JPH06276524A - Device for recognizing vehicle running in opposite direction - Google Patents

Abstract

PURPOSE:To improve the accuracy of recognizing a vehicle running in the opposite direction even when plural spots are in existence by recognizing it to be head lights of the vehicle running in the opposite direction when a couple of bright areas corresponding to a vehicle width in the horizontal direction are in existence and a bright area is in existence under the bright areas. CONSTITUTION:A vehicle running in the opposite direction has a couple of left and right head lights and the light emitted to a concerned vehicle is formed as a couple of bright areas at a predetermined interval corresponding to the vehicle width in the horizontal direction. Thus, when a couple of bright areas A whose centroid coordinates Y1,2 are equal to each other whose X coordinate interval X1, X2 corresponds to a standard head lamp interval are detected in an image 120, it is an object area of the head lights of the vehicle running in the opposite direction. Moreover, the head lights are placed usually at a low position and the light reflected in the driving road is emitted in front of the vehicle. Thus, when the bright areas A are in existence in the image 120 and a bright area 3 is in existence under the areas A, the areas A are recognized to be head lights of the vehicle running in the opposite direction and the presence of the vehicle running in the opposite direction is recognized. Thus, even when plural light spots are in existence, the vehicle running in the opposite direction is recognized with high accuracy.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an oncoming vehicle recognition device, and more particularly to an oncoming vehicle recognition device for detecting an oncoming vehicle traveling in front of the own vehicle while the vehicle is traveling.

[0002]

2. Description of the Related Art Conventionally, the problems to be solved by the invention
In order to improve the front visibility of the driver at night or the like, the vehicle is provided with a headlamp that is provided at a substantially tip end of the vehicle to illuminate a predetermined range.

This headlamp is a light-shielding device for shielding the headlight from the irradiation light in order to change the irradiation direction and the irradiation range in front of the vehicle according to the traveling direction of the vehicle such as the traveling direction according to the steering angle and the vehicle speed. There is also a system in which a plate is provided and the movement of the shading plate is controlled to control a boundary portion (hereinafter, referred to as a cut line) between an irradiation region and a non-irradiation region when light is irradiated on a road.

When the cut line of the headlamp is controlled, there is a vehicle (hereinafter referred to as a preceding vehicle) traveling in the same direction in front of the host vehicle within the irradiation range included in the controlled cut line. In this case, the driver of the preceding vehicle will be given an unpleasant glare.

Similarly, in order not to give glare to a vehicle traveling in the opposite direction in front of the subject vehicle (hereinafter referred to as an oncoming vehicle), it is necessary to recognize the oncoming vehicle as well as the preceding vehicle. Become.

Therefore, when controlling the cut line of the headlamp, it is necessary to recognize the positions and directions of the preceding vehicle and the oncoming vehicle in order to control the cut line without giving glare to the preceding vehicle.

As a method of recognizing the preceding vehicle, the tail lamp (red) of the preceding vehicle traveling in front of the own vehicle is detected by an image device (color camera) equipped with a color CCD or the like,
There is a preceding vehicle recognition device that performs image processing on the detected image to specify the position and direction of the preceding vehicle (Japanese Patent Laid-Open No. 62-1 / 1987).
21599, JP 62-131837 A, JP 6
3-78300 publication).

However, the use of an element capable of detecting a color for detecting a tail lamp cannot be easily utilized because the detecting element itself is very expensive. In addition, since image processing of an image including a color component is complicated and difficult, it imposes a heavy load on the recognition device and is practically difficult. Further, since the substantially white light of the headlamp is directly incident from the oncoming vehicle, there is a difference in color and light amount. Therefore, it is not easy to distinguish between the preceding vehicle and the oncoming vehicle.

In order to recognize the oncoming vehicle to solve this, a light amount detector such as a photo sensor is attached to the own vehicle to detect the light amount of the headlamp of the oncoming vehicle.
Oncoming vehicles can be detected and recognized.

However, the reflected light of the light from the own vehicle may be detected or the light such as the electric light around the side road may be detected. If only the amount of light is detected and it is recognized as an oncoming vehicle, other light is erroneously detected. It may be recognized as the light of the headlamp of the oncoming vehicle.

In consideration of the above facts, the present invention is based on an image in which a plurality of lights such as headlights and outside lights of an oncoming vehicle exist.
An object is to obtain an oncoming vehicle recognition device that can easily recognize an oncoming vehicle.

[0012]

In order to achieve the above object, the oncoming vehicle recognition device of the present invention comprises an image pickup means for picking up an image of the front of the vehicle and a pair of bright parts which are horizontally separated from the picked up image by a predetermined distance. Candidate area extracting means for extracting an area as a candidate area for a light of an oncoming vehicle, and an oncoming vehicle recognition means for recognizing the candidate area as a light of an oncoming vehicle when there is a bright area below the candidate area, It is characterized by having.

