Method for monitoring a track section along a platform

Abstract

The invention comprises a method for monitoring a track section (STA) located at a platform (BS), in which the track section (STA) is captured by several imaging sensors. The sensors are aligned with the track section (STA) such that overlapping images (B1 ... B7) can be generated. As a result of the overlap, every location of the track section (STA) is present in at least two images. The generated images are analyzed by computer for obstacles in the track section (STA). A signal indicating the presence of an obstacle is generated if an obstacle is detected by the analysis. A signal indicating a fault is generated if the obstacle is detected only once by the analysis. Furthermore, the invention comprises a railway system with a platform (BS), a computer program, and a computer-readable storage medium.

Description

Technical field

[0001] The invention comprises a method for monitoring a track section located at a railway platform. Furthermore, the invention comprises a railway system with a platform and a track section located thereon. The invention also comprises a computer program product containing program instructions. Finally, the invention comprises a computer-readable storage medium containing data. Technical background

[0002] According to the state of the art, it is known to monitor platforms using cameras. The camera images are monitored, for example, by the operating personnel of a railway company, so that train traffic can be interrupted if an obstacle is detected on the track section extending along the platform. Obstacles are understood to be objects or living beings whose size would lead to a collision with an approaching train. These are foreign bodies on the relevant track section that can preferably enter the track section from the platform.

[0003] The problem arising from the described state of the art is that manually monitoring train platforms is labor-intensive and therefore costly. Furthermore, misjudgments by staff cannot be ruled out. However, automating the monitoring process presents the challenge that the task at hand is safety-relevant (safety, in the context of this invention description, refers to operational safety, also known as safety), meaning that any automation must meet the safety requirements for railway operations. Summary of the invention

[0004] The object of the invention is to solve the problems described in the prior art. In particular, it is an object to provide a method for monitoring a track section located at a railway platform, with which monitoring can be automated and a high level of safety can be achieved at the same time. Furthermore, it is an object of the invention to provide a vehicle, a computer program product, and a computer-readable storage medium with which the improved method can be implemented.

[0005] According to a first aspect of the invention, a method for monitoring a track section located at a railway platform is described in which the track section is detected by several imaging sensors for monitoring purposes.

[0006] Imaging sensors are defined as sensors whose measured values ​​can be converted into an image of the environment being monitored, primarily the track section and the platform edge. These can include, for example, optical cameras, radar, lidar, and similar devices.

[0007] A device is computer-aided or computer-implemented if it has a computing environment, or a method is computer-implemented if a computing environment performs at least one step of the method.

[0008] A computing environment is an IT infrastructure consisting of functional components such as processors, memory units, programs, and the data to be processed by these programs. This data is used to execute at least one application, which has a specific task to perform. Additional functional components can include sensors and actuators, which enable the computing environment to interact with the outside world. The IT infrastructure can also be organized as a network of these functional components.

[0009] Within a computing environment, computing instances form functional units that can be assigned to applications (defined, for example, by a number of program modules) and can execute them. During application execution, these functional units form self-contained systems, either physically (e.g., computer, processor) and / or virtually (e.g., program module).

[0010] Computers are electronic devices consisting of several functional components and possessing data processing capabilities. For example, computers can be clients, servers, handheld computers, communication devices, and other electronic devices for data processing, which may include processors and memory units and may also be interconnected via interfaces to form a network.

[0011] Processors can be, for example, converters, sensors for generating measurement signals, or electronic circuits. A processor can be a central processing unit (CPU), a microprocessor, a microcontroller, or a digital signal processor, possibly in combination with a memory unit for storing program instructions and data. The term "processor" can also refer to a virtualized processor or a soft CPU.

[0012] Storage units can be implemented on computer-readable storage devices in the form of random-access memory (RAM) or data storage devices (hard disk or data carrier).

[0013] Program modules are individual software functional units that enable a program sequence of process steps according to the invention.

