DE102020112825A1 - Method for detecting relevant static objects within a lane as well as computing device for a driver assistance system of a vehicle - Google Patents

Abstract

The invention relates to a method for detecting relevant static objects in the surroundings of a vehicle with the following steps: receiving sensor data from at least one surroundings sensor of the vehicle, the sensor data describing the surroundings, determining a static occupancy map based on the sensor data, whereby the static occupancy map is not describes moving parts of the environment, the sensor data being continuously received and the static occupancy map being accumulated using the sensor data, determining lane data that describe at least one lane in the environment, and identifying relevant static objects that are at least partially within the at least one Lane, from the static objects based on the accumulated static occupancy map and the lane data.

Description

The present invention relates to a method for detecting relevant static objects in the surroundings of a vehicle. The present invention also relates to a computing device for a driver assistance system of a vehicle. The present invention also relates to a computer program.

In driver assistance systems that enable automated or autonomous driving, it is essential to reliably detect objects in the vicinity of the vehicle. It is known from the prior art that all relevant objects in the vicinity of the vehicle are tracked over time. In most cases, no distinction is made as to whether these objects are static or stationary objects or dynamic or moving objects. If, for example, all vehicles parked on the hard shoulder and in parking lots next to the street as well as standing pedestrians and cyclists who are far from the lane are recorded, this leads to a significantly larger number of relevant objects.

If stationary and moving objects are tracked in a common approach, measurement inaccuracies or errors in object extraction often result in virtual movements, so that parked vehicles are incorrectly recognized as moving or rotating. This can lead to false braking or evasive maneuvers on irrelevant static objects or to incorrect behavior on relevant static objects. Due to the large number of objects, the algorithms also have a high computing power, since properties such as shape, position and / or orientation have to be estimated for all objects. These estimates are mostly based on the current measurement and the extracted properties are averaged over time.

In the present case, the focus is particularly on the detection of relevant static objects. For this purpose, it is known from the prior art to determine corresponding occupancy cards on the basis of sensor data from environment sensors. These occupancy maps can also be referred to as grid-based maps or as grid maps. Occupancy cards of this type are used in most sensor systems or driver assistance systems for recording the vehicle surroundings in order to make the large number of sensor data manageable. Static occupancy maps can also be determined over time, which describe the static objects in the area. With each measurement, the static detections can be entered in the occupancy card so that a detailed image of the visible object shape is obtained in the occupancy card after a few measuring cycles. According to the prior art, however, it is not taken into account whether the detected static objects are in the lane, the relevant lane or next to the lane.

The object of the present invention is to provide a solution for how relevant static objects or road users in the vicinity of a vehicle can be recognized more reliably.

According to the invention, this object is achieved by a method, by a computing device, and by a computer program with the features according to the independent claims. Advantageous developments of the present invention are specified in the dependent claims.

A method according to the invention is used to detect relevant static objects in the surroundings of a vehicle. The method includes receiving sensor data from at least one environment sensor of the vehicle, the sensor data describing the environment. Furthermore, the method includes the determination of a static occupancy card on the basis of the sensor data, the static occupancy card describing non-moving parts of the environment. The sensor data are continuously received and the static occupancy card is accumulated on the basis of the sensor data. The method also includes determining lane data which describe at least one lane in the vicinity. Furthermore, the method includes the identification of relevant static objects, which are located at least in regions within the at least one lane, from the static objects on the basis of the accumulated static occupancy map and the lane data.

With the help of the method, relevant static objects in the vicinity of one's own vehicle which are in a lane are to be recognized. In this context, all objects can be defined as relevant that have an influence on the trajectory of the own vehicle, that is to say that, for example, require braking or evasive action. Static obstacles outside the lanes, which can be part of the infrastructure, plants or the like, should not be taken into account in the following. The method can be carried out with a computing device of a driver assistance system of the vehicle. When the vehicle is in operation or while the vehicle is in motion, the sensor data are continuously provided with the environment sensors of the driver assistance system. For example, measurements or measurement cycles that follow one another in time can be carried out with the environment sensors will. The sensor data are transmitted to the computing device in each measurement cycle. The environment sensors can be radar sensors, lidar sensors, ultrasonic sensors and / or cameras. The sensor data describe the surroundings of the vehicle and in particular objects in the vicinity of the vehicle.

