Method and device for operating a rail vehicle

Abstract

The invention relates in particular to a method for operating a rail vehicle. In the method, according to the invention, an image of the surroundings of the rail vehicle is recorded by means of a camera system to form real image data, which is converted into virtual aerial view data, and if the real image data is converted into the virtual aerial view data, the virtual aerial view data is converted into the virtual aerial view data. The position taken for conversion of the rail vehicle corresponds to the real position and the direction taken for conversion of the rail vehicle corresponds to the real direction, the virtual bird's-eye view data corresponds to the reference bird's-eye view data, the virtual bird's-eye view data and the reference bird's-eye view data are compared, and the direction taken for conversion of the rail vehicle is corrected in the scope of a correction method. And calculating the corrected virtual aerial view data until a termination criterion is met, detecting the direction taken by the rail vehicle when the termination criterion is met as the real direction of the rail vehicle, and generating a corresponding direction description.

Description

TECHNICAL FIELD

[0001] The invention relates to a method for operating a rail vehicle, a direction detection device and a vehicle control device for a rail vehicle, and a rail vehicle itself. BACKGROUND

[0002] When a rail vehicle is in motion, due to the design of the track and the wheels, the direction of the rail vehicle fluctuates to some extent, so in other words the orientation of the longitudinal direction of the vehicle can deviate at least slightly from the respective direction of travel. This deviation can lead to problems in monitoring the driving operation and the vehicle control. SUMMARY

[0003] The technical problem addressed by the invention is to provide a method with which the respective direction of a rail vehicle can be determined easily.

[0004] According to the invention, this object is achieved by a method for operating a rail vehicle.

[0005] According to this, it is provided according to the invention that an image of the surroundings of the rail vehicle is captured using a camera system to form real image data, the real image data are converted into virtual bird's-eye view data, if the position of the rail vehicle taken for the conversion into virtual bird's-eye view data corresponds to the real position and the direction of the rail vehicle taken for the conversion corresponds to the real direction when the real image data are converted into virtual bird's-eye view data, then the virtual bird's-eye view data correspond to reference bird's-eye view data, the virtual bird's-eye view data and the reference bird's-eye view data are compared, and the direction of the rail vehicle taken for the conversion is corrected within the scope of a correction method, and the corrected virtual bird's-eye view data are calculated until a termination criterion is met, and the direction taken by the rail vehicle when the termination criterion is met is detected as the real direction of the rail vehicle, and a respective direction specification is generated.

[0006] The main advantage of the method according to the invention is that the method can be carried out in a very simple manner and with little additional component effort, since only the camera system for capturing the real image data needs to be provided in addition in terms of hardware, provided that such a camera system does not already exist for other operating reasons. The method steps for determining the direction specification can be carried out using hardware in the form of a computing device, which is already present as standard in modern rail vehicles, for example in their vehicle control devices, and can be upgraded by a software retrofit to carry out the method according to the invention. The same applies if the method is to be carried out on the track side using track-side hardware.

[0007] The virtual bird's-eye view data preferably represent the surroundings of a section of a railway track system, which can be travelled by the rail vehicle respectively, in a data manner. The virtual bird's-eye view data preferably depict the environment around the rail vehicle in an angle of 360 degrees.

[0008] The reference bird's-eye view data preferably represent the entire railway track system traversable by a rail vehicle in the form of data, ie all sections of the railway track system traversable by a rail vehicle, and therefore always represent the section traveled by the rail vehicle at any given time.

[0009] The reference bird's eye view data and the fictitious bird's eye view data are preferably each associated with a coordinate system and are preferably referenced to the same coordinate system, eg each referenced to the same global coordinate system.

[0010] With regard to obstacle detection during travel of a rail vehicle, it is considered advantageous if a component suitable for obstacle detection is present and performs a check to determine whether an obstacle is present in the preceding route section in the direction of travel, the direction information being used in the check.

[0011] It is considered advantageous with regard to low hardware expenditure if a component suitable for obstacle recognition forms a component of the camera system and generates at least a subset of the real image data, and the real image data of the component are checked to determine whether an obstacle is present in the road section ahead in the direction of travel, wherein the direction information is used in the checking.

