JP2022122704A - Headlight control device and headlight control method - Google Patents

Abstract

To provide a headlight control device which can improve the detection accuracy of a tunnel and turn on a headlight at timing at which a vehicle arrives an entrance of the tunnel.SOLUTION: A tunnel entrance/exit recognition unit 12 previously machine-learns images of entrances of tunnels and performs image recognition of the entrance of the tunnel included in a photographed image obtained by imaging the front side of the vehicle by an on-vehicle camera 11 on the basis of a pattern obtained by machine learning. A distance calculation unit 13 calculates a distance to the entrance of the tunnel from the current position of the vehicle. A travel speed detection unit 14 detects a travel speed of the vehicle. A light turn-on/turn-off control unit performs control so as to calculate an arrival time at which the vehicle arrives the entrance of the tunnel on the basis of the distance to the entrance of the tunnel and the travel speed, decides timing to turn on a headlight of the vehicle by comparing the current time and the arrival time, and turns on the headlight 20 at the decided timing.SELECTED DRAWING: Figure 1

Description

The present invention relates to a headlight control device and a headlight control method.

An automobile (vehicle) is equipped with an automatic headlight lighting function that automatically turns on the headlights when an illuminometer detects that the surroundings of the automobile are dark. When a car enters a tunnel, the automatic headlight lighting function works and the headlights can be turned on automatically.

JP-A-2001-39210 JP-A-2005-75304 JP-A-2006-36048

However, since many lights are installed near the entrance of the tunnel, it is relatively bright, so the headlights are turned on when the illuminance measured by the illuminometer after traveling a predetermined distance into the tunnel falls below the threshold. There is Also, even when the vicinity of the tunnel entrance is dark, the headlights may be turned on when the illuminance below the threshold continues for a predetermined period of time in order to prevent malfunction or chattering that repeatedly turns on and off in a short period of time.

Therefore, according to the general automatic headlight lighting function, the headlights cannot be turned on at the timing when the car reaches the entrance of the tunnel. time will be delayed. In Patent Documents 1 to 3, in order to eliminate the delay in turning on the headlights, it is described that the front of the vehicle is photographed with a camera, the entrance of the tunnel is detected based on the photographed image, and the headlights are turned on. It is

However, in the headlight control devices described in Patent Documents 1 to 3, the tunnel is detected based on the brightness value (including the proportion of dark areas and average brightness) of the captured image. As mentioned above, if the entrance of the tunnel is bright, it may not be possible to detect the tunnel. It may not be detected. It is desired to improve the detection accuracy of the tunnel and turn on the headlights at the timing when the vehicle reaches the entrance of the tunnel.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a headlight control device and a headlight control method capable of improving the detection accuracy of a tunnel and turning on the headlights at the timing when the vehicle reaches the entrance of the tunnel. .

The present invention conducts machine learning in advance on images of multiple tunnel entrances, and based on the patterns obtained by machine learning, recognizes tunnel entrances included in images captured by an in-vehicle camera in front of the vehicle. a tunnel entrance recognition unit, a distance calculation unit that calculates the current position of the vehicle and a distance to the tunnel entrance image-recognized by the tunnel entrance recognition unit, a running speed detection unit that detects the running speed of the vehicle, and the When the tunnel entrance recognition unit image-recognizes the entrance of the tunnel, the vehicle reaches the entrance of the tunnel image-recognized based on the distance calculated by the distance calculation unit and the running speed detected by the running speed detection unit. a light lighting control unit that calculates an arrival time, compares the current time with the arrival time, determines timing for lighting the headlights of the vehicle, and controls the headlights to light at the determined timing. A headlight control device is provided.

In the present invention, an in-vehicle camera photographs the front of a vehicle, and a tunnel entrance recognition unit performs machine learning in advance on images of a plurality of tunnel entrances. The tunnel entrance included in the photographed image is image-recognized, the distance calculation unit calculates the current position of the vehicle and the distance to the tunnel entrance image-recognized by the tunnel entrance recognition unit, and the traveling speed detection unit detects the vehicle. based on the distance calculated by the distance calculating unit and the traveling speed detected by the traveling speed detecting unit when the tunnel entrance recognition unit recognizes the image of the entrance of the tunnel. to calculate the arrival time at which the vehicle reaches the entrance of the tunnel whose image has been recognized, compare the current time with the arrival time, determine the timing for turning on the headlights of the vehicle, and determine the timing for turning on the headlights of the vehicle at the determined timing. To provide a headlight control method for controlling the lighting of a ride.

