DE102007019531A1 - Lane detection with cameras of different focal lengths - Google Patents

Abstract

The invention relates to an evaluation device for a driver assistance system for a vehicle (1), having an input for receiving first image information (I-K1) which has been detected by a first camera (K1) oriented in the forward direction of the vehicle with a first focal length, and an input for receiving second image information (I-K2) detected by a second forward-facing vehicle (K2) having a second focal length. Furthermore, a determination component is provided for determining a course of a lane lane (SP1) using the first image information (I-K1) and the second image information (I-K2).

Description

The The invention relates to an evaluation device for a driver assistance system for a vehicle.

Under the concept of driver assistance systems (English: ADAS, Advanced Driver Assistance Systems) functions are summarized which to assist the driver of a motor vehicle. The aim of the driver assistance systems is often the increase safety by avoiding dangerous situations their emergence and the support of the driver for accident prevention in critical situations. Other goals are the increase of comfort through stress reduction and relief the driver in standard situations, facilitating orientation by situation-dependent prepared and driver-appropriate mediated Environment information, as well as the increase of driving pleasure.

Examples for driver assistance functions are traction control or traction control such as ABS (antilock braking system), ASR (traction control), ESP (Electronic Stability Program), EDS (Electronic Stability Program) Differential lock), as well as adaptive headlights, fade-in and dimming assist for the driving lights, night vision systems (English: Night Vision), Cruise Control, Park Assist, Brake Assist, ACC (Adaptive Cruise Control) distance control, distance warning, turn assistant, traffic jam assistant, Lane Detection System, Lane Keeping Assist, Lane Keeping Assist, Lane Change Assistant, ISA (Intelligent Speed Adaptation), ANB (Automatic Emergency braking), curve assistant, tire pressure monitoring system, driver condition detection, Traffic sign recognition, Platooning.

Of the Invention is based on the object, an efficient evaluation device for a driver assistance system, and a corresponding Driver assistance system, a method for operating a driver assistance system and to show a computer program product.

These Task is performed by an evaluation device with the characteristics of Claim 1, and by a driver assistance system, by a method and a computer program product with features of siblings Claims solved. Advantageous embodiments and further developments are the subject of dependent claims.

The Inventive evaluation device comprises a Input for receiving first image information, wherein the first Image information from a first forward vehicle-oriented camera with a first focal length and an input for receiving second image information, wherein the second image information is from a second in the forward direction of the vehicle-facing camera with a second opposite captured the first focal length larger focal length were. Furthermore, a determination component is available for determining a course of a lane lane using the first image information and the second image information.

The Evaluation device receives first and second image information. This image information may be the original one act on the respective camera captured data, or by a already processed version of the originally recorded Dates.

The both cameras are facing forward. This means that they at least partially before in a drive in the forward direction the vehicle lying or when driving in the reverse direction Capture the environment behind the vehicle. Thus, both form the first as well as the second picture parts of the before the Vehicle lying environment. Each of the two cameras has one certain focal length, with the focal length of the second camera larger than that of the first camera. is the focal length of one or both cameras variable, so at least for the first and the second image information that the focal length used for the second image information larger than that of the first image information is. The focal length of the cameras determines the dimensions of the Detail of the environment, which the respective camera captures. In a schematic representation of this section as a two-dimensional Circular sector determines the focal length of the opening angle and the radius of the circular sector. According to the ratio the focal lengths correspond to the first image information of a closer vehicle environment and the second image information of a more distant environment of the vehicle.

The Evaluation device uses both data of the first camera as also data from the second camera to the course of a lane to investigate. At the lane track whose course is determined It may be the lane that the vehicle is currently travels, and / or z. B. to an adjacent lane. The course determined by the evaluation device can be determined by the Evaluation device output via a suitable output become. From the evaluation of the course is preferably indicated relative to the vehicle; alternatively, an indication of the absolute Position of the course of the lane track possible. The determined History can z. B. the course of the middle of each lane be, or the right or left boundary of the respective lane track.