[0012] In terms of obstacle detection independent of trackside devices, it is considered advantageous if the camera system is a vehicle-specific camera system and acquires real image data from the vehicle. Alternatively or additionally, real image data acquired by a trackside camera system can also be used.

[0013] The reference bird's-eye view data may be based on a bird's-eye view showing a railway track system on which the rail vehicle can travel; alternatively or additionally, the reference bird's-eye view data may be based on images taken by a surveying vehicle at its known position and orientation during a reference journey on the railway track system on which the rail vehicle can travel.

[0014] It is also considered advantageous if, within the scope of the correction method, the position assumed by the rail vehicle for the switch and the direction assumed by the rail vehicle for the switch are corrected and a position information indicating the position of the rail vehicle is additionally determined.

[0015] The additional position description is preferably determined in the following manner: within the scope of the correction method, the position assumed by the rail vehicle for converting the real image data into virtual bird's-eye view image data is corrected, the corrected virtual bird's-eye view image data are calculated until a termination criterion is met, and the position assumed by the rail vehicle when the termination criterion is met is detected as the real position of the rail vehicle and a corresponding position description is generated.

[0016] The invention also relates to a direction detection device for a rail vehicle.

[0017] According to the invention, in respect of such a direction detection device it is provided that the direction detection device comprises a conversion device designed to convert real image data of a camera system, which display an image of the surroundings of the rail vehicle, into virtual bird's eye view data, and a comparison and correction device designed to compare the virtual bird's eye view data with reference bird's eye view data and to correct the direction assumed by the rail vehicle for the conversion of the real image data within the scope of a correction method and to calculate corrected virtual bird's eye view data until a termination criterion is met and to detect the direction assumed by the rail vehicle when the termination criterion is met as the actual direction of the rail vehicle and to generate a corresponding direction specification.

[0018] With regard to the advantages of the direction detection device according to the invention and advantageous embodiments thereof, reference is made to the statements above in connection with the method according to the invention and advantageous embodiments thereof.

[0019] In a preferred embodiment of the direction detection device it is provided that the direction detection device comprises or integrates a computing device having a computing device which can comprise one or more computing units and a memory, and a computer program product is stored in the memory, which computer program product, when executed by the computing device, forms the conversion device and the comparison and correction device.

[0020] In a particularly preferred embodiment, the conversion device and the comparison and correction device form components of the rail vehicle. Alternatively or additionally, for example for redundancy, the conversion device and the comparison and correction device can be formed by software working in the cloud.

[0021] The invention also relates to a rail vehicle. With regard to the rail vehicle, according to the invention it is provided that it has a direction detection device as described above.

[0022] With regard to the advantages of the rail vehicle according to the invention and advantageous embodiments thereof, reference is made to the statements above in connection with the method according to the invention and advantageous embodiments thereof.

[0023] Advantageously, the rail vehicle has a computing device which can comprise one or more computing units and a memory, and a computer program product is stored in the memory, which computer program product, when executed by the computing device, forms the conversion device and the comparison and correction device of the direction detection device.

[0024] The camera system preferably forms a component of the rail vehicle, the direction of which is to be determined.

[0025] The rail vehicle preferably comprises an obstacle recognition device, which takes into account the direction specification and the real image of the camera system during obstacle recognition.

[0026] The application also relates to a computer program product. According to the application, the computer program product comprises program instructions which, when executed by a computing device, cause the computing device to form the conversion device of the direction detection device and the comparison and modification device and / or to execute the method as described above.

[0027] The application also relates to a vehicle control device for a rail vehicle. According to the application, the vehicle control unit has a computing device, which can comprise one or more computing units, and a memory, in which a computer program product is stored, which, when executed by the computing device, causes the computing device to form the conversion device of the direction detection device and the comparison and modification device and / or to execute the method as described above. BRIEF DESCRIPTION OF DRAWINGS

[0028] The application is explained in more detail below using exemplary embodiments; in the drawings:

[0029] Figure 1 a schematic side view showing components of an exemplary embodiment of a rail vehicle according to the application, in which an embodiment of the method according to the application is explained in terms of a rail vehicle and an embodiment of the direction detection device according to the application,

[0030] Figure 2 a schematic top view showing a rail vehicle according to the application, Figure 1

