VE should follow is determined within the intersection recognition range CR. there is

FIG. 2 is a flow chart showing an overview of the signal determination process executed by the ECU 21. FIG. The traffic light determination process is a process of setting an intersection recognition range CR and determining the display color of the traffic light that the host vehicle VE should follow within the intersection recognition range CR. This process is repeatedly executed by the ECU 21 at a predetermined cycle.

When the traffic light determination process is started, first of all, the ECU 21 acquires an image in front of the own vehicle VE captured by the camera 11 (step S10). Next, the ECU 21 executes entrance position determination processing for determining the entrance position of the intersection based on the target detected from the acquired image (step S20). The target here relates to the passage or stop of the own vehicle VE at the intersection in front of the own vehicle, and in the following description the target is also referred to as the intersection target. In this embodiment, a stop line, a traffic light, and a crosswalk at an intersection are detected as intersection targets. The position determined as the entrance position of the intersection is the front side position of the intersection that exists in front of the vehicle VE.

Next, the ECU 21 executes an intersection recognition range setting process for setting a predetermined range on the far side from the determined entrance position as the intersection recognition range CR (step S30). The end position on the far side of the range set as the intersection recognition range CR is the position on the far side of the intersection that exists in front of the vehicle VE.

Next, the ECU 21 executes signal color determination processing for determining the display color of the traffic light that the host vehicle VE should follow within the intersection recognition range CR (step S50). The display color determined here is displayed on the display 31 . Brake control and accelerator control as driving assistance may be performed according to the display color of the traffic light.

Specific processing contents of the entrance position determination processing in step S20, the intersection recognition range setting processing in step S30, and the signal color determination processing in step S50 will be described below.

FIG. 3 is a flow chart showing the entrance position determination process in step S20 of FIG. In FIG. 3, the ECU 21 determines whether a stop line is included as an intersection target detected from the image in front of the vehicle VE (step S21). When the stop line is included as the intersection target, the ECU 21 determines whether or not the stop line is the stop line on the intersection entrance side (hereinafter also referred to as the entrance side stop line) (step S22).

The specific contents of step S22, that is, the procedure for determining that the stop line is the entrance-side stop line, will be specifically described below. Here, in order to avoid erroneous detection of various road surface markings on the road as stop lines, the stop line on the entrance side is determined based on the positional relationship between the stop line and the traffic light. When the intersection target includes a stop line and a traffic light, the ECU 21 determines whether the stop line is on the entrance side on the condition that the distance between the stop line and the traffic light in the vehicle straight-ahead direction (longitudinal direction) is less than a predetermined distance TH1. Determine the stop line. More specifically, as shown in FIG. 4A, when a stop line 102 and a traffic signal 103 are detected at an intersection 101 ahead of the host vehicle VE, the longitudinal direction of the stop line 102 and the traffic signal 103 is detected. If the directional distance D1 is less than the predetermined distance TH1, it is determined that the stop line 102 is the entrance-side stop line. If the intersection target includes a plurality of stop lines 102, the ECU 21 may determine whether or not the frontmost stop line 102 among the plurality of stop lines 102 is the entrance side stop line. .

Further, when it is determined that the stop line 102 is the entrance-side stop line based on the vertical distance D1 between the stop line 102 and the traffic signal 103, the traffic signal 103 is arranged on the far side of the stop line 102. In some cases, it is arranged on the front side. For example, in FIG. 4(b), the traffic light 103 is arranged on the back side of the intersection 101. In terms of the mutual relationship between the stop line 102 and the traffic light 103, the stop line 102 is on the front side and the traffic light 103 is on the back side. ing. In FIG. 4C, the traffic signal 103 is arranged on the front side of the intersection 101. In terms of the mutual relationship between the stop line 102 and the traffic signal 103, the stop line 102 is on the back side and the traffic signal 103 is on the front side. It's becoming

In this case, the ECU 21 may differ in the predetermined distance TH1 for comparing and judging the vertical distance D1 between the stop line 102 and the traffic light 103 between the case of FIG. 4(b) and the case of FIG. 4(c). Specifically, in the case of FIG. 4B, the predetermined distance TH1 is set to "TH11", and in the case of FIG. 4C, the predetermined distance TH1 is set to "TH12" which is shorter than TH11. TH11 is, for example, 40 m, and TH12 is, for example, 20 m.

The stop line on the entrance side may be determined based on the positional relationship between the stop line 102 and the crosswalk 104. Specifically, as shown in FIG. 4( a ), when a stop line 102 and a crosswalk 104 are detected at an intersection 101 , the ECU 21 detects the longitudinal direction of the stop line 102 and crosswalk 104 . If the distance D2 is less than the predetermined distance TH2, it is determined that the stop line 102 is the entrance side stop line.

When the stop line 102, the traffic light 103, and the pedestrian crossing 104 are detected at the intersection 101, the vertical distance D1 between the stop line 102 and the traffic light 103 is less than the predetermined distance TH1, and the stop line 102 is crossed. It is also possible to determine that the stop line 102 is the entrance-side stop line on the condition that the vertical distance D2 to the sidewalk 104 is less than the predetermined distance TH2.

Further, in step S22, it may be determined by the following procedure that the stop line 102 is the entrance-side stop line. Focusing on the fact that the position of the stop line may be different for each lane on a multi-lane road with multiple lanes running in the same direction, this study focused on the position of the stop line in the own lane and the adjacent lane. , the entrance-side stop line targeted by the own vehicle VE.

　In Figures 5(a) to (c), three lanes are assumed to run in the same direction, and stop lines 102 are provided for each of the lanes L1, L2, and L3. The position of the stop line 102 for each lane is different as shown. As shown in FIG. 5(a), when the own vehicle VE is traveling in the lane L2, the lane L2 is the own lane, and the lanes L1 and L3 are adjacent lanes. The ECU 21 compares the stop line position of the own lane (L2) with the stop line positions of the adjacent lanes (L1, L3), based on the fact that the stop line position of the own lane (L2) is the frontmost side. Let the stop line 102 of (L2) be the entrance side stop line.

Also, as shown in FIG. 5(b), when the own vehicle VE is traveling in lane L3, lane L3 is the own lane and lane L2 is the adjacent lane. The ECU 21 uses the stop line position of the adjacent lane (L2) as a reference based on the fact that the stop line position of the adjacent lane (L2) is on the front side in the comparison between the own lane (L3) and the adjacent lane (L2). , and the selection area RY is defined in a predetermined range on the far side. Then, based on the fact that the stop line position of the own lane (L3) is within the selection area RY, the stop line 102 of the own lane (L3) is set as the entrance side stop line.