[0013]

In the oncoming vehicle recognition apparatus of the present invention, an image in front of the vehicle is picked up by the image pickup means such as a TV camera. As is well known, an ordinary vehicle has a pair of headlamps on the left and right in front of the vehicle. In addition, recently, there are vehicles in which auxiliary lights such as fog lights are arranged around the headlamps.
Therefore, when the oncoming vehicle is imaged by the imaging means at night, one pair of oncoming vehicles are assumed to be present in the image area, and a pair of light spots (areas) are formed in a substantially horizontal direction at predetermined intervals according to the vehicle width. It is formed. Therefore, if a pair of bright areas in a substantially horizontal direction and having a predetermined interval corresponding to the vehicle width are extracted from the input image, the pair of bright areas is not likely to be a headlight or fog lamp of an oncoming vehicle. high. Therefore, the candidate area extraction means extracts a pair of bright areas, which are separated from each other in the horizontal direction by a predetermined distance corresponding to the vehicle width or the like, as the candidate areas for the light of the oncoming vehicle.

Here, in the surrounding environment where the vehicle travels,
There is a wall surface of a building such as an outdoor light or a highly reflective building, and light from these is formed as a light spot region in the captured image. Therefore, if a pair of bright areas with a predetermined interval is recognized as a light of an oncoming vehicle, a light spot area due to reflected light from an external light or a building may be erroneously recognized as a light of an oncoming vehicle. .

By the way, lights such as headlamps and fog lamps of a vehicle illuminate the front of the vehicle. Also, as is well known, the location of the light is close to the road. Therefore, when this oncoming vehicle is imaged when the light is emitted from the light, the reflected light from the road surface is imaged. That is, a bright region is formed below a direct light from a light such as a head lamp or a fog lamp and at a predetermined position (road surface). Therefore, in the present invention, the oncoming vehicle recognition means recognizes the candidate area as a light of the oncoming vehicle when a bright area exists below the candidate area.

By doing so, even when a plurality of light spot regions due to reflection of outside lights or buildings are formed in the captured image, the light reflected from the road surface by the light of the oncoming vehicle is formed. When there is a bright region formed by, the candidate region of the oncoming vehicle is recognized as the light of the oncoming vehicle, so that the oncoming vehicle can be recognized more reliably.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment to which the traveling vehicle recognition device of the present invention is applied will be described in detail below with reference to the drawings. The traveling vehicle recognition device 100 according to the present embodiment applies the present invention when another vehicle traveling in front of the vehicle 10 is obtained from a gradation image obtained by a monochrome TV camera.

As shown in FIG. 1, an engine hood 12 is arranged on the upper surface of a front body 10A of a vehicle 10, and a front bumper 16 is provided at both front and rear ends of the front body 10A in the vehicle width direction. Is fixed. The upper part of the front bumper 16 and the front body 10
In the lower part of A, a pair of left and right headlamps 18 and 20 (both ends in the vehicle width direction) are arranged.

A windshield glass 14 is provided near the rear end of the engine hood 12, and a room mirror 15 is provided above the windshield glass 14 and inside the vehicle 10. This room mirror 15
A TV camera 22 connected to the image processing device 48 (FIG. 4) for photographing the front of the vehicle is arranged in the vicinity.
The TV camera 22 may be disposed near the driver's visual position (so-called eye point) so that the shape of the road ahead of the vehicle can be accurately recognized and the driver's visual sensation can be more closely matched. preferable.

A speedometer (not shown) is arranged in the vehicle 10, and a vehicle speed sensor 66 for detecting the vehicle speed V of the vehicle 10 is attached to a cable (not shown) of the speedometer (not shown).

As shown in FIGS. 2 and 3, the headlamp 18 is a projector-type headlamp and has a convex lens 30, a bulb 32 and a lamp house 34. The lamp house 34 is horizontally fixed to a frame (not shown) of the vehicle 10. The convex lens 30 is fixed to one opening of the lamp house 34, and the optical axis L of the convex lens 30 (convex lens 30) is fixed to the other opening. The bulb 32 is fixed via a socket 36 such that the light emitting point is located on the central axis of the bulb.

The bulb side inside the lamp house 34 is a reflector 38 having an elliptical reflecting surface, and the light reflected by the bulb 38 by the reflector 38 is convex lens 30 and bulb 32.
Is focused between. Actuators 40 and 42 are arranged near this condensing point. This actuator 4
The light of the bulb 32 reflected and condensed by the reflector 38 is blocked by the light blocking cams 40A and 42A of 0 and 42,
The other light is emitted from the convex lens 30.

The actuator 40 includes a light shielding cam 40A,
The actuator 42 is composed of gears 40B and 40C and a motor 40D.
2C and motor 42D. Shading cam 4
0A and 42A are rotatably supported by a rotary shaft 44 fixed to the lamp house 34, and a gear 40B is fixed to the light shielding cam 40A. A gear 40C fixed to the motor 40D meshes with the gear 40B. The motor 40D is connected to the control device 50. The light-shielding cam 40A has a cam shape in which the distance from the rotation shaft 44 to the outer periphery continuously changes, and the control device 50
In accordance with the signal from the light shielding cam 40 in the lamp house 34
When A rotates, the position where the light of the bulb 32 is divided into the passing light and the blocked light changes up and down. Similarly, the light blocking cam 42A is rotatably supported by a rotary shaft 44 fixed to the lamp house 34, and
A gear 42B is fixed to A. A gear 42C fixed to the motor 42D meshes with the gear 42B. The motor 42D is connected to the control device 50.