In addition, lane data are continuously determined by the computing device. These lane data describe one or more lanes in the vicinity of the vehicle. In particular, these lane data can describe lanes and / or lanes. The lane data can also describe a driving envelope which represents the (future) movement of the vehicle or of another road user. This lane data can be determined, for example, on the basis of a digital map and with the aid of a satellite-supported position determination system. Alternatively or additionally, the lane data can be determined on the basis of the sensor data. In this case, lane markings and / or the lanes themselves can be recognized in the sensor data. This applies in particular to the case when a camera or a lidar sensor is used as the environment sensor.

The static occupancy card is determined on the basis of the sensor data. The occupancy map can also be referred to as a grid-based map or grid map. The static occupancy card can represent the non-moving parts of the environment or environment. In a subsequent step, the static objects, for example a parked vehicle, a box on the street or the like, can be extracted from this. This static occupancy map can have a plurality of cells, each cell describing a specific area in the vicinity of the vehicle. Furthermore, each cell describes the probability that the area of the environment which the cell represents is occupied by a static element. The static occupancy card is therefore an image of the static components of the environment around the own vehicle. A subsequent step allows the extraction of static objects from the static occupancy map. Only measurements of static objects and no measurements of dynamic objects are preferably entered in the static occupancy card. As explained above, the sensor data are received for successive measurement cycles. The static occupancy card is accumulated over time. This means that the previously received sensor data are taken into account when determining the static occupancy card. In particular, the static occupancy card is accumulated over at least one time step on the basis of the sensor data from at least one environment sensor. The assignment takes place in particular directly via the static occupancy card, so that no virtual movements and, associated with this, a false classification as a dynamic object occurs.

Furthermore, it is provided that the relevant static objects, which are at least partially located on one of the lanes, are extracted from the static objects. The relevant static objects are to be extracted from the set of all static objects based on the accumulated static occupancy map. On the one hand, stopped road users are to be recognized in the lane, behind which the own vehicle is to be stopped. On the other hand, relevant static objects which have not been stopped and which are located at least in some areas on one of the lanes should be taken into account in the trajectory planning. For example, these can be bypassed or the vehicle can be braked if there is oncoming traffic. This type of static objects, which are relevant static objects but not stopped road users, can be, for example, lost cargo or other objects on the roadway. Furthermore, such objects can be parked vehicles which protrude laterally into the lane. On the basis of the static occupancy map and the lane data, the relevant static objects, including the stopped road users, can be recognized. In addition, the properties of the static objects, such as the shape, position and / or orientation, can then be estimated on the basis of the static occupancy map. Since the position-related and geometric properties of a static object do not change over time, the estimate based on the accumulated static occupancy map is advantageous. In addition, it can be checked, for example, whether the cells of the static occupancy map which describe a static object can be assigned to the lane. In addition, the spatial dimensions of the recognized static objects or the occupied cells and / or the shape of the static objects and / or the lane being covered by the static objects can be taken into account. The shape can be estimated using known methods such as PCA (Principal Component Analysis), for example. Overall, relevant static objects in the vicinity of one's own vehicle can thus be recognized in a reliable manner.

The accumulated static occupancy map preferably has a plurality of cells, the respective cells describing an occupancy by the non-moving part of the environment, and individual cells are selected and / or connected to identify the relevant static objects. As already explained, the respective cells of the static occupancy map can be assigned a probability that these are occupied by a static object. The static occupancy card can first be binarized. In this case, those occupied cells can be selected for which the occupancy probability is greater than a limit value, for example 0.5 or 50%. The selected cells can then be clustered. In this way, for example, gaps between cells that are close together can be closed. To extract the relevant static objects from the static objects, the individual cells can be selected and / or connected. This means that static objects in the vicinity of the vehicle can be reliably identified with little computing effort.

After the cells of the static occupancy map have been selected and / or clustered, these cells can then be assigned to the lanes. The scanline algorithm, for example, can be used for this purpose. As an alternative or in addition, it can be checked for each of the cells whether this is at least partially within one of the detected lanes. Clusters or static objects outside one of the detected lanes cannot be taken into account or deleted. If a cluster or a static object is assigned to a lane, an identifier or ID of the lane can be assigned to this static object.

Stopped road users are preferably identified from the relevant static objects. The aim of the method is, in particular, to differentiate stopping road users and parked road users or static obstacles. In this context, the term stopping road users is to be understood as meaning all objects that are standing on a roadway and can potentially approach them, such as, for example, a stationary vehicle at a red traffic light or vehicles turning. In contrast to parked road users, who should be avoided if necessary, stopping should also be made behind stopping road users. In the case of the relevant static objects recognized, a distinction should be made, in particular, between the parked road users or stationary obstacles (e.g. cardboard boxes on the street) and the temporarily stopped road users (e.g. at a red traffic light).

