Method for calibrating a lidar sensor

Abstract

The invention relates to a method for detecting misalignment of a laser radar sensor (2) of a vehicle (1). According to the invention, a reference measurement is performed with the aid of the laser radar sensor (2), wherein at least one line (S1, S2, S3) is projected onto a road surface (FB) by emitting laser beams in at least one predetermined scanning plane, and wherein the position of the at least one line (S1, S2, S3) relative to the vehicle (1) is determined. At a later time, a calibration measurement is performed, wherein at least one line (S1', S2', S3') is also projected onto the road surface (FB) by emitting laser beams in the at least one predetermined scanning plane, and wherein the position of the at least one line (S1', S2', S3') relative to the vehicle (1) is determined. The determined positions are compared with one another. If the comparison shows that the positions determined during the reference measurement and the calibration measurement are displaced from one another by more than a predetermined degree, it is inferred that the laser radar sensor (2) is misaligned. The reflected radiation can also be detected by a photodetector (3) arranged outside the lidar sensor, for example by a camera having a sensor chip that is sensitive to the wavelength of the laser radiation, usually infrared radiation. The reference measurement and the calibration measurement are preferably carried out while the vehicle is stationary. For example, the reference measurement is carried out when the vehicle is parked and the calibration measurement is carried out when the vehicle is subsequently started. It is thus possible to determine whether an event has occurred during parking that has led to a misalignment of the sensor.

Description

Technical Field

[0001] The invention relates to a method for calibrating a lidar sensor of a vehicle by means of a camera according to the preamble of claim 1 .

[0002] The invention also relates to the use of this method. Background Art

[0003] DE 10 2005 037 094 B3 discloses a method for calibrating a distance measuring sensor having two sensor channels, the radiation of which is in the form of lobes. The method comprises the following steps:

[0004] - by means of the sensor emitting a ray lobe of the sensor channel towards the calibration surface,

[0005] - detect the ray lobes with the camera of the video system,

[0006] - determining a first representative value of the ray lobe in the image coordinates of the camera,

[0007] - change the position of the calibration surface relative to the sensor and camera,

[0008] - detect the ray lobes with a camera,

[0009] - determining a second representative value of the ray lobe in the image coordinates of the camera,

[0010] - Repeat the previous steps for the other sensor channel,

[0011] - the lobe axis of the outgoing ray lobe modeled as a straight line in sensor coordinates,

[0012] - transform the simulated straight line from sensor coordinates to camera image coordinates,

[0013] - comparing this straight line with a representative value determined in image coordinates for each sensor channel,

[0014] - determining a calibration function for compensating for possible deviations of the simulated straight line from a plurality of representative values.

[0015] The sensor is a LiDAR system or LiDAR sensor based on luminescence and light absorption. Summary of the invention

[0016] The object of the present invention is to specify a method for calibrating a lidar sensor of a vehicle which is improved compared to the prior art and the use of such a method.

[0017] This object is achieved according to the invention by a method having the features of claim 1 and a use having the features of claim 7 .

[0018] Advantageous embodiments of the invention are the subject matter of the dependent claims.

[0019] In a method for calibrating a lidar sensor of a vehicle, laser radiation is directed toward a road surface by means of the lidar sensor, and laser radiation reflected toward the lidar sensor is detected by means of the lidar sensor.

[0020] According to the invention, a reference measurement is performed with the aid of a lidar sensor in a calibrated state. In this case, a laser beam is emitted in at least one predetermined scanning plane and thereby projects at least one line onto the road surface. By detecting the reflected laser beam, the position of the at least one line relative to the vehicle is determined and stored for a later evaluation. A reference measurement is performed at a later time. In the reference measurement, a laser beam is also emitted in at least one predetermined scanning plane and thereby also projects at least one line onto the road surface. By detecting the reflected laser beam, the position of the at least one line relative to the vehicle is determined. It is then checked whether the positions determined in the reference measurement and the calibration measurement are displaced from one another. If a displacement exceeding a predetermined degree is determined, a degradation of the lidar sensor is inferred.

