DE102008046545A1 - Method for calibrating assembly for monitoring environment of vehicle, involves detecting environment of vehicle according to multiple image detection units - Google Patents

Abstract

The method involves detecting environment of a vehicle (F) according to multiple image detection units (K1,K2,K3), whose detection area partially overlaps and forms overlapping areas (U12,U13,U23). The image detection units are calibrated by distinctive points or edges (E,MK,FR,DK,KK,KH,A) on the vehicle. The image detection units are calibrated by distinctive points or lines (ML,SL) on a roadway, which are detected by the image detection units. An independent claim is included for a device for executing a method for calibrating an assembly for monitoring an environment of a vehicle.

Description

The The invention relates to a method and a device for calibration an arrangement for monitoring an environment of a vehicle, wherein captures the environment using multiple image capture units is, where from using the image acquisition units captured individual images an overall image is generated on the basis of an image processing unit, which the vehicle and its surroundings from a bird's eye view shows.

commercial vehicles for freight or passenger transport are common due to large vehicle dimensions and a low clarity characterized. Even modern passenger cars are due their outer shape often a very small clarity on. As a result, in particular a shunting with these Vehicles for a driver is very difficult. Therefore It is advantageous to view vehicles and their surroundings from a bird's eye view to display on a screen so that the vehicle and its entire environment are visible to the driver.

Out The prior art includes various methods and devices to such monitoring and representation of an environment a vehicle, wherein a driver of the vehicle in particular a picture of the vehicle and its surroundings from a bird's-eye view is issued. This is an improved for the driver All-round view created this as an assistance function or support during driving, for example, when maneuvering the Vehicle, serves. Furthermore, accidents can be avoided which are often due to poor visibility, especially for large and difficult to manage Vehicles, happen.

Around to ensure proper functioning of these devices Calibration of the individual image acquisition units is required. To a calibration of the image acquisition units are in the state Various methods and devices are known in the art.

Such a method and a device are in the US 5,444,478 disclosed. In this case, a mobile camera is simulated, which represents a wide angle of a view, wherein the overall image is created from individual image acquisition units generated by several imaging units. For this purpose, a brightness of the individual images is fixed to one another, whereby an equal brightness is achieved in the overall image. In addition, during calibration of the image acquisition units, grayscale and color balance adjustments are made to uniformly represent the overall image. The method as well as the device can be used in a vehicle to present it in a bird's-eye view.

Of the Invention is based on the object, a method and an apparatus indicate, by means of which with little instrumental and personal Effort a calibration of individual image acquisition units feasible and high quality of the issued Overall picture is achieved.

The Task is according to the invention in terms of the method in the claim 1 and in terms of the device by the im Claim 6 specified features solved. Advantageous developments The invention are the subject of the dependent claims.

In a method for calibrating a device for monitoring an environment of a vehicle, wherein the environment based on several Image capture units, whose coverage areas at least partially overlap and an overlap area form, is detected, by means of the image acquisition units Captured frames based on an image processing unit an overall picture is generated, which the vehicle and its surroundings from a Bird's eye view, the image acquisition units are using Marked points and / or edges on the vehicle calibrated by positions and / or gradients of these prominent points and / or edges in a current image section of the respective image capture unit with positions and / or gradients of these prominent points and / or edges in a model image are compared, or one A plurality of image acquisition units are identified by means of prominent points and / or Lines calibrated on a roadway, which of this plurality Captured image capture units.

This Method with the two possibilities for calibration The image capture units of the vehicle replaced a previously common Method in which the image acquisition units at the end of a production line calibrated on which the vehicle was manufactured. The Previous method is very complex and can not used without considerable effort, for example in a workshop when, for example, an image capture unit is replaced has been. By means of this method of calibration of the image acquisition units At the end of the production line are the image capture units calibrated only once this time. Possible later Deviations from these calibrated optimal settings can since no further calibration is done, can not be corrected.

The inventive method made possible light a simple calibration of the image acquisition units without additional aids or additional effort. Since this calibration is possible constantly or at regular intervals, in particular during the journey, it is ensured by the method according to the invention that the image acquisition units of the vehicle are optimally adjusted at all times. By means of this method, for example, a recalibration of the image acquisition units is feasible, which are necessary after some time, for example, due to chassis changes that occur in a new vehicle after a long break-in and the state of the art, a complex recalibration of image acquisition units, for example, in a suitable workshop require. Even after similar changes, for example, after a change of a chassis or tires of the vehicle, the image capture devices may need to be recalibrated. This is possible by means of the method according to the invention simple, fast, effective and without time, comfort, and cost burdens. These always optimally calibrated image acquisition units allow optimal monitoring of the environment of the vehicle, which is associated with a significant gain in security.

With the elimination of the complex calibration at the end of the production line is associated with a time and space saving, reducing production costs be reduced. Need image capture units in one Workshop can be replaced, can be used newly Image capture units easily and without much effort be calibrated. This results in a reduction of repair time and repair costs.

embodiments The invention will be explained in more detail with reference to drawings.

there shows:

1 schematically a representation of a vehicle from a bird's eye view with an array of image capture units and their image detail, which are used to represent an overall picture of this bird's eye view.