In one embodiment, the identification of the respective stopped traffic participants is checked for plausibility on the basis of spatial dimensions of the respective relevant static objects and / or a shape of the respective relevant static objects and / or a coverage of the lane by the respective relevant static objects. The stopped road users are to be extracted from the recognized relevant static objects. For this purpose, the spatial dimensions of the static object can be checked. If the object is a small object or if the spatial dimensions of the static object fall below a predetermined limit value, this object can be viewed as a relevant static object but not as a stopped road user. A small object in the lane can be an object, for example. As an alternative or in addition, the lane being covered by the static object can be taken into account. If the static object only covers a small area or if the coverage falls below a minimum value, this object can be viewed as a relevant static object but not as a stopped road user. This can also take into account which area of the roadway the object covers. If the object covers, for example, a side area of the lane, it can be recognized as a parked vehicle which partially protrudes into the lane and not as a temporarily stopped road user. In this way, stopped road users can be reliably distinguished from parked road users or static obstacles.

In a further embodiment, camera data are received and a classification of the relevant static objects within the lane is carried out on the basis of the camera data. Camera data can therefore be received from a camera by means of the computing device. The camera can be an environment sensor of the driver assistance system. In this case the camera data correspond to a subset of the sensor data. The respective static objects can be recognized and / or classified on the basis of the image data or camera data. For example, the objects can be recognized as a vehicle, motorcycle, cyclist, pedestrian or the like. This classification can be taken into account in the plausibility check described above. If an object is recognized as a motorcycle on the basis of the camera data, this can be identified as a stopped road user, even if the spatial dimensions and / or the lane coverage by the object is small. In this case, the vehicle should be stopped behind the road user, even if overtaking is possible.

It is also advantageous if an alignment and / or direction of movement of the respective stopped road users is determined on the basis of the accumulated static occupancy map. When the stopped road users have been extracted from the relevant static objects, an alignment of the stopped Road users are determined. This can be done on the basis of the static occupancy card, especially after binarization and clustering. A rectangular enveloping body or a so-called bounding box can be assigned to the occupied cells of the occupancy cards. For this purpose, different boxes can be determined for the alignment of the detected lanes. It can also be provided that a box is determined on the basis of the orientation of the own vehicle. These different boxes, which have different orientations, can serve as hypotheses. Furthermore, a score function can be applied to differently oriented hypotheses. One of the boxes or enveloping bodies can thus be assigned to the cells or the cluster. The alignment and / or the future direction of movement of the stopped road user can then be determined on the basis of the selected box. In principle, the enveloping bodies can have any shape. When determining the alignment and / or the future direction of movement, the camera data described above can also be used. Using the weighted distance between the box and the cells, a statement can also be made as to how well the object can be represented by a rectangle. This shape estimate can be a further indicator of whether the road user is stopped.

In a further embodiment, the lane data describe at least two lanes in the vicinity and the respective stopped road users are assigned to one of the at least two lanes on the basis of their alignment and / or direction of movement. If several lanes are recognized in the vicinity, it can then be recognized on the basis of the orientation and / or direction of movement of the stopped road user in which of the lanes the stopped road user is and / or will be moving. This information can be used by the driver assistance system to maneuver the vehicle and / or to carry out driving maneuvers.

Provision can also be made for information relating to the objects to be used by a downstream module which is used for path planning. Only those stopped road users who are in a future path of the vehicle can be taken into account. If, for example, no lane change is provided, only stopped road users in the lane or the lane of the vehicle can be taken into account. It can also be provided that only the objects in the future path of the vehicle are taken into account when the stopped road users are recognized. Furthermore, objects on a different lane cannot be taken into account. In this way, the computational effort can be reduced.

As already explained, the lane data can be determined on the basis of a digital map and / or on the basis of the sensor data of the at least one environment sensor. Provision can also be made for odometry data to be received from a position sensor of the vehicle. This odometry data can describe the current speed and the rate of rotation of the vehicle. The odometry data can be taken into account when determining the static occupancy card. In addition, the odometry data can be taken into account as an additional hypothesis in the previously described determination of the orientation of the stopped road users.