According to the headlight control device and the headlight control method of the present invention, it is possible to improve the detection accuracy of the tunnel and turn on the headlights at the timing when the vehicle reaches the entrance of the tunnel.

It is a block diagram showing a headlight control device of one embodiment. 2 is a diagram showing a photographed image of a tunnel entrance, which is part of training data used in machine learning by the tunnel entrance/exit recognition unit 12 of FIG. 1. FIG. FIG. 2 is a diagram showing an example of a method for calculating the distance from the current position of the host vehicle to the entrance of a tunnel by the distance calculation unit 13 of FIG. 1; FIG. 4 is a diagram showing an example of an entrance under an elevated road that is not an entrance to a tunnel; 4 is a flow chart showing the operation of the headlight control device of one embodiment and the headlight control method of one embodiment;

Hereinafter, a headlight control device and a headlight control method according to one embodiment will be described with reference to the accompanying drawings. In FIG. 1 , a headlight control device 10 of one embodiment includes an in-vehicle camera 11 , a tunnel entrance/exit recognition unit 12 , a distance calculation unit 13 , a traveling speed detection unit 14 , and a light ON/OFF control unit 15 . The headlight control device 10 turns the headlights 20 on and off.

For example, when a drive recorder is attached so as to photograph the front of the vehicle, the camera of the drive recorder can be used as a substitute for the in-vehicle camera 11 . Therefore, the headlight control device 10 should include at least the tunnel entrance/exit recognition section 12 , the distance calculation section 13 , the traveling speed detection section 14 , and the light ON/OFF control section 15 . The tunnel entrance/exit recognizing unit 12 has a tunnel entrance recognizing unit and a tunnel exit recognizing unit, but may have only a tunnel entrance recognizing unit.

The in-vehicle camera 11 is attached to the windshield or the like so as to photograph the front of the vehicle. The in-vehicle camera 11 is preferably installed so that the center of the frame of the captured image is at infinity. The image data of the captured image generated by the vehicle-mounted camera 11 is supplied to the tunnel entrance/exit recognition unit 12, the distance calculation unit 13, and the running speed detection unit 14. FIG.

The tunnel entrance/exit recognition unit 12 performs machine learning in advance on images of multiple tunnel entrances and exits, and performs image recognition of tunnel entrances and exits, including unknown tunnels, based on patterns obtained by machine learning. . The tunnel entrance/exit recognizing unit 12 may perform machine learning using about 1,000 photographed images of tunnel entrances and exits in various locations as teacher data. The tunnel entrance/exit recognition unit 12 may use learning called deep learning among machine learning.

FIG. 2 shows photographed images T1 to T6, which are part of the photographed image of the tunnel entrance. The photographed images T1 to T6 and other photographed images of tunnel entrances (not shown) are training data used by the tunnel entrance/exit recognition unit 12 in machine learning. The shape of the entrance of a tunnel is generally semicircular (arc) or square, and there is some variation in brightness near the entrance. The tunnel entrance/exit recognition unit 12 learns in advance the shapes of the entrances of such multiple tunnels and the brightness (darkness) near the entrances, so that even if the tunnel is unknown, the tunnel entrance can be image-recognized. can do. The tunnel entrance/exit recognition unit 12 can similarly perform image recognition of the tunnel exit.

Usually, tunnels go through mountains, so there is some trees or cliffs of the mountains above the tunnels, and there is no sky directly above the tunnels. That is, the brightness above the tunnel is not as bright as the brightness of the sky. The tunnel entrance/exit recognition unit 12 can perform image recognition of the tunnel entrance with higher accuracy by learning including the brightness above the tunnel.

As will be described later, when the headlight control device 10 recognizes the exit of the tunnel based on the image captured by the vehicle-mounted camera 11 and does not turn off the headlight 20, the tunnel entrance/exit recognizing unit 12 recognizes the tunnel entrance/exit of a plurality of tunnels. Only the image of the entrance needs to be machine-learned in advance.

When the tunnel entrance/exit recognition unit 12 recognizes the tunnel entrance, the tunnel entrance/exit recognition unit 12 notifies the distance calculation unit 13 and the light ON/OFF control unit 15 that the tunnel entrance has been recognized.