Therefore, the position above the light shielding cams 40A and 42A is located on the road as a cut line which is a boundary between light and dark in the light distribution in front of the vehicle. That is, as shown in FIG. 16, the light shielding cam 40A forms the cut line 70, and the light shielding cam 42A forms the cut line 72. By turning the light-shielding cam 40A, the cut line 70 is located at a position corresponding to the lowermost position of the upper portion (the position of the cut line 70 in FIG. 16, that is, the position equal to or less than the light portion limit position at the time of so-called high beam). ) To the position corresponding to the uppermost position (the position of the imaginary line in FIG. 16, the so-called bright portion limit position in the case of a low beam). Similarly, the cut line 72 is defined by the shading cam 42.
By turning A, the uppermost position (cut line 72 in FIG.
Position) to the lowest position (position of imaginary line in FIG. 16) in parallel.

The headlamp 20 includes the actuator 4
1, 43 (FIG. 4). The configuration of the headlamp 20 is similar to that of the headlamp 18, and detailed description thereof will be omitted.

As shown in FIG. 4, the control device 50 includes a read only memory (ROM) 52, a random access memory (RAM) 54, a central processing unit (CPU) 56, an input port 58, an output port 60, and these. It is configured to include a bus 62 such as a data bus and a control bus to be connected. The ROM 52 stores a map and a control program described later.

A vehicle speed sensor 66 and an image processing device 48 are connected to the input port 58. Output port 60
Are connected to the actuators 40, 42 of the headlamp 18 and the actuators 41, 43 of the headlamp 20 via the driver 64. Also, the output port 60
Are also connected to the image processing device 48.

The image processing device 48 is a device for image-processing an image taken by the TV camera 22 based on signals input from the TV camera 22 and the control device 50 as described later.

The road shape includes the shape of the running road,
For example, it includes a road shape corresponding to one lane formed by a center line, curbs and the like.

Next, referring to the vehicle recognition traveling control routine shown in FIG. 6, the processing for recognizing the preceding vehicle 11 by image processing and performing cruise control such as constant speed traveling based on the present embodiment will be described. And explain. The position of each pixel on the image formed by the image signal is specified by the coordinates (X _n , Y _n ) of the coordinate system set on the image by the orthogonal X axis and Y axis.

FIG. 5 (1) shows an image 120 that substantially matches the image viewed by the driver when the road 122 on which the vehicle 10 is traveling is photographed by the TV camera 22. The road 122 has white lines 124 on both sides of the lane in which the vehicle 10 travels. The preceding vehicle 11 is recognized by this image 120.

When the image signal of the image 120 is input to the image processing device 48, the image processing is started, the white line candidate point extraction processing and the straight line approximation processing are performed in this order, and the traveling lane of the vehicle 10 is detected. The recognition area W _P is set (step 710). The processing of step 710 will be described.

The white line candidate point extraction process is a process of extracting candidate points estimated to be white lines of the lane in which the vehicle 10 is traveling. First, a predetermined width γ is set with respect to the position of the white line estimation line 126 obtained last time. The area it has is set as the window area W _S (see FIG. 5C). In the case of the first time, the preset value of the white line estimation line 126 set in advance is read and the window area W _S
To set. In addition, in the upper and lower areas of the image 120,
Since the probability that the preceding vehicle 11 exists is low, the upper limit line 128
And a lower limit line 130 are provided, and the range between them is set as the following processing target area. Next, the brightness is differentiated within this window region W _S , and the peak point (maximum point) of this differential value is extracted as an edge point which is a white line candidate point. That is, in the vertical direction (direction of arrow A in FIG. 5 (3)) in the window region W _S , the brightness from the pixel at the lowermost position to the pixel at the uppermost position is differentiated for each pixel in the horizontal direction to change the brightness. A peak point of a large differential value is extracted as an edge point. This continuation of edge points is represented by the dotted line 13 in FIG.
Shown in 2.

In the next straight line approximation processing, the edge points extracted in the white line candidate point extraction processing are subjected to straight line approximation using the Hough transform, and straight lines 134, 1 along the line which is estimated to be the white line.
36 is asked. Obtained straight lines 136 and 138 and lower limit line 13
A region surrounded by 0 and 0 is set as a vehicle recognition region W _P (see FIG. 5 (4)). When the road 122 is a curved road, the area surrounded by the lower limit line 130 having the inclination difference of the above-described straight lines 136 and 138 is the vehicle recognition area W.
_It is set as _P (see FIG. 5 (2)).

Next, when the white line candidate point extracting process and the straight line approximating process are completed, the horizontal edge detecting process and the vertical edge detecting process are performed in this order to determine the presence or absence of the preceding vehicle 11 in the set vehicle recognition area W _P. And the preceding vehicle 1
When there is 1, the inter-vehicle distance ΔV is calculated (step 72).
0). The processing of step 720 will be described.

