Method for calibrating a vehicle headlamp

Abstract

The invention relates to a method (1) for calibrating a vehicle headlamp. A positional deviation of the predefined light distribution is determined by evaluating an image from the at least one vehicle camera. The invention is characterized in that at least three recordings are carried out in a first step (2), a first recording (3) being acquired in the manner of recording the calibration pattern (4), a second recording (5) being acquired without the calibration pattern (4), a third recording (6) being acquired in the manner of recording the inverted calibration pattern (4 '), and a fourth recording (7) being acquired in the manner of recording the inverted calibration pattern (4'). And wherein in a second step (7) the second recording (5) is subtracted from the first and the third recording (3, 6) in order to clear the current scene from the recording (3, 6).

Description

Technical Field

[0001] The invention relates to a method for calibrating a vehicle headlamp / headlamp of the type defined in more detail in the preamble of claim 1. Background Art

[0002] In principle, methods for calibrating vehicle headlights are known in the prior art. Such methods are necessary because vehicle headlights often deviate. Even in vehicles with automatic light range adjustment, mechanical stresses, thermal expansion or contraction, and step losses / step loss of the built-in stepper motors occur over time. All these factors lead to a gradual position deviation of the headlights.

[0003] In particular, deviations in the vertical position are a safety-related problem, since they can dazzle oncoming drivers or reduce their own field of vision. Therefore, the headlight setting should be checked regularly.

[0004] In this regard, a method for generating a three-dimensional depth information map of an environment is known from DE 10 2020 007 613 A1. In this case, a light pattern is projected into the environment by a projector and recorded by a camera. The camera image can then be evaluated, the respective positions of feature points being identified on their corresponding epipolar lines, and depth information about the three-dimensional map is obtained by determining the displacement of the feature points on these lines.

[0005] DE 10 2017 222 708 A1 shows other prior art. The patent describes a camera device for a vehicle that can perform 3D environment acquisition. To this end, at least two camera modules are required, which have at least partially overlapping acquisition areas. Via a control unit, an evaluation unit and a point light source projector, a "pseudo-noise mode" can be used for 3D environment acquisition.

[0006] DE 10 2016 118801A1 describes a method for setting vehicle headlights. To this end, the individual light-emitting units of the headlights are controlled in a time-staggered manner in order to illuminate a scene. The scene is photographed by a camera included in the vehicle within a predetermined time period. By controlling a plurality of light-emitting units in a time-staggered manner, a brightness distribution pattern is generated, from which deviations from a reference pattern can be calculated. The headlights can then be set according to the deviations. For more prior art, reference can be made to DE 10 2012 007908A1, DE 10 2016 006 391A1, DE 10 2011 109 440A1, DE 10 2015 203 889A1, DE 102014 117 845A1, DE 10 2017 117 594A1 and DE 10 2020 000 292A1.

[0007] In order to actively adjust the position of the headlights, for example for lighting range adjustment or for turn signal functions, stepper motors are used. These stepper motors can adjust the light modules in the headlights to a target angle. In this regard, there are also methods for automatically and on-site readjustment of the headlight settings. For example, the light distribution of the headlights in the area in front of the vehicle can be captured and evaluated via a driver assistance camera. For this evaluation, salient points of the cut-off line, such as the H0V0 point, are usually used. People try to identify these salient points in the image. In another variant, where appropriate, a light distribution specially designed for calibration can also be emitted, for example during startup or sleep.

[0008] Based on the distances of the pixels in the camera image to the calibrated reference points, the vertical and horizontal difference angles can then be estimated and compared with the currently implemented adjustment. Vertical and horizontal position deviations can thus be compensated. However, the difficulty lies in uncontrolled or uncontrollable environmental conditions. These can involve light conditions and the structure, shape and positioning of the illuminated surface. Different vehicles can also have different lighting systems. Differences in color and brightness distribution, individual pixel errors or even inaccuracies in the optical path can occur, which can lead to blurring and color shifts.

[0009] Evaluation algorithms can be based in particular on edge and maximum detection, but this can lead to inaccurate feature extraction and variable imprecise calibration. In particular, the safety-related accuracy of 0.1% stipulated by the ECE standard cannot be maintained robustly. And with the introduction of high-resolution light systems based on, for example, LCD, DMD or μLED technology in vehicles, increasingly complex light distributions must be projected. Summary of the invention

[0010] The object of the present invention is to provide a method for calibrating vehicle headlights which overcomes the above-mentioned disadvantages.