Also, as shown in FIG. 5(c), when the own vehicle VE is traveling in the lane L1, the lane L1 is the own lane and the lane L2 is the adjacent lane. The ECU 21 uses the stop line position of the adjacent lane (L2) as a reference based on the fact that the stop line position of the adjacent lane (L2) is on the front side in the comparison between the own lane (L1) and the adjacent lane (L2). , and the selection area RY is defined in a predetermined range on the far side. Then, if there is no stop line position of the own lane (L1) within the selection area RY, the stop line 102 of the adjacent lane (L2) is set as the entrance side stop line.

4 and 5, when a plurality of stop lines 102 are detected as intersection targets in front of the vehicle VE, the ECU 21 determines which of the plurality of stop lines 102 the vehicle VE A configuration may be adopted in which the stop line 102 closest to is determined to be the entrance-side stop line.

Returning to the description of FIG. 3, when it is determined that the stop line as the intersection target is the entrance-side stop line (step S22: YES), the ECU 21 determines the entrance position of the intersection based on the stop line ( step S23). In this embodiment, the position itself of the stop line in the straight-ahead direction of the vehicle VE is set as the entrance position of the intersection. However, it is also possible to set a position shifted to the front side or a position shifted to the back side from the position of the stop line as the entrance position of the intersection.

It should be noted that the process of determining the stop line on the entrance side based on the positional relationship between the stop line and the traffic light (Fig. 4 (a) to (c)), and the process of determining the stop line on the entrance side based on the stop line position for each lane in one-way multiple lanes Either one of the line determination processes (FIGS. 5A to 5C) may be performed, or both of them may be performed. When both processes are performed, for example, one of the stop lines of each lane of a plurality of lanes on one side is determined as a temporary entrance-side stop line. It is preferable to make a final determination that the vehicle is on the side stop line.

On the other hand, if the stop line is not included in the intersection target (step S21: NO) or if it is not determined that the stop line is the entrance-side stop line (step S22: NO), the ECU 21 (step S24). Then, if the traffic signal is included as the intersection target (step S24: YES), the ECU 21 determines the entrance position of the intersection based on the traffic signal (step S25). At this time, if a plurality of traffic lights are included in the intersection target, the intersection entrance position should be determined based on the traffic light closest to the vehicle (the side closest to the own vehicle VE) among the plurality of traffic lights.

Also, if the traffic signal is not included as the intersection target (step S24: NO), the ECU 21 determines whether or not the position of the traffic signal is estimated to be the extrapolated position (step S26). If the position of the traffic signal is estimated to be the extrapolated position (step S26: YES), the ECU 21 determines the entrance position of the intersection based on the traffic signal (step S25).

In short, when a traffic signal is detected as an intersection target, if the traffic signal is out of the field of view of the image captured by the camera 11, or is hidden by trees, signboards, etc., and cannot be physically detected, the signal cannot be detected. The state traffic light (ie, the traffic light used to determine the entrance location) is lost. In this case, the position of the traffic light is estimated as an extrapolated position, and the extrapolated position of the traffic light is used to continue determining the entrance position. The extrapolation position estimation result is preferably held until a predetermined period defined by the travel distance or travel time elapses.

In addition, when a stop line is detected as an intersection target and the entrance position of the intersection is determined based on the stop line, the detected stop line may be lost. In this case, the extrapolated position of the stop line may be estimated and the determination of the entrance position may be continued using the extrapolated position of the stop line.

The ECU 21 should determine whether the traffic signal is on the front side of the intersection or on the back side of the intersection, and determine the entrance position of the intersection based on the position of the traffic signal, taking into consideration the determination result. In this case, if the traffic signal is on the front side of the intersection, the position of the traffic signal may be set as the entrance position, or a position shifted forward or backward by a predetermined distance may be set as the entrance position. In the case of a traffic signal on the back side of an intersection, the entrance position may be set at a position shifted forward from the traffic signal by an assumed distance corresponding to the size of the intersection. It should be noted that the determination of whether the traffic signal is on the front side of the intersection or on the back side of the intersection may be made based on the positional relationship with the intersection center display and the positional relationship with the intersecting road at the intersection.

Although not shown, if all of steps S21, S22, and S24 are negative, it is determined whether or not a crosswalk is included as an intersection target, and if a crosswalk is included as an intersection target , the intersection entrance position may be determined based on the crosswalk.

FIG. 6 is a flow chart showing the intersection recognition range setting process in step S30 of FIG. In FIG. 6, the ECU 21 determines whether or not the entrance position of the intersection has been determined (step S31). If the intersection entrance position has already been determined (step S31: YES), the ECU 21 sets the intersection recognition range CR based on the intersection entrance position (steps S32 to S41).

The outline of the procedure for setting the intersection recognition range CR will be described below with reference to FIGS. 7(a) to 7(c). 7A to 7C show two intersections 101 in front of the vehicle VE. Of the two intersections 101, the intersection 101 on the lower side of the drawing is the next intersection through which the vehicle VE passes. is the intersection 101 of . In addition, in FIGS. 7A to 7C, P1 is the entrance position of the intersection in the vertical direction (the direction in which the vehicle travels straight). Here, first, a position a predetermined distance away from the entrance position P1 to the far side (rear) is provisionally set as the exit position P2, and the range from P1 to P2 in the vertical direction is set as the provisional setting range CRa. The distance between P1 and P2 is, for example, 80 m. Then, the intersection recognition range CR is set by appropriately correcting the provisional set range CRa according to the position of the back side target existing on the back side of the entrance position P1.

In FIG. 7(a), the traffic lights 103a and 103b in the same intersection are detected as the front side target on the front side of the intersection and the back side target on the back side of the intersection, respectively, in the temporary set range CRa. The intersection recognition range CR is set by correcting the exit position P2 (end position of the temporary set range CRa) based on the vertical distance D11 between the traffic lights 103a and 103b. At this time, since the distance D11 between the traffic lights 103a and 103b has a relatively high degree of reliability in the vertical direction, it is possible to appropriately set the intersection recognition range CR. The exit position P3 of the intersection is determined by the traffic light 103b on the far side of the intersection, and the range from P1 to P3 is set as the intersection recognition range CR. Note that the exit position P2 may be corrected based on the distance D11 on condition that the display colors of the traffic signals 103a and 103b are the same.

