WO2018103796A1

WO2018103796A1 - Method and apparatus for camera-based control of a lighting unit of a motor vehicle

Info

Publication number: WO2018103796A1
Application number: PCT/DE2017/200123
Authority: WO
Inventors: Karsten Breuer; Tobias Schwarz; Sascha Semmler
Original assignee: Conti Temic Microelectronic Gmbh
Priority date: 2016-12-06
Filing date: 2017-12-04
Publication date: 2018-06-14
Also published as: DE102016224185A1; DE112017005280A5

Abstract

The invention relates to a method for camera-based control of a lighting unit of a motor vehicle, including steps in which a scene seen in front of the vehicle is imaged in an image by means of a camera system in an imaging step (101) and the image is analysed by virtue of recognizing structures (2-5) in the image and assigning the recognized structures to object classes in an analysis step (102), and the lighting unit of the vehicle is controlled on the basis of the analysis depending on the identified object classes in a control step (106). The invention also relates to a corresponding driver assistance system.

Description

Method and apparatus for camera-based control of a lighting unit of a motor vehicle

Field of the invention

The present invention relates to a lighting system of a motor vehicle, which is automatically controlled in the context of a driver assistance system. By this illumination system, a driving light automatic is realized.

Background of the invention

The field of application of the invention extends to

Motor vehicles, in particular motor vehicles with camera-based driver assistance systems. Through a camera system can

Information about a scene in the area, which is preferably arranged in front of a motor vehicle, are detected and used for example for controlling a cruise control.

From the well-known prior art further driver assistance systems emerge, in which regardless of

Camera system, a light or combined rain light sensor is provided, which is aligned substantially vertically upwards and over which the brightness of the environment can be determined. If poor visibility conditions are determined, a lighting unit can be activated and switched on, for example, to illuminate the roadway. The above-mentioned method has among other things the disadvantage that an additional sensor system has to be provided. This increases the manufacturing complexity and the complexity of

Maintenance work. In addition, a light sensor provides only a limited amount of information and can lead to erroneous results. Also, only the environment of the vehicle is examined; a foresighted analysis of future

Conditions, for example, the lighting before The emergence of the corresponding need to turn on is not possible.

Summary of the invention

It is therefore an object of the present invention to provide a method in which the control of the

Lighting unit with minimal technical effort and can be realized in advance.

The object is based on a method according to claim 1 and a use according to claim 9 and a

Driver assistance system according to claim 10 solved. The respective following dependent claims give advantageous embodiments of the invention.

The invention includes the technical teaching to provide a method for camera-based control of a lighting unit of a motor vehicle, which has steps in which in an imaging step by means of a camera system a

Scene is mapped onto an image in front of the vehicle, and in an analysis step, the image is analyzed, and in a control step based on the analysis, the lighting unit of the vehicle is controlled.

The invention includes the advantage that with only one sensor system in the form of the camera of the driver assistance system, a forward-looking control of the lighting unit is made possible. In particular, can be dispensed with additional sensors. In particular, no step is needed in which the

Brightness of the environment affects the control of the lighting unit.

A preferred embodiment of the invention provides for structures in the image to be recognized and object classes to be assigned, and the lighting unit to be dependent on the

Object classes is controlled. Such a classification can be realized for example by means of semantic segmentation and, if necessary, environmental recognition.

The structures can be assigned to the corresponding object classes, for example by using pattern recognition methods-in particular by the use of neural networks.

The advantage of this embodiment includes that objects located at any distance from the vehicle can be used to proactively assess a need to turn on an additional lighting unit.

The structures are particularly preferred as

Road structures detected and then

Road structure object classes associated with bridges, tunnels, underpasses and / or building facades.

Thus, it is advantageously possible to recognize certain structures on the road ahead, the one

Darkening and therefore directly affect a need to turn on the lighting unit. For example, a tunnel or an underpass may lead to an analysis, the result of which leads to the illumination unit being switched on even before an entrance to the tunnel or the beginning of the underpass is reached.

The invention can be further improved by the expected duration of a passage along the recognized

Road structure is calculated and the lighting unit is turned on only for a calculated duration above a street structure associated threshold. Thus, unnecessary controls of the lighting unit are advantageously avoided, which can cause irritation of the driver depending on the light source in addition to increased wear and energy consumption. For example, it is recognized that the motor vehicle under a narrow bridge

is driving through, so can turn on the

Lighting unit are suppressed. Even if another control, which is based for example on the ambient brightness, which is detected for example by an additional sensor or via the camera, detects a darkening and thus trigger the switching on of the lighting unit, this mechanism would, due to the expected temporal shortness of Darkening, suppress switching on. The threshold may be, for example, a few seconds, for example one or five or ten seconds. When calculating the duration, the current or expected speed of the vehicle may also be included. An advantageous embodiment of the invention provides that structures are also recognized as lighting units and

Luminaire unit categories that include headlamps and / or streetlamps. Such external lighting units can be used as a measure of the

Probability to be assumed that a control of the own lighting unit is required. In doing so, a correlation is assumed between the frequency of occurrence of these lighting units and the need for their use, which has proven surprisingly true in practice.

