JP2002092796A - Lane recognizing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a lane recognizing device capable of discriminating the occurrence of the misrecognition, and surely recognizing a lane. SOLUTION: This lane recognizing device for recognizing a lane 6 where a self-car 5 drives, on the basis of the image information, comprises plural cameras 11, 12 mounted on a vehicle 5 for photographing different parts (area A, area B) of road compartment lines on 61, 62 at left and right of the lane 6, plural road compartment lines recognizing parts respectively mounted corresponding to these cameras 11, 12 and recognizing the road compartment lines 61A, 61B or 62A and 62B in an image, and a determining part for determining the success or failure of the recognition by collating a result of the recognition of the road compartment line recognizing part (extended position 61A2 and 61B2, 62A2 and 62B2).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a lane recognition device used in an automatic driving system or a lane departure warning system.

[0002]

2. Description of the Related Art In an automatic driving system or a lane departure warning system, the lane (traveling lane) in which the vehicle is traveling.
Need to recognize. Such a lane is usually demarcated by a demarcation line such as a white line or a yellow line drawn on a road surface by painting. Therefore, it is possible to recognize a lane by recognizing such lane markings, and various apparatuses and methods for recognizing such lane markings from image information acquired by a camera have been put to practical use.

The technique disclosed in Japanese Patent Application Laid-Open No. 9-91594 is also an example of such a technique, in which a result of recognition of a front white line and information on the behavior of the own vehicle can be obtained without installing a dedicated camera for recognition of a rear white line. There is disclosed a technique for estimating rear white line position information on the basis of the information.

[0004]

The recognition of a white line by an image cannot completely eliminate the possibility of erroneous recognition. Factors of this erroneous recognition include backlight, road surface reflection, reflection of lights from other vehicles, and the like. In the above-described technique, when such erroneous recognition occurs, not only the front white line recognition result but also the rear white line position information is shifted, and it is difficult to recognize the occurrence of erroneous recognition.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a lane recognizing device capable of judging occurrence of erroneous recognition and performing reliable lane recognition.

[0006]

In order to solve the above-mentioned problems, a lane recognizing device according to the present invention is a lane recognizing device for recognizing a lane in which a vehicle is traveling from image information.
(1) A plurality of cameras installed on the vehicle and photographing different portions of the road lane markings provided on the left and right of the lane, and (2) a road lane marking in the image provided for each of these cameras. It is characterized by comprising: a plurality of road lane marking recognition units to be recognized; and (3) a judging unit for judging the success or failure of recognition by collating each recognition result of the road lane marking recognition unit.

According to the present invention, different portions of the road lane marking are photographed using a plurality of cameras, and the position information of the road lane markings is obtained by image recognition for each portion. If the respective image recognitions are successful, no inconsistency should occur between the acquired position information, and if there is an inconsistency, it can be determined that one of the recognition results has an error. For example, 3 for the same road division line
If there are no inconsistencies in the recognition results for the majority of cameras when more than one camera is installed, the results can be used as correct recognition results.

[0008] The determination unit may determine the success or failure of recognition by extending the lane markings recognized by each of the lane marking recognition units to the vicinity of the vehicle and comparing the extended positions. This is because if there is a discrepancy in the recognition result, it is considered that a mismatch also occurs in the extension position itself.

It is preferable that one of the plurality of cameras is a rearward imaging means for parking assistance. By using the rear imaging means for parking assistance for lane recognition during normal traveling, the rear imaging means for parking assistance can be effectively utilized.

[0010]

Preferred embodiments of the present invention will be described below in detail with reference to the accompanying drawings. In order to facilitate understanding of the description, the same constituent elements are denoted by the same reference numerals as much as possible in each drawing, and redundant description will be omitted.

FIG. 1 is a block diagram showing a configuration of a main part of a lane recognition device according to the present invention. The present apparatus includes two cameras 11 and 12 arranged on the front and rear sides, and white line recognition ECUs 21 and 22 dedicated to the cameras 11 and 12, respectively.
It is composed of an erroneous recognition determination ECU 3 connected to these ECUs.