In FIG. 7(a), only the signal 103b on the far side of the signals 103a and 103b in the same intersection is detected, and along with this, the signal 103b is detected as the target on the front side, and the signal 103b is detected as the target on the back side. It is also conceivable that 103c is detected. In this case, since the traffic signals 103b and 103c belong to different intersections, their display colors may differ. The exit position P2 of the provisional setting range CRa may be corrected based on. However, if the display colors of the traffic lights 103b and 103c are different in "red/yellow", the exit position P2 may not be corrected. In addition, the exit position P2 may be configured to be corrected forward by a predetermined processing distance.

In FIG. 7(b), the stop line 102a is detected as the near side target and the traffic light 103b is detected as the far side target in the provisional setting range CRa. The intersection recognition range CR is set by correcting the exit position P2 based on the vertical distance D12 between the stop line 102a and the traffic light 103b. At this time, it is preferable to correct the exit position P2 when there is a traffic light 103b having a distance difference of a predetermined distance or more from the stop line 102a. The exit position P3 of the intersection is determined by the traffic light 103b on the far side of the intersection, and the range from P1 to P3 is set as the intersection recognition range CR.

In FIG. 7(c), stop lines 102a and 102b at different intersections are detected as the near side target and the far side target in the temporary set range CRa, respectively. The intersection recognition range CR is set by correcting the exit position P2 based on the vertical distance D13 between the stop lines 102a and 102b. The exit position P3 of the intersection is determined by the position a predetermined distance in front of the stop line 102b on the far side, and the range from P1 to P3 is set as the intersection recognition range CR.

As described above, in this embodiment, the exit position P3 of the intersection 101 is determined based on the target detected behind the entrance position P1, and the intersection recognition range CR is defined from the entrance position P1 to the exit position P3. is set. The setting of each intersection recognition range CR in FIGS. 7A to 7C described above gives top priority to the setting of FIG. ) may be appropriately performed in the order of setting.

It should be noted that the lateral range of the intersection recognition range CR may include the road on which the own vehicle VE travels. For example, the width of the intersection recognition range CR may be a predetermined value. If the road width is known from map information or the like, the width of the intersection recognition range CR may be set from that road width. The width of the intersection recognition range CR may be set accordingly. A configuration may be used in which the road edge in the lateral direction of the road is recognized from the camera image or the detection information of the radar device 12, and the width of the intersection recognition range CR is set according to the recognition result.

Returning to the description of FIG. 6, the ECU 21 sets a predetermined range on the far side from the entrance position P1 as a provisional set range CRa (step S32). After that, the ECU 21 determines whether or not the front side traffic signal and the back side traffic signal within the same intersection are detected in the provisional setting range CRa (step S33). If the front traffic signal and the rear traffic signal are detected (step S33: YES), the ECU 21 determines whether or not the display colors of the front traffic signal and the rear traffic signal are the same (step S34). If the display color is the same between the front traffic signal and the rear traffic signal (step S34: YES), the ECU 21 corrects the exit position P2 of the temporary set range CRa based on the vertical distance D11 between the traffic signals. Then, the intersection recognition range CR is set (step S35). This processing corresponds to the processing described with reference to FIG.

Further, when the display colors are different between the front traffic signal and the rear traffic signal (step S34: NO), the ECU 21 corrects the exit position P2 of the provisional setting range CRa based on the position of the front traffic signal, and corrects the intersection recognition range CR. is set (step S36). In this case, in FIG. 7A, the intersection recognition range CR is set with the position of the traffic light 103b as the exit position.

Also, if the front side traffic signal and the rear side traffic signal are not detected (step S33: NO), the ECU 21 determines whether the front side stop line and the rear side traffic signal are detected (step S37). If the near side stop line and the far side traffic signal are detected (step S37: YES), the ECU 21 determines the temporary set range CRa based on the vertical distance D12 between the near side stop line and the far side traffic signal. The exit position P2 is corrected and the intersection recognition range CR is set (step S38). This processing corresponds to the processing described with reference to FIG.

If the near side stop line and the far side traffic light are not detected (step S37: NO), the ECU 21 determines whether or not the near side stop line and the far side stop line are detected (step S39). ). If the near side stop line and the far side stop line are detected (step S39: YES), the ECU 21 temporarily sets the range based on the vertical distance D13 between the near side stop line and the far side stop line. The exit position P2 of CRa is corrected, and the intersection recognition range CR is set (step S40). This process corresponds to the process described in FIG. 7(c).

If the near side stop line and the far side stop line are not detected (step S39: NO), the ECU 21 does not correct the exit position P2 of the temporary set range CRa, and sets the temporary set range CRa as the intersection recognition range CR. Set (step S41).

8 and 9 are flowcharts showing the signal color determination process in step S50 of FIG. In FIG. 8, the ECU 21 determines whether or not the intersection recognition range CR has been set (step S51). If the intersection recognition range CR has been set (step S51: YES), the ECU 21 determines whether or not the own vehicle VE is traveling within the intersection recognition range CR (step S52).

Then, if the host vehicle VE is not within the intersection recognition range CR but is traveling in front of the intersection recognition range CR (step S52: NO), the ECU 21, in steps S53 to S63 of FIG. The display color of the traffic lights to be followed by the own vehicle VE among the traffic lights inside is determined. If the own vehicle VE is traveling within the intersection recognition range CR (step S52: YES), the ECU 21 restricts determination of the intersection entrance position in steps S71 to S75 of FIG. The display color of the traffic lights to be followed by the own vehicle VE among the traffic lights in the CR is determined.

10(a) to (c) and FIGS. will be explained based on FIGS. 10(a) to (c) and FIGS. 11(a) to (c) are diagrams showing, in chronological order, changes in the detection state of the traffic signals 103a and 103b in front of the vehicle in the intersection recognition range CR. Among them, FIGS. 10A to 10C show, in the intersection recognition range CR, first, a traffic light 103a on the front side of the intersection (front side of the own vehicle VE) and a traffic light 103a on the back side of the intersection (the front side of the own vehicle VE) as intersection targets. The traffic light 103b on the front rear side) is detected (FIG. 10(a)), and the case where a time-series change occurs from that state is shown. 11(a) to (c), only the traffic signal 103a on the front side of the intersection is detected as an intersection target in the intersection recognition range CR (FIG. 11(a)). It shows the case where a series change occurs. In FIG. 11A, for example, the traffic signal 103b on the far side of the intersection is hidden by a preceding vehicle or the like, and only the traffic signal 103a on the front side of the intersection is detected.