[0014] These software functional units can be implemented in a single computer program or in several communicating computer programs. The interfaces implemented here can be implemented in software within a single processor or in hardware if multiple processors are used.

[0015] Interfaces can be implemented using hardware, for example wired or wireless connections, or software, for example as interaction between individual program modules of one or more computer programs, and serve to exchange data, preferably in the form of digital data sets or analog signals.

[0016] To avoid misunderstandings, it should be noted that individual claim features are numbered with lowercase Latin letters, without regard to the claim numbering. This means that each letter appears only once in the entire claim set, allowing for unambiguous addressing of the relevant claim features without mentioning the claim number. Therefore, the order of the letters is irrelevant.

[0017] According to the invention, it is provided that a) the sensors are aligned with the track section in such a way that images with an overlap can be generated, whereby as a result of the overlap every location of the track section is present in at least two images, b) the generated images are analyzed by computer for obstacles in the track section, c) a signal indicating the presence of an obstacle is generated if an obstacle has been detected by the analysis, and d) a signal indicating a fault is generated if the obstacle has only been detected once by the analysis.

[0018] The terms used in this description of the invention have the following meanings.

[0019] When referring to a track section, this means the track system, comprising a superstructure consisting primarily of the rails and sleepers (and also, for example, fastenings for the rails), and a substructure that provides the foundation for the sleepers, such as fill or a concrete base. It is readily apparent that the aforementioned components of the track section form part of the visible surface and are therefore also captured in the imaging process. However, if these components are obscured by an object, this can be determined in a known manner by evaluating the images generated by the imaging process. For this purpose, the process can, for example, utilize computer-generated artificial intelligence.

[0020] When this description of the invention refers to "detection," it means that the imaging sensor scans the track section or measures any radiation emanating from the track section. This can be done optically, for example, using an image sensor, but also by scanning, for example, using radar, whereby an image can also be generated using radar scanning.

[0021] For the purposes of this invention, obstacles are defined as all objects whose presence is detected in the track section, even though these objects are not intended to be there. Examples include people, animals, or larger inanimate objects. Plants or parts of plants are also considered inanimate objects, as they cannot move independently. Objects that are part of the railway infrastructure in a broader sense and cover the track section, such as balise housings, are not considered obstacles.

[0022] A signal indicating an obstacle is to be interpreted, based on its characteristics, as meaning that a foreign object has been detected in the track section. This will also be referred to as an obstacle signal in the following. A signal indicating a fault is to be interpreted, based on its characteristics, as meaning that a fault has occurred, resulting in a foreign object being detected only once. A fault must be present because the overlap of the generated images is sufficient for every location in the track section to be represented in at least two images. This also means that a detected foreign object influences the generation of at least two images. Therefore, if a foreign object is detected only once, a fault exists with regard to the proper functioning of the procedure, which can have various causes. 1) A foreign object might be positioned at the edge of the image in such a way that it is not fully captured and therefore not detected. 2) Image recognition might fail even though the foreign object is completely within the image. 3) An imaging sensor might be defective, for example, completely failed.

[0023] Thus, the error signal can be interpreted in multiple ways and may require different actions. The hardware and software involved in the process can be checked for proper function (maintenance can be requested, for example). Furthermore, the presence of an obstacle is likely, which is why the obstacle signal is generated in this case as well, according to process step c). In other words, the fact that an error signal is generated in process step d) is not interpreted as meaning that an obstacle signal generated according to process step c) was generated unjustifiably, even though this possibility cannot be ruled out. In the interest of operational safety, in case of doubt, the process proceeds as if the obstacle signal was generated correctly.

