Patent Application
20150332099
The present invention relates to a sensor system for a motor vehicle. In particular, the present invention relates to an apparatus and a method for detecting precipitation for a motor vehicle.
WO 2012/092911 A1 describes a method for detecting rain comprising a camera and a lighting source. The camera is disposed behind a pane, in particular in the interior of a vehicle behind a windscreen and focused onto a remote region that lies in front of the pane. The lighting source for generating at least one light beam that is directed at the pane directs the at least one light beam towards the pane such that at least one beam that is reflected from the outer face of the pane impinges on the camera. The light quantity of the at least one beam that impinges on the camera can be measured by the camera.
In order to detect raindrops also at night, it is described in WO 2010/072198 A1 to couple light via a coupling element into the windscreen and to guide it via total reflection in the pane. By a decoupling element the totally reflected light is decoupled in the direction of the camera.
When there are water drops on the windscreen, a part of the light is decoupled and is no longer totally reflected to the decoupling element. It is again disadvantageous here that for each modified pane inclination the integrated camera lighting unit must be mechanically adapted to the modified installation condition.
In known active methods for rain detection a light source specifically required for this purpose is used, by which the reduction of the reflected light intensity is measured and used for rain detection. The projected light, however, encloses only a relatively small surface on the windscreen, which is used for the detection. Raindrops outside the detection surface remain undetected due to the light sources or LED-based detection method. In case of light or moderate rain it can happen that the windscreen is wetted uniformly with drops, whereas no drops fall onto the position of the detection surface. For a stable rain detection, in particular with light precipitation it is, therefore, desirable to increase the detection surface.
The object of the invention is to increase the efficiency of sensor systems.
This object is achieved by the subject-matters of the independent claims. Embodiments and further developments can be taken from the dependent claims, the description and the drawings.
A first aspect of the invention relates to an apparatus for detecting precipitation on a windscreen for a motor vehicle, the apparatus comprising a camera device, which is designed such that it records an image at a first moment and a second image at a second moment following the first moment, a comparison device, which is designed such that it produces a comparison image by comparing the first image with the second image, and an image detection device, which is designed such that it identifies objects on the first comparison image.
According to the second aspect of the present invention a method for detecting precipitation on a windscreen for a motor vehicle is provided, the method comprising the following steps: recording a first image at a first moment and recording a second image at a second moment following the first moment, producing a comparison image from the first image and the second image and identifying objects on the comparison image.
Advantageous embodiments of the invention are characterized in the dependent claims.
The present invention is characterized by the fact that for increasing the detection surface the entire image of a camera device or of a camera is used. The field of view of a camera, which is used for traffic sign detection, for a lane change assistant or for a blind spot monitoring or for any other driver assistance function in the vehicle, provides a detection surface which is increased compared to an LED-based reflection sensor, so that a probability of occurrence of a drop in the entire image is clearly higher than the probability of occurrence of a drop in the detection surface covered by the LEDs.
The camera is located for example at the windscreen of the vehicle and additionally focused onto the far range.
The present invention permits advantageously that the camera can continue to be used as an environment sensor for driver assistance systems. The objects to be detected such as e.g. lane markings, traffic signs, other road users, etc. lie in the far range of the camera compared to the windscreen and can be imaged by the camera in a focused manner, so there is no more need for focusing the camera to different distance regions.
This, however, leads to a blurred representation without contours of the drops in the images recorded by the camera device. The drops adhering to the windscreen or any other impurities can be seen in the image as small defects or intensity fluctuations or distortions, as the optical image path is slightly changed by the drops.
In addition, the drops separated on the windscreen can lead to brightness changes in the recorded image by light coupling caused by refraction or any other optical effects.
Using the entire image for detecting precipitation on the windscreen can increase the detection surface many times over, however, for this purpose a method adapted to the present invention for reliable detection of raindrops under unfocussed conditions is necessary.
The present invention permits advantageously to provide rain detection without any lighting of the windscreen.
The present invention uses a comparison of two single images recorded at different times. If as a result of the prevailing rain a drop is incident on the viewing range of the camera device within the time offset of the two recorded images, then by the described intensity change the drop will become visible in the difference image of the two images. The advantage of this method of the present invention is the reduction of the computing effort needed to detect the drop.
Admittedly, also moving objects will become visible in the difference image. Other changes in the camera image, not resulting from the precipitation, lead to disturbances of the rain detection and must be distinguished from potential drops.
Advantageously, horizontally moving objects can be ignored in the rain detection. In particular with stationary objects and light rain the detection probability is strongly reduced with a small detection surface.
Typically, with a stationary vehicle the observed movements of other traffic users run at right angles to the vehicle or comprise at least one transverse component of its motion vector, i.e. in case the vehicle stands at the traffic light and pedestrians cross the road. Here, it is assumed that only horizontally moving objects are classified to be disturbing in the rain detection.