1 schematically shows a representation of a vehicle F from a bird's eye view with an array of image acquisition units K1, K2, K3 and their image sections B1, B2, B3, which are used to represent an overall picture G of this bird's eye view. The image capture units K1, K2, K3 in the exemplary embodiment shown here are, for example, wide-angle cameras, hereinafter referred to as cameras K1, K2, K3 for reasons of clarity. The respective image section B1, B2, B3 is exactly predetermined for each camera K1, K2, K3 and represents the part of a recording area of the respective camera which is used as a single image of the respective camera K1, K2, K3 for the overall image G from a bird's-eye view ie, these image sections B1, B2, B3 of each camera K1, K2, K3 are processed in an image processing unit (not illustrated here) in such a way that an overall image G is produced from a bird's eye view of the vehicle F and its surroundings.

Especially in overlapping areas U12, U13, U23, d. H. areas in which image sections B1, B2, B3 of several cameras K1, K2, K3 overlap, must be made of areas of frames several cameras K1, K2, K3 an image of the overlap area U12, U13, U23 as a part of the overall image G determined and displayed become. Are the cameras K1, K2, K3 not installed exactly, so that their reception areas and thus also for the overall picture G bird's-eye view used single frames are shifted so the image processing unit may not work determine exact overall image G, since it is especially in the overlapping areas U12, U13, U23 can lead to misrepresentations, such as a double or shifted representation of imaged objects or, for example, a lane marking SL, ML, which in the overall picture G bird's-eye view as several shifted or slanted to each other running lines is shown. Furthermore you can by using shifted frames dead angles in the overall picture G bird's-eye view, d. H. Areas of the environment of Vehicle F, which is not in the overall picture G bird's eye view be imaged, which involves great security risks are. Ie. optimal calibration of the cameras K1, K2, K3 must be ensured, and this both directly after installation or, for example, a repair-related replacement of a Camera K1, K2, K3 and during operation of the vehicle F. A single frame of the cameras used must be exactly the same Image section B1, B2, B3 correspond.

Therefore, an image of the respective predefined image detail B1, B2, B3 for each camera K1, K2, K3 is stored in the image processing unit as a comparison image. To calibrate the respective camera K1, K2, K3, the current image detail is compared with the stored comparison image, which corresponds to the image detail B1, B2, B3, which is a single image of the respective camera K1, K2, K3 for processing an overall image G of a bird's-eye view is to be used, in particular positions of prominent points and / or lines E, MK, FR, DK, KK, KH, A are compared. Are these positions in the stored comparison image and the current image section of the respective camera K1, K2, K3, the current image section is shifted in the image processing unit until it completely overlaps the stored comparison image, ie until the predetermined points and / or lines E, MK, FR, DK, KK, KH, A are exactly in the same position on the current image section as on the saved comparison image.

such distinctive points and / or lines E, MK, FR, DK, KK, KH, A are for both cameras K1, K2 arranged on the front of the vehicle F, for example, as in the embodiment shown here, striking Edges MK on a hood of the vehicle F, a side windscreen frame FR, a roof side edge DK and / or for example a striking on the bonnet arranged manufacturer emblem E. Für the rear camera K3, these are for example such striking Edges KK on a trunk lid, a rear edge KH the Boot lid and / or for example an antenna base A. This type of calibration can be done both with the vehicle F as well as being done while driving because these distinctive points and / or lines neither their position nor change their shape.

A Another possibility of calibration is a use of prominent points and / or lines ML, SL on a carriageway, on which the vehicle F is moving, for example lane markings. However, a prerequisite for this type of calibration is that the vehicle F moves on a straight lane, why vor a start and during such a calibration Steering angle and yaw rate of the vehicle F monitors must be, d. H. only if the steering angle and the yaw rate against Go zero, such a calibration is feasible. In this type of calibration, such prominent points and / or Lines ML, SL on the road to calibrate the cameras K1, K2, K3 is used, each in a superposition area U12, U13, U23 several, d. H. of at least two cameras K1, K2, K3 are, d. H. be imaged by these at the same time.

are These cameras K1, K2, K3 properly calibrated, will be such a line or such a point is shown in bird's eye view exactly da from the image processing unit in the overlay areas U12, U13, U23 an overall picture G from the individual pictures of the cameras K1, K2, K3 is determined. Is at least one camera K1, K2, K3 not exactly calibrated, for example, a median strip ML a lane twice, so slightly shifted a second Sometimes shown or not exactly straight, so that two shown Cross median stripe ML.

in the Embodiment shown here for calibrating the two front cameras K1, K2 both the median strip ML of the roadway as well as a side strip SL are used since both are in the overlapping area U12 of both cameras K1, K2 are located. The rear camera K3 can be with the front left Camera K1 with the help of the center strip ML and with the front camera K2 be calibrated using the side strip SL. to Calibration in this embodiment, however, is the median stripe ML more suitable, because this is based on the individual sections, the cameras K1, K2, K3 both across the can also be calibrated in the direction of travel of the vehicle F, d. H. if on the overall picture G bird's eye view two stripes are displayed side by side, so the respective image sections B1, B2, B3 calibrated. until the stripes completely overlap and only one strip is displayed.