[0024] One advantage of the invention is that the redundant generation of obstacle signals provides an additional failure indicator for the hardware and software involved in the process, which is firmly integrated into the process flow. In other words, if the process is running correctly, an obstacle signal for the foreign object in question must always be generated twice, i.e., redundantly. If this is not the case, it indicates one of the aforementioned errors 1) to 3). A partial failure of the system therefore means that it can still be used to detect foreign objects in the track section. This is achieved because even the single detection of a foreign object as an obstacle, as explained above, triggers the generation of an obstacle signal. At the same time, the process can be repaired in time, so that a complete failure does not normally occur. For example, a defective camera can be replaced.This increases the overall operational reliability of the process, making it advantageously suitable for use in safety-relevant applications such as the monitoring of a track section adjacent to a platform, as presented here. Ideally, the safety levels SIL-1 and SIL-2 applicable to such tasks in railway operations can be achieved.

[0025] The requirements for the certification of safety-relevant applications, for example in railway technology, are very high. According to the international standard IEC 61508, and specifically for the railway sector according to the European standard EN 50129, four Safety Integrity Levels (SILs) are distinguished for safety functions to ensure the required functional safety. Safety Integrity Level 4 represents the highest and Safety Integrity Level 1 the lowest level of safety integrity. The respective Safety Integrity Level influences the confidence interval of a measured value; the higher the Safety Integrity Level that the respective device must meet, the smaller the confidence interval.The dimension of functional safety for the various Safety Integrity Levels (SILs) can be clearly described by the expected frequency of a failure of the safety-relevant system, MTBF (Mean Time Between Failures), which is expressed in years (a). For SIL-1, this ranges from 10 to 100 years, for SIL-2 from 100 to 1000 years, for SIL-3 from 1000 to 10000 years, and for SIL-4 from 10000 to 100000 years.

[0026] According to a further aspect of the invention, a railway system with a platform and a track section located thereon is described, wherein a plurality of imaging sensors are installed in the railway system for monitoring the track section.

[0027] According to this aspect, the invention provides that the railway system has a computing environment configured to carry out a method according to one of the preceding claims. The advantages associated with this aspect of the invention have already been explained above, and reference is made to these advantages.

[0028] According to a further aspect of the invention, a computer program product is described, containing program instructions that can be executed by a computing environment. According to this aspect, the invention provides that at least steps b), c), and d) of the method according to any one of claims 1-9 are executed.

[0029] According to the invention, a computer program product containing program modules with program instructions is described, wherein the program modules can run in the same computing instance or in several computing instances of the computing environment. The computer program product, which can comprise one or more computer programs, can be used to carry out the method according to the invention and / or its exemplary embodiments, and the advantages described above are achieved through its implementation.

[0030] According to a further aspect of the invention, a computer-readable storage medium containing data, which is stored as data records on the storage medium, is described. According to this aspect, the invention provides that the data records make the computer program product described above, according to the last preceding claim, executable.

[0031] Furthermore, a provisioning device for storing and / or providing the computer program in the form of a computer-readable storage medium is described. The provisioning device is, for example, a storage unit that stores the computer program and makes it available for retrieval. Alternatively or additionally, the provisioning device is a network service, a computer system, a server system, in particular a distributed computer system, such as a cloud-based system or virtual computer system, which stores the computer program on a computer-readable storage medium and preferably makes it available in the form of a data stream.

[0032] The provision of the computer program product takes the form of program modules describing program data sets as a file, in particular as a download file, or as a data stream, in particular as a download data stream. The computer program product is transferred, for example, using the provisioning device to a computing environment so that the method according to the invention can be executed in one or more computing instances of this computing environment. Embodiments of the invention

[0033] Further developments of the invention, describing variants, are explained below without limiting the basic idea of ​​the invention.

[0034] According to one variant, the aspects of the invention explained above are determined by the fact that each of the sensors belongs either to a first group of sensors or to a second group of sensors, wherein the generated images of the first group each border on each other in a boundary area with adjacent images and the generated images of the second group also each border on each other in a boundary area with adjacent images.