The horizontal edge detection process is performed in the vehicle recognition area W _P.
First, a horizontal edge point is detected by the same processing as the above-mentioned white line candidate point detection processing. Next, the detected horizontal edge points are laterally integrated to detect a peak point E _P at a position where the integrated value exceeds a predetermined value (see FIG. 5 (5)).

The vertical edge detection processing, when the peak point E _P of the integrated value of the horizontal edge points are multiple, the peak point located below on the image E _P (closer point distance) in order, the peak point E _P The window regions W _R and W _L for detecting the vertical lines are set so as to include both ends of the horizontal edge points included in (see (6) in FIG. 5). This wind area W _R ,
Detect vertical edges in W _L and detect vertical lines 138R,
When 138L is stably detected, it is determined that the preceding vehicle 11 exists. Next, the vehicle width is obtained by obtaining the lateral distance between the vertical lines 138R and 138L detected in each of the window regions W _R and W _L , and the position of the horizontal edge of the preceding vehicle 11 and the obtained vehicle The inter-vehicle distance ΔV between the preceding vehicle 11 and the host vehicle 10 is calculated from the width. The vertical interval between the vertical lines 138R and 138L is
It can be calculated from the difference between the respective representative X coordinates of R and 138L (for example, average coordinate values or frequent coordinate values).

When the above processing is completed, the set traveling processing is executed (step 730). Step 730 is an example of processing for feedback-controlling the presence of a preceding vehicle in set traveling such as constant speed traveling control and inter-vehicle distance control.
For example, when the calculated inter-vehicle distance ΔV exceeds a predetermined value, constant speed traveling is continued, or when the inter-vehicle distance ΔV becomes equal to or less than the predetermined value, constant speed traveling is canceled. When controlling the inter-vehicle distance to a predetermined value, the vehicle speed and the like are controlled so that the inter-vehicle distance ΔV between the host vehicle 10 and the preceding vehicle 11 maintains the predetermined distance.

The operation of this embodiment will be described below. When the driver turns on a light switch (not shown) to turn on the headlamps 18 and 20, the control main routine shown in FIG. 7 is executed every predetermined time. In this control routine, the preceding vehicle is recognized in step 200 (FIG. 8), the gain of the actuator for the light distribution control for the preceding vehicle is set in the next step 300 (FIG. 9), and the oncoming vehicle is determined in the next step 400. 11A is recognized (FIG. 11), the gain of the actuator for the light distribution control for the oncoming vehicle is set in the next step 500 (FIG. 10), and the headlamp 18, based on the gain set in the next step 600, 20 is light distribution controlled.

Next, the details of step 200 will be described.
As shown in FIG. 8, in step 202, the white line detection window area W _sd is set in the same manner as above. When traveling at night, it is possible to detect only an image up to approximately 40 to 50 m in front of the vehicle 10, and it is not necessary to detect an image over 60 m in front of the vehicle 10. For this reason, in the present embodiment, the white line detection window region W _sd is detected by detecting the region up to 60 m in front of the vehicle 10, and therefore the white line detection window region W _{sd is obtained} by removing the region of the predetermined horizontal line 140 or more from the window region W _S. Is set (see FIG. 12).

In the next step 204, the approximation along the white line
Find a straight line. That is, the white line detection window area W_sdWithin
Edge points are detected, Hough transform is performed, and a linear approximation is performed.
Approximate straight lines 142 and 144 along the white line of the road 122
(See FIG. 12, see step 710 above). Next
At step 206, the intersection P of the approximate straight lines 142 and 144 _N
(X coordinate, X_N), The intersection P_NAnd the standard straight road
Intersection P of time₀(X coordinate, X₀) And horizontal displacement A
(A = X_N-X₀). This displacement A is for the road 1
It corresponds to the degree of 22 curved roads.

Next, the vehicle speed V of the host vehicle 10 is read (step 208), and the left and right correction widths α _R , α for correcting the position of the approximate straight line according to the vehicle speed V and the degree of the curved road (displacement amount A). Set _L (step 210). For example,
The degree of the curved road is determined to be a straight road, a right curved road, or a left curved road, and the correction widths α _R and α _L are set according to the respective curved road degrees. The determination of the straight road, the right curved road, and the left curved road can be made by presetting a predetermined threshold value of the displacement amount A that can be regarded as a straight road.

On a road regarded as a straight road, the radius of curvature of the road on which the vehicle can turn when traveling at a high speed is large, and it can be considered that the road is running on a substantially straight road. On the other hand, when traveling at a low speed, even if the road in front of the vehicle is close to a straight line, the radius of curvature of the road may be small in the distance, and the preceding vehicle may not be included in the recognition area by the white line approximation up to 60 m ahead. is there.
Therefore, by increasing both the correction widths α _R and α _L during low speed traveling and decreasing both during high speed traveling (see FIG. 19).
When the vehicle travels at low speed, the vehicle recognition area is made larger than that at the time of high speed travel, and the recognition area of the preceding vehicle 11 is made larger (see FIG. 1).
4).

