DE102006060893A1 - Device and method for determining a free space in front of a vehicle - Google Patents

Abstract

The invention relates to a device (14) for determining a free space in front of a vehicle (10), comprising: - an imaging sensor (16) which is designed to record an image of an observation area (22) in front of the vehicle (10); an image processing unit (18) arranged to process the captured image to obtain a processing image in which image edges in the image are marked; and - a pattern recognition unit (20), which is designed to determine as free space in front of the vehicle (10) that part of the observation area (22) which corresponds to an image edge-free area of the processing image starting from a lower image edge of the image.

Description

The The invention relates to a device for determining a free space in front of a vehicle according to claim 1 and a method for determining a clearance in front of a vehicle according to claim 15th

Future driver assistance systems for stop and go support and Systems for driverless Operating transport vehicles such as subways, trains or Suspended runways will realize a fully automatic restart. For this purpose, the approach room directly in front of the vehicle (motor vehicle, Subway, train) from the restart after a stop, for example in a traffic jam, a vehicle in front is waiting in front of a traffic light or a stop of the subway in a station, to be monitored. For a sure Going backwards it is important to realize if the approach room is free is and in the meantime no persons or other objects into the space in front of the vehicle. This can be, for example with the usual Beam sensors such as radar, lidar, ultrasound systems implemented become. Due to its location in the vehicle front, in particular in the grille However, for example, radar or lidar systems are limited in able to monitor a space immediately in front of the vehicle. Therefore, it is preferable to mount behind a windshield Sensor system, for example, a camera system that a lot can capture wider space in front of the vehicle. They are, for example stereobased camera systems known to be a three-dimensional Allow detection of obstacles. The effort for such However, stereo cameras is costly and requires relatively a lot of space in the windscreen in the vehicle, as two Cameras must be placed in this area.

task The present invention is therefore an apparatus and a To provide a method for determining a clearance in front of a vehicle which have improved properties over the prior art.

These The object is achieved by a device for determining a free space in front of a vehicle with the features of claim 1 and a corresponding Method with the features of claim 17 solved. Further embodiments The invention will become apparent from the dependent claims.

One essential idea of the invention is that with a single imaging sensor a picture is taken in front of the vehicle and the edges analyzed therein, for example, by Boundary lines of shadows or lane markings in the Image are included (image edges). By evaluating these image edges, the one conclusion to give up obstacles in front of the vehicle, then can the free space in front the vehicle is to be considered as the space starting from a lower edge of the image corresponds to an image edge-free area. This means that the area in front of the vehicle as a free space is considered, in which no image edge-causing objects are included. This review of the free space in front of the vehicle works both when the vehicle is stationary (for example, in a traffic jam or before the start) and when the vehicle is moving and therefore offers the Possibility, in almost all on the road occurring situations to recognize an object in front of the vehicle.

Besides, too two consecutive pictures of an area to be monitored in front of the vehicle taken, compared these two pictures and the different elements in these two images by moving edges be mapped in a difference image. Then afterwards pattern recognition is applied to these motion edges, to determine whether it is in a supervised area before the Move vehicle objects into it. Will only a few movement edges found, it can be assumed that no object located in front of the vehicle, due to the entry of this object in the area in front of the vehicle in a period between the two Images a moving edge for this object would arise in the difference image. Because continued by the inventive approach only an imaging sensor is used is a lower cost Implementation as for example a stereo camera system possible. In addition, enough a smaller space for the arrangement of such a sensor in the vehicle.

• – einen bildgebenden Sensor, der ausgebildet ist, um ein Bild eines Beobachtungsbereichs vor dem Fahrzeug aufzunehmen;
• – eine Bildverarbeitungseinheit, die ausgebildet ist, um das aufgenommene Bild zu verarbeiten, um ein Verarbeitungsbild zu erhalten, in dem Bildkanten in dem Bild markiert sind; und
• – eine Mustererkennungseinheit, die ausgebildet ist, um als Freiraum vor dem Fahrzeug denjenigen Teil des Beobachtungsbereichs zu bestimmen, der einem Bildkanten-freien Bereich des Verarbeitungsbildes ausgehend von einem unteren Bildrand des Bildes entspricht.

