JPH1115952A - Travel lane recognition device - Google Patents

Abstract

(57) [Problem] To provide a traveling lane recognition device capable of correctly recognizing a traveling lane at a low cost by avoiding the influence of a preceding vehicle when recognizing a white line by image processing. SOLUTION: Image information picked up by an image pickup means 2 is processed by an image information processing means 4 and an estimating means 5 based on left and right white line positions on a front road obtained from the processed information.
In the traveling lane recognition device for estimating the traveling lane position with respect to the vehicle, there is provided an obstacle position detecting means 3 for detecting an obstacle on the road ahead, and the estimating means 5 detects the obstacle by the obstacle position detecting means 3. Then, when the obstacle is unevenly distributed on either the left or right in the travel lane obtained and estimated, the white line information on the side of the left and right white lines separated from the obstacle position is displayed. The driving lane position is configured to be mainly estimated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a driving method for determining a driving lane defining shape specified by a driving lane marking represented by a white line on a road from image information obtained by photographing with a television camera or the like. More particularly, the present invention relates to a travel lane recognition apparatus that performs travel lane recognition in consideration of the presence of an obstacle such as a preceding vehicle existing on a road ahead.

[0002]

2. Description of the Related Art In recent years, a television camera has been installed in an automobile to grasp the shape of a running road and the attitude of the vehicle with respect to the road, and to use it for automatic driving control of the automobile.
Techniques have been developed that are used for various warnings to the driver. For example, of image information from a camera installed in a vehicle, a portion having high brightness is a white line candidate point, and a point closest to the white line position (white line recognition information) obtained in the previous screen information is recognized as a white line. Then, based on this road white line, the traveling lane was grasped and steering was performed,
A technique relating to road white line recognition (running lane recognition) is disclosed in Japanese Patent Application Laid-Open No. 3-137798.

[0003] Also, JP-A-7-85249 discloses that
A brightness change is searched in the horizontal direction in the white line search area based on the image information, and a portion where the brightness change is extremely small is determined to be a road portion, and the brightness change is adjacent to a road portion where the brightness change is extremely small. There is disclosed a technique for recognizing a portion where is large as a white line. Such a road white line recognition device (traveling lane recognition device) includes a camera (imaging means) and an ECU for processing image information from the camera.
(Electronic control unit), an image (actual image) is first obtained by a camera, and this image is processed by the ECU. At this time, the ECU converts the image (original image) into a two-dimensional image in plan view. Then, the road white line of the original image is converted into the road white line of the planar view image, and white line recognition is performed.

[0004]

When recognizing a white line (a lane marking) on a road, it is assumed that there is no obstacle in the space between the camera and the white line. That is, if there is an obstacle on the optical axis from the white line to the camera, the white line cannot be captured by the camera, and the white line cannot be recognized naturally. There may be a preceding vehicle running or a stopped vehicle in front of the road on which the vehicle actually travels.Therefore, other vehicles in front of the road may become obstacles, making it difficult to recognize the white line. There is a problem that a white line is erroneously recognized, such as an edge portion of a vehicle being recognized as an edge of a white line.

Therefore, when recognizing a white line (a lane marking), it is desired to avoid the influence of a preceding vehicle on the road so that the white line can be recognized more correctly. Japanese Patent Application Laid-Open No. Hei 8-30770 discloses that an image area obtained by using a television camera and which includes an obstacle such as a preceding vehicle is processed in an image area excluding the prohibited area. In this way, a technique for avoiding the influence of a preceding vehicle on a road ahead in recognizing a lane marking is disclosed. However, in this technique, the logic for setting the processing prohibited area or the like is likely to be complicated, and a high processing capability is required for the image processing apparatus, resulting in an increase in cost.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-described problems, and recognizes the influence of a preceding vehicle on a road ahead without imposing a large burden on an image processing system when recognizing a white line (traveling lane) by image processing. It is an object of the present invention to provide a traveling lane recognition device which can avoid the traveling lane and can recognize the traveling lane more accurately at low cost.

[0007]

According to a first aspect of the present invention, there is provided a driving lane recognizing apparatus, comprising: an image pickup means for picking up an image of a road ahead of a vehicle; and an image for processing image information picked up by the image pickup means. Information processing means; and estimating means for estimating a traveling lane position for the vehicle based on left and right white line positions on the front road obtained from information processed by the image information processing means. The captured image information is processed by the image information processing means, and the estimating means estimates a travel lane position for the vehicle based on left and right white line positions on the front road obtained from the processed information.