First, FIGS. 10(a) to (c) will be explained. In FIG. 10(a), traffic lights 103a and 103b are detected within the intersection recognition range CR. In this case, the traffic lights 103a and 103b are considered to be the traffic lights to which the host vehicle VE should follow, and the display colors of the traffic lights 103a and 103b are determined.

In FIG. 10(b), the own vehicle VE has moved forward from the state of FIG. 10(a), and the traffic light 103a on the near side is out of the angle of view of the camera, causing the traffic light 103a on the near side to be in a lost state. ing. In this case, the extrapolation position is estimated by extrapolation processing for the traffic signal 103a on the lost front side. In the extrapolation process, the relative position of the traffic signal 103a with respect to the vehicle VE is estimated based on the position of the traffic signal 103a immediately before the vehicle is lost and the vehicle speed of the vehicle VE. Then, the display color is determined for the traffic signal 103b, which remains in the detected state, out of the traffic signals 103a and 103b.

In FIG. 10(c), from the state of FIG. 10(b), that is, the state where the traffic light 103a on the front side is lost, the traffic light 103b on the far side is also in the lost state. In other words, the traffic lights 103a and 103b detected within the intersection recognition range CR are all lost. In this case, the display color determined immediately before the traffic light 103b on the far side is lost is maintained.

In addition, as a situation in which the traffic light is lost from the state of FIG. 10A, as shown in FIG. It is conceivable that the traffic signal 103b on the far side is hidden by the preceding vehicle or trees and lost. In this case, it is preferable to estimate the extrapolated position of the lost traffic signal 103b on the far side and to determine the color of the signal based on the traffic signal 103a on the near side that remains in the detected state.

Also, in FIG. 11(a), only the traffic light 103a on the front side of the intersection is detected as the intersection target within the intersection recognition range CR. In this case, the traffic light 103a is considered to be the traffic light that the host vehicle should follow, and the display color of the traffic light 103a is determined.

In FIG. 11(b), the own vehicle VE has moved forward from the state of FIG. 11(a), and the traffic light 103a on the near side is out of the angle of view of the camera, so that the traffic light 103a on the near side is in a lost state. ing. In this case, the display color output of the traffic light is temporarily interrupted.

In FIG. 11(c), the state of FIG. 11(b) (that is, the state in which the traffic light 103a on the near side is lost) is switched to the state in which the traffic light 103b on the far side is newly detected. In this case, the display color of the traffic light 103b on the far side is determined.

Returning to the description of FIG. 9, the ECU 21 determines whether or not the intersection target detected within the intersection recognition range CR includes at least one of the front side traffic light and the back side traffic light (step S53). If at least one of the front traffic signal and the rear traffic signal is included (step S53: YES), the ECU 21 determines whether or not both the front traffic signal and the rear traffic signal are in the detection state (step S54). . If both the front side traffic light and the far side traffic light are in the detection state (step S54: YES), the ECU 21 determines the display color that the host vehicle VE should follow based on the respective traffic lights (step S55). In the situation shown in FIG. 10(a), the display colors of the traffic lights 103a and 103b are determined in step S55.

If both the front and rear traffic lights are not in the detection state (step S54: NO), the ECU 21 determines whether or not only the front traffic light is in the detection state (step S56). If only the front traffic light is in the detection state (step S56: YES), the ECU 21 determines the display color that the host vehicle VE should follow based on the front traffic light (step S57). In the situation shown in FIG. 11(a), the display color of the traffic signal 103a is determined in step S57.

If the front traffic light is not in the detection state (step S56: NO), in other words, if only the back traffic light is in the detection state, the ECU 21 determines whether or not the front traffic light is in the extrapolation estimation state (step S56: NO). S58). If the near side traffic light is in the extrapolation estimation state (step S58: YES), that is, if the far side traffic light is in the detection state and the near side traffic light is in the extrapolation estimation state, the ECU 21 , the display color to be followed by the own vehicle VE is determined (step S59). In the situations shown in FIGS. 10(b) and 11(c), the display color of the traffic signal 103b is determined in step S59.

If the near side traffic light is not in the extrapolation estimation state (step S58: NO), in other words, if the far side traffic light is in the detection state and the near side traffic light is in the non-detection state and not in the extrapolation estimation state, the ECU 21 It is determined whether or not the front side traffic signal has been detected in the previous process, that is, whether or not the far side traffic signal has been lost this time (step S60). Then, if the front traffic signal was detected in the previous process (step S60: YES), the ECU 21 estimates the extrapolated position of the front traffic signal (step S61). Also, the display color to be followed by the own vehicle VE is determined based on the far side traffic light (step S59). When shifting from the state shown in FIG. 10(a) to the state shown in FIG. 10(b), extrapolation processing in step S61 and color determination of the traffic light 103b in step S59 are performed.

If the front side traffic signal was not detected in the previous process (step S60: NO), in other words, only the back side traffic signal was detected, and the front side traffic signal was not detected last time, and extrapolation estimation is performed. If it is not in the state either, the ECU 21 terminates this processing as it is.

If neither the front traffic signal nor the back traffic signal is originally detected within the intersection recognition range CR (step S53: NO), the ECU 21 determines in the previous process that there is a back traffic signal and the front traffic signal is outside. It is determined whether or not it is in an interpolation estimation state (step S62). If there is a traffic signal on the far side and the traffic signal on the front side is in the extrapolation estimation state in the previous process (step S62: YES), the ECU 21 waits for a predetermined period defined by the travel distance or the travel time after entering that state. The previous signal color output is maintained until elapses (step S63). In the situation shown in FIG. 10(c), the display color of the traffic signal 103b is determined in step S59.

Also, if the host vehicle VE is traveling within the intersection recognition range CR (step S52 in FIG. 8: YES), the process proceeds to step S71. The ECU 21 determines whether or not the intersection target detected within the intersection recognition range CR includes a traffic signal (step S71). If the intersection target includes a traffic signal (step S71: YES), the ECU 21 determines whether or not the rear traffic signal is detected (step S72). Then, if the far side traffic signal is detected (step S72: YES), the ECU 21 determines whether or not there is a lost history of the far side traffic signal. The lost history of the far side traffic signal here means that the far side traffic signal was detected before the own vehicle VE entered the intersection recognition range CR, and the own vehicle VE entered the intersection recognition range CR. This is the history indicating that the far side traffic light is no longer detected.

If there is a lost history of the far side traffic light (step S73: YES), the ECU 21 does not determine the entrance position based on the far side traffic light (step S74). In other words, when the far side traffic signal that was detected before entering the intersection recognition range CR is once lost after entering the intersection recognition range CR and then redetected, the ECU 21 , the entrance position is not determined using the rear traffic light.