The present invention relates in one embodiment to an apparatus for determining a clearance in front of a vehicle, comprising:

• An imaging sensor configured to capture an image of an observation area in front of the vehicle;
• An image processing unit configured to process the captured image to obtain a processing image in which image edges in the image are marked; and
• A pattern recognition unit which is designed to determine as free space in front of the vehicle that part of the observation area which corresponds to an image edge-free region of the processing image starting from a lower image edge of the image.

In a further embodiment, the Image processing unit to be formed to recognize lane markers and / or a shadow profile and mark as an image edge in the image. This offers the advantage that in this way a special adaptation to typical driving situations can take place, so that, for example, already frequently occurring edge courses on the roadway in the image processing unit can be preprogrammed and thus the evaluation can be facilitated. Alternatively or additionally, a minimum length of shadow-related edges can also be provided so that an estimate of the size of the corresponding object can be estimated in advance.

• – den bildgebenden Sensor umfassen, der ausgebildet ist, um zu einem ersten Zeitpunkt das Bild und zu einem nach dem ersten Zeitpunkt liegenden zweiten Zeitpunkt ein zweites Bild eines Beobachtungsbereichs vor dem Fahrzeug aufzunehmen;
• – die Bildverarbeitungseinheit ausgebildet ist, um durch Bilden einer Differenz von Bilddaten des Bildes und Bilddaten des zweiten Bildes ein Differenzbild zu erhalten, welches Bewegungskanten aufweist, die durch sich unterscheidende Bildpunktwerte der Bilddaten des Bildes gegenüber den Bilddaten des zweiten Bildes definiert sind; und
• – eine Mustererkennungseinheit, die ausgebildet ist, um als Freiraum vor dem Fahrzeug denjenigen Teil des Beobachtungsbereichs zu bestimmen, der einem Bewegungskanten-freien Bereich des Differenzbildes ausgehend von einem unteren Bildrand des Differenzbildes entspricht.

Also, in one embodiment, the apparatus may be for determining a clearance in front of the vehicle

• Comprise the imaging sensor, which is designed to record the image at a first point in time and to record a second image of an observation area in front of the vehicle at a second point in time lying after the first point in time;
• - the image processing unit is adapted to obtain, by taking a difference of image data of the image and image data of the second image, a difference image having motion edges defined by differing pixel values of the image data of the image versus the image data of the second image; and
• - A pattern recognition unit which is designed to determine as free space in front of the vehicle that part of the observation area corresponding to a moving edge-free area of the difference image, starting from a lower image edge of the difference image.

This allows not only to recognize an obstacle in front of the vehicle with individual pictures and thus a clear flexibilization of the deployment scenario Offer. In particular, the safety of a moving vehicle clearly increased when this dynamic aspect of obstacle detection is taken into account becomes.

According to one another embodiment According to the present invention, the image processing unit can be formed be a threshold on pixel values of the difference image apply such that in the difference image only motion edges are included whose pixel values are above the threshold. This offers the advantage that minor disturbances in the difference image not taken into account in the pattern recognition and thus a pattern recognition works much more trouble-free.

In a further embodiment the direction of movement of the edges is analyzed. Edges that are move into the observed area are weighted more heavily as edges that move out of the observed area. In other words For example, the pattern recognition unit may be configured to have a movement direction to analyze the edges and to assist in determining the free space the vehicle those motion edges that are in the observation area move in stronger to weight as movement edges, resulting from the observation area move out, with the heavier weights the observation edge moving into the observation area such that with respect this observation edge moving into the movement area a safe distance in the space to be determined in front of the vehicle is provided. This will ensure that objects that are move into the observed area treated as particularly vulnerable to collision become.

Of the Imaging sensor can be present in a further embodiment of the present Be formed to the image and the second image during a Stop the vehicle. In particular, shortly before restarting is it important, possible To recognize objects in the approach area in front of a vehicle safely. Farther This feature allows the driver to be psychologically often In addition, for example visually considered as monotone traffic situation supports can be. At the same time, the evaluation of images in the standing Condition of the vehicle can be significantly relieved, because of the usually minor changes from image to image the image processing less resources in shape claimed by computing power.