[0031] Figure 3 a preferred embodiment of a direction detection device is shown in more detail, which can be used in a rail vehicle according to the application and can execute the method according to the application, Figure 1 and Figure 2 a preferred embodiment of a direction detection device is shown in more detail, which can be used in a rail vehicle according to the application and can execute the method according to the application,

[0032] Figure 4 a further preferred exemplary embodiment of a direction detection device is shown in more detail, which can be used in a rail vehicle according to the application and can execute the method according to the application, and Figure 1 and Figure 2 a further preferred exemplary embodiment of a direction detection device is shown in more detail, which can be used in a rail vehicle according to the application and can execute the method according to the application, and

[0033] Figure 5 an embodiment of a vehicle control device according to the application is shown, which can advantageously be used in a rail vehicle according to the application and comprises a direction detection device which can execute the method according to the application. Figure 1 and Figure 2 an embodiment of a vehicle control device according to the application is shown, which can advantageously be used in a rail vehicle according to the application and comprises a direction detection device which can execute the method according to the application.

[0034] For the sake of clarity, in the drawings, the same reference signs are always used for the same or similar parts. DETAILED DESCRIPTION

[0035] Figure 1 ​Components of an exemplary embodiment of a rail vehicle 10 according to the present invention are shown in a schematic side view, which travels on a railway track system 20 in a travel direction F predetermined by the rails of the railway track system in the direction of an obstacle 30. The rail vehicle 10 comprises, in particular, a camera system 100, a direction detection device 120, and an obstacle detection device 130.

[0036] The camera system 100 includes one or more cameras 110 , such as a plurality of cameras 110 , which detect the surroundings of the rail vehicle 10 , preferably within a 360-degree range around the rail vehicle 10 .

[0037] The obstacle recognition device 130 is connected to a component suitable for obstacle recognition, such as the camera 110 of the camera system 100 oriented forward in the direction of travel or in a Figure 1 Another component not shown in the figure is directed forward, such as a radar or laser system.

[0038] The direction detection device 120 is designed to convert real image data RB from the camera 110 (which shows the surroundings of the railway track system 20 on which the rail vehicle 10 is traveling) into virtual bird's-eye view image data FVPD, i.e., bird's-eye view image data representing the surroundings of the rail vehicle 10. This conversion is preferably performed taking into account the position and orientation of the rail vehicle. As a starting value for performing the conversion or as a first cycle of the conversion within the scope of the optimization method, the direction detection device 120 can use a position information Xs of the position of the rail vehicle 10, which is provided by a positioning device 140 (e.g., a GPS system or an odometer). As a starting value for the orientation of the rail vehicle 10, the direction detection device 120 can set an error angle relative to the track orientation to zero.

[0039] The direction detection device 120 also compares the virtual bird's-eye view data FVPD with reference bird's-eye view data RVPD (hereinafter also referred to as reference bird's-eye view data). The reference bird's-eye view data RVPD is bird's-eye view image data of the surroundings of the railway track system 20, which is also displayed in a bird's-eye view and is traversable by the rail vehicle 10. The reference bird's-eye view data RVPD is related to a fixed reference coordinate system (e.g., the Earth's geographic coordinate system or global coordinate system) and can therefore also be referred to as "real" bird's-eye view image data of the surroundings of the railway track system 20.

[0040] The virtual bird's eye view data FVPD calculated by the direction detection device 120 taking into account the real image data RB preferably relate to the same coordinate system as the reference bird's eye view data RVPD in order to be able to compare the virtual bird's eye view data FVPD with the reference bird's eye view data RVPD as easily as possible.

[0041] The "real" bird's eye view data is considered to be correct and can thus be used as a reference for evaluating the fictitious bird's eye view data FVPD. The "real" bird's eye view image data can be based on, for example, a bird's eye view showing the railway track system 20 drivable by the rail vehicle 10, measurement data of the railway track system 20 or construction planning data of the railway track system 20. The reference bird's eye view data RVPD can also be based on measurements recorded with a measurement vehicle during a reference trip on the railway track system 20 drivable by the rail vehicle 10.