In Further development of the invention is radiation from the first camera detected in the wavelength range of visible light. It is possible that the first camera exclusively is sensitive in the wavelength range of visible light, or mainly in this wavelength range, or u. a. in this wavelength range. additionally or alternatively, radiation from the first camera may be in the wavelength range the near infrared are detected. Here it is in particular possible that the first camera mainly in the Wavelength range of visible light and in smaller To begin in the wavelength range of the near infrared sensitive is. It is understood that the first image information is that of the vehicle environment in the respective wavelength range, which is detectable by the first camera emitted radiation depict.

In Further development of the invention is radiation from the second camera detected in the wavelength range of the near infrared. It is possible that the second camera exclusively is sensitive in the wavelength range of the near infrared, or mainly in this wavelength range, or u. a. in this wavelength range. It goes without saying that the second image information is that of the vehicle environment in the respective wavelength range, which of the second Camera is detectable, imaging emitted radiation.

one Further development of the invention according to the evaluation device designed for determining the course of the lane track based at least a lane marking and / or at least one lane edge. Road markings are also called road marking or ground mark; it's on the road surface attached characters. Examples are continuous or interrupted Centerlines, or two or more parallel lines. In which Road edge can be z. B. a strip of grass and / or Earth, a guardrail or a wall act. By a suitable image evaluation algorithm may be included in the image information after for a lane marking and / or a roadside edge characteristic features are sought.

In Embodiment of the invention form the first image information different from the second image information Parts of the at least one lane marking and / or at least a roadside edge. It is possible that overlaps between the ones in the first image information and the one in the second Image information contained parts of road marking and / or Roadside edge exist.

however There are parts of road markings and / or roadside which only in the first image information and not in the second image information are included, and / or parts, which only in the second Image information and not included in the first image information are.

It it is possible that the first image information first from the second image information not shown parts of the at least a lane marking and / or the at least one lane edge map, and that the second image information is second of the first image information not pictured opposite the first parts further removed from the vehicle parts of at least a lane marking and / or the at least one lane edge depict. Each of the road marking and / or the roadside containing image information can then be used to determine the course of the lane track are used.

The Evaluation device can be designed to determine the course a lane track within the first image information and the Determining the course of a lane track within the second Image information, as well as to combine using the first image information and using the second image information determined courses. When combining can the courses z. B. juxtaposed, if they refer to different parts of the course. Be progressions determined at matching positions, z. Legs Averaging done.

The inventive driver assistance system comprises an evaluation device of the type explained, the first and the second camera, and an application unit for use the course of the lane track determined by the evaluation device as part of a driver assistance function, in particular in the context a lane departure warning or a lane departure warning assistant.

The The vehicle according to the invention comprises a driver assistance system of the type explained, as well as headlights, which in terms of their broadcast to one of the first camera and / or the second Camera detectable wavelength range are adjusted. For example, infrared lights may be present when the second camera is an infrared camera.

In the method for operating a driver assistance system, first image information is acquired by a first camera oriented in the forward direction of the vehicle, a second camera aligned in the forward direction of the vehicle with a second focal length greater than the first focal length, second image information is acquired, and Use of the first and second image information becomes a course of a lane lane averages.

The Computer program product according to the invention a driver assistance system has the functionality an input for receiving first image information, which from a first in the forward direction of the vehicle aligned Camera with a first focal length were captured, an input for receiving second image information, which is from a second in the forward direction of the vehicle aligned second Camera with a second opposite the first focal length larger focal length were recorded, as well as one Determination component for determining a course of a lane using the first image information and the second image information. Under a computer program product may be related to the present invention in addition to the actual computer program (with his about the normal physical interaction between Program and computing unit going beyond technical effect) in particular a record carrier for the computer program, a file library, a configured arithmetic unit, as well For example, a storage device or a server on the or the files associated with the computer program are to be understood.