[0008] Further, there is provided an obstacle position detecting means for detecting the position of the obstacle when there is an obstacle on the road ahead, and the estimating means includes an obstacle position detecting means for detecting the position of the obstacle. When the obstacle and its position above are detected, if the obstacle is unevenly distributed to the left or right in the travel lane that has been estimated and obtained, the obstacle in the left and right white lines The travel lane position is estimated mainly using the white line information on the side separated from the object position.

That is, if the obstacle position is unevenly located on the left white line side, the travel lane position is estimated mainly based on the right white line information, and if the obstacle position is unevenly located on the right white line side. The travel lane position is estimated mainly based on the left white line information. Thus, the travel lane position can be estimated without being affected by the obstacle.

In the driving lane recognizing device according to the present invention, the estimating means determines whether or not the obstacle approaches the vehicle within a white line recognition distance based on the detection information of the obstacle position detecting means. Since the travel lane position is estimated mainly based on the white line information on the side separated from the obstacle position, the obstacle is located beyond the white line recognition distance, that is, the obstacle does not affect the white line recognition. Then, the travel lane position is estimated based on the left and right white line information.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described below with reference to the drawings. FIGS. 1 to 5 show a traveling lane recognition apparatus as an embodiment of the present invention and an automatic lane recognition apparatus using the traveling lane recognition apparatus. 3 shows a steering device. FIG.
As shown in FIG. 1, the vehicle 1 has a camera 2 as an imaging means for imaging the road condition ahead of the vehicle, and an obstacle position detection for detecting the position of an obstacle such as a preceding vehicle if there is an obstacle ahead of the vehicle. Means 3, image information processing means 4 for appropriately processing image information from image information from the camera 2 to recognize left and right white line positions on the front road, and white line position image information and obstacles by the image information processing means 4. Position detecting means 3
Traveling lane estimating means (estimating means) 5 for estimating the position and characteristics of the traveling lane and the positional relationship with the vehicle from the obstacle position information from the vehicle, a steering actuator 21 for turning the steered wheels 22, and traveling A controller 6 for controlling the steering actuator 21 based on the recognition result of the lane estimating means 5 is provided. In addition, 20 is a steering wheel.

The obstacle position detecting means 3, the image information processing means 4, the traveling lane estimating means 5, and the controller 6
The electronic control unit includes a CPU, an input / output interface, a ROM, a RAM, and the like. The traveling lane recognition device includes a camera 2 as an imaging unit, an image information processing unit 4, an obstacle position detection unit,
And estimating means 5.

Note that the camera 2 is installed slightly downward in front of the front of the vehicle, and is capable of photographing in a predetermined photographing range on the road ahead of the vehicle. Obstacle position detection means 3
Has a laser radar 3A and a radar information processing unit 3B that processes information from the laser radar 3A.
In particular, the laser radar 3A detects whether or not there is any object such as a preceding vehicle on the road ahead of the vehicle 1 (these are referred to as obstacles), and if an obstacle exists, the position of the obstacle and the obstacle. To detect the distance to.

That is, the detection direction of the laser radar 3A is directed to the front of the vehicle body so as to correspond to the camera 2,
The distance of the object captured by the laser radar 3A from the radar can be detected, and the position of the object captured by the laser radar 3A on the radar is grasped in correspondence with the image of the traveling lane position captured by the camera 2. Is available.

The radar information processing unit 3B associates the position on the radar of the object captured by the laser radar 3A with the image related to the travel lane position captured by the camera 2, and compares the position of the object in front with the travel lane position already grasped by the object ahead. To determine which one is left or right. As shown in FIG. 2, the image information processing means 4 takes in the original image 41 from the camera 2, extracts a road white line from the original image 41, and looks at the extracted image of the road white line from above vertically. The image is converted into a simple planar view image 42. The extraction of the road white line will be described later.

The estimating means 5 estimates the position of the driving lane based on the position information of the white lines 12L and 12R as roadside lines at the left or right end of the driving lane recognized by the image information processing means 4 in the plan view image 42. And the area defined by the extracted left and right road white lines 12L and 12R, or the extracted left and right road white lines 12L and 12R.
An area defined from one of the 12Rs and a lane width recognized in advance is estimated as a traveling lane.