In a further embodiment, depending on the identification of the relevant objects but not stopped traffic participants (i.e. parking or stationary obstacle) within the lane, an overtaking process is decided from the static objects. If a static object is identified as a relevant static object but not a short-term traffic participant, the driver assistance system can be used to carry out a driving maneuver to overtake the static object or to drive past the static object. If overtaking is currently not possible because the evasive trajectory is occupied, for example due to oncoming traffic, a certain distance can be maintained in the direction of travel from the static object. In this way, overtaking is possible when the trajectory is clear again. If a static object is recognized as a stopped road user, the driver assistance system can bring the driver's vehicle to a stop behind the stopped road user. In this case, the distance is significantly smaller in order to keep the distances short, e.g. in front of traffic lights or in a traffic jam. A road user who has stopped briefly should not be overtaken. Overtaking can also be decided as a function of the lane allocation and / or the lane occupancy of the relevant static object or the stopped road user. This applies, for example, to the case when a road user who has stopped turning is still partially protruding into his own lane.

Overall, a distinction can be made between three types of static objects. On the one hand, there are static objects that are not relevant, such as obstacles or road users outside of your own driving path or the road. There should be no reaction to these irrelevant objects. There are also the relevant static objects described above and those that have been stopped Road users, the stopped road users being a subset of the relevant static objects.

A computing device according to the invention for a sensor system of a vehicle is designed or set up to carry out a method according to the invention with advantageous embodiments thereof. The computing device can in particular be formed by an electronic control unit of a vehicle.

A driver assistance system according to the invention for a vehicle comprises a computing device according to the invention. Furthermore, the driver assistance system can have a plurality of environment sensors with which the sensor data can be provided. The driver assistance system can also include a position sensor for providing the odometry data. A digital map can be stored in a memory of the vehicle or of the driver assistance system. In addition, the position of the vehicle can be determined by means of a satellite-supported position determination system. The vehicle can be maneuvered automatically or autonomously by means of the driver assistance system. Depending on whether the static objects in the vicinity are recognized as relevant stopped traffic participants or as relevant parking traffic participants or static obstacles, a driving maneuver can be planned with the driver assistance system.

A vehicle according to the invention comprises a driver assistance system according to the invention or a computing device according to the invention. The vehicle is designed in particular as a passenger car. It can also be provided that the vehicle is designed as a utility vehicle.

A further aspect of the invention relates to a computer program, comprising commands which, when the program is executed by a computing device, cause it to execute a method according to the invention and the advantageous refinements thereof. The invention further relates to a computer-readable (storage) medium, comprising instructions which, when executed by a computing device, cause the computing device to execute a method according to the invention and the advantageous embodiments thereof.

The preferred embodiments presented with reference to the method according to the invention and their advantages apply accordingly to the computing device according to the invention, for the driver assistance system according to the invention, for the vehicle according to the invention, for the computer program according to the invention and for the computer-readable (memory) medium according to the invention.

Further features of the invention emerge from the claims, the figures and the description of the figures. The features and combinations of features mentioned above in the description, as well as the features and combinations of features mentioned below in the description of the figures and / or shown alone in the figures, can be used not only in the specified combination, but also in other combinations or on their own, without the frame to leave the invention.

• 1 eine schematische Darstellung eines Fahrzeugs, welches ein Fahrerassistenzsystem aufweist;
• 2 eine schematische Darstellung einer Verkehrssituation, bei welcher sich eine Mehrzahl von statischen Objekten in einer Umgebung des Fahrzeugs befindet;
• 3 eine schematische Darstellung von Messpunkten eines statischen Objekts, wobei die Messpunkte zu unterschiedlichen Zeitpunkten aufgenommen werden;
• 4 ein schematisches Ablaufdiagramm eines Verfahrens zum Erfassen von statischen Objekten in der Umgebung des Fahrzeugs gemäß einer ersten Ausführungsform;
• 5 eine schematische Darstellung einer statischen Belegungskarte, bei welcher belegte Zellen ausgewählt werden und die ausgewählten Zellen geclustert werden;
• 6 eine schematische Darstellung von mehreren Hypothesen-Boxen, welche den belegten Zellen einer statischen Belegungskarte zugeordnet werden; und
• 7 ein schematisches Ablaufdiagramm eines Verfahrens zum Erfassen von statischen Objekten in der Umgebung des Fahrzeugs gemäß einer zweiten Ausführungsform.