On the road which is regarded as a right curve road,
The left and right regions where the preceding vehicle is present vary depending on the degree of the curved road (see FIG. 15). Therefore, the correction values α _R ′ and α _L ′ are determined according to the vehicle speed V (corresponding to the correction amount in FIG. 19), and the left and right gains GL and GR are determined according to the degree of the curved road (displacement amount A). (See FIGS. 20 and 21). By setting the final correction width with the correction value and the gain, the left and right correction widths α _R and α _L are set to independent values. Therefore, when the radius of curvature is small on the right curved road (the displacement amount A is large) and the probability that the preceding vehicle 11 exists on the right side is high, the correction width α _R on the right side becomes large and the correction width α _L on the left side becomes small. Become. Further, when the radius of curvature is large (the displacement amount A is small) on the right curved road, the correction width α on the right side
_R becomes smaller and the correction width α _L on the left side becomes larger.

Incidentally, the road which is regarded as the left curve road,
The characteristics are opposite to those of the road that is regarded as a right curve road. That is, the radius of curvature is small on the left curved road (the displacement amount A is large),
When the probability that the preceding vehicle 11 exists on the left side is high, the correction width α _R on the right side becomes small and the correction width α _L on the left side becomes large.

In the next step 212, the lower limit line 130,
Approximate straight lines 142 and 144 and the set left and right correction width α
A vehicle recognition area W _P for recognizing the preceding vehicle 11 is determined using _R and α _L (see FIG. 13).

When the vehicle recognition area W _P is determined as described above, the process proceeds to step 214, and horizontal edge point integration is performed within the vehicle recognition area W _P determined in the same manner as the preceding vehicle detection processing of step 720. Thus, the existing preceding vehicle is recognized and the inter-vehicle distance ΔV is calculated (step 216).

As described above, since the recognition area of the preceding vehicle 11 is changed according to the vehicle speed and the degree of the curve of the road, the obtained vehicle recognition area is a range in which there is a high probability that the preceding vehicle actually exists. It can be surely included, and the preceding vehicle can be recognized with high accuracy.

In this embodiment, when the white line cannot be detected, the vehicle recognition area based on the position of the previously detected white line is used.

Next, the details of step 300 will be described.
Step 300 is a subroutine for setting the gain (rotation amount of the light shielding cam) for moving the actuator to the position of the cut line that does not apply the grain to the preceding vehicle 11 (see FIG. 9).

First, in step 302, the inter-vehicle distance ΔV is read and the process proceeds to step 304. In step 304, the gains DEG _L and DEG _R of the actuators 40 and 42 are determined according to the inter-vehicle distance ΔV and the degree of the curve.

To determine the gains DEG _L and DEG _R of the left and right actuators 40 and 42 for the road regarded as a straight road, the gain is increased as the inter-vehicle distance ΔV increases (see FIG. 22). In this embodiment, a map 1 in which the relationship between the inter-vehicle distance and the gain in the case of a straight road is a table
Is stored in the ROM 52 as.

Further, when the road is regarded as a right-curved road, the gain D is increased according to the degree of the right-curved road (displacement amount A).
Determine EG _L and DEG _R. If the radius of curvature of the right curve of the road is large, consider it as a straight line and map 1 (Fig. 22)
Actuators 40, 4 according to the inter-vehicle distance ΔV
The gains DEG _L and DEG _R of ₂ are determined. If the radius of curvature of the right curve is small, a high probability of the presence of the preceding vehicle to the right, well control of the right light distribution (cut line 70), while the gain the DEG _R of the actuator 42 to a predetermined value, the right The gain DEG _L of the actuator 40 corresponding to the cut line control is set to a value according to the inter-vehicle distance ΔV. That is, the gain DEG _R is a predetermined value regardless of the inter-vehicle distance ΔV, and the gain DEG _L
Increases as the inter-vehicle distance ΔV increases (Fig. 2
3). In the present embodiment, the relationship between the inter-vehicle distance and the gain on the right-curved road is set as a map 2 which is a table RO.
It is stored in M52. If the inter-vehicle distance is less than a predetermined value (for example, 70 m), it is a short distance from the preceding vehicle 11, and the influence of glare due to the inter-vehicle distance is large. Therefore, the gains DEG _L and DEG _R are set to be the same as those on the straight road. decide.

When the road is regarded as a left-curved road, it is set with the characteristic opposite to that of the road regarded as a right-curved road. That is, when the probability that the preceding vehicle 11 is present on the left side is high, the gain DEG _L is set to a predetermined value regardless of the inter-vehicle distance ΔV, and the gain DEG _R is increased as the inter-vehicle distance ΔV increases. In this embodiment, this relationship is stored in the ROM 52 as the map 3 which is a table.

The degree of the straight road, the right curve road, or the left curve road can be judged as large or small by storing a judgment reference value in advance.

As described above, in this embodiment, the vehicle recognition area for recognizing the preceding vehicle existing on the road ahead of the vehicle is set from the image taken by the TV camera, and the vehicle recognition area is set according to the vehicle speed and the shape of the road. Since the vehicle recognition area is changed to recognize the preceding vehicle and the light distribution of the headlamp is changed, the optimum light of the headlamp of the own vehicle 10 can be obtained without giving glare to the driver of the preceding vehicle 11. Irradiation can be performed.