In particular, the estimating means 5 uses the obstacle information such as the preceding vehicle detected by the obstacle position detecting means 3 to determine the distance from the vehicle to the object such as the preceding vehicle and the lateral position of the object (already detected). The travel lane position is estimated in accordance with the travel lane information based on the travel lane information. In the controller 6, the left-right deviation (lateral deviation) of the own vehicle, the direction of the own vehicle (declination angle β), The curvature of the traveling lane (road curvature) is calculated, and automatic steering control is performed based on the calculation results.

As described above, the travel lane position is estimated by
The recognition is performed based on the position information of the white lines 12L and 12R as roadside lines at the left end or the right end of the traveling lane recognized by the image information processing means 4. Here, recognition of the white lines 12L and 12R will be described. Here, the recognition of the white line 12L as the roadside line at the left end of the traveling lane will be described. However, the same applies to the case where the white line 12R at the right end of the traveling lane is used as a reference. I will call it.

In the image information recognizing means 4, first, FIG.
As shown in FIG. 1A, a camera 2 provided on a vehicle 1 uses a camera 2 to cover a flat area (eg, 5 m to 30 m).
m), the image in the vertical direction on the screen is partially omitted from the image information. Then, a plurality of horizontal lines 11 are set at equal intervals on this screen. The capture of the black-and-white image information is updated every minute control cycle. As shown in FIG. 3B, on each horizontal line 11, the left and right positions of the white line position on the previous screen are required. (Here, the left and right 50 pixels [do
t]) is set as a white line search area (processing target area) 10. For the first screen, the white line position on the straight road is used as the previous screen data.

Then, as shown in FIG. 3C, the brightness of each horizontal line is differentiated in the horizontal direction from the left. Reference numeral 14 in the figure is a guardrail. By the way, a normal road surface has low luminance and a small change in luminance. On the contrary,
Since the brightness of the white line 12 is much higher than that of the normal road surface, when the brightness of the road is differentiated in this way, the luminance change is positive at the boundary point from the normal road surface to the white line 12, and the white line 12 changes to the normal road surface. Differential data is obtained such that the luminance change becomes negative at the boundary point of. FIG. 3D shows an example of such differential data.

Then, for each data of each horizontal line 11, peaks of the differential values appear in the order of plus and minus from the left, and the interval between the peaks is the white line 12
As a white line candidate, a combination that fits within a range that is considered appropriate (the interval from the positive peak to the negative peak is, for example, within 30 dots) is extracted, and FIG.
As shown in (e), the midpoint is stored as a white line candidate point 15.

Then, among these white line candidate points 15,
Only the point closest to the screen center is left as the final candidate point.
This means that, for example, when the vehicle 1 is traveling on the left side, the search area 1
The right side of 0 is a road surface with little change in normal luminance, and the white line candidate point 15 closest to this normal road surface can be determined as the white line 12. Therefore, further to the left of the white line 12,
Objects that cause noise (for example, guardrails 14)
Is present, the white line 12 can be reliably recognized from the image information captured by the camera 2.

Finally, as shown in FIG. 3F, the vertical continuity of the white line candidate points 15 in each horizontal line data is sequentially verified from the bottom of the screen. First, the inclination between the upper and lower ends of the white line 12 on the previous screen is calculated in advance.
Then, assuming that the lowest point 15A is the white line 12, it is verified whether or not the candidate point 15B on the upper horizontal line 11 is within the range of ± 50 dots of the inclination of the previous white line 12.

If the candidate point 15B falls within this range, it is regarded as a white line. If not, the candidate point 15B is rejected, and the coordinates interpolated from the above-mentioned inclination are regarded as the white line position. By performing the same operation for each horizontal line for this detection, a continuous white line 12 can be recognized. Such white line recognition work is continuously performed at a required cycle, and the recognition of the white line 12 is updated each time.

White line 12 as the roadside line at the right end of the traveling lane
The recognition of R is performed in the same manner. The image information processing means 4 converts the white lines 12R and 12L on the original image 41 recognized in each recognition cycle in this way into a planar view image 42, and estimates the road center line LC from the white line 12L at the left end of the traveling lane. based on the road centerline LC _R which may be estimated from _L and the traveling lane right end of the white line 12R, and performs the estimation of the road center line LC according to the detection information from the obstacle position detection means 3.