[0035] When this description of the invention refers to images being adjacent, it means that the contents of neighboring images align in such a way as to create a seamless overall image. The focus of this overall image is the track section to be monitored, which is thus depicted completely and without gaps in each group of images. The boundary of the images is therefore located at their respective edges. In other words, the boundary of the images is a connecting area between adjacent images, through which the contents of the images can be seamlessly joined. Furthermore, the entire sensor data set is not necessarily used to generate the images. Due to the positioning of the sensors, a sensor axis defining the sensor's orientation can be arranged obliquely opposite the track section.In other words, the track segment forms a surface that approximates a horizontal plane, with the sensor axis oriented obliquely, i.e., not perpendicularly, to this horizontal plane. Due to the resulting perspective distortion in the sensor image, the area covered by the sensor must be larger than the area relevant to the part of the track segment being captured. To generate the image, it is essentially extracted from the sensor image and then contains only the area relevant for carrying out the process. Computer-aided algorithms that enable such a combination of individual images into a complete image are generally known and therefore do not need to be explained in detail here.

[0036] The boundary region can thus be defined as the edge area of ​​the respective adjacent images where they border the respective neighboring image. For practical reasons (alignment of the cameras in the train stations), an overlap area may also occur in this edge region. However, this embodiment of the invention is essentially designed not to use such an overlap area to generate redundant pixels in the overall image of a group of images (i.e., the images generated by one group of sensors). Rather, the other group of images is used to generate the redundancy.

[0037] Because there are two groups of images, each of which provides a complete representation of the track section, the redundancy of individual image elements, as explained above, is ensured – once in the first group and once in the second group. This does not, of course, preclude the possibility of generating a third or further groups of images, thus creating triple or even multiple redundancies.

[0038] One advantage of this approach is that by assigning sensors to the first and second groups respectively, the generation of redundancy in object detection within the images can be effectively controlled. These objects are then imaged exactly once in each group of images, meaning exactly twice when using both a first and a second group. This simplifies the process for determining the necessity of error signals. Furthermore, it allows for the detection of an error with a comparatively high degree of certainty.

[0039] According to one variant, the aspects of the invention explained above are determined by the fact that each image is divided into a central area and a peripheral area that at least partially surrounds this central area, wherein e) a warning signal is generated as an obstacle signal if an obstacle is detected in the perimeter area and f) an alarm signal is generated as an obstacle signal if an obstacle is detected in the central area.

[0040] A warning signal is a signal generated to indicate a potential hazard. The warning signal is intended to trigger further checks or simply heighten alertness, for example, for the train driver. In contrast, an alarm signal is a signal that, with regard to the necessary safety of railway operations, requires immediate safety action. This can preferably be achieved through an automatic emergency braking of the vehicle or an emergency stop initiated by the train driver.

[0041] It is easy to understand that the intrusion of a potential obstacle into the edge area does not (yet) provoke a safety-critical incident. This could happen, for example, if a passenger's hand is standing too close to the platform edge and extends over it. In contrast, a person on the track in the safety-relevant central area would be detected, and the resulting alarm signal would initiate or require immediate action.

[0042] The examples given illustrate that a perimeter area must cover a (currently) non-safety-critical section of the platform, including the track, while the central area must cover the entire area where detected obstacles require immediate safety action. A perimeter area is not necessarily present at all edges of the generated image. As the example above clearly demonstrates, a perimeter area that captures the platform edge is particularly advantageous. The side of the track section opposite the platform edge can also be additionally protected by a perimeter area if, for example, there is another track with a platform edge opposite. After all, obstacles can also enter the track from there.If the edge opposite the platform is secured by a fence or if platform screen doors are provided on the platform, the aforementioned edge areas in the images can be omitted.

[0043] A boundary area, which includes the boundary area already explained above, can be defined to reduce the probability of triggering false alarm signals when neighboring images overlap.

[0044] One advantage of this variant is that setting up edge zones allows the method according to the invention to react more sensitively, i.e., earlier, to (potential) obstacles in the track section, thus, for example, putting the train driver on alert. On the other hand, the higher sensitivity of the method does not always immediately trigger an alarm and consequently a safety measure, which in the case of a "false alarm" would unnecessarily disrupt train traffic.