[0011] According to the invention, this object is achieved by a method having the features of claim 1, here in particular the features of the characterizing part of claim 1. Advantageous embodiments and developments are given in the dependent claims.

[0012] In the core of the method according to the invention, at least three recordings are performed in a first step, wherein a first recording is acquired by recording a calibration pattern, a second recording is acquired without the calibration pattern, a third recording is acquired by recording an inverted calibration pattern, and in a second step the second recording is subtracted from the first recording and the third recording in order to remove the current scene from the recording, wherein the calibration pattern comprises a circular / ring-shaped pattern, the position of which in the recording is known in advance, the centers of which are arranged on a horizontal line, and each center is calculated by a straight-line approximation, wherein it is determined whether the calibration pattern is projected on a suitable surface suitable for carrying out the calibration. By subtracting the images, it is obtained that only the calibration pattern is visible in the image at any time. This significantly simplifies the subsequent evaluation algorithm and can be performed in a significantly more robust manner. Thus, a system is provided for the automatic and high-precision calibration of the settings of a high-resolution headlight system in a vehicle.

[0013] With this method for calibrating vehicle headlights, position deviations from a predefined light distribution are determined by evaluating images from at least one vehicle camera. These position deviations can then be compensated, for example, by an installed stepper motor and redefinition of the zero position.

[0014] During calibration, the method can detect the structure of the area in front of the vehicle and, in particular, output whether there is a wall in front of the vehicle. In this case, the direction of the wall can also be determined. All this information can be used to adjust the projection of the headlights according to the current area in front.

[0015] Preferably, the first photographing record and the third photographing record may be compared pixel by pixel in the third step so as to identify brightness differences between pixels at the same position.

[0016] According to an advantageous embodiment, it can be provided that areas with very small brightness differences are represented as gray areas. Areas that are also present in the camera image, i.e. in the recording, outside the calibration pattern are marked as gray areas so that they do not have to be evaluated further. This has the advantage that the background scene is represented with gray values ​​that are clearly separated from the projection area of ​​the headlights. Furthermore, the influence of optical crosstalk of activated pixels due to imperfect optical channels can be eliminated by fusing the two inverted / opposite images. This can improve the subsequent positioning of the circle, for example.

[0017] The calibration pattern comprises circles, wherein the positions of these circles in the recordings are known in advance. With a high-resolution system, completely different projections can be used for calibration. Thus, for example, circles can be used, which are advantageous. For example, circles have a constant center point, independent of the focus of the illumination system on the projection surface. Here, circles are also robust to distortions.

[0018] The center of the circle is arranged on a horizontal line. Advantageously, the position of the circle in the illumination area of ​​the headlight is known in advance. This can be achieved, for example, via a known distance between parallel lines and a maximum permissible distance or a known distance between the center of the circle.

[0019] Each center point is calculated by a straight line approximation, wherein it is determined whether the calibration pattern is projected on a suitable surface for carrying out the calibration. For example, the projection on a plane can be ensured by the required maximum twist and the distance of parallel lines from each other. Only in this case is it advantageous to continue with the calibration.

[0020] According to a very advantageous improvement of this concept, it can be provided that the vehicle camera is pre-calibrated. This is advantageous for carrying out the above-mentioned straight line approximation method. The pre-calibrated vehicle camera is also advantageous for determining the position and rotation, ie the attitude, of the projection surface.

[0021] According to an advantageous embodiment, it can be provided that the attitude of the projection plane is determined by twisting the calibration pattern horizontally and vertically and calculating the attitude of the projection plane via the displacement of the center point of the circle in the camera image. In particular, this is done by assuming a flat projection surface. The orientation of the headlights can thus be determined via the attitude of the plane, the position of the center point of the circle in the camera image and the known angular position of the relevant headlight pixels and the associated position deviation can be derived. The known angular position can be a vertical angle and / or a horizontal angle relative to / with respect to the center point of the circle.

[0022] Likewise or alternatively, the headlights can be calibrated in advance so that the circle is used as a characteristic point for triangulation of the three-dimensional coordinates. Advantageously, an additional 2×3 recording of the two offset calibration patterns is also performed here. In this case, in particular, a corresponding pair of camera headlights is used as a structured light system.

[0023] It is also conceivable to model the expansion of the projected circle in space as a cone profile. Here, too, the position of the ring in three-dimensional space can be precisely determined, wherein the setting of the headlights can be directly inferred, since the installation position of the headlights is known, and the light beam can therefore be reconstructed.