That is, in the obstacle position detecting means 3,
When the laser radar 3A captures an object (obstacle) 16 such as a preceding vehicle on the road ahead, it detects the distance from the vehicle to the object 16 and the position of the object 16 on the radar. Also, as shown in FIG. 4, the radar information processing unit 3B associates the position on the radar of the object 16 captured by the laser radar 3A with the image of the traveling lane position already grasped from the camera 2 information. Then, it is determined whether the object 16 is unevenly located left or right with respect to the traveling lane position already grasped.

In this case, since the traveling lane in a predetermined range (for example, 5 m to 30 m) ahead of the vehicle has already been grasped, the radar information processing unit 3B places the captured object 16 in this range (that is, for example, 30 m from the vehicle). ), The position of the object 16 with respect to the traveling lane position (road center) at this point (around 30 m from the vehicle) is determined. If the center 16A in the left and right direction of the object 16 captured by the radar 3A is separated from the already-estimated traveling lane center LCF at a predetermined distance (30 m) by a predetermined distance (L0) or more from the vehicle, the corresponding direction Is determined to be unevenly distributed.

In the example shown in FIG. 4, the center 16A of the object 16 in the left-right direction is more than the distance L1 (L
1> L0), and it can be said that the object 16 is unevenly distributed on the left white line 12L side. The estimating means 5 estimates the road center line LC according to the detection information from the obstacle position detecting means 3.

1. If there is no obstacle such as a preceding vehicle ahead of the vehicle or if there is an obstacle such as a preceding vehicle ahead of the vehicle but this obstacle is not unevenly distributed on the left or right on the road ahead, each road center line LC _L, the center line of the road by averaging the LC _R (e.g. simple average) LC (= LC _L + _L
C _R ) is calculated. 2. Case, information of better obstacle is not unevenly distributed among the road center line LC _L, LC _R vehicle forward there is an obstacle such as a preceding vehicle the obstacle are unevenly distributed in the right or left on the road ahead The road center line LC is calculated according to.

That is, if the center position of the obstacle position is to the left of the road center line LC calculated so far (see FIG. 4), the road center line LC based on the white line on the left side where the obstacle position exists is located. in order that _L be unreliable, the remaining road centerline LC _R, based on the right of the white line to adopt a road center line LC. Conversely, if the center position of the obstacle position is on the right side of the road center line LC calculated so far, the road center line LC _R based on the white line on the right side where the obstacle position exists.
Since the reliability decreases, the remaining road centerline LC _L based on the left of the white line to adopt a road center line LC.

However, 1. (When there is no obstacle such as a preceding vehicle in front of the vehicle and when there is an obstacle in front of the vehicle but it is not unevenly distributed on either side) (When there is an obstacle such as a preceding vehicle in front of the vehicle and this obstacle is unevenly distributed on either side), the road center line estimated up to that point LC
It is necessary to ensure continuity with For this reason, a low-pass filter 43 capable of smoothing the output of the estimated road center line LC information is provided.

When the road center line LC is estimated as described above, the lateral deviation and the deflection angle β are calculated based on the road center line LC, and automatic steering is performed.
By the way, the controller 6 controls the steering actuator 21 based on the traveling lane position information for the vehicle 1 estimated by the estimating means 5 to perform automatic steering.

That is, the vehicle 1 is an auto-steering vehicle capable of automatically turning the steered wheels 20 based on the image taken by the camera 2, and the controller 6 uses the vehicle estimated by the estimating means 5. A steering angle to be taken by the vehicle (this is referred to as a target steering angle) is set from the positional relationship between the vehicle 1 and the traveling lane, and the operation of the steering actuator 21 is controlled in accordance with the set target steering angle to steer the steered wheels 20. It is supposed to.

Here, the control by the automatic steering device provided in the vehicle 1 will be described. This automatic steering device performs automatic steering control following artificial steering operation (driver's steering operation). I have. For example, the driver performs the steering operation (change of the steering angle) mainly when the traveling direction of the vehicle 1 is corrected to be inconsistent with the direction of the traveling road (traveling lane). This is also performed to correct this when the vehicle 1 is going out of the running lane to the left or right. The reason that the running direction does not match the running lane direction is mainly during running on a curved road, but even when running on a straight running path, the running direction deviates from the running lane direction due to the attitude of the vehicle itself moving in the yaw direction In some cases.