The invention will now be explained in more detail using preferred exemplary embodiments and with reference to the accompanying drawings. Show:

• 1 a schematic representation of a vehicle having a driver assistance system;
• 2 a schematic representation of a traffic situation in which a plurality of static objects are in the vicinity of the vehicle;
• 3 a schematic representation of measuring points of a static object, the measuring points being recorded at different times;
• 4th a schematic flow diagram of a method for detecting static objects in the vicinity of the vehicle according to a first embodiment;
• 5 a schematic representation of a static occupancy map, in which occupied cells are selected and the selected cells are clustered;
• 6th a schematic representation of several hypothesis boxes, which are assigned to the occupied cells of a static occupancy card; and
• 7th a schematic flow diagram of a method for detecting static objects in the vicinity of the vehicle according to a second embodiment.

Identical or functionally identical elements are provided with the same reference symbols in the figures.

1 shows a vehicle in a schematic representation 1 , which is designed here as a passenger car, in a plan view. The vehicle 1 includes a driver assistance system 2 , by means of which the vehicle 1 can be maneuvered automatically or autonomously. The driver assistance system 2 comprises a computing device 3 which, for example, by an electronic control unit of the vehicle 1 can be formed.

It also includes the driver assistance system 2 two environment sensors 4th , which in the present case are designed as radar sensors and / or lidar sensors. The driver assistance system also includes 2 another environment sensor 5 , which is designed as a camera. In principle, the driver assistance system 2 a plurality of environment sensors 4th , 5 exhibit. With the respective environment sensors 4th , 5 sensor data are provided, which an environment 6th of the vehicle 1 describe. This sensor data can from the environment sensors 4th , 5 to the computing device 3 be transmitted. It also includes the driver assistance system 2 a motion sensor 7th , by means of which the current speed and / or rate of rotation of the vehicle 1 can be determined by means of odometry.

2 shows, in a greatly simplified representation, a traffic situation in which a plurality of static objects are located 8th in the neighborhood 6th of the vehicle 1 are located. The vehicle moves 1 along a direction of movement 9 . The vehicle 1 is in a first lane L1 . In the neighborhood 6th of the vehicle 1 is also a second lane L2 available. The respective lanes L1 , L2 can be determined using a digital map. Alternatively or additionally, corresponding lane markings of the lanes can be used L1 , L2 can be recognized on the basis of the sensor data or image data of the camera.

The respective static objects 8th can be recognized on the basis of the sensor data, which with the environment sensors 4th , 5 to be provided. There are also areas here 10 of the objects 8th shown, which with the environment sensors 4th , 5 can be recognized. In the traffic situation shown, static objects 8th irrelevant obstacles 11th outside the lanes L1 , L2 . These obstacles 11th should not be taken into account, as these are stationary and not on the lanes L1 , L2 can move. In the lane L1 of the vehicle 1 there is also a relevant static object in the form of a static obstacle 12th . With this static obstacle 12th For example, it can be a lost load, for example a box. This static obstacle 12th covers only a small area of the lane L1 . In this case, by means of the driver assistance system 2 an evasive maneuver to drive past the static obstacle 12th to be determined. It can also be provided that the vehicle 1 by means of the driver assistance system 2 at a greater distance from the static obstacle 12th is stopped so that the driver can decide or carry out driving maneuvers. In addition, there is a relevant static object 8th a parked vehicle 13th in the neighborhood 6th of the vehicle 1 . From this parked vehicle 13th only one side protrudes into the lane L1 so that a small area of the lane L1 is covered. At this parked vehicle 13th can the vehicle 1 be maneuvered past.

In the lane L1 of the vehicle 1 is located as a static object 8th also a first stopped road user 14a . At the stopped road user 14a it can be a stopped vehicle, for example. This stopped first road user 14a the vehicle approaches 1 from behind. Based on the sensor data of the environment sensors 4th , 5 can the rear area 10 of the stopped first road user 14a are recorded. This is where the vehicle should be 1 behind the stopped first road user 14a be stopped. It is also located as a static object 8th a second stopped road user 14b in the neighborhood 6th of the vehicle 1 . Also this second stopped road user 14b the vehicle approaches 1 from behind. With the help of the environment sensors 4th , 5 can be used as an area 10 the stern and part of the side of the second stopped road user 14b be recognized. The goal is to recognize that this second stopped road user 14b want to turn right and that the vehicle 1 on the second stopped road user 14b can drive past.