The processing of steps S72 to S74 will be specifically described using FIGS. 12(a) to (c). FIGS. 12(a) to 12(c) are diagrams showing, in chronological order, changes in the detection states of the traffic lights 103a and 103b when the own vehicle VE enters the intersection recognition range CR.

In FIG. 12(a), the traffic lights 103a and 103b on the front side and the back side of the intersection are detected in the own vehicle VE before entering the intersection recognition range CR. After that, in FIG. 12B, the traffic signal 103a on the near side of the intersection is out of the angle of view of the camera and is lost, and the position of the traffic signal 103a is estimated by extrapolation. In this state, the traffic signal 103b on the far side of the intersection is also lost. For example, when the traffic signal 103b on the far side is hidden by the preceding vehicle, the traffic signal 103b is lost. At this time, a history is stored indicating that the traffic signal 103b on the far side of the intersection is no longer detected within the intersection recognition range CR.

After that, in FIG. 12(c), the traffic light 103b on the far side of the intersection is redetected while the own vehicle VE has entered the intersection recognition range CR. However, in this case, if an attempt is made to determine the entrance position of the next intersection based on the traffic light 103b, the entrance position of the intersection may be erroneously recognized. Therefore, the entrance position is not determined by the traffic light 103b re-detected after the own vehicle enters the intersection recognition range CR.

Returning to the description of FIG. 8, when the intersection target includes a traffic light (step S71: YES), after step S72 is denied or after step S74, the ECU 21 Of the traffic lights, the display color of the traffic lights that the host vehicle VE should follow is determined (step S75).

When multiple traffic lights are detected within the intersection recognition range CR, the display color of the traffic lights to be followed by the own vehicle VE (hereinafter also referred to as the own vehicle traffic light) should be determined based on the plurality of traffic lights. For example, in step S55 of FIG. 8 or step S75 of FIG. 9, the signal color may be determined by the following processing.

FIG. 13 is a diagram showing traffic lights that exist at the same intersection. In FIG. 13, a plurality of traffic lights 103_1 to 103_6 are detected within the intersection recognition range CR, and the ECU 21 classifies each of these traffic lights 103_1 to 103_6 according to their positions and forms, and according to their display colors. Determine the display color of traffic lights. In this case, the ECU 21 weights the traffic signals 103_1 to 103_6 according to their positions and display colors, and determines the display color of the vehicle traffic signal based on the weighted evaluation results.

　Explain the classification according to the position and form of traffic lights.

(1) Weighting is performed according to the position in the lateral direction with respect to the own vehicle VE (the lateral position with respect to the own vehicle center line LC), and the own vehicle traffic signal is specified. In other words, if the lateral position of the traffic signal 103 is excessively far from the own vehicle VE, there is a possibility that the traffic signal 103 is not the own vehicle signal. do. In FIG. 13, the traffic lights 103_1 to 103_4 are given a large weight, and the traffic lights 103_5 to 103_6 are given a small weight.

Each traffic signal 103 may be weighted according to whether the left lateral distance to the vehicle VE is greater than the reference distance DL or whether the right lateral distance to the vehicle VE is greater than the reference distance DR. . In this case, on the left side of the host vehicle VE, the weight is increased if the lateral distance to the traffic light is smaller than the reference distance DL, and the weight is decreased if the lateral distance to the traffic light is greater than the reference distance DL. On the right side of the vehicle VE, the weight is increased if the lateral distance to the traffic light is smaller than the reference distance DR, and the weight is decreased if the lateral distance to the traffic light is greater than the reference distance DR. More specifically, assuming the shape of the traffic lane on the road, ie, the road shown in the drawing, for example, is left-hand traffic, the reference distance DL on the left side of the vehicle VE should be smaller than the reference distance DR on the right side. That is, for example, when comparing the traffic lights 103_5 and 103_3, the lateral distance to the vehicle VE is shorter for the traffic lights 103_5 than for the vehicle VE, but since DL<DR, the weight of the traffic lights 103_5 is lowered, and the traffic lights 103_3 are weighted. weight is increased.

(2) Weighting is performed according to the vertical distance from the own vehicle VE to specify the own vehicle traffic signal. In other words, if the distance in the vertical direction to the own vehicle VE is excessively large, there is a possibility that the traffic signal is farther away than the next intersection. .

(3) Weighting is performed according to the height of the traffic signal, and the own vehicle's traffic signal is specified. In other words, a traffic signal whose height position is lower than the predetermined first height may be a traffic signal for pedestrians, not for automobiles. . In FIG. 13, the pedestrian traffic light 105 is detected in front of the own vehicle VE, and the weight of the pedestrian traffic light 105 is reduced. In addition, a traffic signal whose height position is higher than the predetermined second height may not be a traffic signal but a luminous object such as a street light or a light-emitting signboard, so the weight is reduced.

(4) Weighting is performed according to the direction of the traffic signal with respect to the own vehicle VE, and the own vehicle traffic signal is specified. That is, as shown in FIG. 13, each of the traffic lights 103_1 to 103_6 in the intersection has different orientations (specifically, the orientation of the front of the traffic lights) with respect to the own vehicle VE. In this case, the weight of the traffic signal that does not face the vehicle VE is reduced to reduce the possibility of being recognized as the traffic signal of the vehicle.

(5) Determine whether or not a housing having a lighting part is detected as the shape recognition of the traffic light. In this case, if the housing of the traffic light is not detected and only the light emission (lamp) is detected, the traffic light may be misrecognized, so the weight is reduced.

The ECU 21 performs at least one of the above weightings (1) to (5), and classifies traffic signals according to their position and form.

In addition to the above, it is also possible to determine whether or not the traffic signal is the vehicle's traffic signal based on the map or locator results, and increase the weight if it is determined to be the vehicle's traffic signal.

Next, we will explain the separation based on the display color of traffic lights.

Traffic lights (main lights of traffic lights) basically light up in either red, yellow, or green. Therefore, for each traffic signal detected within the intersection recognition range CR, the display color is determined in a predetermined period within a predetermined period of time, and the number of times of determination is counted for each color. Then, the number of determinations for each color counted by each traffic light and the evaluation value indicating the degree of certainty of color determination in the image are summed for each color to calculate the score value for each display color. Also, the display color with the highest score value is determined as the display color of the vehicle traffic signal. The degree of certainty of color determination may be calculated, for example, by comparison with a predetermined color reference.