Therefore, in this automatic steering system, steering is mainly performed so that the traveling direction of the vehicle 1 matches the direction of the traveling lane, and a steering element for correcting the lateral position of the vehicle 1 is added to this steering. It is supposed to. In the case of artificial steering (driver steering), the relative operation between the vehicle and the road is determined based on information visually obtained by the driver, and the steering operation is performed. That is, the driver
Based on the information obtained from the eyes, the relative relationship (deviation) between the traveling direction of the vehicle 1 and the traveling lane direction and the lateral displacement (lateral deviation) of the vehicle 1 are determined and corrected. The steering operation is being performed.

By organizing the information visually obtained by such a driver, it can be classified into three elements: the radius of the curve of the road, the vehicle speed, and the riding comfort (lateral acceleration, lateral jerk). Further, in general, when the vehicle is traveling on a steady circle on a curved road, a portion close to the circle, a straight running road, or the like (during steady traveling), the steering angle is to be kept constant. The steering is performed so as to maintain the steering angle δ _{0 corresponding} to the curvature.

On the other hand, when an attempt is made to enter a curved road from a straight road (at the time of transitional traveling), the steering angle δ is gradually reduced to a steering angle δ ₀ corresponding to the curvature of the curved road from a point short of the curved road. To increase. In this case, the driver determines how far before the point where the vehicle enters the curved road (ie, how much the steering start distance D is to be set), and at what ratio (steering speed δ _V ) Whether to increase the angle δ is set in consideration of the vehicle speed at that time, the expected riding comfort (the degree of application of lateral acceleration), and the like.

That is, when the steering elements determined by the driver at the time of the steering operation are classified, they are divided into a steering angle δ ₀ , a steering start distance (steering start timing) D, and a steering speed δ _V. In the automatic steering device according to the present embodiment, automatic steering is performed by a method similar to such driver steering. For this reason, the estimation means 5 corresponding to the driver's visual system and the respective amounts required for steering, that is, the steering angle δ ₀ , the steering start distance D, and the steering speed δ _V are set to correspond to the driver's data processing system. It has a controller 6 to be set.

Then, based on the road center line LC in the planar view image 42 estimated by the estimating means 5, a point separated from the vehicle 1 by a predetermined distance (ie, a point at a predetermined height in the planar view image 4B). Is calculated. The declination β is, as shown in FIG. 5, a curved road center line L
The reference line position information at a first detection point (shown as perigee in the figure) which is a predetermined distance away from the vehicle, and the angle between the tangent to C and the direction of the vehicle center line. It can be calculated from reference line position information at a second detection point (shown as an apogee in the figure) separated by a predetermined amount.

That is, the declination β is calculated as an angle between a straight line connecting the first detection point and the second detection point and the center line of the vehicle 1. The declination calculated in this way is a declination at an intermediate point (x mark in the figure) between the first detection point (perigee) and the second detection point (apogee), and is at least a certain distance or more in front of the vehicle 1. The declination of a point. In this example, the point closest to the vehicle on the road center line LC based on the image information from the camera 2 is set as the first detection point, and the lateral deviation is calculated based on the first detection point. The curve radius R is estimated based on the deviation angle β.

Since the driving lane recognizing device as one embodiment of the present invention is configured as described above, the white line recognition by the present device is performed as follows. In the present apparatus, white line recognition is performed on the white line 12L at the left end of the running lane and the white line 12R at the right end of the running lane. Based on the white line recognition, the position of the running lane in which the vehicle is running with respect to the vehicle is determined. (In other words, the position of the vehicle with respect to the driving lane) is estimated. First, recognition of the white lines 12L and 12R will be described using the left white line 12L as an example.

First, as shown in FIG.
On a flat ground, the area in front of the vehicle (for example,
The monochrome image information m) is fetched at every minute control cycle, and a plurality of horizontal lines 11 are set at regular intervals on this screen at each cycle. Then, as shown in FIG.
On each horizontal line 11, a required range of the left and right positions of the white line position on the previous screen (for example, left and right 50 pixels [do
t]) is set as a white line search area (processing target area) 10. In the initial screen, a white line position on a straight road is used as previous screen data.