Advantageously, a switching state of the

Luminous units determined and calculated a comparison value of a number of switched on lighting units against a number of switched-off lighting units, and then the lighting is switched on at a comparison value above a predefined threshold value.

The comparison value can be, for example, a quotient or a difference. If, for example, a sufficient percentage of oncoming vehicles has switched-on headlights, or a sufficient percentage of the street lamps arranged in the field of view of the camera system is switched on, this can be determined by analyzing the image, whereby switching on the

Lighting unit is effected. This is special

advantageous, since this method can be applied even in poor visibility, which is not yet noticeable by low brightness.

An embodiment of the invention further provides that a switching state of the lighting units is determined and a number of switched on lighting units within a predefined period of time is determined and the lighting is turned on at a number above a predefined threshold.

This will provide an analysis with increased safety and security

Reduced error rate created, since the result of the analysis and the control of the lighting unit is thus independent of the number of switched off lighting units. As soon as a sufficient number of switched on lighting units is detected within a certain time interval, the own lighting unit is controlled.

A further advantageous embodiment of the invention provides that the scene is imaged by means of a wide-angle camera, by means of which a lower area below the horizon and an upper area above 45 ° above the horizon are detected, and wherein the brightness of the upper area of the image is analyzed is turned on, and the lighting is turned on when a critical brightness is below. The wide-angle camera can be used as a camera, for example

Fisheye lens or be designed as a camera with increased vertical viewing angle.

Advantageously, it is thus possible with only one camera, in addition to the road and other road users, for example, to capture the sky and its brightness. Usually, wide-angle cameras are only used to capture the widest possible horizontal image area, for example as many horizontally distributed as possible

Traffic participants or as many horizontally arranged sections to be able to capture. Because the camera also detects vertical sections, in particular above 45 ° above the horizontal, a brightness-related control of the lighting unit can be realized by a simple analysis of the corresponding image section.

In other words, therefore, a wide-angle camera of a driver assistance system is additionally used as a light sensor.

A further advantageous embodiment provides that the brightness of a central region of the image is analyzed, and a comparison value between the brightness in the upper region and the brightness in the middle region is calculated and the illumination is controlled as a function of the comparison value.

The comparison value can be, for example, a quotient or a difference. The center area may be near the vanishing point of the image, and the top area may be at the top of the image. This makes it possible, for example, to suppress the switching on of the lighting unit when only an underpass is passed, in which the middle area is comparatively light and the upper area is comparatively dark.

Brief description of the figures Other measures that improve the invention will be discussed below together with the description of preferred

Embodiments of the invention with reference to three figures shown in more detail. Show it

Fig. 1 is a diagram of a first embodiment of the invention, and Fig. 2 is a diagram of a second embodiment of the invention, and

Fig. 3 is a diagram of a third embodiment of the invention.

Detailed description of the invention

According to FIG. 1, in a method according to a first exemplary embodiment of the invention, a scene in front of the vehicle is imaged on an image in an imaging step 101 by means of a camera system, and the image is analyzed in an analysis step 102.

Within the analysis step 102, structures in the image are first recognized in an object recognition step 103, for example by means of semantic segmentation, and the recognized structures are assigned to object classes.

The structures are assigned in the present embodiment Lichtinheitsobj Ektklassen and Straßenstrukturobj ektklassen.

If, in a first assignment step 104, at least one lighting unit, such as at least one headlight, rear light and / or a street lamp, is detected, a first comparison step 105 is continued. Otherwise, if no such object classes are detected, then continue with mapping step 101.

In the first comparison step 105, the switching state of the light units is determined. It is therefore determined whether the respective lighting unit is on or off. Subsequently, in the first comparison step 105, a comparison value of a number of switched on lighting units is calculated against a number of switched-off lighting units. This comparison value can be designed, for example, as quotient or difference. If the comparison value is smaller than a predefined threshold value, then step 101 continues. Otherwise, in the case of a comparison value above a predefined threshold value, the illumination is switched on in a control step 106.

The object recognition step 103 is followed, for example, by a second assignment step 104 in addition to the first assignment step 104

Assignment step 105.