[0045] According to one variant, the aspects of the invention explained above are determined by the fact that the sensors are aligned with the track section in such a way that the respective edge areas of the images extend between the respective central areas of the images and the platform edge.

[0046] If all images capture the space between the track and the platform edge as a perimeter zone, it is advantageously possible to monitor the entire platform edge for potential hazards caused by obstacles entering from the platform. It is also advantageous to integrate the platform edge itself, and even an adjacent safety strip of the platform, into the perimeter zone (provided that the aforementioned image elements are also detected by the sensors).

[0047] According to one variant, the aspects of the invention explained above are determined by the fact that for each boundary region at least one vector is defined which indicates a direction in which an obstacle can move into the part of the track section covered by the image.

[0048] One advantage of this variant is that the vectors make it easier and faster to assess whether identified objects should be considered a hazard in the form of a (potential) obstacle or not.

[0049] According to one variant, the aspects of the invention explained above are determined by the fact that obstacles found are supplemented by an enclosing body through computer-aided processing of the image data, and g) the volume of the body is calculated and h) the position of a projection of the body onto the track section is calculated.

[0050] It is assumed that, based on the camera image and any additional sensor information (e.g., lidar, IR), an enclosing shape can be found for each object located near the area. For example, a cylinder for stationary people, a cuboid or similar shape for other objects. This implies a rough classification of the object. Similarly, a vector can be assigned to each object with possible directions of movement. For example, a train can move with or against its direction of travel, but not to the right or left relative to its direction of travel.

[0051] Calculating the volume of an object helps determine whether it must be classified as an obstacle. For example, a small and / or flat object on the track will likely be easily traversed by an approaching vehicle without a collision. This applies, for instance, to packaging waste that has been thrown onto the track.

[0052] Projecting the object onto the track section serves to assess the object's position on the track section. Evaluating the projection is simpler than evaluating the object itself. This allows the evaluation to be performed within a horizontal, approximately two-dimensional area defined by the track section.

[0053] One advantage of this variant is that the hazard potential of an identified object can be better assessed, and thus its classification as an obstacle can be carried out more easily and with greater certainty.

[0054] According to one variant, the aspects of the invention explained above are determined by the fact that a test routine is carried out when a vehicle is on the track section.

[0055] One advantage of this approach is that vehicles regularly use the track section at a platform to facilitate boarding and alighting. A test routine can be performed each time this occurs. During this routine, the vehicle must be detected. Simultaneously, it can be verified whether the vehicle is being detected redundantly by the relevant sensors. This allows for regular verification of the procedure's functionality and the hardware components involved, such as the sensors. This significantly increases the safety of the procedure, as the test routine also provides evidence of component failures.

[0056] According to one variant, the aspects of the invention explained above are determined by the fact that, during the execution of the test routine, it is checked whether the vehicle was recognized as the object.

[0057] One advantage of this variant is that no warning or alarm signal needs to be triggered when a train approaches and the detected object is identified as a vehicle. This is, of course, a non-safety-critical process that occurs regularly during normal operation. If the object can be identified as a train, the test routine can advantageously be carried out simultaneously with the ongoing operation of the method. When a train approaches, it is necessary to simultaneously run the normal operation of the method according to the invention in order to detect obstacles on the track section. Alternatively, it would, of course, also be possible to forgo monitoring by the method according to the invention during test operation (running the test routine) and instead perform manual monitoring, for example, by the train driver or station staff who analyze the images.

[0058] According to one variant, the aspects of the invention explained above are determined by the fact that during the execution of the test routine i) counts how often all sensors have detected the object as a vehicle, where m is the number of detections of the vehicle, j) only generates an error signal if the number m is not a multiple of a number n, where, as a result of the overlap, every location of the track section is present in the number of n images.