In a further embodiment According to the present invention, the pattern recognition unit can be formed be to signal a blocking of starting the vehicle, if not at least one free space of a predefined size in front of the Vehicle was detected. This ensures that no rear-end collision if the distance to a preceding vehicle is too short is caused. Also, it can be ensured that no startup is done by persons who may while Waiting for the vehicles on the road and stepped between the vehicles have run.

According to a further embodiment of the present invention, the pattern recognition unit may also be designed to signal a blocking of the startup of the vehicle if an object having a predefined minimum size was detected in a predefined area in front of the vehicle. This allows the approach to be blocked especially when large objects, such as For example, people are detected in the area in front of the vehicle. However, if smaller objects are detected in front of the vehicle, such as garbage or newspapers, a start of the vehicle can still be done without a risk to the vehicle or a person is to be feared.

According to one embodiment of the invention, the pattern recognition unit can be designed to perform a geometric calculation of the free space in front of the vehicle from the determined values of the differential image by the distance d point to the imaging sensor from the installation height h of the imaging sensor in the vehicle, the sensor pitch angle α Image line y of the point whose distance is to be determined, the pixel size η and the focal length f is determined according to the following formula:

In a further embodiment According to the present invention, the imaging sensor can be formed to take more pictures after the second time, and the image processing unit may be further configured to to save the first and second picture as well as the other pictures and from the second picture and further pictures further difference pictures to absorb with corresponding movement edges. This offers itself the possibility, while standstill to continuously monitor the space in front of the vehicle.

Also may be the pattern recognition unit according to another embodiment be formed of the present invention to prevent penetration of a Object in a space to be kept in front of a stationary vehicle by a comparison of the movement edges in different difference pictures to recognize. This not only allows the static state before the vehicle by evaluating a single difference image, but also a movement of an object can be detected. in this connection For example, a distinction is made between movements in the observation room into and out of the observation room. A movement out of the observation room, e.g. due to an arriving Vehicle should be as possible do not block your own restart. In contrast, a movement into the observation room classified as relevant for collision and, for example, as a result, an automatic restart be blocked. The pattern recognition unit can be realized by integrating the edge pixels into a profile in columns become. Are in this profile detected peaks above one certain threshold, so for example, a flag is set and thus signaling that something is in the observation area has moved

In a further embodiment For example, the pattern recognition unit may be designed to be composed of a Plurality of difference images a movement of one to be kept in the Space in front of the vehicle penetrating object in this to be kept Space to track. This provides the opportunity to estimate whether the object soon leave the area in front of the vehicle, so that already preparatory activities can be taken to restart.

In a further embodiment In the present invention, the imaging sensor may be a camera be. this makes possible besides the use of an ultrasonic sensor, a radar sensor or a lidar sensor is a cost effective implementation of the Invention. The image of a camera can also be used for other functions of the driver assistance system be used.

Farther can in one embodiment In the present invention, the device has another imaging Have sensor and the pattern recognition unit be formed, to the image data of the first or second image with image data of the other to verify the plausibility of the imaging sensor. This ensures that an error in the imaging sensor can be detected easily and quickly can and a danger of property or persons due to accidental re-entry of the vehicle can be avoided.

• – Aufnehmen eines Bildes eines Beobachtungsbereichs vor dem Fahrzeug;
• – Verarbeiten des aufgenommenen Bildes, um ein Verarbeitungsbild zu erhalten, in dem Bildkanten in dem Bild markiert sind; und
• – Bestimmen desjenigen Teils des Beobachtungsbereichs als Freiraum vor dem Fahrzeug, der einem Bildkanten-freien Bereich des Verarbeitungsbildes ausgehend von einem unteren Bildrand des Bildes entspricht.

Also, in one embodiment, a method for determining a clearance in front of a vehicle may be provided, comprising the steps of:

• - taking an image of an observation area in front of the vehicle;
• Processing the captured image to obtain a processing image in which image edges in the image are marked; and
• - Determining that part of the observation area as a free space in front of the vehicle, which corresponds to a picture edge-free area of the processing image from a lower edge of the image.