[0042] The orientation detection device 120 also performs a correction and optimization process using the fictitious bird's eye view data FVPD and the reference bird's eye view data RVPD. Here, the orientation detection device 120 assumes that the fictitious bird's eye view data FVPD corresponds to the reference bird's eye view data RVPD if the position taken by the rail vehicle 10 for the conversion of the real image data RB into the fictitious bird's eye view data FVPD corresponds to the real position and the orientation taken by the rail vehicle 10 for the conversion corresponds to the real orientation; a deviation between the fictitious bird's eye view data FVPD and the reference bird's eye view data RVPD occurs if the position taken for the conversion deviates from the real position of the rail vehicle 10 and / or the orientation taken by the rail vehicle 10 for the conversion deviates from the real orientation.

[0043] Within the scope of the correction and optimization method, the orientation detection device 120 can consistently correct or frequently correct the parameters used for the conversion of the real image data RB into the fictitious bird's eye view data FVPD, i.e. the orientation taken by the rail vehicle 10 and the position taken, until the deviation of the fictitious bird's eye view data FVPD from the reference bird's eye view data RVPD in the area of the position taken by the rail vehicle 10 for the conversion is minimized and / or is below a predetermined deviation threshold.

[0044] Once the minimum value and / or the predetermined deviation threshold is reached, the orientation detection device 120 can consider that it has taken into account the correct position and the correct orientation of the rail vehicle 10 for the conversion of the real image data RB into the fictitious bird's eye view data FVPD and can generate an orientation specification OAt indicating the actual orientation of the rail vehicle 10, which can be used by the obstacle detection device 130.

[0045] The orientation specification OAt output by the orientation detection device 120 can be absolute angle information based on a coordinate system of the reference bird's eye view data RVPD, i.e. for example a global coordinate system, or alternatively, an error angle quantifying the orientation relative to the track of the railway track system 20 at the position of the rail vehicle 10.

[0046] Based on the orientation specification OAt and preferably also taking into account the real image data RB of the camera system 100, the obstacle recognition device 130 determines whether there is a collision risk when the rail vehicle 10 continues to travel. If a collision risk is determined, the obstacle recognition device 130 preferably generates a warning signal WS.

[0047] Figure 2 The rail vehicle 10 according to Figure 1 is shown in a simplified top view. It can be seen that due to the design of the tracks of the railway track system 20 and the wheels of the rail vehicle 10, the orientation of the rail vehicle 10 - compared to the respective longitudinal direction of the track and thus the direction of travel F - can fluctuate during travel and the orientation of the rail vehicle 10 can differ slightly from the longitudinal direction of the track; in other words, an error angle can occur between the longitudinal direction of the track and the longitudinal direction of the rail vehicle 10, which is marked in Figure 2 by the reference sign φ. The orientation specification OAt output by the orientation detection device 120 can for example quantify this error angle φ.

[0048] In Figure 2 , an obstacle 30 can also be seen, which is also visible in the real image data RB of the camera system 100 and which is located at a relatively large distance A from the rail vehicle 10. Due to the large distance A, the above-mentioned error angle φ between the orientation of the rail vehicle 10 and the orientation of the track can lead to a misinterpretation of the real image data RB of the camera system 100 if this error angle φ is not taken into account in the obstacle detection. Figure 2 The obstacle 30 is shown, in order to visualize the problem, in a position next to the track where there is no collision risk (dotted line) and in the ballast where there is a collision risk (solid line).

[0049] By taking into account the orientation specification OAt of the orientation detection device 120 when evaluating the real image data RB of the camera system 100, the obstacle recognition device 130 can reduce the risk of misinterpretation due to a possible error angle between the rail vehicle 10 and the direction of travel F set by the track and reduce the risk of an obstacle 30 located in the ballast (see the obstacle 30 marked in solid line in Figure 2 being incorrectly classified as being located outside the ballast (see the obstacle 30 marked in dotted line in Figure 2 ).

[0050] Figure 3 The rail vehicle 10 according to Figure 1Fig. 2 shows an embodiment of the direction detection device 120. The direction detection device 120 comprises a conversion device 121 which determines virtual perspective data FVPD from real image data RB of the camera system 100, i.e. perspective data which describe the surroundings of the rail vehicle 10. The perspective image data FVPD can be determined using known image processing software, e.g. using artificial intelligence which has been trained to e.g. recognize tracks and track lines in the image data RB. The determination of perspective image data from real image data is known e.g. in the context of assistance systems for passenger cars, so that the software of such systems can also be used in the conversion device 121.