When the tunnel entrance/exit recognition unit 12 recognizes the entrance of the tunnel, the distance calculation unit 13 calculates the distance from the current position of the vehicle to the entrance of the tunnel. As shown in FIG. 3, the distance calculation unit 13 calculates the distance V from the current position of the vehicle based on the vertical distance V from the center of the frame indicated by the white circle at infinity of the photographed image to the lower end of the entrance of the tunnel. Calculate the distance to the tunnel entrance. The longer the distance from the current position of the vehicle to the entrance of the tunnel, the shorter the distance V, and the shorter the distance, the longer the distance V. The distance calculation unit 13 supplies the calculated distance to the light ON/OFF control unit 15 .

The distance calculation unit 13 may calculate the distance from the current position of the vehicle to the entrance of the tunnel using a method other than the distance V of the image captured by the vehicle-mounted camera 11 . If the in-vehicle camera 11 is a stereo camera, the distance calculation unit 13 can calculate the distance from the current position of the vehicle to the entrance of the tunnel based on the left and right captured images. The distance calculation unit 13 repeatedly calculates the distance from the current position of the vehicle to the tunnel entrance while the tunnel entrance/exit recognition unit 12 recognizes the tunnel entrance.

The running speed detector 14 detects the running speed of the own vehicle. Here, the traveling speed detection unit 14 detects the traveling speed of the own vehicle based on the motion vector of the image captured by the vehicle-mounted camera 11 . Instead of the traveling speed detecting unit 14 detecting the traveling speed of the own vehicle based on the photographed image, the traveling speed of the own vehicle may be detected by acquiring the vehicle speed pulse of the vehicle. In the configuration in which the traveling speed detection unit 14 detects the traveling speed of the own vehicle based on the captured image, wiring for supplying the vehicle speed pulse to the headlight control device 10 is not required, and the configuration of the headlight control device 10 is simplified. It is preferable because it is

The traveling speed detection unit 14 repeatedly detects the traveling speed of the own vehicle while the image data of the photographed image is supplied from the vehicle-mounted camera 11 . The traveling speed detection unit 14 supplies the traveling speed of the own vehicle to the light ON/OFF control unit 15 .

When notified by the tunnel entrance/exit recognition unit 12 that the tunnel entrance has been recognized in front of the vehicle, the light lighting/extinguishing control unit 15 calculates the time required for the vehicle to reach the tunnel entrance. . If the distance from the current position of the vehicle supplied from the distance calculation unit 13 to the entrance of the tunnel is divided by the running speed of the vehicle supplied from the running speed detection unit 14, the time required to reach the entrance of the tunnel is calculated. Calculated. The light ON/OFF control unit 15 calculates the time of arrival at the entrance of the tunnel.

The headlight control device 10 has a clock (not shown), and the light ON/OFF control unit 15 compares the current time and the arrival time to determine the timing to turn on the headlights 20 . The light on/off control unit 15 controls to turn on the headlight 20 at the timing when the current time coincides with the arrival time.

The light turn-on/turn-off control unit 15 may finely adjust the timing of turning on the headlights 20 according to the running speed of the host vehicle. When the traveling speed of the host vehicle is faster than a predetermined speed, the light lighting/lighting-off control unit 15 may determine the timing of turning on the headlights 20 as a predetermined time before the current time coincides with the arrival time. good.

Specifically, if the traveling speed is less than 80 km/h, the light ON/OFF control unit 15 determines the timing at which the current time coincides with the arrival time as the lighting timing of the headlights 20 . When the traveling speed is 80 km/h or more and less than 100 km/h, the light turn-on/turn-off control unit 15 sets the time 0.5 seconds before the current time coincides with the arrival time as the turn-on timing of the headlight 20. decide. If the traveling speed is 100 km/h or more, the light turn-on/turn-off control unit 15 determines the time one second before the current time coincides with the arrival time as the turn-on timing of the headlight 20 .

It is not preferable that there is an error in the arrival time when the vehicle reaches the entrance of the tunnel, and the timing of turning on the headlights 20 is later than the timing when the vehicle actually enters the tunnel. Therefore, it is preferable to set the lighting timing of the headlight 20 to be slightly earlier than the calculated timing in which the current time coincides with the arrival time as the traveling speed increases.