At this time, for the traffic lights (traffic lights after weight adjustment) that have undergone classification by weighting in (1) to (5) above, it is preferable to perform score determination for each color of the traffic lights. In this case, the score value is added to the weighting result to determine the display color of the vehicle's traffic signal. In other words, the score value of each display color is corrected according to the weighting of (1) to (5) above, and the display color of the vehicle traffic light is determined based on the score value after the correction. It should be

　Figs. 14(a) and 14(b) show the results of counting the number of determinations by color for the traffic lights 103_1 to 103_3 in Fig. 13 . In FIG. 14A, it is determined that the current signal color is "red" from the score value obtained by accumulating the number of determination times for each color of the traffic lights 103_1 to 103_3. In addition, in FIG. 14(b), the color-specific score values are the same for “yellow” and “green”. In this case, it is preferable that the current signal color is determined to be the same as the previous determination. Alternatively, when the score values are the same as in FIG. 14B, the signal color may be determined according to the priority of green>yellow>red. However, in order to prevent misjudgment of the signal color, when the signal color is switched, it is preferable to determine the final signal color when it is determined that the signal color has changed more than a predetermined number of times in succession.

In addition to the main lights of three colors (red, yellow, and green), some traffic lights are equipped with arrow lights that display arrows. It's not what it is. Therefore, when judging an arrow light, for each traffic light, it is judged whether the arrow light is turned on at a predetermined cycle within a predetermined time, and the number of lighting determinations is counted. Then, the total value of the number of lighting decision times of the arrow lights at each traffic light is taken as a score value, and if the score value is greater than a predetermined value, it is determined that the arrow lights are on. Alternatively, the total value (score value) of the lighting determination times of the arrow lights is divided by the number of traffic lights having the arrow lights, and if the calculated value is greater than a predetermined value, it is determined that the arrow lights are lit. good too.

According to the embodiment described above, the following excellent effects can be obtained.

From the image captured by the camera 11, the intersection target related to the passage or stop of the vehicle VE at the intersection in front of the vehicle is detected, and the entrance position of the intersection is determined based on the intersection target. According to this configuration, since the entrance position of the intersection can be properly grasped, the information can be used for various kinds of driving support, and the technology for driving support can be improved regarding traveling at the intersection.

　The stop line and traffic lights at the intersection are detected as intersection targets, and the entrance position is determined based on the detection results of the stop line and traffic lights. In this case, even if which of the stop line and the traffic signal can be detected changes depending on the shape of the road and the running condition of the own vehicle, the entrance position of the intersection can be determined appropriately.

At intersections, a stop line is often set up on the front side of the intersection, and it is conceivable that the stop line directly indicates the entrance position of the intersection. However, assuming that there are various road markings on the road in addition to the stop line, there is concern that such road markings may be erroneously detected as the stop line. In this regard, when a stop line and a traffic light are detected at an intersection, the distance between the stop line and the traffic light in the straight-ahead direction of the vehicle VE is less than a predetermined distance. The entrance position is determined (see FIGS. 4(a) to 4(c)). Thereby, the entrance position of the intersection can be determined appropriately.

Also, when a stop line and a pedestrian crossing are detected at an intersection, the intersection is detected based on the stop line on the condition that the distance between the stop line and the pedestrian crossing in the straight ahead direction of the vehicle VE is less than a predetermined distance. is determined (see FIG. 4(a)). Thereby, the entrance position of the intersection can be determined appropriately.

The camera image in front of the vehicle may contain multiple targets of the same type. For example, multiple stop lines and multiple traffic lights may be reflected. In this case, the entrance position of the intersection is determined based on the closest target from the own vehicle VE among a plurality of targets of the same type, so the entrance position of the intersection can be determined appropriately.

Assuming that the position of the stop line may be different for each lane on a multi-lane road with multiple lanes running in the same direction, it is possible to determine whether the position of the stop line is detected in the own lane or in the next lane. The position of the entrance to the intersection is determined based on the position of the stop line while taking into account the position of the stop line (see FIGS. 5(a) to 5(c)). In other words, if the stop line position of the own lane is closer to the user than the stop line position of the own lane and the stop line position of the adjacent lane, the entrance position of the intersection is determined based on the stop line position of the own lane. I made it In addition, when the stop line position of the next lane is on the front side, a selection area is defined in a predetermined range on the back side with reference to the stop line position of the next lane, and the stop line of the own lane is set in the selection area. If so, determine the intersection entrance position based on the own lane's stop line position; otherwise, determine the intersection entrance based on the next lane's stop line position if there is no stop line position for the own lane within the selected area. determined the position. As a result, even if the stop line position is detected in either the own lane or the adjacent lane on a road with multiple lanes on one side, the entrance position of the intersection can be determined appropriately.

When the intersection target used to determine the intersection entrance position is lost and the position of the intersection target is estimated as an extrapolated position, the extrapolated position is used to determine the intersection entrance position. made it continue. As a result, even if the detected intersection target is lost due to the lack of the camera 11 or the like, the entrance position of the intersection can be continuously grasped.

　In order to properly determine whether the own vehicle VE should pass or stop at an intersection, it is desirable to properly understand the display color of the traffic light that the own vehicle VE should follow. In this regard, a predetermined range on the back side from the determined entrance position of the intersection is defined as an intersection recognition range CR, and the display color of the traffic light to be followed by the vehicle VE is determined within the intersection recognition range CR. As a result, it is possible to properly grasp the display color of the traffic signal while properly grasping the traffic signal to which the own vehicle VE should follow.

A configuration in which the exit position of the intersection is determined based on the intersection target existing on the back side of the determined entrance position of the intersection (back side target), and the range from the entrance position to the exit position is set as the intersection recognition range CR. (see FIGS. 7(a) to (c)). Then, the display color of the traffic signal to which the own vehicle VE should follow is determined within the intersection recognition range CR. As a result, it is possible to properly grasp the display color of the traffic signal while properly grasping the traffic signal to which the own vehicle VE should follow.

When the own vehicle VE enters the intersection, the traffic signal on the back side of the intersection (back traffic signal) and the traffic signal on the front side of the intersection (front traffic signal) are detected within the intersection recognition range CR. It is conceivable that the signals are lost due to different factors (see FIGS. 10(a) to 10(c)). For example, it is conceivable that each traffic light is lost due to physical factors such as being cut off outside the angle of view of the camera or being hidden by other objects such as trees or other vehicles. In this regard, from the state in which the far side traffic light and the near side traffic light are detected within the intersection recognition range CR, the traffic light position is estimated as an extrapolated position due to the loss of one of the traffic lights. , the display color of the traffic signal determined immediately before the loss of the other traffic signal is maintained. Thereby, regardless of the loss of each traffic signal, the display color of the traffic signal can be continuously recognized.