The Computer program product according to the invention and the Inventive methods are particularly suitable for the evaluation device according to the invention, this also applies to the embodiments and developments can. For this purpose, they may comprise further suitable means or steps.

in the The following is the invention with reference to an embodiment explained in more detail. Showing:

1 : a motor vehicle with a driver assistance system,

2 : a section of a driver assistance system.

This in 1 illustrated motor vehicle 1 drive currently on the right lane SP1, in 1 to the right. At the right edge of the right lane SP1 is the lane boundary MR. This may be z. B. to act a lane marking in the form of a line or the edge of a strip of grass and / or earth. Between the right lane SP1 and the left lane SP2 is the lane marker MM.

The vehicle 1 has a video camera K1, z. B. a CMOS camera. The camera K1 captures images mainly in the wavelength range of the visible spectrum. This can be accomplished by placing a suitable filter in front of the sensor of the camera K1, or by the sensor being sensitive only in said wavelength range. The camera K1 can be a black and white camera, which is less expensive than a color camera. Moreover, the resolution of a black-and-white camera is larger for a given number of pixels than for a color camera. For the driver of the vehicle 1 the use of a black and white camera is not disadvantageous in that it is not displayed the image captured by the camera 1.

The camera K1 is directed in the direction of travel, thus taking pictures of the front of the vehicle 1 lying on the environment. For this purpose, it may, for. B. near the roof and the windshield of the vehicle 1 be mounted. B1 designates the region which is shown schematically by the camera K1 and which is determined by the focal length of the camera K1. This is so large that the detection area B1 up to about 50 meters in front of the vehicle 1 extends.

Using an image processing or lane detection algorithm suitable for this purpose, the images recorded by the camera K1 are evaluated and the lane boundary MR and / or the lane marking MM are thereby detected. Because of this detection, the course of the road boundary MR and / or road mark MM and thus the course of the road lane SP1 with respect to the own vehicle 1 known. For this purpose, known methods can be used, for. As described in the document WO 2007/002964 A2 ,

Of the detected course of the lane SP1 can, for. For example the driver assistance function lane departure warning (English: Lane Departure Warning) or Lane Keep Assist (English: Lane Keep Assist) be used. Here, the driver in an unintentional Being warned to leave his lane.

A stable lane detection using the camera K1 is in good visibility conditions during the day to about 50 meters in front of the vehicle 1 possible, at night to about 35 meters. The range limitation at night is in particular due to the limited illumination of the vehicle surroundings by the low beam of the vehicle 1 conditionally. For improved performance of the driver assistance functions Lane Departure Warning, Lane Keeping Assist or Adaptive Cruise Control (ACC), it is desirable to extend the course of the lane SP1 beyond the area B1 detected by the camera K1 to know. This is especially true in the presence of a lane curvature. For this purpose, it is possible, by extrapolation, to extend the course of the traffic lane SP1 determined within the detection range of the camera K1 beyond this range. However, this is associated with great inaccuracies; the extrapolated lane course may differ significantly from the actual one. In particular, the extrapolated lane curvature may be opposite to the actual one if the inflection point of the lane lane is outside the range B1 detected by the camera K1.

In addition to the camera K1, the vehicle has 1 via another video camera K2. The K2 camera captures images mainly in the Near Infrared (NIR) wavelength range, ie 0.7-1.4 μm. This can be effected by placing a suitable filter in front of the sensor of the camera K2, or by the sensor being sensitive only in said wavelength range. The camera K2 is directed in the direction of travel, thus taking pictures of the front of the vehicle 1 lying on the environment. For this purpose, it may, for. B. near the roof and the windshield of the vehicle 1 be mounted.

The camera K2 can z. As in the context of the driver assistance function Night Vision serve to increase driving safety at night or fog rides. For this purpose, the images taken by the camera K2 infrared images to the driver of the vehicle 1 be displayed on a display. This exploits the fact that the range of the near-infrared portion of radiation emitted by an incandescent lamp of the vehicle headlight is higher than in the visible range. In addition, infrared lights S can be provided, which illuminate the environment with infrared radiation.