When capturing static objects 8th it is intended that the measurements with the environment sensors 4th , 5 are accumulated over time. For this purpose, measurements or measuring cycles that follow one another in time are carried out. This is illustrated below using 3 explained. It is assumed that the vehicle is 1 a static object 8th approaching from behind and to the left of this static object 8th drives past. With the static object 8th For example, it can be the parked vehicle 13th Act. Doing so are on the left of 3 first measuring points 15a shown, which are determined at a first point in time. These first measuring points 15a describe the back of the static object 8th .

To the right of 3 are second measuring points 15b shown, which are determined at a second point in time, which follows the first point in time. The second measuring points 15b describe the back and part of the left side of the static object 8th . In addition to the second measuring points 15b are also the first measuring points 15a shown. This simplified example is to the right of 3 all previously recorded measuring points 15a , 15b of the static object 8th shown. The first measuring points 15a the corresponding second measuring points 15b assigned or the corresponding measuring points 15a , 15b can be merged. Those at different times recorded measuring points 15a , 15b , are offset from one another. This can be due to the fact that the viewing angle has changed between the points in time. When these measuring points 15a , 15b viewed independently of one another, it can be mistakenly assumed that the static object is 8th moves and / or rotates. This is based on the respective box 16 can be seen which is the orientation of the object 8th describes. By accumulating the measurement points 15a , 15b this can be prevented. As explained in more detail below, a static occupancy card M can be used.

4th shows a schematic flow diagram of a method for detecting static objects 8th in the neighborhood 6th of the vehicle 1 according to a first embodiment. This method can be carried out by means of the computing device 3 be performed. A computer program can be used for this purpose using the computing device 3 are executed. In one step S1 data on the lanes are generated on the basis of a digital map and with the help of a satellite-supported positioning system L1 , L2 determined. The data of the lanes L1 , L2 the spatial dimensions or the geometry of the lanes L1 , L2 describe. Alternatively or additionally you can do this in one step S2 the data on the lanes L1 , L2 based on the sensor data of the environment sensors 4th , 5 and in particular the camera. In one step S3 the sensor data from the environment sensors 4th , 5 receive. In one step S4 becomes odometry data from the position sensor 7th receive. Using the sensor data and the odometry data, which shows the current speed and rate of rotation of the vehicle 1 describe, the static occupancy map M can then be determined.

Such a static occupancy card M is in section a) of 5 shown schematically. The static occupancy card M has a plurality of cells 16 , 17th on, with each of the cells 16 , 17th a certain area in the area 6th of the vehicle 1 describes. The occupancy card M includes occupied cells 16 , which is a static object 8th in the neighborhood 6th describe, as well as unoccupied cells 17th which is not a static object 8th in the neighborhood 6th describe. The static occupancy map M shown describes the second stopped road user 14b according to 2 . As previously in connection with 3 has been explained, this static occupancy card M can be accumulated over time. In particular, measurements with the environment sensors that follow one another in time 4th , 5 are carried out and the static occupancy card M can be derived from this. In addition, describe the respective occupied cells 16 the static occupancy map M with what probability the area in the area 6th which one of this cell 16 is associated with a static object 8th is occupied. In the present case, the probabilities are illustrated by different hatching.

In one step S5 the procedure according to 4th a binarization is first carried out on the basis of the static occupancy card M. This binarization is in section b) of 5 shown schematically. The static occupancy map M becomes selected cells 18th certainly. The selected cells 18th are those occupied cells 16 the occupancy card M, which is likely to be a static object 8th are occupied. For this purpose, those occupied cells 16 at which the probability is above a predetermined threshold value, for example 0.5, is selected. It is then based on these selected cells 18th carried out a clustering. This is in section c) of 5 made clear. There are more cells 19th between the selected cells 18th inserted. When clustering, those selected cells 18th determine which are close to each other. For example, the so-called closing can be used for this purpose. Closing is a basic morphological operation in digital image processing. As an alternative or in addition, so-called connected-component labeling can be used. The result of the clustering is in section d) of 5 shown and shows an optimized occupancy map M 'with occupied cells 16 ' and blank cells 17 ' . That way, those cells can 16 ' be determined which to the same static object 8th belong.