[0059] The advantage is that a test routine can be executed very simply and at high speed if the algorithm for generating an error signal only needs to check whether the number m of vehicle detection events is a multiple of the redundant coverage n of the track section by the sensors. In other words, all redundant sensors must detect the vehicle. If the number m of detection events is not a multiple of the redundancy n, this is a clear indication that detection by one of the sensors was unsuccessful. This then triggers the generation of an error signal, which can then be used, for example, to perform maintenance on the safety system that executes the procedure. Exemplary embodiments of the drawing

[0060] Further details of the invention are described below with reference to the drawing. Identical or corresponding drawing elements are provided with the same reference numerals in each figure and are only explained more than once to the extent that differences arise between the individual figures.

[0061] The exemplary embodiments described below are preferred embodiments of the invention. In these exemplary embodiments, the described components each represent individual variants of the invention, which can be considered independently of one another. Each of these variants further develops the invention independently and can therefore be regarded as part of the invention individually or in a combination other than that shown. Furthermore, the described components can also be combined with the variants of the invention described above. Figure 1The figure shows a schematic embodiment of the railway system device according to the invention, showing the interactions between the functional components used. Figure 2 shows an exemplary embodiment of a computing environment for the device according to Figure 1 as a block diagram of the individual functional components and the interfaces formed between them, wherein individual computing instances execute program modules that can each run in one or more of the exemplary computers shown, and wherein the interfaces shown can accordingly be implemented in software in one computer or in hardware between different computers. Figure 3 and Figure 4 show schematic examples of images of a track section on a platform that overlap or border each other, in a top view. Figure 5schematically shows an image with a central area and a peripheral area, and people as obstacles in a three-dimensional representation. Figure 6 An embodiment of the method according to the invention is shown as a flowchart, wherein the process steps shown can be implemented individually or in groups by program modules, and wherein the computing instances and interfaces are defined according to Figure 2 are indicated by example. Detailed description of the exemplary implementations

[0062] Regardless of the grammatical gender of terms, persons with male, female or other gender identities are equally included.

[0063] According to Figure 1The diagram shows a platform BS and a track GL, on which a vehicle FZ is entering platform BS. There is also a control center LZ, where, for example, the current timetable can be monitored. Communication between the control center LZ, the vehicle FZ, and platform BS is possible via antennas AT through a fifth interface S5, a sixth interface S6, and a seventh interface S7.

[0064] Platform BS is equipped with a first sensor SN1, a second sensor SN2, a third sensor SN3 and a fourth sensor SN4, which in the example according to Figure 1 These are surveillance cameras. They monitor a track section STA of track GL, located in front of platform BS. If a living obstacle HD1 in the form of a person or an inanimate obstacle HD2 in the form of an object is present on track GL, this is detected by the sensors.

[0065] A computing environment RU, in which the inventive method takes place, can be considered jointly by Figure 1 and Figure 2 can be extracted. The computing instances and functional components used interact with each other via interfaces. It can be seen that a first computer CP1 (compare) is located on platform BS. Figure 2 ) communicates with the sensors. A third computer, CP3, is used in the control center LZ, and a second computer, CP2, is provided in the vehicle FZ. Naturally, the computing instances formed by the control center LZ, platform BS, and vehicle FZ can also contain multiple computers, of which the in Figure 2 The images shown are merely examples to illustrate the communication connections.

[0066] According to Figure 2The computers and sensors that form the respective computing instances are described in more detail below. In the first computer, CP1, a first processor, PR1, is connected to a first memory unit, SE1, via an eleventh interface, S11. In the second computer, CP2, a second processor, PR2, is connected to a second memory unit, SE2, via a twelfth interface, S12. In the third computer, CP3, a third processor, PR3, is connected to a third memory unit, SE3, via a thirteenth interface, S13. In the first sensor, SN1, a fifth processor, PR5, is connected to a fifth memory unit, SE5, via a fifteenth interface, S15. In the second sensor, SN2, a sixth processor, PR6, is connected to a sixth memory unit, SE6, via a sixteenth interface, S16. In the third sensor, SN3, a seventh processor, PR7, is connected to a seventh memory unit, SE7, via a seventeenth interface, S17.In the fourth sensor SN4, an eighth processor PR8 is connected to an eighth memory unit SE8 via an 18th interface S18.