Next, details of step 400 will be described (see FIG. 11). When the oncoming vehicle recognition subroutine shown in FIG. 11 is executed, the routine proceeds to step 402, where an oncoming vehicle recognition area is set. In detail, similar to the setting of the preceding vehicle recognition area, the vehicle speed V and the degree of the curved road (displacement amount A)
The oncoming vehicle recognition area W _PO for recognizing the oncoming vehicle 11A is determined according to the above (see FIG. 17). That is, the right side area of the approximate line 144 and the oncoming vehicle recognition region W _PO, it corrects the left limit position of the oncoming vehicle recognition region W _PO set by approximate line 144.

When the road is regarded as a substantially straight road, the correction width α _RO for correcting the position of the approximate straight line is made large at low speed traveling and made small at high speed traveling as in the case of recognition of the preceding vehicle. (See FIG. 24). In this case, the oncoming vehicle recognition area W _PO during low speed traveling is wider than that during high speed traveling.

When it is regarded as a right curve road,
Correction value α according to vehicle speed V_RO'(Figure 24), degree of curved road
Gain GR according to the match (displacement A)_O(Fig. 2
5), this correction value α_RO'And gain GR_OBased on
(Multiplication) The correction on the right side that finally corrects the position of the approximate straight line
Positive width α_ROTo decide. This determined correction width α_ROUsing
Oncoming vehicle recognition area W for recognizing and processing the oncoming vehicle 11A_PO
To decide. On the other hand, if it is regarded as a left curve road,
Correction value α according to vehicle speed V_RO', The degree of the left curve road (change
Gain GR according to the unit A)_O(See Fig. 26)
), Correction value α_RO'And gain GR_OFinal based on
Correction width α_ROOncoming vehicle recognition area W _PODetermine
It

It should be noted that the characteristic of the gain for the right curved road has a gentler slope than the characteristic for the left curved road. This is because the on-vehicle presence probability is high on the right side.

When the oncoming vehicle recognition area W _PO is determined as described above, the process proceeds to step 404, and the image 120 (see FIG. 18 (1)) which is the input image is binarized. That is, the light from the headlamp of the oncoming vehicle is direct light,
Since it is relatively easy to specify the light amount, a predetermined threshold value of the image 120 (for example, 90% of the peak brightness value)
The above area is binarized as a bright area (for example, data 1) and an area less than the threshold is a dark area (for example, data 0) (see FIG. 18 (2)). Next, the expansion / contraction process is repeated a predetermined number of times (three times in this embodiment) to remove the unevenness (step 406). That is, all the boundary pixels in the bright area are deleted, and the contraction processing for removing one skin is performed, and conversely, the expansion processing for growing the boundary pixels in the background direction and making one skin thicker is performed, thereby weakly connecting areas. At the same time as separating, minute irregularities at the boundary between the bright region and the dark region are removed.

In the next step 408, labeling is performed for each bright region from which the minute irregularities have been removed (see reference numerals 1 to 3 in FIG. 18 (3)). Then, step 410
The center-of-gravity position and the area of each pixel are calculated for each of the labeled bright regions. This barycentric position can be calculated from the X coordinate value and Y coordinate value of each pixel included in the bright area, and the area can be calculated by counting the number of pixels included in the bright area. In this case, as shown in FIG. 18 (3), the bright area with the code number 1 has the barycentric value (X ₁ , Y ₁ ),
The area is S ₁ . Similarly, the light area with the reference number 2 has a barycentric value (X
₂ , Y ₁ ), the area S ₂ , and the bright region of reference number 3 has the center of gravity value (X ₃ , Y ₃ ), the area S ₃ .

Here, the oncoming vehicle 11A is usually provided with a pair of left and right headlamps, and the light emitted by the oncoming vehicle 11A toward the host vehicle 10 is a pair in a substantially horizontal direction and corresponds to the vehicle width. It is formed as bright areas at predetermined intervals. Therefore, if a pair of bright areas in the substantially horizontal direction and at predetermined intervals according to the vehicle width are detected from the image 120, the pair of bright areas is highly likely to be a headlamp of an oncoming vehicle. Therefore, in the next step 412, the coordinates of the center of gravity are substantially equal,
The oncoming vehicle 1 detects all bright area pairs whose X-coordinate distance is equal to or less than a predetermined value corresponding to the headlamp distance of a standard vehicle.
1A headlamp candidate area. In this case, the bright area pair A corresponds (see FIG. 18 (3)).