• – Aufnehmen eines ersten Bildes eines Beobachtungsbereichs vor dem Fahrzeug zu einem ersten Zeitpunkt und eines zweiten Bildes des Beobachtungsbereichs vor dem Fahrzeug zu einem nach dem ersten Zeitpunkt liegenden zweiten Zeitpunkt durch einen bildgebenden Sensor;
• – Bilden einer Differenz von Bilddaten des ersten Bildes und Bilddaten des zweiten Bildes, um ein Differenzbild zu erhalten, welches Bewegungskanten aufweist, die durch sich unterscheidende Bildpunktwerte der Bilddaten des ersten Bildes gegenüber den Bilddaten des zweiten Bildes definiert sind; und
• – Bestimmen desjenigen Teils des Beobachtungsbereichs als Freiraum vor dem Fahrzeug, der einem Bewegungskanten-freien Bereich des Differenzbildes ausgehend von einem unteren Bildrand des Differenzbildes entspricht. Dieses Bestimmen kann beispielsweise dadurch erfolgen, dass ein Setzen eines Flags ausgeführt wird, wenn die Musterekennungseinheit Kantenbewegungen in den Beobachtungsbereich hinein detektiert. Die Mustererkennungseinheit kann realisiert werden, indem spaltenweise die Kantenpixel zu einem Profil aufintegriert werden. Werden in diesem Profil Peaks erkannt die oberhalb eines gewissen Schwellwertes liegen, so wird das Flag gesetzt und damit signalisiert, dass sich etwas in den Beobachtungsbereich bewegt hat.

In a further embodiment of the present invention, a method for determining a clearance in front of a vehicle may be provided, comprising the following steps:

• - taking a first image of an observation area in front of the vehicle at a first time and a second image of the observation area in front of the vehicle at a second time lying after the first time by an imaging sensor;
• - forming a difference of image data of the first image and image data of the second image to obtain a difference image having motion edges defined by differing pixel values of the image data of the first image versus the image data of the second image; and
• - Determining that part of the observation area as a free space in front of the vehicle, which corresponds to a moving edge-free area of the difference image, starting from a lower edge image of the difference image. This determination can be made, for example, by the fact that a setting of a flag is carried out when the pattern recognition unit detects edge movements in the observation area. The pattern recognition unit can be realized by integrating the edge pixels into a profile in columns. If peaks are detected in this profile which are above a certain threshold value, then the flag is set and thus signaled that something has moved into the observation area.

Further Advantages and applications The present invention will become apparent from the following description in conjunction with the embodiments illustrated in the drawings.

In the description, in the claims, in the abstract and in the drawings are those in the listed below List of reference symbols used terms and associated reference numerals used.

The Drawings show in

1 a representation of the use of an embodiment of the present invention;

2 a representation of the detection of movement edges in an image;

3 a representation of the geometric calculation of the actual free space in front of the vehicle from the values determined in the difference image; and

4 a representation of the weighting of the movement edges for assessing the free space in front of a vehicle depending on a direction of movement of an object in front of the vehicle.

in the Following can identical and / or functionally identical elements with the same reference numerals be provided. The absolute values given below and dimensions are only exemplary values and do not represent a limitation of Invention on such dimensions.

1 shows a representation of the use of an embodiment of the present invention. Here is a vehicle 10 shown that, for example, in a traffic jam behind a truck 12 located. The vehicle 10 is with an embodiment of the device according to the invention 14 fitted.

This in 1 illustrated embodiment of the device 14 includes an imaging sensor 16 which may be, for example, an optical camera, an ultrasonic sensor, a radar sensor or lidar sensor. Furthermore, the device comprises 14 an image processing unit 18 and one with the image processing unit 18 coupled pattern recognition unit 20 , The imaging sensor 16 Can be in the area of the windshield of the vehicle 10 be arranged. This results in an observation area 22 due to the increased arrangement of the imaging sensor 16 is substantially larger than is possible in a mounting of the sensor, for example in the front, in particular radiator area of a motor vehicle.

In this observation area 22 can be an object 24 be recognized, for example, a person during the traffic jam between the vehicle 10 and the truck 12 has gone. The imaging sensor 16 takes a first picture, in which the object 24 not yet between the vehicle 10 and the truck 12 located. Thereafter, a second image is taken, in which the object 24 already between the vehicle 10 and the truck 12 located. The first and second images are then compared with each other by producing a difference image from the first and second images. The through the object 24 Differences caused in the second image lead to movement edges in the difference image resulting from the presence of the object 24 in the observation area 22 result. There will now be a distance between through the object 24 caused bottom trailing edge and a lower edge of the difference image measured, a conclusion on the position of the object 24 in the observation area 22 or a conclusion to a distance of the object 24 from the vehicle 10 allows. Here, it is assumed that the lower edge of the difference image or the first and second image of the front position of the vehicle 10 equivalent.