If at least one road structure such as at least one bridge, a tunnel, an underpass and / or a building facade is detected in the second allocation step 107, then a second comparison step 108 is continued. Otherwise, if no such object classes are recognized, then continue with mapping step 101. In the second comparison step 108, the expected duration of a transit along the recognized road structure is determined. This can be done, for example, by a recognition of structural characteristics such as the length of the road structure and a comparison with the current or expected speed of the vehicle. If the duration is less than a predefined threshold, then

Imaging step 101 continued. Otherwise, in the case of a duration above a predefined threshold, and the lighting in one

Control step 106 is turned on.

According to FIG. 2, a driver assistance system according to a second exemplary embodiment of the invention recognizes a number of lighting units 1 in the form of headlights 2, rear lights 3 and street lamps 4. In an analysis step 102, we will recognize them

Switching status determined. If the quotient or the difference of switched-on lighting units 1 with respect to switched off lighting units 1 is greater than a predefined threshold value, the driver assistance system switches on the lighting unit of the vehicle - not shown here - independently. Alternatively, the driver assistance system can

Lighting unit already then independently

turn on when an absolute number of light units, for example within a predefined duration, is detected.

According to Figure 3 recognizes a driver assistance system according to a third embodiment of the invention, a road structure in the form of a tunnel 5. Based on geometric characteristics that can be removed from the image, also allows the spatial length of the tunnel, and thus by means of the speed of the vehicle in which Driver assistance system is arranged to determine the duration of transit. If the time duration is greater than a predefined threshold value, the lighting unit of the vehicle (not shown here) is switched on automatically by the driver assistance system. This switching can already happen before the vehicle enters the shadow area 6 in front of the tunnel 5, that is, before a darkening would be measurable, for example by measuring the brightness of the vehicle environment.

The invention is not limited to the above

described preferred embodiments. There are Rather, modifications thereof are conceivable, which of

Protection scope of the following claims are included. For example, it is also possible that the images are captured using a wide-angle camera, by means of which

at the same brightness information in the range above the vehicle or above an angle 45 ° above a horizontal can be determined, the brightness information is used, for example, to weightings within the analysis step with.

Claims

claims

1. Method for camera-based control of a

Lighting unit of a motor vehicle, comprising steps in which

in an imaging step (101) by means of a camera system a scene in front of the vehicle is imaged onto an image, and in an analysis step (102) the image is analyzed by recognizing structures in the image and assigning the recognized structures to object classes, and

in a control step (106), the lighting unit of the vehicle is controlled on the basis of the analysis as a function of the recognized object classes.

2. The method of claim 1, wherein the structures are recognized as road structures.

3. The method of claim 1 or 2, wherein the structures are assigned by using pattern recognition method the corresponding object classes.

4. The method according to any one of the preceding claims, wherein the structures are assigned by using neural networks the corresponding object classes.

5. The method according to any one of claims 2, 3 or 4, wherein the classification of the road structures by means of semantic segmentation.

6. The method of claim 2, wherein the structures

Road structural object classes, the bridges, tunnels (5), underpasses and / or building facades

lock in .

7. The method according to any one of claims 2 to 6, wherein the expected duration of a passage along the recognized

Road structure is calculated and the lighting unit only is turned on for a calculated duration above a threshold associated with the road structure.

8. The method according to any one of the preceding claims, wherein the structures are recognized as lighting units (1).

9. The method of claim 8, wherein the structures

Luminaire unit categories which include headlamps (2), taillights (3) and / or street lamps (4).

10. The method according to any one of claims 8 or 9, wherein a switching state of the lighting units (1) is determined and a comparison value of a number of switched on

Lighting units (1) compared to a number of switched off lighting units (1) is calculated and the lighting is turned on at a comparison value above a predefined threshold.

11. The method according to any one of claims 8 or 9, wherein a switching state of the lighting units (1) is determined and a a

Number of switched on lighting units (1) within a predefined period is determined and the

Lighting unit at a number above one

pre-defined threshold is turned on.

12. Use of a wide-angle camera system as a camera system in a method according to one of claims 1 to 11, wherein the scene is imaged by means of a wide-angle camera, by which a lower region below a horizontal and an upper region is detected above 45 ° above the horizontal, and wherein the brightness of the upper portion of the image is analyzed, and the illumination is turned on when it falls below a critical brightness.

13. A driver assistance system of a motor vehicle, comprising a camera for imaging a scene in an area in front of the motor vehicle; a lighting unit for lighting the scene; a control unit for analyzing the images captured by the camera and for controlling the lighting unit;

wherein the control unit is adapted to perform a method in which