In one step S6 the procedure according to 4th the cells are extracted 16 ' the optimized static occupancy card M 'or the selected and clustered cells of the static occupancy card M. For this purpose, the data about the lanes L1 , L2 used to provide information about the relevant lanes L1 , L2 in the neighborhood 6th of the vehicle 1 to obtain. This information can, for example, be polygons, which are defined by a center line in the middle of the respective lane L1 , L2 and two boundary lines are described. In doing so, those occupied cells 16 ' selected which are within the relevant lanes L1 , L2 lie. The scanline algorithm, for example, can be used for this purpose. When the cells 16 ' are assigned to a cluster, the cluster can have an identification number of the lane L1 , L2 be assigned. A cluster can also have multiple lanes L1 , L2 be assigned. For the extraction of the cells 16 ' A brute force method can also be used, in which for each of the occupied cells 16 ' it is checked individually whether this is a lane L1 , L2 can be assigned. If a cell 16 ' within a lane L1 , L2 is arranged, this can be kept and this can be an identification number of the lane L1 , L2 be assigned. If a cell 16 ' outside the lanes L1 , L2 it can be deleted.

If this procedure is appropriate to the traffic situation 2 applied, the static obstacles 11th not be taken into account as this is not a lane L1 , L2 can be assigned. The first road user to be stopped 14a can the first lane L1 be assigned. The second stopped road user 14b , the static obstacle 12th and the parked vehicle 13th can use both the first lane L1 as well as the second lane L2 be assigned.

Subsequently, the remaining static objects 8th which one lane L1 , L2 were assigned to the standing road users 14a , 14b be identified. For this purpose, according to the procedure of 4th in one step S7 a plausibility check carried out. With small static objects 8th or with static objects 8th showing the lane L1 , L2 cover only a little, an evasive maneuver can be carried out. In the present case, the stopped road users should 14a , 14b can be determined behind which the vehicle 1 should come to a stop or which should be braked. This can be done for the respective cluster, which the static objects 8th describe, at least one of the following criteria can be checked: The size or the spatial dimensions of the cluster can be checked. If the spatial dimensions are small or fall below a predetermined limit value, the cluster or the static object can 8th not as a temporarily stopped road user 14a , 14b must be taken into account. Thus, in the traffic situation of 2 illustrated static obstacle 12th not as a temporarily stopped road user 14a , 14b must be taken into account.

The coverage of the lane can be used as a further criterion L1 , L2 through the object 8th can be used. If the static object 8th only a small, predetermined area of the lane L1 , L2 covered, the object can 8th also not as a temporarily stopped road user 14a , 14b must be taken into account. For example, it can be checked whether the overlap is so small that the vehicle 1 on the object 8th can be moved past. This can be done in the traffic situation of 2 depicted parked vehicle 13th which is the lane L1 only slightly covered, not as a briefly stopped road user 14a , 14b must be taken into account. As a further criterion, a classification can be carried out on the basis of the sensor data and in particular the image data of the camera. For example, if a static object 8th is classified as a stationary motorcycle, it should not be overtaken. Overall, the stopped road users can 14a , 14b be identified.

In one step S8 the procedure according to 4th are for the relevant stopped road users 14a , 14b Enveloping body, for example rectangular enveloping body (bounding box), determined. In this way, the orientation and the (future) direction of movement of the stopped road users can be determined 14a , 14b to be determined. For this purpose, for example, a score function can be applied to differently oriented hypotheses. The hypothesis can be the alignment or orientation of the lane L1 , L2 be at the object position as most objects are 8th move parallel to a lane. The hypothesis can also include the orientation of the vehicle 1 or the vehicle's camera 1 be in case the cells 16 ' associated with a camera object. Based on these hypotheses, hypothesis boxes can then be created B1 , B2 , B3 will be generated that all cells 16 ' of the object 8th contain.

This is exemplified in 6th illustrated. Here are the cells 16 ' the optimized static occupancy map M 'shown. There is also a hypothesis box B1 for the first lane L1 , a hypothesis box B2 for the second lane L2 and a hypothesis box B3 for the direction of movement 9 of the vehicle 1 shown. The score function optimizes that the static occupancy is as close as possible to the limit of the hypothesis box B1 , B2 , B3 lies. One solution is the center of the cells 16 ' into the respective coordinate system of the hypothesis box B1 , B2 , B3 to transform. Then the minimum distance to the visible edge is the minimum of the x and y values. The score of the score function is the sum of these values. In the present case, the cluster is given the hypothesis box B2 assigned. From this it can then be deduced that the second stopped road user 14b into the second lane L2 want to turn.