2 shows an illustration of the detection of image edges by way of example with reference to two images. In the left partial representation of 2 is a picture of a camera reproduced, which is shown as the vehicle with the device according to the invention on a road behind a load dare 12 located. As a basic assumption, it is assumed that an empty lane, except for the lane markings, does not contain any strong (image) edges. Shadow edges formed by "raised" objects stand out from the roadway in their contrast and suggest objects or obstacles on the roadway (or at least directly next to it). In the right partial representation of 2 It can be seen how image edges can be determined and marked in the (static) evaluation of a single image. In particular, it can be clearly seen that the area in front of the vehicle was analyzed in such a way by the lane markings and the image edges of the shadow course behind the truck that in the area 30 No movement edges were detected in front of the vehicle and that this area 30 as a free space in front of the vehicle can be considered. In contrast, the side strips are detected by edge detection in an edge region of the road and evaluated as a limit to the free space in front of the vehicle. In the right part diagram the 2 is clearly recognizable by the edge detection that there is no object between the vehicle and the truck. This also allows the use of the invention for tracking. If the distance to the preceding truck does not change during the journey, there is no moving image edge in the moving image, so that there is no change in the distance to the truck and thus no restriction of the free space in front of the vehicle.

bestimmt.In 3 It is shown how a geometric calculation of the free space in front of the vehicle is carried out from the determined values of the differential image, which was generated from images taken by a camera mounted behind the rearview mirror on the windshield of a motor vehicle. Hereby poses 3 the distance is determined from the camera installation height h in the vehicle, the camera pitch angle α, the image line y of the point whose distance is to be determined, the pixel size η and the camera focal length f according to the formula

certainly.

in the The following describes an exemplary method which is a automatic free space determination realized by means of monocular camera, for example for the restart of a stopped vehicle. It is intended that an automatic restart only takes place when there is a free space before the vehicle is sized so that it is larger than one before as sufficient large defined free space is and in the sensor system or the associated evaluation unit is deposited. Can not this be done by one or more sensor systems (camera, Ultrasound, radar, lidar, etc.) is an active one Intervention, for example, the driver, necessary for the Ride is continued.

in the Below is an advantageously configured device and a just described a method that the space in front of a vehicle determined before a return. The vehicle is with at least a sensor system, with associated evaluation unit and a storage device. In the storage device When the vehicle is at a standstill, information about moving objects is displayed in the detection range of the sensor or sensors in saved. Optionally, moving objects are tracked during standstill, for more information about, for example, the relative speed or predict a probable behavior of the object (Transverse movement, object moves to the vehicle to or from the vehicle path).

Come Several sensor systems are used, so their data for plausibility of the result (Sensor data fusion). Possible is also a limitation on objects that are on the road or in a previously defined relevant Area are located. The method thus selects objects that are after stand still in the space in front of the vehicle. In front of you Restart is checked whether these objects are still in front of the vehicle or this Have already left the area. A criterion of whether an active intervention, e.g. the driver of the vehicle, is intended to continue the journey or an automatic Restart is done, depending on that one of the previously described selected objects still in stops the given free space. Other criteria can be e.g. the object properties (size, relative speed, Human, object, vehicle etc.).

• 1. Es wird ein Differenzbild aus zwei aufeinander folgenden Bildern berechnet. Es entsteht ein Bild mit Bewegungskanten.
• 2. Es wird ein Schwellenwert auf das Differenzbild angewendet, so dass nur starke Bewegungskanten verbleiben.
• 3. Alle Bewegungskanten werden während des Stillstands gespeichert.
• 4. Vor einer Wiederanfahrt wird der Freiraum berechnet, zum Beispiel werden ausgehend vom unteren Bildrand die nächsten gespeicherten Bewegungskanten bestimmt. Die entstehende Kontur begrenzt den Freiraum.