[0021] The determination of the position of the line projected onto the road is advantageously based on the determination of the distance between the respective line and the vehicle.

[0022] In order to use the lidar sensor, it needs to be calibrated. In order to quickly provide the data acquired by the lidar sensor to an operating vehicle system, for example, a vehicle system for implementing automated driving, in particular highly automated driving or autonomous driving or driving without a driver, it is very advantageous to quickly check whether the lidar sensor is still calibrated. With the method, such a very fast and at the same time reliable check can be achieved without much effort.

[0023] In one possible embodiment of the method, the displacement is determined only if the reference measurement and the calibration measurement are carried out at the same location, i.e. at the same vehicle position. Due to road bulges, the line projected onto the road surface may be distorted. If the reference measurement and the calibration measurement are carried out at the same location, the line projected onto the road surface is distorted in the same way. This makes it possible to avoid measurement errors caused by road bulges.

[0024] In another possible embodiment of the method, the reference measurement and the calibration measurement are performed while the vehicle is stationary. For example, the reference measurement is performed when the vehicle is parked and the calibration measurement is performed when the vehicle is subsequently started. It is thus possible to determine whether an event has occurred during parking or in the time between performing the reference measurement and the calibration measurement that has led to a sensor misalignment. Such an event can be, for example, a change in the vehicle load or a change in the weight distribution within the vehicle or a tire change. With this method, it is possible to quickly check whether the lidar sensor is calibrated, for example during a short stop in a parking lot or at a gas station. If the calibration shows that the lidar sensor is calibrated, the taught calibration parameters can be used for operation. Otherwise, an online calibration can be started to update the calibration parameters.

[0025] In another possible embodiment of the method, more than one line is projected onto the road surface during the reference measurement and the calibration measurement. The lines projected onto the road surface during the reference measurement form a reference pattern, while the lines projected onto the road surface during the calibration measurement form a calibration pattern. Such a pattern with a line structure can be generated, detected and evaluated very simply and without much effort.

[0026] In another possible embodiment of the method, the lines are generated by means of a plurality of lasers of a lidar sensor arranged one above the other. Such generation can be carried out particularly simply and precisely.

[0027] In another possible embodiment of the method, during the calibration measurement, a defined pattern is projected onto the road surface for a predetermined period of time, for example a period of 30 seconds. Such a period of time allows the influence of road undulations to be mitigated by averaging or integration.

[0028] In the use according to the invention, the above-described method is used in a vehicle that can be operated automatically, in particular highly automatically or autonomously. Since the data acquired by means of the lidar sensor are provided very quickly to the vehicle system for performing the automated driving, such automated driving can be adopted very quickly and reliably after the vehicle stops.

[0029] In this case, for example, when a misalignment of the lidar sensor is detected, automated driving, in particular highly automated driving or autonomous driving of the vehicle is stopped and / or inhibited. Traffic safety can thus be significantly increased. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The embodiments of the present invention are explained in detail below with reference to the accompanying drawings, wherein:

[0031] Figure 1 schematically shows a perspective view of a vehicle and a road surface at a first moment in time,

[0032] Figure 2 Schematically shows the Figure 1 A perspective view of the vehicle and the road surface at a second moment in time.

[0033] Corresponding parts are provided with the same reference symbols in all the figures. DETAILED DESCRIPTION

[0034] Figure 1 A perspective view of the vehicle 1 and the road surface FB is shown at a first instant in time, at which a reference measurement is performed.

[0035] Vehicle 1 includes a lidar sensor 2 which emits laser radiation in a conventional manner and detects reflected laser radiation. Data of the vehicle surroundings acquired at least by means of lidar sensor 2 are used for automated driving, in particular highly automated driving or autonomous driving or driverless driving of vehicle 1 .