FIG. 2 is a diagram for explaining the concept of erroneous recognition determination in the lane recognition device. As shown in FIG. 2, the lane 6 is divided into left and right lane markings 61 and 62.
Consider the case where the vehicle is traveling. The front camera 11 installed in front of the vehicle, for example, in the center of the front of the vehicle compartment, acquires an image of the hatched area A. This front camera 1
Reference numeral 1 denotes a camera for recognizing a road condition in the traveling direction of the vehicle when the vehicle travels forward. Therefore, the area A is set so as to cover a position relatively far from the vehicle.
On the other hand, the rear camera 12 installed at the rear of the vehicle, for example, at the center of the rear part of the vehicle compartment, acquires an image of the hatched area B. The rear camera 12 functions as a camera for confirming the rear when the vehicle is parked backward and the like. Therefore, the area B is set to cover a position closer to the vehicle than the area A with a wide angle.

The front white line recognition ECU 21 can recognize the left and right division line portions 61A and 62A in the area A, while the rear white line recognition ECU 22 can recognize the left and right division line parts in the area B. 61B and 62B.
The erroneous recognition determination ECU 3 extends the white lines recognized by the white line recognition ECUs 21 and 22 to obtain the positions directly below the respective vehicles (61A ₂ , B ₂ , 62A ₂ and B ₂ in FIG. ₂ ). From these deviations, the presence or absence of erroneous recognition is determined.

FIG. 3 is a flowchart of the process.
This process is repeated from the time the ignition key is turned on to the time the ignition key is turned off.

In step S1, frame images (corresponding to the images of the areas A and B shown in FIG. 2) acquired by the front camera 11 and the rear camera 12 are sent to the front white line recognition ECU 21 and the rear white line recognition ECU 22, respectively. In step S2, the position information of the white lines 61 and 62 is acquired from the frame images acquired by the white line recognition ECUs 21 and 22 by image recognition processing such as feature amount extraction processing. More specifically, the front white line recognition ECU 21 recognizes the white line portions 61A and 62A in the area A as the rear white line recognition E.
CU22 includes 61B and 62B corresponding to white line portions in area B.
And the position information thereof is erroneously recognized and determined by ECU3.
Notify.

In step S3, it is first determined whether or not the recognition processing in each of the white line recognition ECUs 21 and 22 in step S2 has been successful. If any of the white lines cannot be recognized because the white line is thin or interrupted, and the position information is not notified to the misrecognition determination ECU 3, the processing is successfully performed by one of the ECUs 21 and 22. It is determined that there is no recognition, the process proceeds to step S4, the recognition is unsuccessful, and the process ends without outputting the remaining recognition result.

If both recognition processes have succeeded in step S3, the process proceeds to step S5, in which each of the recognized white lines is extended to obtain a position immediately below the extension destination vehicle. For example, when taking the case of the white line 61A as an example, determine the position of the extended position 61A ₂ in the lateral center of gravity W of the vehicle to extend from the nearest position 61A ₁ to the vehicle of the recognized white line 61A the white lines 61A . Similarly, white line 6
Extension position 62A _{2 of} 2A, extension position 61 of white line 61B
Obtaining B _2, the extended position 62B ₂ of the white line 62B, respectively.

In step S6, each white line thus obtained is determined.
Position immediately below the vehicle, that is, 61A _TwoAnd 61B_Two, 62A
_TwoAnd 62B_TwoIt is determined whether or not each matches.
If the lane is a straight line, it is acceptable if each recognition is successful.
If these are the lane marking position 61C just below the actual vehicle,
Should match 62C. However, as shown in FIG.
Like lane 6 itself is curved or white line image recognition
It is impossible to find a perfect match due to the effect of
Therefore, both are within a predetermined threshold (for example, 20 cm).
If so, it is determined that they match, and the process proceeds to step S7.
And migrate. In step S7, the front camera 11
Recognized by the recognized white lines 61A and 62A and the rear camera 12.
White lines recognized by connecting the recognized white lines 61B and 62B
By further correcting the line position, the actual vehicle position 6
The position of the white line is output closer to 1C and 62C.

On the other hand, if the distance is greater than the threshold value,
It is considered that any of the white line recognitions has failed, and the process proceeds to step S4 to make the recognition unsuccessful, and the process ends without outputting the remaining recognition results.

The white line position information recognized by the front white line recognition ECU 21 from the image obtained by the front camera 1 and the rear white line recognition ECU 22 from the image obtained by the rear camera 2
By estimating the position immediately below the vehicle based on the recognized white line position information and comparing the results, it is possible to verify the recognition result and suppress erroneous recognition.