The method described above is suitable, inter alia, for a camera sensor. Assuming that an empty lane does not contain strong edges, the edges in the image data are extracted. Shadow edges that move into the picture form shadow edges and stand out from the roadway in their contrast. A method of detecting moving objects with a camera could be as follows:

• 1. A difference image is calculated from two successive images. The result is a picture with moving edges.
• 2. A threshold value is applied to the difference image, so that only strong Bewegungsungskan remain.
• 3. All motion edges are stored during standstill.
• 4. Before a restart, the free space is calculated, for example, starting from the lower edge of the image, the next stored movement edges are determined. The resulting contour limits the free space.

• 1. Es wird ein Differenzbild aus zwei aufeinander folgenden Bildern T1 und T1 + dt berechnet. Es entsteht ein (Differenz-)Bild mit Bewegungskanten.
• 2. Es wird ein Schwellenwert auf das Differenzbild angewendet, so dass nur starke Bewegungskanten verbleiben.
• 3. Die Bewegungskanten werden spaltenweise aufintegriert, so dass ein Kantenprofil entsteht.
• 4. Überschreitet das Profil einen Schwellwert wird ein Freiraum-Flag gesetzt, das eine Bewegung in den Freiraum anzeigt
• 5. Vor einer Wiederanfahrt wird das Freiraum-Flag abgefragt. Beispielsweise kann das Wiederanfahren blockiert werden, wenn das Freiraum-Flag gesetzt ist und das Wiederanfahren erst dann freigegeben werden, wenn sichergestellt ist, dass sich kein Objekt mehr vor dem Fahrzeug befindet. Dies kann durch ein Rücksetzen des Flags erfolgen, wenn sich das Objekt aus dem Beobachtungsbereich hinausbewegt und durch eine Auswertung diese sich aus dem Beobachtungsbereich hinausbewegenden Bewegungskanten ein Löschen des Flags bewirken.

Alternatively, the method according to an exemplary embodiment can also proceed as follows 4 is shown in more detail:

• 1. A difference image is calculated from two successive images T1 and T1 + dt. The result is a (difference) image with motion edges.
• 2. A threshold value is applied to the difference image so that only strong motion edges remain.
• 3. The movement edges are integrated on a column-by-column basis so that an edge profile is created.
• 4. If the profile exceeds a threshold, a free space flag is set that indicates movement into the free space
• 5. Before a restart, the free space flag is queried. For example, restarting may be blocked if the clearance flag is set and restarting is not enabled until it is certain that there is no object in front of the vehicle. This can be done by resetting the flag when the object moves out of the observation area and causes an erasure of the flag by evaluating this movement edge moving out of the observation area.

Farther also become static image edges when analyzing the difference image such as. the lane markings filtered out, as these in both Images occur. Thus, the focus can only be directed to the "moving" image edges, which makes the image processing much easier.

10

vehicle

12

truck

14

contraption for determining a free space in front of a

vehicle

16

imaging Sensor, camera

18

Image processing unit

20

Pattern recognition unit

22

observation area

24

object

30

when Free space recognizable area

Claims (17)