Also, when the own vehicle VE enters the intersection, it is conceivable that the front side traffic light is lost within the intersection recognition range CR, and then the rear side traffic light is detected ( (see FIGS. 11(a) to 11(c)). In this regard, when a traffic signal on the front side is detected within the intersection recognition range CR, the position of the traffic signal is estimated as an extrapolated position due to the loss of the traffic signal on the front side, and furthermore, a traffic signal on the back side is detected under that situation. Second, the display color of the traffic signal on the far side is determined as the display color of the traffic signal to which the own vehicle VE should follow. In this case, it is highly probable that the display color indicated by the far side traffic light is the display color of the traffic light that the vehicle VE should follow when passing through the intersection. Therefore, when the rear traffic signal is detected while the position of the traffic signal is estimated as an extrapolated position due to the loss of the front traffic signal, the display color indicated by the rear traffic signal is regarded as the display color of the traffic signal that the vehicle VE should follow. Therefore, it is possible to properly grasp the display color that the own vehicle VE should follow.

A rear traffic signal is detected before the own vehicle VE enters the intersection recognition range CR, and after the rear traffic signal is temporarily lost while the own vehicle VE has entered the intersection recognition range CR. It is conceivable that the far side traffic light may be detected again (see FIGS. 12(a) to (c)). In such a case, if the entrance position is determined based on the far side traffic light, there is a high possibility that the entrance position is not appropriate for setting the intersection recognition range CR. In this regard, in such a case, the proper entrance position can be maintained by not determining the entrance position based on the rear traffic light.

Within the intersection, there may be traffic lights at positions separated laterally or longitudinally from the own vehicle VE, for example. In addition, it is conceivable that luminous objects such as street lights and luminous signboards that may be mistaken for traffic lights exist in intersections. Furthermore, it is conceivable that the plurality of traffic signals may include traffic signals for different routes, or that signal colors may differ due to variations in operation or the like. In consideration of these factors, all the traffic signals existing within the intersection recognition range CR are extracted, and the traffic signals to be followed by the own vehicle VE are determined from the results of classifying the extracted traffic signals by at least one of position, form, and display color. The display color is now judged. As a result, even if there are many traffic lights within the intersection recognition range CR, the display color of the traffic lights to be followed by the own vehicle VE can be properly determined.

The above embodiment may be modified, for example, as follows.

Detecting the road surface arrow display on the near side of the intersection, matching the detection result of the road surface arrow display with the traffic signal detected within the intersection recognition range CR, and including the arrow light for the traffic light that the vehicle VE should follow. A configuration may be adopted in which signal display determination is performed. FIG. 15 is a diagram showing the road in front of the intersection. The road surface of the left lane is marked with a straight arrow indicating straight ahead, and the road surface of the right lane is marked with a right turn arrow indicating a right turn. That is, the left lane is the straight lane, and the right lane is the right-turning vehicle. The traffic light 103 is provided with an arrow light in addition to the main light. In this example, the arrow lights include a straight-ahead light and a right-turn light. Here, for example, when the vehicle VE travels in the right lane (right-turn lane), a right-turn arrow is detected as a road surface arrow display, and a right-turn light is detected as an arrow light of a traffic signal that the vehicle VE should follow. be.

FIG. 16 is a flow chart showing determination processing for arrow lights. This process is preferably executed, for example, when step S53 in FIG. 9 is affirmative.

The ECU 21 determines whether or not a road surface arrow display is detected on the road surface of the own lane on which the own vehicle VE travels (step S81). When the road arrow display is detected (step S81: YES), the ECU 21 determines whether or not the arrow light in the lighting state is detected at the traffic signal in front of the vehicle (step S82). Then, if an arrow light in a lit state is detected at the traffic signal in front of the vehicle (step S82: YES), the ECU 21 displays the arrow light having the same arrow direction as the road arrow display as a signal display to be followed by the vehicle VE. (step S83). For example, if the road arrow display is a right-turn arrow, the right-turn light of the traffic light is determined as the display of the own vehicle's traffic light, and if the road arrow display is a left-turn arrow, the left-turn light of the traffic light is determined as the display of the own vehicle's traffic light, If the road surface arrow display is a straight arrow, the straight light of the traffic signal is determined as the display of the vehicle traffic signal.

When the road arrow display is detected, if the traffic light in front of the vehicle is not provided with an arrow light corresponding to the direction of the road arrow, the traffic light to which the vehicle VE should follow is determined according to the main light of the traffic light. display color of is determined.

After the road arrow display is detected, if the road arrow display is lost due to the camera's angle of view being cut off or another vehicle, etc., the position of the road arrow display before being lost should be estimated by extrapolation processing. Estimation of the road surface arrow display by extrapolation processing is either that the vehicle VE has traveled and reached an intersection, that another road surface arrow display has been detected in front of the vehicle, or that the vehicle VE has traveled a predetermined distance. should be terminated by

In addition to the display of the road surface arrow of the own lane, it may be configured to detect the display of the road surface arrow of the next lane. Further, when the own vehicle VE changes lanes during extrapolation of the road surface arrow display, the extrapolation information of the destination of the lane change may be recognized as the road surface arrow display of the own lane after the lane change. For example, when the vehicle changes lanes to the adjacent lane on the right, the extrapolation information of the road surface arrow of the adjacent lane on the right is used as the road surface arrow information of the own lane.

If there are multiple road arrow indications in the same lane, each of the multiple road arrow indications should be detected. If a plurality of road surface arrow indications in the same lane are lost, all of the road surface arrow indications should be estimated by extrapolation processing. Extrapolation should not be performed if there is another roadmark indication with a vertical distance difference.

In the above-described embodiment, when setting the intersection recognition range CR, the provisional setting range CRa is set with reference to the entrance position P1, and the provisional setting range CRa is set to the back side target on the back side of the entrance position P1. (See FIGS. 7A to 7C), but this may be changed. For example, the provisional setting range CRa is not set. Specifically, the ECU 21 determines the exit position of the intersection based on the position of the back side target on the back side of the entrance position after determining the entrance position from the stop line or the traffic signal, and determines the distance from the entrance position to the exit position. is set as the intersection recognition range CR.

· When the entrance position of the intersection is determined, it is possible to display the entrance position on the display 31 or to activate the warning device 32 according to the distance from the own vehicle VE to the entrance position.