[0036] Due to changes in the loading state of vehicle 1, changes on the vehicle and external influences, the orientation of lidar sensor 2 relative to the vehicle surroundings may change, which affects the data collected by means of lidar sensor 2. For this reason, an accurate calibration of lidar sensor 2 is always necessary, in particular online calibration.

[0037] When the vehicle 1 is parked, it may happen that the lidar sensor 2 is misaligned due to improper parking. When the parking process is completed and the vehicle 1 is put into use again, it is important to quickly identify whether the lidar sensor 2 is still calibrated or misaligned. With an misaligned lidar sensor 2, functional safety during automated driving or driverless driving can no longer be guaranteed. In this case, automated driving or driverless driving of the vehicle 1 is always prohibited, at least until the lidar sensor 2 is recalibrated, for example by online calibration.

[0038] In order to detect the misalignment of the lidar sensor 2, a reference measurement is first performed. Here, one or more lines S1, S2, S3, S4 or a reference pattern RM consisting of a plurality of lines is projected onto a road surface FB, which is in particular a flat structure, based on the laser beam. In this case, the lidar sensor 2 is designed for so-called line scanning, in which a plurality of lines S1, S2, S3, S4, so-called layers, are generated on the road surface FB at different distances A1 to A4 by means of a plurality of superimposed lasers of the lidar sensor 2. In this case, the distances A1 to A4 from the vehicle 1 increase as the arrangement height of the laser increases. In this case, the first three lines S1, S2, S3 form the reference pattern RM.

[0039] The reference pattern RM is acquired by the lidar sensor 2 and used as Figure 2 The reference pattern RM shown is stored. In other possible embodiments, the reference pattern RM can be formed by means of another number of lines S1 , S2 , S3 .

[0040] Figure 2 Show according to Figure 1 A perspective view of the vehicle 1 and the road surface FB at a later second moment, in particular after the vehicle 1 has been restarted or during a new ignition process. A calibration measurement is performed at the second moment. Here, the lines S1′, S2′, S3′ shown in dashed lines or a calibration pattern M formed by these lines are projected onto the road surface FB and the position of the lines S1′, S2′, S3′ or the position of the calibration pattern M relative to the vehicle 1 is determined.

[0041] Before restarting, for example, when the load of the vehicle 1 has changed significantly or the orientation of the lidar sensor 2 has been changed due to mechanical effects (for example, when hitting an obstacle), a misalignment of the lidar sensor 2 occurs. The change causes a displacement of the lines S1′, S2′, S3′ projected onto the road surface FB at the second moment relative to the lines S1, S2, S3 projected onto the road surface FB at the first moment, or a displacement between the reference pattern RM and the calibration pattern M.

[0042] From said displacement, which is determined, for example, by determining the distances A1 ′- A3 ′ between the lines S1 ′- S3 ′ and the vehicle 1 and subsequently comparing them with the distances A1 to A3 contained in the reference pattern RM, a misalignment of the lidar sensor 2 can be determined.

[0043] That is, in order to identify a misalignment of the lidar sensor 2, during a calibration measurement, for example during vehicle start-up, a calibration pattern M is projected onto the road surface FB by means of the lidar sensor 2 and detected by means of the lidar sensor 2. Then, the displacement between the calibration pattern M and the reference pattern RM is determined, and then, if the displacement exceeds a predetermined degree of, for example, + / - 0.1 m, it is inferred that the lidar sensor 2 is misaligned.

[0044] When such a misalignment of the lidar sensor 2 is detected, the automated driving, in particular highly automated driving or driverless driving, of the vehicle 1 is stopped or prohibited in one possible design. Preferably, an online calibration is then started and, after its completion, automated driving, in particular highly automated driving or driverless driving of the vehicle 1 is allowed again.

[0045] When the determined displacement does not exceed a predetermined degree, it is assumed that the lidar sensor 2 is not misaligned, and in order to Figure 1 The stored calibration parameters are used for operating the lidar sensor 2 and for evaluating the data acquired with the aid of the lidar sensor 2 .

[0046] The predetermined degree is determined and predetermined, in particular, individually for different vehicle types.