Contraption ( 14 ) for determining a free space in front of a vehicle ( 10 ), comprising: - an imaging sensor ( 16 ) which is adapted to form an image of an observation area ( 22 ) in front of the vehicle ( 10 ); An image processing unit ( 18 ) configured to process the captured image to obtain a processing image in which image edges in the image are marked; and a pattern recognition unit ( 20 ), which is designed to be free in front of the vehicle ( 10 ) that part of the observation area ( 22 ) corresponding to an image edge free area of the processing image from a lower image edge of the image. Contraption ( 14 ) according to claim 1, characterized in that the image processing unit ( 18 ) is formed to recognize lane markers and / or a shadow course and mark as an image edge in the image. Contraption ( 14 ) according to claim 1, characterized in that - the imaging sensor ( 16 ) is formed, at a first time the image and at a second time after the second time a second image of the observation area ( 22 ) in front of the vehicle ( 10 ); The image processing unit ( 18 ) is adapted to obtain, by taking a difference of image data of the image and image data of the second image, a difference image having moving edges defined by differing pixel values of the image data of the image versus the image data of the second image, the difference image being a processing image provided; and - the pattern recognition unit ( 20 ) is designed to be free space in front of the vehicle ( 10 ) that part of the observation area ( 22 ), which corresponds to a moving edge-free region of the difference image starting from a lower image edge of the difference image. Device according to claim 3, characterized in that the image processing unit ( 18 ) is adapted to apply a threshold value to pixel values of the difference image such that in the difference image only motion edges are included whose pixel values are above the threshold value. Device according to one of claims 3 or 4, characterized in that the imaging sensor ( 16 ) is formed to the image and the second image during a standstill of the vehicle ( 10 ). Device according to claim 5, characterized in that the pattern recognition unit ( 20 ) is configured to block the starting of the vehicle ( 10 ), if not at least one clearance of a predefined size in front of the vehicle ( 10 ) was detected. Device according to claim 5 or 6, characterized in that the pattern recognition unit ( 20 ) is configured to block the starting of the vehicle ( 10 ) to signal when an object ( 24 ) with a predefined minimum size in a predefined area in front of the vehicle ( 10 ) was detected. Device according to one of claims 3 to 7, characterized in that the pattern recognition unit ( 20 ) is designed to perform a geometric calculation of the free space in front of the vehicle from the determined values of the differential image by the distance d point to the imaging sensor ( 16 ) from the installation height h of the imaging sensor ( 16 ) in the vehicle, the sensor pitch angle α, the image line y of the point whose distance is to be determined, the pixel size η and the focal length f are determined according to the following formula:

Device according to one of claims 3 to 8, characterized in that the imaging sensor ( 16 ) is adapted to take further images after the second time and the image processing unit ( 18 ) is further configured to store the image and the second image and the further images and to record from the second image and further images further difference images with corresponding movement edges. Device according to claim 9, characterized in that the pattern recognition unit ( 20 ) is designed to prevent intrusion of an object ( 24 ) in a free space in front of a stationary vehicle ( 10 ) by comparing the movement edges in different difference images. Device according to claim 10, characterized in that the pattern recognition unit ( 20 ) is configured to analyze a direction of movement of the edges and to determine the free space in front of the vehicle ( 10 ) those edges of motion that are in the observation area ( 22 Move inwards to weight more than movement edges, resulting from the observation area ( 22 ), whereby the heavier weights of the observation area ( 22 ) in such a way that with respect to this in the range of motion ( 22 ) moving in the observation edge a safety distance in the free space to be determined in front of the vehicle ( 10 ) is provided. Device according to one of claims 10 or 11, characterized in that the pattern recognition unit ( 20 ) is formed in order to be stationary when the vehicle is stationary ( 10 ) to analyze a direction of movement of the edges and to determine the free space in front of the vehicle ( 10 ) those edges of motion that are in the observation area ( 22 Move inwards to weight more than movement edges, resulting from the observation area ( 22 ), whereby the heavier weights of the observation area ( 22 ) in such a way that with respect to this in the range of motion ( 22 ) moving in observation edges and a restart of the vehicle ( 10 ) in response to the signal. Device according to claim 12, characterized in that the pattern recognition unit ( 20 ) is formed in order to aufintegrieren the moving edges of the difference images column by column, so that an edge profile is formed and to set when exceeding the profile above a threshold signal, to indicate a movement in the free space. Device according to one of claims 9 to 13, characterized in that the pattern recognition unit ( 20 ) is adapted to move from a plurality of differential images a movement of a space to be kept clear in front of the vehicle ( 10 ) penetrating object ( 24 ) in this room to be kept free. Device according to one of claims 1 to 14, characterized in that the imaging sensor ( 16 ) is a camera. Device according to one of claims 3 to 15, characterized in that the device ( 14 ) has a further imaging sensor and the image processing unit ( 18 ) is designed to check the image data of the image or the second image with image data of the further imaging sensor for plausibility. Procedure ( 14 ) for determining a free space in front of a vehicle ( 10 ), comprising the following steps: - taking an image of an observation area ( 22 ) in front of the vehicle ( 10 ); Processing the captured image to obtain a processing image in which image edges in the image are marked; and Determining the part of the observation area ( 22 ) as clearance in front of the vehicle ( 10 ) corresponding to an image edge-free region of the processing image from a lower image edge of the image.