With B2 is the schematically represented detected by the camera K2 area designated by the focal length of the camera K2. This is such that the detection area B2 is up to about 200 meters in front of the vehicle 1 extends. The focal length of the camera K2 is larger than that of the camera K1. Accordingly, the field of view of the camera K2 is narrower than that of the camera K1, but extends further forward. The larger focal length of the camera K2 is particularly suitable for the aforementioned driver assistance function Night Vision, since in this case of the dipped beam of the vehicle 1 not illuminated area should be made visible to the driver.

2 shows a section of the driver assistance system of the vehicle 1 , The evaluation component S is provided by the camera K1 with the information I-K1 and the camera K2 with the information I-K2. The information I-K1 and I-K2 may be the image data captured by the respective camera K1 or K2, or data resulting from a processing of this image data. The evaluation component S evaluates the information I-K1 and I-K2 and determines the course of the lane SP1 relative to the own vehicle on the basis of this evaluation 1 , The evaluation unit S outputs to the application unit A the information IS which can be taken from the course of the lane SP1. The application unit A uses the information IS in terms of the respective driver assistance function, in a lane departure warning driver assistance function z. B. to determine an unintended lane change of the vehicle 1 ,

The application unit A is connected to an interface I. This may be a man-machine interface and / or an interface for the autonomous control of the vehicle 1 act. About a man-machine interface can be issued in an audible, visual or haptic manner to the driver information or warning, z. B. due to leaving the lane SP1. Through an interface to the autonomous control of the vehicle can be used in the driving behavior of the vehicle 1 be intervened without the need for the involvement of the driver, z. B. by slowing down or steering the vehicle 1 ,

As already explained, determines the evaluation component S using the information I-K1 within the coverage area B1 of the camera K1 lying course of the lane SP1. As well determines the evaluation component S using the information I-K2 lying within the detection range B2 of the camera K2 Course of the lane SP1. The determination of the course takes place each based on a detection of road boundary MR and / or Lane marker MM. The detection of lane boundary MR and / or Road marking MM is both within the images of the visible Wavelength range of the camera K1, as well as within the images of the near infrared range of the camera K2 possible, because these pictures are very similar to each other. Especially appear in the visible wavelength range bright or dark objects also in the near infrared as light or dark.

The evaluation component S merges the lane tracks obtained from the two image information items I-K1 and I-K2, so that the entire course of the lane track SP1 within the coverage areas B1 and B2 results therefrom. This can be done by connecting the courses in non-overlapping areas. In overlapping areas, however, either the course obtained from the image information I-K1 or the history obtained from the image information I-K2, or a combination of both.

This means that the evaluation component S the course of the lane SP1 of the own vehicle 1 until the end of the detection range B2 of the camera K2 can determine. This is not possible using only the image information I-K1, since the detection range B1 of the camera K1 is significantly smaller due to the smaller focal length. Even using only the image information I-K2 this is not possible because the detection range B2 of the camera K2 is narrower due to the larger focal length, so that in the nä heren environment of the vehicle 1 the road boundary MR and / or road mark MM as in 1 can not be detected by the camera K2.

to Capturing the image information I-K2 can be the infrared headlights S be used so that the camera K2 reflected from the environment Radiation of the infrared headlights S detected. Detects the camera K1 in addition to the visible light at least partially radiation of the near infrared range, so also in terms of the Camera K1 the infrared headlights S for the determination the course of the lane SP1 be used. As already mentioned is the detection range B1 of the camera K1 at Night severely limited due to the low range of the low beam. This restriction can be removed or at least reduced when the infrared lights S are turned on.