[0067] A combination of Figure 1 and Figure 2The following can be deduced. The fifth processor PR5 of the first sensor SN1 and the first processor PR1 of the first computer CP1 are connected via a first interface S1. The sixth processor PR6 of the second sensor SN2 and the first processor PR1 of the first computer CP1 are connected via a second interface S2. The seventh processor PR7 of the third sensor SN3 and the first processor PR1 of the first computer CP1 are connected via a third interface S3. The eighth processor PR8 of the fourth sensor SN4 and the first processor PR1 of the first computer CP1 are connected via a fourth interface S4. The first processor PR1 and the second processor PR2 are connected via a fifth interface S5. The second processor PR2 and the third processor PR3 are connected via a sixth interface S6.The seventh interface S7 connects the first processor PR1 and the second processor PR2.

[0068] If, within the scope of this invention description, only computers, processors, memory units, sensors, images or interfaces are mentioned, the information generally refers to all of the computers, processors, memory units, sensors, images and other functional components named above in detail, which, connected via the interfaces, contribute to the formation of the computing environment RU.

[0069] In the Figures 3 and 4The platform BS can be schematically identified, and its length defines track section STA on track GL. This track section STA can be represented in several images B1 ... B9, although the actual edge of the original images captured by the sensors is not shown in figures three and four. It becomes clear that, due to the field of view 'a' resulting for the sensors, a larger original image is captured by the sensors than is shown in the figures. Figures 3 and 4As indicated, only the section of the original image displayed as an image, suitable for monitoring track section STA, is of interest. Algorithms for selecting the appropriate section in the original image and for correcting the resulting image are known and will not be explained in detail here. Images B1 ... B9 are generated using sensors SN1 ... SN9 (thus, an additional fifth sensor SN5, a sixth sensor SN6, a seventh sensor SN7, an eighth sensor SN8, and a ninth sensor SN9 are provided, which are constructed in the same way as those in Figure 1 and Figure 2 (sensors shown).

[0070] In Figure 3The first image B1, the third image B3, the fifth image B5, and the seventh image B7 form a first group, while the second image B2, the fourth image B4, and the sixth image B6 form a second group. Adjacent images in the first group and adjacent images in the second group directly abut each other with their relevant image sections. Thus, complete monitoring of track section STA is possible with both the first and second groups of images. This creates redundancy, as every point of interest on track section STA is therefore depicted in two images. The images from the first group and the images from the second group each overlap by half.

[0071] In Figure 4The first image B1, the fourth image B4, and the seventh image B7 form a first group; the second image B2, the fifth image B5, and the sixth image B6 form a second group; and the third image B3, the sixth image B6, and the ninth image form a third group. This ensures that every point of track section STA is represented in three images. Unlike in Figure 3 do the images overlap according to Figure 4 No. This is because the representations of images 1-3, images 4-6, and images 7-9 are exactly superimposed and therefore in Figure 4 They cannot be individually identified. In order to obtain the same image section α for the superimposed images of each group, the first sensor SN1, the second sensor SN2 and the third sensor SN3, as well as the fourth sensor SN4, the fifth sensor SN5 and the sixth sensor SN6, as well as the seventh sensor SN7, the eighth sensor SN8 and the ninth sensor SN9 are arranged at one point on platform BS.