From the image information, as shown in FIG. 3C, the brightness of each horizontal line is differentiated in the horizontal direction from the left, and the differential data of each horizontal line is obtained as shown in FIG. 3D.
From the left], the peaks of the differential values appear in the order of plus and minus from the left, and the interval between the peaks is considered to be appropriate as the white line 12 (the interval from the plus peak to the minus peak is, for example, within 30 dots). ) Are extracted as white line candidates, and the midpoint thereof is stored as a white line candidate point 15 (see FIG. 3E).

Then, among these white line candidate points 15,
Only the point closest to the screen center is left as the final candidate point.
By limiting the white line candidate point 15 to the one closest to the center of the screen in this way, an object that causes noise (for example, the guardrail 14 or a vehicle in another traveling lane) exists outside the white line 12. Even in this case, the white line 12 can be reliably recognized from the image information obtained by the camera 2.

Finally, as shown in FIG. 3 (f), the continuity of the white line candidate points 15 in each horizontal line data in the vertical direction is sequentially verified from the bottom of the screen. First, the inclination between the upper and lower ends of the white line 12 on the previous screen is calculated in advance. If the lowermost point 15A is the white line 12, the candidate point 15B on the upper horizontal line 11 is calculated by the inclination ± 5 of the previous white line 12.
0 dot is compared, and candidate point 1
If 5B is within this range, it is regarded as a white line. If not, the candidate point 15B is rejected, and the coordinates calculated by interpolation from the above-mentioned inclination are regarded as the position of the white line.

By performing such an operation for each horizontal line, continuous white lines 12 can be recognized. Such white line recognition work is continuously performed at a required cycle, and the recognition of the white line 12 is updated each time.
In this manner, the white lines 12L, 1 on the left and right of the traveling lane are periodically repeated.
Recognition of 2R is performed in the same manner.

[0047] and, with the center line of the road LC _L estimated from the white line 12L of the traveling lane leftmost lane right edge of the white line 12
R based on the road center line LC _R estimated from the road center line LC _R based on the detection information from the obstacle position detecting means 3.
Is estimated. That is, in the obstacle position detection means 3, when the laser radar 3A captures an object such as a preceding vehicle on the road ahead, it detects the distance from the vehicle to the object and the position of the object on the radar. . In the radar information processing unit 3B, the laser radar 3
The camera 2 captures the position of the object on the radar on the radar 2 and associates it with the image of the traveling lane position already grasped, and determines whether the object is skewed left or right with respect to the traveling lane position already grasped. judge.

Here, since the traveling lane in a predetermined range (for example, 5 m to 30 m) ahead of the vehicle has already been grasped, the radar information processing unit 3B places the captured object within this range (for example, within 30 m) from the vehicle. )
It is determined which position the object is at relative to the traveling lane position. If the center of the object captured by the radar 3A is separated from the already estimated running lane center line by a predetermined distance to the left or right, it is determined that the object is unevenly distributed in the corresponding direction.

The estimating means 4 determines whether or not there is an obstacle such as a preceding vehicle in front of the vehicle and there is an obstacle such as a preceding vehicle in front of the vehicle according to the detection information from the obstacle position detecting means 3. obstacle when not unevenly distributed to the right or left on the road ahead calculates the road center line LC _L, the center line of the road on average LC _R LC a _{(= LC L + LC R)} . Further, when the front of the vehicle there is an obstacle such as a preceding vehicle the obstacle are unevenly distributed in the right or left on the road ahead, each road center line LC _L, better obstacle of LC _R is not unevenly distributed The center line LC is calculated according to the information. That is, the center position of the obstacle position is the road center line LC calculated so far.
If the left than the remaining road centerline LC _R, based on the right of the white line to adopt a road center line LC. Conversely, if the right of the center line of the road LC center position is calculated to its obstacle position, the remaining left road centerline LC _L based on the white line to adopt a road center line LC.

Based on the road center line LC estimated in this way, the lateral deviation is calculated by the lateral deviation calculating means 7, the deflection angle β is calculated by the deflection angle calculating means 8, and the curvature of the traveling lane is calculated by the curvature state estimating means 9. (Curvature) is calculated, and automatic steering control can be performed based on the calculated information. In this way, in the traveling lane recognition device,
The road center line LC is estimated based on highly reliable white line recognition information without an obstacle position, without being based on the white line recognition information with reduced reliability due to the presence of an obstacle. There is an advantage that the estimation of the center line LC, that is, the estimation of the traveling lane can be performed with higher accuracy.