Here, an example in which the recognition result is corrected when the recognition is determined to be successful has been described. However, when the recognition is determined to be successful, it is possible to obtain a wider range of images and to realize accurate recognition. Alternatively, only the recognition result of the front camera 11 may be output. If only one of the lanes matches, only one of the lanes may be output normally. When it is determined that the recognition is erroneous, the output may not be stopped, and each or one of the position information may be output together with the information indicating the possibility of the erroneous recognition.

In the above-described embodiment, an example has been described in which the road marking line recognized by each of the front camera and the rear camera is extended to the vicinity of the vehicle to compare the extension positions. Based on the behavior of the vehicle (vehicle speed, direction of travel, travel trajectory, etc.), the position of the lane marking recognized by the front camera should be recognized by the rear camera as the host vehicle advances, and the actual rearward position is estimated. You may make it compare with the recognition position in a camera. In this way, since it is not necessary to extend the road lane marking recognized by the rear camera, the recognition result by each camera can be accurately verified even when the road lane marking recognized by the rear camera is very short.

In the above description, the case where two cameras, a front camera and a rear camera, are used has been described as an example, but a side camera may be used instead of either one. When a side camera is used, only one of the left and right white lines is detected, so that at least one camera must be installed on each of the left and right. As the rear camera, it is preferable to use a camera that performs parking assist at the time of retreat.

Alternatively, the white lines may be recognized by acquiring images of the left and right white lines at three or more different regions (which may partially overlap) using three or more cameras. When three or more cameras are used for recognition, the extension positions of the white lines recognized from the images of the cameras to the area directly below the vehicle do not need to match, and even if there is a partial mismatch, the majority will match. In such a case, the recognition result may be used using the matched result.

The position information of the white line recognized in this way can be suitably used in a lane departure warning system or an automatic vehicle traveling system.

[0026]

As described above, according to the present invention, a plurality of (partially overlapping) portions of a lane marking (white line) that divides the lane in which the vehicle is traveling are captured by different cameras. Detecting white line position information by image recognition, collating the detected position information and verifying each recognition result to determine the success or failure of recognition and suppressing erroneous recognition to perform reliable lane recognition Can be.

[Brief description of the drawings]

FIG. 1 is a block diagram schematically showing a lane recognition device according to the present invention.

FIG. 2 is a diagram illustrating the principle of the present invention.

FIG. 3 is a flowchart illustrating an operation of the apparatus of FIG. 1;

[Explanation of symbols]

3: misrecognition determination ECU, 5: vehicle, 6: lane, 11: front camera (camera), 12: rear camera (camera), 2
1 ... front white line recognition ECU (white line recognition ECU), 22 ... rear white line recognition ECU (white line recognition ECU), 61, 62 ... white line (partition line).

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) B60R 21/00 624 B60R 21/00 624G 626 626A 628 628D G06T 1/00 330 G06T 1/00 330A 7/60 2007/60 200J (72) Inventor Makoto Nishida 1st Toyota Town, Toyota City, Aichi Prefecture Inside Toyota Automobile Co., Ltd. (72) Inventor Toshiaki Kakinami 2-1, Asahimachi, Kariya City, Aichi Prefecture Aisin Seiki Co., Ltd. (72) Inventor Shoko Kojima 1-41, Chukuji-Yokomichi, Nagakute-cho, Aichi-gun, Aichi Prefecture F-term in Toyota Central R & D Laboratories, Inc. 5B057 AA16 BA02 BA11 CA12 DA08 DB02 DC01 DC32 5H180 AA01 CC04 LL01 LL02 LL04 5L096 BA04 BA18 CA05 DA03 FA67 FA69 FA76 HA01 JA26

Claims (3)

[Claims] 1. A lane recognizing device for recognizing a lane in which a vehicle is traveling from image information, comprising: a plurality of cameras installed on the vehicle for photographing different portions of a lane marking provided on the left and right of the lane; A plurality of road lane marking units provided for each of the cameras for recognizing road lane markings in an image, and a recognition result is determined by comparing each recognition result of the road lane marking recognition unit. A lane recognition device comprising: a determination unit. 2. The lane recognition according to claim 1, wherein the determination unit determines the success or failure of the recognition by extending the road lane marking recognized by each of the road lane marking recognition units to the vicinity of the vehicle and comparing the extension positions. apparatus. 3. The lane recognition device according to claim 1, wherein one of the plurality of cameras is a rearward imaging unit for parking assistance.