A possible extension can be that not only the exact hypotheses are selected. For example, the hypotheses can be expanded in both directions by several hypotheses with less deviation. For example, if the orientation is 20 °, hypotheses can be evaluated for the angles 10 °, 15 °, 18 °, 20 °, 22 °, 25 °, and 30 °. Another strategy could be to directly adapt a hypothesis box B1 , B2 , B3 for all cells 16 ' apply. It can also be provided that an occupancy value is determined in the original occupancy card M in order to store the cells 16 ' weighted based on the probability of occupancy. In one step S9 the procedure according to 4th a plausibility check is then carried out for the object 8th be performed.

7th shows a schematic flow diagram of a method for detecting static objects 8th in the neighborhood 6th of the vehicle 1 according to a second embodiment. Compared to the method according to 4th this procedure also includes one step S10 , in which objects based on the sensor data or image data of the camera 8th be recognized. In another additional step S11 the camera objects can be the detections of the environment sensors 4th , which are designed as radar sensors and / or lidar sensors, assigned. Each detection is given an object ID of the camera object as a designation. This object ID can then be stored in addition to the static occupancy card M or the optimized static occupancy card M '. On the basis of this information, it can then be determined in a subsequent step which cells 16 ' to the same object 8th belong. Furthermore, there is the option of defining properties of the camera object, such as the class of the object 8th or the orientation of the object 8th , To determine.

With this information, the following improvements can be made: Clustering can be performed based on the object ID. Furthermore, a cluster can be divided if it is assigned to two different camera objects. In addition, clusters that have no assignment to a camera object can be ignored. Very small clusters can also be taken into account if they are classified, for example, as cyclists, motorcycles, pedestrians or the like. Furthermore, the alignment can be used as a hypothesis for adapting the hypothesis box B1 , B2 , B3 be used. In addition, the object class and other properties, such as an activated hazard warning system or the like, can be extracted.

Claims (11)

Method for detecting relevant static objects (12, 13, 14a, 14b) in an environment (6) of a vehicle (1) with the following steps: - Receiving sensor data from at least one environment sensor (4, 5) of the vehicle (1), the sensor data describing the environment (6), - Determination of a static occupancy card (M) on the basis of the sensor data, the static occupancy card (M) describing non-moving parts of the environment (6), - The sensor data are continuously received and the static occupancy card (M) is accumulated on the basis of the sensor data, - Determination of lane data which describe at least one lane (L1, L2) in the environment (6), and - Identifying the relevant static objects (12, 13, 14a, 14b), which are at least partially within the at least one lane (L1, L2), from the static objects (8) using the accumulated static occupancy map (M) and the lane data . Procedure according to Claim 1 , characterized in that the accumulated static occupancy map (M) has a plurality of cells (16), the respective cells (16) describing an occupancy by the non-moving part of the environment (6), and individual cells (16) for identification of the relevant static objects (14a, 14b) are selected and / or connected. Procedure according to Claim 1 or 2 , characterized in that stopped road users (14a, 14b) are identified from the relevant static objects (12, 13, 14a, 14b). Procedure according to Claim 3 , characterized in that the identification of the respective stopped road users (14a, 14b) on the basis of spatial dimensions of the respective relevant static objects (12, 13, 14a, 14b), a shape of the respective relevant static objects (12, 13, 14a, 14b ) and / or if the lane (L1, L2) is covered by the respective relevant static objects (12, 13, 14a, 14b). Method according to one of the preceding claims, characterized in that camera data are received and a classification of the relevant static objects (12, 13, 14a, 14b) within the lane (L1, L2) is carried out on the basis of the camera data. Method according to one of the Claims 3 until 5 , characterized in that an orientation and / or direction of movement of the respective identified stopped road users (14a, 14b) is determined on the basis of the accumulated static occupancy map (M). Procedure according to Claim 6 , characterized in that the lane data describe at least two lanes (L1, L2) in the vicinity (6) and the respective identified stopped road users (14a, 14b) based on their alignment and / or direction of movement of one of the at least two lanes (L1, L2) be assigned. Method according to one of the preceding claims, characterized in that the lane data are determined on the basis of a digital map and / or on the basis of the sensor data of the at least one environment sensor (4, 5). Method according to one of the preceding claims, characterized in that overtaking is decided as a function of the identification of the relevant static objects (12, 13, 14a, 14b) from the static objects (8). Computing device (3) for a driver assistance system (2) of a vehicle (1), the computing device (3) being set up to carry out a method according to one of the preceding claims. Computer program, comprising instructions which, when the program is executed by a computing device (3), cause them, a method according to one of the Claims 1 until 9 to execute.

Images