[0024] Another advantageous embodiment can provide that the pose is determined by the known angular position of a calibrated vehicle camera, wherein flat surfaces and unstable scenes can be detected in order to adapt the projection and / or the animation for the presentation to the projection plane. Advantageously, this allows an exact measurement of the projection surface in front of the vehicle, wherein not only flat surfaces such as walls or driveways can be detected, but also more complex and unstable scenes. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Further advantages of the method according to the invention are given further by the remaining dependent claims and become clearer from the exemplary embodiments which are described in more detail below with reference to the attached drawings.

[0026] in:

[0027] Figure 1 A possible implementation of the calibration pattern is shown;

[0028] Figure 2 Another possible implementation of the calibration pattern is shown;

[0029] Figure 3 A possible sequence of the method is shown. DETAILED DESCRIPTION

[0030] exist Figure 1 A possible embodiment of a calibration pattern 4 is shown in the illustration of FIG. The calibration pattern 4 has individual circles 9, which have respective centers 10. The surface represents, for example, a surface 13 of a headlight.

[0031] exist Figure 2 Another variant embodiment of method 1 can be seen in FIG. Figure 1 The circles 9 are arranged differently. Furthermore, horizontal lines 11 can be seen which extend along the circle centers 10. In such a calibration pattern 4, vertical lines 11' can also be drawn for evaluation, which likewise connect two circle centers 10 to one another.

[0032] exist Figure 3 A possible flow of method 1 is shown in FIG. In a first step 2, three individual recordings are generated. A first recording 3 is obtained by recording a calibration pattern 4. A second recording 5 is obtained without a calibration pattern 4. A third recording 6 is obtained by recording an inverted / reversed calibration pattern 4'. In a second step 7, the second recording 5 is subtracted from the first recording 3 and the third recording 6. Individually existing scenes can thus be removed from the recordings 3 and 6. For the sake of clarity, the second step 7 is shown repeatedly. In a third step 8, the two recordings 3 and 6 are compared with each other pixel by pixel in order to find brightness differences between pixels at the same position. Slight brightness differences are represented as gray areas. This relates to the area outside the headlights 13, which is represented as a slight dotted line.

[0033] The method thus makes it possible to calibrate the headlights simply and automatically without additional hardware components. Advantageously, a high degree of accuracy can be achieved. Furthermore, the method is robust to environmental conditions and can be carried out autonomously, so that no intervention by the driver or other operator is required. This leads to an increase in safety and comfort. Likewise, the contour of the projection space calculated during operation can be used to adapt the projection of the vehicle to the existing projection surface, for example for a startup animation for display.

Claims

1. A method (1) for calibrating a vehicle headlight, wherein position deviations of a predefined light distribution are determined by evaluating images from at least one vehicle camera, It is characterized in that In a first step (2), at least three shots are performed, wherein a first shot record (3) is obtained by shooting a calibration pattern (4), a second shot record (5) is obtained without the calibration pattern (4), and a third shot record (6) is obtained by shooting an inverted calibration pattern (4'), and wherein in a second step (7), the second shot record (5) is subtracted from the first shot record and the third shot record (3, 6) in order to remove the current scene from the shot records (3, 6), The calibration pattern (4) comprises a circle (9), the position of which in the recording is known in advance, and the centers (10) of the circles (9) are arranged on a horizontal line (11), each center (10) being calculated by a straight line approximation method, wherein it is determined whether the calibration pattern (4) is projected onto a suitable surface for performing the calibration.

2. The method (1) according to claim 1, It is characterized in that In a third step (8), the first recording and the third recording (3, 6) are compared pixel by pixel to identify brightness differences between pixels at the same position.

3. The method (1) according to claim 2, It is characterized in that Areas with very small differences in brightness are represented as gray areas.

4. The method (1) according to any one of claims 1 to 3, It is characterized in that The vehicle camera is pre-calibrated.

5. The method (1) according to any one of claims 1 or 4, It is characterized in that The attitude of the projection plane is determined by causing the calibration pattern (4) to be twisted horizontally and vertically and calculating the attitude of the projection plane via the displacement of the center of the circle (10) in the camera image (12).

6. The method (1) according to claim 4 and 5, It is characterized in that The pose is determined by the known angular position of the calibrated vehicle camera, wherein flat surfaces and unstable scenes can be detected in order to adapt the projection and / or the animation for the presentation to the projection plane.

Images