Further, the headlamp is usually arranged at a low position, and the light from the headlamp reflected on the road surface such as a road or a traveling road is also emitted to the front of the vehicle. Therefore, when the oncoming vehicle 11A is present, a bright region is formed in the image 120 below the direct light (bright region pair) from the headlamp and at a predetermined position (road surface). Therefore, if the bright region exists below the bright region pair, the presence of the oncoming vehicle 11A can be recognized with high accuracy. Further, the formation state of this bright region differs depending on the state of the road surface. For example,
On a paved road or the like, light from a pair of headlamps is scattered on the road surface to form one bright area (Fig. 1).
8). If the road surface has a high reflectance due to rain or the like, 1
Each light from the pair of headlamps is reflected on the road surface, and two bright regions are formed on the road surface (see FIG. 30 (1)). Therefore, in the next step 414, of the detected candidate regions (bright region pairs) of the oncoming vehicle 11A, one or two bright regions having an area of a predetermined value or more correspond to the lower side of the bright region pair. If there is an area pair, it is recognized as the headlamp of the oncoming vehicle 11A, and it is recognized that the oncoming vehicle 11A exists. That is, in the case of FIG. 18, the bright area pair A (reference numeral 1,
The presence of the oncoming vehicle 11A is recognized by recognizing the bright region pair A as the headlamp of the oncoming vehicle 11A by the presence of the bright region (reference numeral 3) corresponding to the second bright region. In the case of rain or the like, as shown in FIG. 30 (2), the bright area pair B (the bright areas of the numerals 4 and 5) corresponding to the bright area pair B (the numerals 6 and 7) exists. B oncoming vehicle 11A
Oncoming vehicle 11
A recognizes the existence.

When this oncoming vehicle 11A is recognized, it is recognized as the oncoming vehicle 11A in the next step 416,
The inter-vehicle distance ΔV from the own vehicle to the oncoming vehicle is calculated based on the Y coordinate of the bright area pair A (the average value of the coordinate values Y ₁ and Y ₂ ). That is, as the inter-vehicle distance increases, the coordinate value shifts upward in the image 120, and the ratio is TV.
It is proportional to the shooting magnification of the camera 22. In addition, inter-vehicle distance ΔV
Is the interval of the bright area pair A (difference between coordinate values X ₁ and X ₂ ),
X for standard vehicle width (headlamp spacing)
It can also be calculated from the ratio of the distance on the coordinates.

In this way, the oncoming vehicle 11A is recognized in the oncoming vehicle recognition area W _PO determined according to the degree of the curved road and the vehicle speed. During this recognition process,
In the image (image) taken by the TV camera 22,
Even when a plurality of light spots are formed by reflected light from other than the external light or the vehicle, each bright area of the pair of headlamps is detected, and the road surface existing below the bright area pair is detected. When there is a bright region in the reflection part of the above, it is recognized as the headlamp of the oncoming vehicle and the oncoming vehicle is recognized. As described above, in the present embodiment, it is possible to extract only a bright region that is an oncoming vehicle with high accuracy, and it is possible to more reliably recognize the oncoming vehicle.

Next, the details of step 500 will be described.
Step 500 is a subroutine for setting the gain (rotation amount of the light shielding cam) for moving the actuator to the position of the cut line that does not give the grain to the oncoming vehicle 11A (see FIG. 10).

First, in step 502, the inter-vehicle distance ΔV
Is read and the process proceeds to step 504. In step 504, the gains DEG _L and DEG _R of the actuators 40 and 42 are determined according to the inter-vehicle distance ΔV and the degree of the curve.

When the oncoming vehicle 11A is considered to be a straight road or a right-curved road, it is highly likely that the oncoming vehicle 11A is present on the right side of the screen, and therefore gray is not generated by changing the actuator 42 with respect to the left cut line. Absent. Therefore, in determining the gains DG _L and DG _R , as shown in FIG. 27, only the gain DG _L for the actuator 40 with respect to the cut line on the right side is increased as the inter-vehicle distance ΔV increases. In this embodiment, the relationship between the inter-vehicle distance and the gain shown in FIG. 27 is stored in the ROM 52 as the map 4 which is a table.

Further, in the case of a left-curved road, even if the left-side cut line 72 is changed on the left-curved road with a small curvature, the contribution to Glen is small, so that the gain DG _L , regardless of the inter-vehicle distance ΔV, DG _R is determined to be a predetermined value (Fig. 2
8). In this embodiment, the ROM 5 is used as the map 5 which is a table showing the relationship between the inter-vehicle distance and the gain shown in FIG.
It is stored in 2.

When the degree of curve on the left curved road is small, the gains DG _L and DG _R are set to values according to the inter-vehicle distance ΔV (see FIG. 29). In this embodiment, FIG.
The relationship between the inter-vehicle distance and the gain shown in is stored in the ROM 52 as a map 6 which is a table.

Next, the details of step 600 will be described.
When the determination of the gains of the actuators 40, 42 for the preceding vehicle 11 and the oncoming vehicle 11A is completed as described above, in step 516, one of the gains DEG _L , DG _L for the actuator 40 and the gain DEG _R for the actuator 42, Select a small gain for either DG _R. By controlling the actuators according to the gains of the selected actuators 40 and 42, the light shielding cams of the actuators 40 and 42 are moved to change the light distribution of the headlamp 18.

As described above, in this embodiment, the vehicle recognition area for recognizing the preceding vehicle is further added with the oncoming vehicle recognition area for recognizing the oncoming vehicle to recognize the vehicle existing in the road ahead of the vehicle. As a result, the headlamp of the host vehicle 10 can optimally irradiate the light without giving glare to the oncoming vehicle.