Across from the use of only the camera K1 results in the explained Proceeding an enlargement of the area, within which the course of the lane SP1 is known. Out the measurement of the course of the lane SP1 in larger Distances result in a significant improvement over an extrapolation method in which only the Near-range determined lane data on the course of the lane is closed at a greater distance. hereby is an earlier and more precise reaction of the application unit A possible within the scope of the respective driver assistance function.

The The invention has been described in an embodiment described. It is understood that many changes and modifications are possible without the scope of the Is left invention.

QUOTES INCLUDE IN THE DESCRIPTION

This list The documents listed by the applicant have been automated generated and is solely for better information recorded by the reader. The list is not part of the German Patent or utility model application. The DPMA takes over no liability for any errors or omissions.

Cited patent literature

• WO 2007/002964 A2 [0028]

Claims (12)

Evaluation device (S) for a driver assistance system for a vehicle ( 1 ) having an input for receiving first image information (I-K1) detected by a first forward-facing camera of the vehicle (K1) at a first focal length, an input for receiving second image information (I-K2) detected by a second forward camera of the vehicle (K2) having a second focal length greater than the first focal length, a determination component for determining a lane course (SP1) using the first image information (I-K1) and the second image information ( I-K2). Evaluation device (S) according to claim 1, wherein of the first camera (K1) radiation in the wavelength range of visible light is detected. Evaluation device (S) according to claim 1 or 2, wherein from the first camera (K1) radiation in the wavelength range the near infrared is detected. Evaluation device (S) according to one of the claims 1 to 3, wherein from the second camera (K2) radiation in the wavelength range the near infrared is detected. Evaluation device (S) according to one of the claims 1 to 4, wherein the determination component is designed to determine the course of the lane track (SP1) based on at least one lane marking (MM) and / or at least one roadside edge (MR). Evaluation device (S) according to claim 5, wherein the first image information (I-K1) relative to the second image information (I-K2) different parts of the at least one road mark (MM) and / or the at least one lane edge (MR). Evaluation device (S) according to claim 5 or 6, wherein the first image information (I-K1) first of the second image information (I-K2) not shown parts of the at least one lane marking (MM) and / or the at least one lane edge (MR) , and the second image information (I-K2) second from the first image information (I-K1) not shown in the first parts of the vehicle ( 1 ) depict remote parts of the at least one lane marking (MM) and / or of the at least one lane edge (MR). Evaluation device (S) according to one of the claims 1 to 7, wherein the determination component is adapted to determine the course of a lane track (SP1) within the first image information (I-K1) and for determining the course of a lane (SP1) within the second image information (I-K2), as well as for combining the using the first image information (I-K1) and using the second image information (I-K2) determined gradients. Driver assistance system for a vehicle, comprising an evaluation device (S) according to one of the claims 1 to 8, the first camera (K1) and the second camera (K2), one Application unit (A) for using the of the evaluation device (S) determined course of the lane (SP1) in a Driver assistance function. Vehicle ( 1 ) with a driver assistance system according to claim 9, and headlights (S), which are adapted in terms of their radiation to one of the first camera (K1) and / or the second camera (K2) detectable wavelength range. Method for operating a driver assistance system for a vehicle, taking from a first in the forward direction of the vehicle-oriented camera (K1) with a first focal length first image information (I-K1) is captured, from a second in the forward direction of the vehicle-oriented camera (K2) with a second opposite the first focal length larger focal length second image information (I-K2) be detected using the first image information (I-K1) and the second image information (I-K2) is a course of a lane lane (SP1) is determined. Computer program product for a driver assistance system for a vehicle, with an entrance to receive of first image information (I-K1), which is from a first in Forward direction of the vehicle-oriented camera (K1) with recorded at a first focal length, an entrance to receive of second image information (I-K2), which is from a second in Forward direction of the vehicle-oriented camera (K2) with a second larger compared to the first focal length Focal length were recorded, an investigative component of the Determining a course of a lane lane (SP1) using the first image information (I-K1) and the second image information (I-K2).