[0051] In addition, the direction detection device 120 comprises a readout device 122 which reads the stored reference perspective data RVPD from the memory 50. The readout device 122 preferably limits the reading of the reference perspective data RVPD to the position area which comprises the position at which the conversion device 121 uses the real image data RB for the conversion into virtual perspective data FVPD. In other words, the readout device 122 preferably extracts only the position-dependent data area from the entire reference perspective data set DS stored in the memory 50.

[0052] The reference perspective data set DS can be based e.g. on a perspective showing the railway track system 20 drivable by the rail vehicle 10, on measurement data of the railway track system 20 or on construction planning data of the railway track system 20. The reference perspective data set DS can also be based on measurements made with a measurement vehicle during a reference trip.

[0053] The reference perspective data RVPD can be two-dimensional only, i.e. without taking topology into account, or can be three-dimensional and comprise height information describing the system topology of the railway track system 20.

[0054] If the position at which the rail vehicle 10 takes the conversion of the real image data RB into virtual perspective data FVPD corresponds to the actual position of the rail vehicle 10 and the direction at which the rail vehicle 10 takes the conversion corresponds to the actual direction of the rail vehicle 10, ideally or theoretically a perfect match between the virtual perspective data FVPD and the reference perspective data RVPD can be determined. The direction detection device 120 makes use of this fact for determining the actual position and the actual direction of the rail vehicle 10.

[0055] The direction detection device 120 further comprises a comparison and correction device 123 which compares the virtual perspective data FVPD and the reference perspective data RVPD, respectively.

[0056] If the virtual perspective data FVPD and the reference perspective data RVPD match, it can be concluded that the conversion of the real image data RB into the virtual perspective data FVPD is based on the actual position and the actual orientation of the rail vehicle 10: In this case, the comparison and correction device 123 can output on the output side an orientation specification OAt which indicates that the orientation used for said conversion is now considered to be actually correct, and a position specification Xt which indicates that the position used for said conversion is considered to be the actually correct position.

[0057] If the virtual perspective data FVPD and the reference perspective data RVPD do not match or at least not well enough, the comparison and correction device 123 can within the scope of a correction method slightly correct the orientation taken by the rail vehicle 10 and preferably also the position taken by the rail vehicle 10 to form a corrected orientation specification OAm and a corrected position specification Xm, and cause the conversion device 121 to calculate a corrected virtual perspective data FVPD based on the corrected orientation specification OAm and the corrected position specification Xm. Preferably, the readout device 122 provides the reference perspective data RVPD based on the corrected position specification Xm.

[0058] The comparison and correction device 123 can then again perform a comparison between the virtual perspective data FVPD and the reference perspective data RVPD. The direction detection device 120 can frequently implement the cycle of correcting the orientation and the position, recalculating the fictitious perspective data FVPD, and comparing the fictitious perspective data FVPD with the reference perspective data RVPD until a termination criterion is met.

[0059] The comparison and correction device 123 can detect the orientation taken by the rail vehicle 10 when the termination criterion is met as the actual orientation of the rail vehicle 10 and generate a corresponding orientation specification OAt. The same applies to the corresponding position of the rail vehicle 10: The comparison and correction device 123 can detect the specification Xm of the position taken by the rail vehicle 10 when the termination criterion is met as the actual position of the rail vehicle 10 and generate a corresponding position specification Xt which indicates the actual position of the rail vehicle.

[0060] As a starting value for the first cycle of the method or optimization method, the direction detection device 120 can use a position specification Xs about the position of the rail vehicle 10 which is provided by a positioning device 140, for example a GPS system or an odometer. The direction detection device 120 can set the error angle φ with respect to the orientation of the track to zero as a starting value OAs for the orientation of the rail vehicle 10. The corresponding orientation of the track can be read, for example, from a route atlas present in the rail vehicle, which can be stored, for example, in the memory 50 and which is related to the already mentioned global coordinate system.

[0061] Figure 4Shown according to Figure 3 The preferred embodiment of the direction detection device 120 is shown. The direction detection device 120 comprises a computing device 11a and a memory 11b. The computing device 11a may comprise one or more computing units. The computer program product CPP is stored in the memory 11b. When executed by the computing device 11a, it forms the conversion device 121, the reading device 122 and the comparison and correction device 123 of the direction detection device 120, and preferably also forms the obstacle recognition device 130, as described above in conjunction with Figures 1 to 3 Explained exemplarily.