Horizontal movements of identified objects can be detected and filtered from recorded pictures or images via conventional methods of image processing, such as e.g. tracking over several images, and thus can be distinguished from drops, which either newly impinge on the pane, hence are new or which, as a rule, move from top to bottom in the camera image.
Advantageously, objects running diagonally or moving can be ignored. The method of ignoring horizontally running objects can be extended to a moving ego-vehicle, in which the apparatus according to the present invention is used.
By the proper motion of the ego-vehicle additional motion vectors of the objects appear in the image, which run transversely and diagonally from the vanishing point to the side. These motions can also be filtered to distinguish them from changes in the image caused by raindrops.
Specifically with a very light precipitation an increased detection surface and a high detection quality is desirable, as in this case only occasional raindrops reach the detection surface.
Advantageously, blurred edges can be observed. The blurred representation of the drops leads to a softening of the edges in the background. Sharp edges in the background can be considered temporally to detect a possible blurring by drops.
Further, clear features in the difference image can be classified. Another option is to classify objects into drops and non-drops. For this purpose, features which are clear and to be verified by the image detection device for a drop can be defined in the difference image or picture.
A corresponding classifier then distinguishes in the difference image whether the features of a shown object corresponds to the drop features or not. This would have the advantage that objects from all directions can be ignored.
The invention represents a simple but reliable option to detect rain with a vehicle camera, in particular with a driver assistance camera.
The present invention permits advantageously to measure the wetting of a measuring field defined by the field of view of the camera and further permits to estimate the wetting of the pane and for example to activate a windscreen wiper of the vehicle, when an appropriate wetting is reached.
In other words, a core idea of the present invention can be that image regions are observed, in which impinging of moving objects is unlikely. These can be e.g. sky regions, engine covers or corners of the image.
With this the present invention allows advantageously an increase of the detection surface by using the entire camera image.
The present invention permits advantageously that a robust distinction between drop and other background objects can be made.
According to an advantageous further development the other, still sharply shown background objects could, however, be considered for one or more driver assistance function(s).
In an advantageous embodiment of the present invention it is provided that the image detection device is designed such that it determines a movement of at least one object from the amount of identified objects. This can be made e.g. when comparing the first image with the second image by the comparison device. The comparison image thus can provide information which image points contain an object point in the first image and no longer in the second image or vice versa. In the first case the object would have moved away from this image point, in the reverse case the object would have moved towards this image point or it newly appeared at this image point.
This permits in a comparably easy way to detect the movements of objects.
In an advantageous embodiment of the present invention it is provided that the comparison device is designed such that it produces a difference image as a first comparison image.
This permits in an easy way to detect changes of objects and appearance of new objects.
In an advantageous embodiment it is provided that the camera device is designed such that it records subsequent to the second image one or more images as a third image or as further images.
This results advantageously in an increased reliability of a motion detection of the identified objects and the determination of a motion direction of the identified objects.
In an advantageous embodiment of the present invention it is provided that the comparison device is designed such that it produces a second comparison image by comparing the second image with the third image.
This permits in an easy way to identify objects reliably.
In an advantageous embodiment of the present invention it is provided that the image detection device is designed such that it detects a movement of an object from the amount of identified objects by comparing the first comparison image with the second comparison image.
In an advantageous embodiment of the present invention it is provided that the image detection device is designed such that it detects one or that movement of an object from the amount of identified objects by comparing the first image with the second image.
Here, different methods for motion detection can be combined or used for cross-checkings.
This permits advantageously to detect the movement of the objects with a reduced computing effort of the image detection device or of the comparison device.
In an advantageous embodiment of the present invention it is provided that the image detection device is designed such that it assigns a motion direction to the detected motion of the object and identifies the object on the basis of the associated motion direction.
This increases advantageously robustness or reliability of the method for raindrop detection as performed by the apparatus for detecting precipitation.
In an advantageous embodiment of the present invention it is provided that the image detection device is designed such that it classifies the identified objects on the first comparison image.
The objects to be classified, for example raindrops or non-raindrops, such as motor vehicles or pedestrians, are described by their features or characteristics.
These features are recognized by the image detection device and span an n-dimensional feature space. An object, for example a raindrop, thus is represented by a certain expression of the features 1, . . . , n, i.e. by a feature vector in the feature space.
In an advantageous embodiment of the present invention it is provided that the image detection device is designed such that it classifies the identified, objects on the first comparison image on the basis of the form of the identified objects.
In an advantageous embodiment of the present invention it is provided that the camera device is designed such that it records the images with a predefined depth of field and that the image detection device is designed such that it identifies the objects within and outside the predefined depth of field.