When the tunnel entrance/exit recognition unit 12 recognizes that the vehicle has passed through the tunnel exit, the tunnel entrance/exit recognition unit 12 notifies the light ON/OFF control unit 15 that it has recognized that the tunnel exit has passed. do. If the tunnel entrance/exit recognizing unit 12 performs machine learning using a photographed image of the timing at which the vehicle passes through the tunnel exit as training data, the tunnel entrance/exit recognizing unit 12 can recognize that the vehicle has passed through the tunnel exit. can recognize.

When the tunnel entrance/exit recognizing unit 12 notifies that the passage of the tunnel exit has been recognized, the light lighting/extinguishing control unit 15 controls the headlight 20 to turn off.

There is no problem even if the timing of turning off the headlight 20 at the exit of the tunnel is delayed. Therefore, when it is recognized that the illuminance measured by the illuminance meter exceeds the threshold value and the host vehicle has passed through the exit of the tunnel, the headlight 20 may be controlled to be turned off. The light ON/OFF control unit 15 has a configuration including a light ON control unit and a light OFF control unit. The light extinguishing controller may extinguish the headlight 20 based on the illuminance measured by the illuminometer.

FIG. 4 is not the entrance of a tunnel, but the entrance under an elevated road provided to intersect the road on which the vehicle is traveling. The entrance under the viaduct is square like a tunnel, but it is bright near the entrance and very bright above. The tunnel entrance/exit recognition unit 12 recognizes that the photographed image shown in FIG. 4 deviates from the pattern of the tunnel entrance obtained by machine learning, so the entrance under the overpass may be mistakenly recognized as the entrance of the tunnel. no. Therefore, the headlight control device 10 does not erroneously turn on the headlights 20 under the overpass of the elevated road.

In addition to the underpass entrance of the elevated road shown in FIG. 4, similarly, the headlight 20 is not erroneously turned on under the elevated railway at the entrance under the elevated railway.

By the way, the tunnel entrance/exit recognition unit 12 recognizes a white line on the road, a center line, or a vehicle running ahead by recognizing the photographed image, and calculates the direction in which the own vehicle travels. can. Therefore, the tunnel entrance/exit recognizing unit 12 can image-recognize the entrance of a tunnel that exists in the direction in which the vehicle is traveling, and misidentifies the entrance of the tunnel that exists in a direction different from the direction in which the vehicle is traveling. No image recognition.

The operation of the headlight control device 10 and the headlight control method executed by the headlight control device 10 will be described again using the flowchart shown in FIG. In FIG. 5, the headlight control device 10 captures an image of the front of the vehicle with the vehicle-mounted camera 11 in step S1. When the camera is provided outside the headlight control device 10, the headlight control device 10 acquires a photographed image from the camera in step S1.

The headlight control device 10 (tunnel entrance/exit recognition unit 12) analyzes the captured image in step S2. The headlight control device 10 (tunnel entrance/exit recognition unit 12) determines whether or not the tunnel entrance is recognized in step S3. If the tunnel entrance is not recognized (NO), the headlight control device 10 (tunnel entrance/exit recognition unit 12) repeats the processes of steps S2 and S3.

If the tunnel entrance is recognized in step S3 (YES), the headlight control device 10 (distance calculator 13) calculates the distance from the host vehicle to the tunnel entrance in step S4.

The headlight control device 10 (running speed detector 14) detects the running speed of the own vehicle in parallel with steps S2 to S4.

In step S6, the headlight control device 10 (light ON/OFF control unit 15) calculates the arrival time at the tunnel entrance based on the distance to the tunnel entrance and the traveling speed. In step S7, the headlight control device 10 (light ON/OFF control unit 15) determines whether or not the host vehicle has reached the entrance of the tunnel. If the own vehicle has not reached the entrance of the tunnel (NO), the headlight control device 10 (the light ON/OFF control section 15) repeats the processes of steps S6 and S7.

If the own vehicle has reached the entrance of the tunnel in step S7 (YES), the headlight control device 10 (light on/off control unit 15) controls to turn on the headlight 20 in step S8. . As described above, the timing at which the vehicle reaches the entrance of the tunnel includes the timing when the current time matches the arrival time, and the time a predetermined time before the current time matches the arrival time.