[0062] For this purpose, the computer program product CPP comprises a position detection software module SW120 having a conversion software module SW121 , a readout software module SW122 and a comparison and correction device software module SW123 , as well as an obstacle detection software module SW130 .

[0063] In accordance with Figure 4 In the exemplary embodiment of the present invention, a reference bird's-eye view data set DS with all reference bird's-eye view data RVPD and preferably also the aforementioned road map set is stored in the storage section SA of the memory 11 b. The readout device 122 extracts only the location-related data areas from the entire reference bird's-eye view data set DS using the corresponding position specifications Xm or location specifications Xs, which are also used by the conversion software module SW121 to generate the virtual bird's-eye view data FVPD.

[0064] Figure 5 Shown according to Figure 1 A particularly preferred embodiment of a rail vehicle 10 is shown. The rail vehicle 10 is equipped with a vehicle control device 11 having a computing device 11a and a memory 11b. The computing device 11a may include one or more computing units.

[0065] The vehicle control software FSW is stored in the memory 11 b and controls the rail vehicle 10 when executed by the computing device 11 a , as is known from conventional rail vehicles.

[0066] The reference bird's-eye view data set DS (see storage section SA) and the computer program product CPP are also stored in the memory 11b. When executed by the computing device 11a, the computer program product CPP forms the conversion means 121, the reading means 122, the comparison and correction means 123 of the direction detection means 120 and the obstacle recognition means 130, as described above in combination with Figures 1 to 3The above-mentioned exemplary embodiments are explained by way of example. For this purpose, the computer program product CPP comprises an orientation detection software module SW 120 having a conversion software module SW 121 for providing the function of the conversion device 121, a readout software module SW 122 for providing the function of the readout device 122 and a comparison and correction device software module SW 123 for providing the function of the comparison and correction device 123 as well as a barrier detection software module SW 130 for providing the function of the barrier recognition device 130.

[0067] In the exemplary embodiment according to Figure 5 It is also preferably provided in the exemplary embodiment according to

[0068] Finally, it should be mentioned that the features of all the above-mentioned exemplary embodiments can be combined with one another in any way in order to form further other exemplary embodiments of the application.

[0069] All the features of the dependent claims can also be combined individually with each independent claim, either individually or in any combination with one or more other dependent claims, in order to obtain further different exemplary embodiments.

[0070] List of reference signs

[0071] 10 rail vehicle

[0072] 11 vehicle control device

[0073] 11a computing device

[0074] 11b memory

[0075] 20 railway track system

[0076] 30 barrier

[0077] 50 memory

[0078] 100 camera system

[0079] 110 camera

[0080] 120 orientation detection device

[0081] 121 conversion device

[0082] 122 readout device

[0083] 123 comparison and correction device

[0084] 130 obstacle detection device

[0085] 140 positioning device

[0086] A distance

[0087] CPP computer program product

[0088] DS reference aerial view data set

[0089] F direction of travel

[0090] FSW vehicle control software

[0091] FVPD fictitious aerial view data

[0092] OAm corrected directional indication

[0093] OAs initial value

[0094] OAt directional indication

[0095] RB real image data

[0096] RVPD reference aerial view data

[0097] SA storage section

[0098] SW 120 orientation detection software module

[0099] SW 121 conversion software module

[0100] SW 122 readout software module

[0101] SW 123 comparison and correction device software module

[0102] SW 130 obstacle detection software module

[0103] WS warning signal

[0104] Xm corrected position indication

[0105] Xs location indication

[0106] Xt position indication

[0107] φ error angle

Claims

1. A method for operating a rail vehicle (10), It is characterized by: - recording images of the surroundings of the rail vehicle (10) using the camera system (100) to form real image data (RB), - converting the real image data (RB) into the virtual bird's eye view data (FVPD), the virtual bird's eye view data corresponding to the reference bird's eye view data (RVPD) if, when converting the real image data (RB) into the virtual bird's eye view data (FVPD), the position of the rail vehicle (10) assumed for the conversion corresponds to the real position and the orientation of the rail vehicle (10) assumed for the conversion corresponds to the real orientation, - comparing the virtual bird's eye view data (FVPD) with the reference bird's eye view data (RVPD) and correcting the orientation of the rail vehicle (10) for the transition within the scope of a correction method and calculating the corrected virtual bird's eye view data (FVPD) until a termination criterion is met, and - detecting the direction taken by the rail vehicle (10) when the termination criterion is met as the actual direction of the rail vehicle (10) and generating a corresponding direction indication (OAt).