This permits advantageously to determine whether and if so, how much rain or how much other precipitation in form of flies, midges and other insects or dust lies on the windscreen.
In an advantageous embodiment of the present invention it is provided that the image detection device is designed such that it classifies the identified objects on the first comparison image on the basis of a local frequency of the identified objects. The local frequency or spatial frequency is the reciprocal vale of the spatial period length of an object.
In an advantageous embodiment of the present invention it is provided that the camera device is designed such that it records the first image and the second image without any lighting. For the apparatus according to the invention for detecting precipitation on the windscreen of a vehicle no active lighting of the windscreen is required in accordance with this preferred embodiment. By comparing two or more images also in the unlighted condition changes based on precipitation on the pane can be reliably detected. It can be assumed that in darkness the headlights of the own vehicle are activated, so that the camera can record images with sufficient intensity.
This permits advantageously to do without an additional light source and permits a power-saving raindrop detection. Further, this permits a higher traffic security, as unnecessary light sources are avoided, which could confuse other traffic users.
The described embodiment and further developments can be combined in any way.
Further possible embodiments, further developments and implementations of the invention comprise also not explicitly mentioned combinations of features of the invention described above or in the following in relation to the exemplary embodiments.
The enclosed drawings shall provide further understanding of the embodiments of the invention. The enclosed drawings demonstrate the embodiments and serve in context with the description of the explanation of concepts of the invention.
Other embodiments and many of the mentioned advantages result from the drawings. The shown elements of the drawings are not necessarily shown to scale.
In the figures of the drawings identical and/or functionally identical elements, parts, components or method steps are referred to by the same reference numerals, unless otherwise indicated.
An apparatus 100 for detecting precipitation for a motor vehicle comprises for example a camera device 101, a comparison device 102, and an image detection device 103.
The camera device 101 can be designed for example such that it records a first image at a first moment and a second image at a second moment following the first moment.
The camera device 101 can be embodied as an optical camera and is disposed for example behind a pane or windscreen 20, in particular in the interior of a vehicle, e.g. behind a windscreen, and focused onto a remote region that lies in front of the pane or windscreen 20. The camera preferably comprises an objective used for focusing and an image sensor, e.g. a CCD sensor or CMOS sensor and can record images within an image angle region or spatial angle referred to as image angle 11.
The comparison device 102 can be designed for example such that it produces a first comparison image by comparing the first image with the second image.
The image detection device 103 can be designed such that it identifies objects on the first comparison image.
The camera device 101 can be designed such that it records the images with a predefined depth of field and the image detection device 103 can be designed such that it identifies the objects within and outside the predefined depth of field.
For example the depth of field or focal depth can be defined as a measure for the expansion of the sharp region in the object space of a copied optical system, the camera device. The focal depth describes the quantity of the distance range, within which an object appears sufficiently sharp in the image of the camera optics.
For example a large depth of field is obtained by small camera apertures or objectives with short focal lengths.
In this case the focal depth of the camera device 101 can reach from approx. 5 m to 100 m or from 8 m to 200 m or from 12 m to infinity.
The method for detecting precipitation for a motor vehicle comprises the following steps:
A first step of the method comprises recording S1 of a first image at a first moment and recording of a second image at a second moment following the first moment.
A second step of the method comprises producing S2 a comparison image from the first image and the second image.
A third step of the method comprises identifying S3 of objects on the comparison image.
The method steps can be repeated iteravely or recursively and in any sequence.
A detection surface 40 of an LED-based detection apparatus is shown in explanation on the windscreen 20 shown in
The raindrops 31 outside the detection surface 40 remain undetected by the LED-based detection method. With a light to moderate rain it can happen that the windscreen is uniformly wetted with drops, whereas no drops fall onto the position of the detection surface 41, as is shown in
For a stable rain detection in particular with light precipitation it is, therefore, desirable to increase the detection surface.
The increased detection surface of the camera device 101, which is defined by the field of view or image angle 11 of the camera device 101, permits already with a reduced precipitation intensity to detect raindrops 31.
Although the present invention was described above on the basis of preferred exemplary embodiments, it is not restricted thereto, but is modifiable in many ways. In particular, the invention can be changed or modified in various ways without deviating from the core of the invention.
In addition, it may be pointed out that “including” and “comprising” do not exclude other elements or steps and “one” does not exclude a plurality.
Further, it may be pointed out that the features or steps, which were described with reference to one of the above exemplary embodiments, can be used also combined with other features and steps of other exemplary embodiments described above. Reference numerals in the claims are not to be considered as restrictions.
| Number | Date | Country | Kind |
|---|---|---|---|
| 10 2014 209 197.7 | May 2014 | DE | national |