There the median strip ML but not continuous, but steadily can be interrupted, beginnings or ends of individual Sections of the center strip for the calibration of the image sections B1, B2, B3 are used in the direction of travel, since these beginnings and ends each represent prominent points that are also in the overlap area U12, U13, U23 of the respective image sections B1, B2, B3 match need to be accurately represented in the overall picture G of bird's-eye view to be otherwise otherwise, for example, on the overall picture G is a solid Line can be seen. In the embodiment shown here, in which the vehicle F moves on this straight lane, is therefore checked by the image processing unit, whether the center line ML before, beside and behind the vehicle F uniform than straight, broken line and the lateral line SL uniform before, beside and behind the vehicle as a uniform straight line in the overall picture G bird's eye view is presented. If not, will be the respective image sections B1, B2, B3 of the respective cameras K1, K2, K3 adjusted until the representation is exact. These Image analysis, adaptation of image sections B1, B2, B3 and subsequent verification requires several processing cycles of the processing unit until a exact calibration of all cameras K1, K2, K3 is reached.

An adaptation of parameters for a color intensity of the individual images of the cameras K1, K2, K3 and a white balance for these individual images by the image processing unit can also be carried out by means of the two calibration methods, so that an overall image G of the bird's-eye view has a uniform color progression and no color jumps between individual regions of the overall image G, which each originate from individual images of different cameras K1, K2, K3. This is made possible by a comparison of the color intensities of the respective overlapping regions U12, U13, U23, since these are imaged by two cameras K1, K2, K3 in each case and their color representation must therefore match. If this is not the case, then the corresponding parameters of the respective camera K1, K2, K3 or the image processing of the individual images of the respective camera K1, K2, K3 can be correspondingly changed and adapted by the image processing unit during the creation of the overall image G, so that a color uniform overall picture G can be shown from a bird's eye view. During the journey, for example, occurring larger intensity or color jumps are ausfilterbar by means of appropriate filters.

In addition to those in the 1 Positions shown are also possible for the cameras K1, K2 and K3 any other positions. In a further preferred embodiment, the cameras 1 and 2 are located, for example, in the left or right vehicle exterior mirrors, so that advantageously a field of view of 180 degrees on both sides of the vehicle can be detected. Furthermore, the camera 3, for example, in the trunk lid, designed as a rear-view camera, be arranged.

F

vehicle

K1, K2, K3

Image capture units, cameras

B1, B2, B3

picture details

G

overall picture

U12, U13, U23

overlapping areas

e

Manufacturer emblem

MK

edge on bonnet

FR

lateral Windscreen frame

DK

Roof side edge

KK

edge on boot lid

KH

rear Edge boot lid

A

antenna

ML

central reservation

SL

shoulder

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

• US 5444478 [0005]

Claims (6)

Method for calibrating an arrangement for monitoring an environment of a vehicle (F), the environment being detected on the basis of a plurality of image acquisition units (K1, K2, K3) whose detection ranges at least partially overlap and form an overlap region (U12, U13, U23) in which individual images obtained by means of the image acquisition units (K1, K2, K3) are generated from an image processing unit to produce an overall image (G) showing the vehicle (F) and its surroundings from a bird's-eye view, characterized in that the image acquisition units (K1, K2 , K3) by means of prominent points and / or edges (E, MK, FR, DK, KK, KH, A) on the vehicle (F) are calibrated by positions and / or gradients of these prominent points and / or edges (E, MK , FR, DK, KK, KH, A) in a current image section of the respective image acquisition unit (K1, K2, K3) with positions and / or gradients of these prominent points and / or edges (E, MK, FR, DK, KK, KH , A) in a model image, or that a plurality of image acquisition units (K1, K2, K3) are calibrated by means of salient points and / or lines (MI, SL) on a lane which detects from this plurality of image acquisition units (K1, K2, K3) become. Method according to claim 1, characterized in that that the calibration of the image acquisition units (K1, K2, K3) at stationary vehicle (F) or while driving the vehicle (F) is performed. Method according to claim 1 or 2, characterized that calibration of a plurality of image acquisition units (K1, K2, K3) while driving on a straight road is performed with straight lane markings. Method according to claim 3, characterized that before and / or during the calibration, a steering angle and a yaw rate of the vehicle (F) are monitored, wherein Calibration only occurs when the steering angle and the yaw rate tend towards zero. Method according to one of the preceding claims, characterized in that by means of a comparison of the Image acquisition units (K1, K2, K3) shown overlapping areas (U12, U13, U23) a color intensity and / or white balance calibration the individual image acquisition units (K1, K2, K3) performed becomes. Apparatus for carrying out the method according to one of claims 1 to 5, characterized by a Arrangement for monitoring an environment of a vehicle (F), comprising a plurality of image acquisition units (K1, K2, K3) whose Coverage areas overlap at least partially, one Image processing unit for generating an overall image (G) acquired individual images of the image acquisition units (K1, K2, K3), the overall image (G) representing the vehicle (F) and its surroundings a bird's eye view, and a display device for display of the overall picture (G).