2. The method according to claim 1, It is characterized by: - a component suitable for obstacle detection is present and is used to check whether an obstacle (30) is present in the road section ahead in the direction of travel, - wherein said orientation specification (OAt) is used in said examination.

3. The method according to claim 2, It is characterized by: - components suitable for obstacle recognition constitute a component of the camera system (100) and generate at least a subset of the real image data (RB), and - checking the real image data (RB) of the component to determine whether there is an obstacle (30) in the road section ahead in the direction of travel, - wherein said orientation specification (OAt) is used in said examination.

4. The method according to any one of the preceding claims, It is characterized by: - the camera system (100) is the vehicle's own camera system (100), and -Collect real image data (RB) from the vehicle side.

5. The method according to any one of the preceding claims, It is characterized by: The reference bird's-eye view data (RVPD) is based on a bird's-eye view showing a railway track system (20) traversable by a rail vehicle (10).

6. The method according to any one of the preceding claims, It is characterized by: The reference bird's eye view data (RVPD) are based on images taken by a measuring vehicle at its known position and orientation during a reference journey on a railway track system (20) traversable by a rail vehicle (10).

7. The method according to any one of the preceding claims, It is characterized by: - within the scope of the correction method, the position of the rail vehicle (10) assumed for the transition and the orientation of the rail vehicle (10) assumed for the transition are corrected, and - In addition, a position information (Xt) is determined which indicates the position of the rail vehicle (10).

8. The method according to claim 7, It is characterized by: A position information (Xt) indicating the position of a rail vehicle (10) is determined by: - within the scope of the correction method, the position of the rail vehicle (10) assumed for the conversion of the real image data (RB) into virtual bird's-eye view data (FVPD) is corrected and the corrected virtual bird's-eye view data (FVPD) are calculated until a termination criterion is met, and - detecting the position assumed by the rail vehicle (10) when the abort criterion is met as the actual position of the rail vehicle (10) and generating a corresponding position information (Xt).

9. A direction detection device (120) for a rail vehicle (10), It is characterized by: The direction detection device (120) comprises: a conversion device (121) designed to convert real image data (RB) of the camera system (100) into fictitious bird's-eye view image data (FVPD), the real image data showing an image of the surroundings of the rail vehicle (10), and - A comparison and correction device (123) designed to compare the virtual bird's-eye view data (FVPD) with the reference bird's-eye view data (RVPD) and to correct the orientation of the rail vehicle (10) for the transformation within the scope of a correction method, and to calculate the corrected virtual bird's-eye view data (FVPD) until a termination criterion is met, and to detect the orientation of the rail vehicle (10) when the termination criterion is met as the actual orientation of the rail vehicle (10) and to generate a corresponding orientation specification (OAt).

10. The direction detection device (120) according to claim 9, It is characterized by: - the direction detection device (120) comprises a computing device (11a) and a memory (11b), said computing device can comprise one or more computing units, and A computer program product (CPP) is stored in the memory (11b), which, when executed by the computing device (11a), forms the conversion device (121) and the comparison and correction device (123).

11. The direction detection device (120) according to any one of the preceding claims 9 to 10, It is characterized by: The camera system (100) is a component of a rail vehicle (10), the direction of which is to be determined.

12. The direction detection device (120) according to any one of the preceding claims 9 to 11, It is characterized by: The conversion device (121) and the comparison and correction device (123) are components of the rail vehicle (10).

13. The direction detection device (120) according to any one of the preceding claims 9 to 11, It is characterized by: The conversion device (121) and the comparison and correction device (123) are formed by software operating in the cloud.

14. A computer program product (CPP) for a direction detection device (120), It is characterized by: The computer program product (CPP) comprises program instructions which, when executed by a computing device (11a), cause the computing device to form a conversion device (121) and a comparison and correction device (123) of a direction detection device (120) configured according to any one of the preceding claims 9 to 13 and / or to execute a method according to any one of the preceding claims 1 to 8.

Images