In the above embodiment, the light distribution cam in front of the vehicle is controlled by the light shielding cam, but the light of the headlamp may be shielded by a light shielding plate or a shutter. Further, although the light distribution is controlled by blocking the light of the headlamp, the emission optical axis of the headlamp may be deflected.

Further, in the above embodiment, the case where the oncoming vehicle is a vehicle which runs on the left side of the road ahead of the host vehicle and runs under the road regulation is described. However, the present invention is not limited to this, and the right side It can be easily applied to vehicles passing by.

Further, in the above embodiment, the data of the white line of the road, which is the initial data, is stored as the data when the vehicle runs on the straight road where the flat line and the line of the predetermined width are provided on both sides of the vehicle. Even if the white line cannot be detected at the time of detection, the standard recognition area can be set. Further, by storing a plurality of patterns of this data and selecting them, the recognition area can be determined by the driver's setting.

Although the oncoming vehicle is recognized by the headlamp of the oncoming vehicle in the above embodiment, the present invention is not limited to this, and an auxiliary light such as a fog lamp is detected to recognize the oncoming vehicle. You may do it.

[0078]

As described above, according to the present invention, the oncoming vehicle is recognized by recognizing the light of the oncoming vehicle such as the headlight or fog lamp of the oncoming vehicle and the light reflected from the road surface by the light. Because
Even if there are multiple light spots in the image, it is possible to extract a highly accurate area that is the light of the oncoming vehicle,
There is an effect that the oncoming vehicle can be accurately recognized.

[Brief description of drawings]

FIG. 1 is a perspective view of a front portion of a vehicle used in the present embodiment as seen obliquely from the front of the vehicle.

FIG. 2 is a schematic perspective view of a headlamp to which the present invention is applicable.

FIG. 3 is a schematic sectional view of a headlamp (II in FIG. 2;
Line).

FIG. 4 is a block diagram showing a schematic configuration of a control device.

FIG. 5 is an image diagram for explaining a process of recognizing a preceding vehicle based on an image signal output by a TV camera photographed in the daytime.

FIG. 6 is a flowchart showing a preceding vehicle recognition processing routine based on an image signal of a TV camera photographed in the daytime.

FIG. 7 is a flowchart showing a control main routine of this embodiment.

FIG. 8 is a flowchart showing a preceding vehicle recognition processing routine of the present embodiment.

FIG. 9 is a flowchart showing a gain setting subroutine of a preceding vehicle.

FIG. 10 is a flowchart showing a gain setting subroutine for an oncoming vehicle.

FIG. 11 is a flowchart showing an oncoming vehicle recognition processing routine of the present embodiment.

FIG. 12 is a diagram showing a window area at the time of recognizing a white line.

FIG. 13 is a diagram showing a vehicle recognition area.

FIG. 14 is an image diagram for explaining that the vehicle recognition area is changed according to the vehicle speed.

FIG. 15 is an image diagram showing window regions and correction widths for curved roads having different curvatures.

FIG. 16 is an image diagram for explaining a cut line displaced by an actuator.

FIG. 17 is an image diagram showing an oncoming vehicle recognition area of the present embodiment.

FIG. 18 is an image diagram showing a process of recognizing an oncoming vehicle according to the present embodiment.

FIG. 19 is a diagram showing the relationship between the vehicle speed and the correction width (correction value) of the window region in the present embodiment.

FIG. 20 is a diagram showing the relationship between the degree of a right curve road and the gain that determines the correction width on the right side of the window.

FIG. 21 is a diagram showing the relationship between the degree of a right curve road and the gain that determines the correction width on the left side of the window.

FIG. 22 is a diagram showing a relationship between an inter-vehicle distance and a control gain of an actuator.

FIG. 23 is a diagram showing a relationship between an inter-vehicle distance and a control gain of an actuator.

FIG. 24 is a diagram showing the relationship between the vehicle speed and the correction width (correction value) of the window region in the present embodiment.

FIG. 25 is a diagram showing the relationship between the degree of a left curved road and the gain that determines the correction width on the right side of the window.

FIG. 26 is a diagram showing the relationship between the degree of a right curve road and the gain that determines the correction width on the right side of the window.

FIG. 27 is a diagram showing a relationship between an inter-vehicle distance and a control gain of an actuator.

FIG. 28 is a diagram showing a relationship between an inter-vehicle distance and a control gain of an actuator.

FIG. 29 is a diagram showing a relationship between an inter-vehicle distance and a control gain of an actuator.

FIG. 30 is an image diagram showing a picked-up image of an oncoming vehicle and a bright region by a headlamp in case of rain or the like.

[Explanation of symbols]

 18, 20 Headlamp 40, 42 Actuator 22 TV camera 48 Image processing device 50 Control device 66 Vehicle speed sensor 100 Traveling vehicle recognition device

Claims (1)

[Claims] 1. An image pickup means for picking up an image in front of a vehicle, a candidate area extracting means for extracting a pair of bright areas horizontally separated from the picked-up image by a predetermined distance as a candidate area for a light of an oncoming vehicle, the candidate. When there is a bright area below the area,
An oncoming vehicle recognition device comprising: an oncoming vehicle recognition unit that recognizes the candidate area as a light of an oncoming vehicle.