Subsequently, in step S9, the headlight control device 10 (tunnel entrance/exit recognition unit 12) determines whether or not the exit of the tunnel has been recognized. If the tunnel exit is not recognized (NO), the headlight control device 10 (the light on/off control section 15 and the tunnel entrance/exit recognition section 12) repeats the processes of steps S8 and S9. If the exit of the tunnel is recognized (YES), the headlight control device 10 (light on/off control unit 15) controls the headlight 20 to be turned off in step S10.

In step S11, the headlight control device 10 determines whether or not a power supply (not shown) has been turned off. In the case of an engine-equipped vehicle, turning off the power means stopping the engine of the own vehicle and turning off the accessory power supply. In the case of an electric vehicle, this is a power-off state in which power to the headlight control device 10 is stopped. If the power is not turned off (NO), the headlight control device 10 repeats the processing of steps S1 to S11. If the power is turned off (YES), the headlight control device 10 terminates the process.

As described above, the tunnel entrance/exit recognizing unit 12 (tunnel entrance recognizing unit) does not simply detect the tunnel entrance based on the luminance value (including the ratio of dark areas and average luminance) of the captured image. In the headlight control device and headlight control method of the present embodiment, the shape of the tunnel entrance and the brightness near the entrance (in addition, the brightness above the tunnel) are machine-learned in advance, and based on the pattern obtained by machine learning. image recognition of the tunnel entrance. Therefore, according to the headlight control device and the headlight control method of the present embodiment, it is possible to improve the detection accuracy of the tunnel and turn on the headlights at the timing when the own vehicle reaches the entrance of the tunnel.

The present invention is not limited to the embodiments described above, and various modifications can be made without departing from the gist of the present invention. In FIG. 1, the tunnel entrance/exit recognition unit 12, the distance calculation unit 13, the traveling speed detection unit 14, and the light ON/OFF control unit 15 may be configured by a microprocessor or a microcomputer. The headlight control device 10 may include a non-temporary storage medium that stores a computer program (headlight control program) that causes a microcomputer (CPU) to execute the processing shown in FIG.

10 Headlight control device 11 In-vehicle camera 12 Tunnel entrance/exit recognition unit (tunnel entrance recognition unit)
13 distance calculation unit 14 running speed detection unit 15 light lighting/extinguishing control unit (light lighting control unit)
20 head ride

Claims (3)

Machine learning is performed in advance on images of multiple tunnel entrances, and based on the patterns obtained by machine learning, the tunnel entrance recognition unit recognizes images of tunnel entrances included in the images captured by the in-vehicle camera in front of the vehicle. When,
a distance calculation unit that calculates the current position of the vehicle and the distance to the tunnel entrance image-recognized by the tunnel entrance recognition unit;
a traveling speed detection unit that detects the traveling speed of the vehicle;
When the tunnel entrance recognition unit image-recognizes the entrance of the tunnel, the vehicle reaches the tunnel entrance image-recognized based on the distance calculated by the distance calculation unit and the running speed detected by the running speed detection unit. a light lighting control unit that calculates an arrival time, compares the current time with the arrival time, determines a timing to turn on the headlights of the vehicle, and controls the headlights to turn on at the determined timing;
A headlight control device comprising: wherein, when the traveling speed of the vehicle is faster than a predetermined speed, the light lighting control unit determines a time before a predetermined time when the current time coincides with the arrival time as the timing for turning on the headlights of the vehicle. 2. The headlight control device according to 1. In-vehicle camera shoots the front of the vehicle,
A tunnel entrance recognition unit that performs machine learning in advance on images of multiple tunnel entrances, based on the pattern obtained by machine learning, recognizes the tunnel entrance included in the captured image generated by the on-vehicle camera,
A distance calculation unit calculates the distance between the current position of the vehicle and the tunnel entrance image-recognized by the tunnel entrance recognition unit,
A traveling speed detection unit detects the traveling speed of the vehicle,
The light lighting control unit
When the tunnel entrance recognition unit image-recognizes the entrance of the tunnel, the vehicle reaches the tunnel entrance image-recognized based on the distance calculated by the distance calculation unit and the running speed detected by the running speed detection unit. to calculate the arrival time of
comparing the current time and the arrival time to determine the timing to turn on the headlights of the vehicle;
A headlight control method for controlling to turn on the headlight at a determined timing.

Images