[0047] In order to be able to take into account possible road irregularities that may occur during the calibration measurement, a defined pattern M is projected onto the road surface FB and during the calibration measurement, the pattern is detected over a predetermined period of time, for example 30 seconds. Thus, measurement errors can be reduced by integration or averaging.

[0048] Furthermore, in one possible embodiment, a comparison of the test pattern M with the reference pattern RM is performed in order to determine whether the test pattern M is distorted relative to the reference pattern RM. Since such a distortion of the test pattern M is an indicator of a road bump and measurement errors can result from road bumps, the displacement is determined in particular only if it has been determined in advance that the pattern generated in the test measurement is not distorted relative to the reference pattern or is distorted below a predetermined error level.

[0049] The laser radiation reflected toward the lidar sensor 2 is detected in the present exemplary embodiment by means of a photodetector arranged in the lidar sensor 2. However, it is also conceivable to detect the reflected radiation by means of a photodetector 3 arranged outside the lidar sensor, for example by means of a camera having a sensor chip that is sensitive to the wavelength of the laser radiation, usually the wavelength of infrared radiation.

Claims

1. A method for identifying misalignment of a lidar sensor (2) of a vehicle (1), characterized in that: - performing a reference measurement with the aid of the lidar sensor (2), wherein at least one line (S1, S2, S3) is projected onto the road surface (FB) by emitting laser radiation in at least one predetermined scanning plane, and wherein the position of the at least one line (S1, S2, S3) relative to the vehicle (1) is determined, - performing a calibration measurement at a later time with the aid of the lidar sensor (2), wherein at least one line (S1′, S2′, S3′) is scanned by emitting a laser beam in the at least one predetermined scanning plane. projected onto the road surface (FB), and wherein the position of the at least one line (S1′, S2′, S3′) projected in the calibration measurement relative to the vehicle (1) is determined, and - if it is determined that the positions determined during the reference measurement and the calibration measurement are displaced from one another by more than a predetermined degree, it is inferred that the lidar sensor (2) is misaligned, To determine the position of a line projected onto the road surface, the distance between the corresponding line and the vehicle is determined, comparing the distance (A1′, A2′, A3′) between the at least one line (S1′, S2′, S3′) projected in the calibration measurement and the distance (A1, A2, A3) between the at least one line (S1, S2, S3) projected in the reference measurement and the vehicle, If the displacement determined by comparing the distances exceeds a predetermined extent, it is concluded that the lidar sensor (2) is misaligned.

2. The method according to claim 1, characterized in that: The reference measurement and the calibration measurement are performed at the same vehicle position.

3. The method according to claim 1 or 2, characterized in that: The reference measurement and the calibration measurement are performed with the vehicle at a standstill, wherein the reference measurement is performed when the vehicle is parked, and wherein the calibration measurement is performed when the vehicle is subsequently started again.

4. The method according to claim 1 or 2, wherein During the reference measurement and the verification measurement, more than one line (S1, S2, S3 or S1′, S2′, S3′) is projected onto the road surface (FB), wherein the lines (S1, S2, S3) projected onto the road surface (FB) during the reference measurement form a reference pattern (RM), and the lines (S1′, S2′, S3′) projected onto the road surface (FB) during the verification measurement form a verification pattern (M), and the displacement between the positions determined in the reference measurement and the verification measurement is determined by comparing the reference pattern (RM) and the verification pattern (M).

5. The method according to claim 4, characterized in that: The more than one line (S1, S2, S3 or S1', S2', S3') is generated by means of a plurality of superposed lasers of the lidar sensor (2).

6. Use of the method according to one of claims 1 to 5 in a vehicle (1) which can be driven automatically.

7. The use according to claim 6, characterized in that: Autonomous driving is highly automated driving or driverless driving.

8. The use according to claim 6 or 7, characterized in that: When misalignment of the lidar sensor (2) is detected, the autonomous driving of the vehicle (1) is stopped and / or prohibited.

Images