The controller and techniques described in this disclosure may be implemented by a dedicated computer provided by configuring a processor and memory programmed to perform one or more functions embodied by the computer program. may be Alternatively, the controls and techniques described in this disclosure may be implemented by a dedicated computer provided by configuring the processor with one or more dedicated hardware logic circuits. Alternatively, the control units and techniques described in this disclosure can be implemented by a combination of a processor and memory programmed to perform one or more functions and a processor configured by one or more hardware logic circuits. It may also be implemented by one or more dedicated computers configured. The computer program may also be stored as computer-executable instructions on a computer-readable non-transitional tangible recording medium.

Although the present disclosure has been described with reference to examples, it is understood that the present disclosure is not limited to those examples or structures. The present disclosure also includes various modifications and modifications within the equivalent range. In addition, various combinations and configurations, as well as other combinations and configurations, including single elements, more, or less, are within the scope and spirit of this disclosure.

Claims (14)

1. A vehicle control device (21) for recognizing traffic conditions in front of the own vehicle based on an image captured by a camera (11) for imaging the front of the own vehicle,
   a target detection unit that detects, from the image, a target related to the passage or stop of the own vehicle at an intersection in front of the own vehicle;
   an entrance determination unit that determines an entrance position of the intersection based on the target detected by the target detection unit;
   A vehicle control device.
2. The target detection unit detects at least one of a stop line and a traffic light at the intersection as the target,
   The vehicle control device according to claim 1, wherein said entrance determination unit determines said entrance position based on a detection result of at least one of said stop line and said traffic light.
3. When the targets detected by the target detection unit include the stop line and the traffic light, the entrance determination unit determines that the distance between the stop line and the traffic light in the straight-ahead direction of the vehicle is a predetermined distance. 3. The vehicle control device according to claim 2, wherein the entrance position is determined based on the stop line on condition that the distance is less than the stop line.
4. The target detection unit detects, as the target, a stop line and a pedestrian crossing at the intersection,
   The entrance determination unit determines the entrance position based on the stop line on condition that the distance between the stop line and the pedestrian crossing in the straight-ahead direction of the vehicle is less than a predetermined distance. 4. The vehicle control device according to any one of 3.
5. When a plurality of targets of the same type are included as the targets detected by the target detection section, the entrance determining section determines which of the plurality of targets of the same type is closest to the own vehicle. The vehicle control device according to any one of claims 1 to 4, wherein the entrance position of the intersection is determined based on the landmark.
6. The target detection unit detects a stop line as the target for each lane in a plurality of lanes on one side having a plurality of lanes in which the traveling direction is the same,
   The entrance determination unit
   When the stop line position of the own lane in which the own vehicle travels is compared with the stop line position of the adjacent lane adjacent to the own lane, if the stop line position of the own lane is on the near side, determining the entry location based on the stop line location;
   When the stop line position of the adjacent lane is on the front side, a selection area is defined in a predetermined range on the back side with reference to the stop line position of the adjacent lane, and the own lane is stopped within the selection area. If there is a line position, the entrance position is determined based on the stop line position of the own lane, and if there is no stop line position of the own lane within the selected area, based on the stop line position of the adjacent lane. The vehicle control device according to any one of claims 1 to 5, wherein the entrance position is determined.
7. The target detection unit estimates the position of the target as an extrapolation position when the target in the detected state is lost,
   When the target used to determine the entrance position is lost and the position of the target is estimated as an extrapolated position, the entrance determination unit uses the extrapolated position to determine the entrance position. 7. The vehicle control device according to any one of claims 1 to 6, further comprising a determination of .
8. A predetermined range on the back side from the entrance position determined by the entrance determination unit is used as an intersection recognition range to be recognized as being in the intersection, and the display color of the traffic light to be followed by the vehicle is determined within the intersection recognition range. The vehicle control device according to any one of claims 1 to 7, comprising a signal display determination section.
9. a back side target detection unit configured to detect, as a back side target, the target existing on the back side of the entrance position determined by the entrance determination unit;
   an exit determination unit that determines an exit position of the intersection based on the far side target detected by the far side target detection unit;
   The signal display determination unit defines the intersection recognition range from the entrance position determined by the entrance determination unit to the exit position determined by the exit determination unit, and the vehicle follows within the intersection recognition range. 9. The vehicle control device according to claim 8, wherein the display color of the signal is determined.
10. The target detection unit detects at least a traffic signal as the target, and estimates the position of the traffic signal as an extrapolated position when the detected traffic signal is lost,
    The signal display determination unit detects a signal on the back side of the intersection and a front signal on the front side of the intersection by the target detection unit within the intersection recognition range. , the signal position is estimated as an extrapolated position due to the loss of one of these signals, and if the other signal is lost under that situation, the display determined immediately before the loss of the other signal. 10. The vehicle control device according to claim 8 or 9, which maintains color.
11. The target detection unit detects at least a traffic signal as the target, and estimates the position of the traffic signal as an extrapolated position when the detected traffic signal is lost,
    The signal display determination unit extrapolates the position of the traffic signal from the state in which the front traffic signal, which is the traffic signal on the front side of the intersection, is detected by the target detection unit within the intersection recognition range, based on the loss of the front traffic signal. Further, when a rear traffic signal, which is a traffic signal on the rear side of the intersection, is detected under the situation, the display color of the traffic signal to be followed by the own vehicle is set as the display color of the traffic signal. The vehicle control device according to any one of claims 8 to 10, which determines the
12. The target detection unit detects at least a traffic light as the target,
    The entrance determination unit determines that, before the vehicle enters the intersection recognition range, the target detection unit detects a back-side traffic signal that is a back-side traffic signal of the intersection, and the vehicle enters the intersection recognition range. Claims 8 to 11, wherein the entrance position is not determined based on the far side traffic signal when the far side traffic signal is once lost and the far side traffic signal is re-detected while the own vehicle has entered the vehicle. The vehicle control device according to any one of claims 1 to 3.
13. The traffic signal display determination unit extracts all traffic signals existing within the intersection recognition range, and classifies the extracted traffic signals according to at least one of position, form, and display color, and determines the traffic signals to be followed by the vehicle. 13. The vehicle control device according to any one of claims 8 to 12, wherein the display color of is determined.
14. An arrow display detection unit that detects a road surface arrow display attached to the road surface on the front side of the intersection in the straight ahead direction of the vehicle,
    The signal display determining unit associates the detection result of the road surface arrow display with the traffic light detected within the intersection recognition range, and determines the signal display including the arrow light for the traffic light to be followed by the own vehicle. 14. The vehicle control device according to any one of items 8 to 13.

Images