Further, since the radar laser 3A, which is the main component of the obstacle position detecting means 3, is already installed in a vehicle that performs inter-vehicle distance control or the like, there is no need to newly install the radar laser 3A. In addition, the driving lane can be recognized more accurately without imposing a large burden on the image processing system, at a low cost, and without being affected by the preceding vehicle on the road ahead.

Further, the road center line LC information estimated by the estimating means 5 is smoothed and output by the low-pass filter 43, so that the position of the road center line LC is smoothly continuous without abrupt change. Thus, control based on the position of the road center line LC can be stabilized. In addition,
In the present embodiment, the left and right white lines 12L, recognizes 12R, road center line LC _L based on each, after estimating the LC _R, the road center line LC on the basis of the detection information of the obstacle position detection means 3 Is estimated, but the left and right white lines 12
From the stage of recognizing L, 12R, the obstacle position detecting means 3
Based on the detection information, the white line on the side where the obstacle position exists is not recognized, and only the white line on the side where the obstacle position does not exist is recognized. Based on the recognized white line position, The configuration may be such that the road center line LC is estimated. In this case, the load on the image processing system is further reduced.

[0053]

As described above in detail, according to the driving lane recognizing device of the first aspect of the present invention, the image information picked up by the image pickup means is processed by the image information processing means, and further, the image information is picked up by the estimating means. From the detection information of the obstacle on the road ahead detected by the obstacle position detection means, if the obstacle is unevenly distributed to the left or right in the traveling lane already recognized, the side separated from the obstacle position Since the travel lane position is estimated mainly based on the white line information, it is not based on the white line recognition information whose reliability has been reduced due to the presence of an obstacle, but based on the highly reliable white line recognition information without an obstacle position. Thus, there is an advantage that the road center line LC can be estimated, and the estimation of the road center line LC, that is, the estimation of the traveling lane can be performed more accurately.

According to the driving lane recognizing device of the present invention, when an obstacle approaches within the white line recognition distance from the vehicle, the white line information on the side separated from the obstacle position is mainly used. When the travel lane position is estimated and the obstacle is located at a position that does not affect the recognition of the white line beyond the white line recognition distance, the travel lane position can be estimated based on the left and right white line information. The lane position estimation accuracy can be improved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a traveling lane recognition device as one embodiment of the present invention.

FIG. 2 is a schematic configuration diagram of a traveling lane recognition device as one embodiment of the present invention.

FIG. 3 is a schematic diagram illustrating road white line recognition by a driving lane recognition device as one embodiment of the present invention.

FIG. 4 is a schematic diagram of a plan view image of a road for explaining detection of an obstacle position in the traveling lane recognition device as one embodiment of the present invention.

FIG. 5 is a schematic diagram of a plan view image of a road explaining a declination and a lateral deviation calculated from an estimation result by the traveling lane recognition device as one embodiment of the present invention.

[Explanation of symbols]

 Reference Signs List 1 vehicle 2 camera as imaging means 3 obstacle position detection means 3A laser radar 3B radar information processing unit 4 image information processing means 5 traveling lane estimation means (estimation means) 12, 12L, 12R white line 16 object (obstacle) LC road Center line

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Takashi Ota Mitsubishi Motors Corporation

Claims (2)

[Claims] 1. An imaging means for imaging a road ahead of a vehicle,
Image information processing means for processing image information captured by the image capturing means; and a traveling lane position for the vehicle based on left and right white line positions on the front road obtained from the information processed by the image information processing means. A traveling lane recognizing device having an estimating means for estimating, wherein an obstacle position detecting means for detecting a position of the obstacle when an obstacle is present on the road ahead is provided; When the obstacle position and its position on the road ahead are detected by the object position detection means, if the obstacle is unevenly located on either the left or right of the traveling lane obtained by estimation, The travel lane recognition device is characterized in that the travel lane position is estimated mainly based on white line information on a side of the left and right white lines separated from the obstacle position. 2. The method according to claim 1, wherein the estimating means includes a white line on a side separated from the obstacle position when the obstacle approaches the vehicle within a white line recognition distance based on the detection information of the obstacle position detecting means. The travel lane recognition apparatus according to claim 1, wherein the travel lane position is estimated mainly based on the information.