[0072] In Figure 5 is one of the in Figures 3 and 4 The images shown are three-dimensional (and thus before any distortion correction). It becomes clear that the image can be divided into different areas, namely a central area ZB and a peripheral area RB. The peripheral area RB borders the central area ZB on the outside, but does not have to completely surround it (even if this is shown in the diagram). Figure 5(This is the case in the example shown). Rather, the edge area RB can be further divided into several segments, not all of which need to be included in the image. There is a platform-side edge segment BRS and an edge segment ARS facing away from platform BS. These two edge segments run parallel to the platform edge. In addition, there are two transition-side edge segments URS, which are aligned at right angles to the platform edge. The transition-side edge segments URS define a boundary area within which adjacent images, which should actually blend seamlessly into one another, are allowed to overlap. Therefore, these edge segments are referred to as transition-side edge segments URS. They define, so to speak, a tolerance range for the alignment of the images, which may not overlap beyond the width specified by the transition-side edge segment URS.The tolerance range is preferably dimensioned such that predefined obstacles to be detected are larger (preferably at least twice as large) than the width of the transition-side edge segment URS defining the tolerance range, measured parallel to the platform edge.

[0073] In Figure 5 Two people are also depicted who can form living obstacles HD1. It can be seen that one of the people is standing in the center of the central area ZB. The other person is standing at the platform edge with their arm extending over the edge, so that it is located in the platform-side edge segment BRS. Therefore, the person in the central area ZB is intended to trigger an alarm signal to interrupt train traffic at platform BS, while the person at the platform edge is only intended to trigger a warning signal, for example, to increase the train driver's attention.

[0074] In Figure 5The diagram also schematically illustrates how the assessment can be performed using a computer-aided algorithm. The living obstacles HD1 are surrounded by an enclosing body UKP, into which they each fit, using image processing. In the example according to... Figure 5 The enclosing bodies UKP are each cylinders. It is now possible to perform a projection PRJ of the enclosing body UKP onto the image and thus the background. This is in Figure 5 The area is shown hatched. Using the PRJ projection, the algorithm can assess where the obstacle is located. Specifically, as shown in Figure 5 shown in one case with the entire projection area PRJ completely in the central area ZB and in the other case with only a part of the projection area PRJ in the platform-side edge area RB.

[0075] The following describes the method according to the invention by way of example, as shown in the flowchart according to Figure 3 The functional components and computing instances are presented and explained step by step. Figure 1 and 2 These processes can be carried out. Computer-aided steps take place in the processors, which are not shown in detail.

[0076] In the first step 1, the process is started (abbreviated: START).

[0077] In a second step, the STA track section is recorded using the sensors (SCN for short). Image processing can already be performed at this stage to obtain relevant images from the original images generated by the sensors (selection of image sections, possibly distortion correction). However, the original images can also be used.

[0078] In a third step, the images generated by the sensors (ANL for short) are analyzed to detect obstacles on track section STA. If necessary, enclosing bodies and projection surfaces can be generated (see Figure 5).

[0079] In a fourth step, step 4, a check is performed to see if an obstacle has been detected on track section STA (abbreviated as OBS). If not, the process recursively repeats step 2. If an obstacle has been detected, the process continues to step 5. For the purpose of obstacle detection, the images are subjected to computer-aided image processing in a known manner, and the image content is analyzed for the presence of obstacles (application of known image processing algorithms).

[0080] In a fifth step (5), a further check is performed to determine whether the detected obstacle was identified in a central area (ZB) of the image (abbreviated: CEN?). During image processing, an area within the image is thus defined and interpreted as the central area (ZB). If the object is at least partially located within the central area (ZB), there is a risk of collision between approaching trains and the obstacle, and the process continues to step 7. If the obstacle is not detected in the central area (ZB), this means that the obstacle lies entirely within a peripheral area (RB) of the image, where the peripheral area (RB) seamlessly adjoins the central area (ZB), and the process continues to step 6. Thus, the central area (ZB) and the peripheral area (RB) of the image combine to form the evaluation area within the image. This evaluation area does not necessarily have to encompass the entire image content if the original images are used.The original images may also contain elements that do not need to be checked for obstacles (for example, a background behind track section STA or the foreground of platform BS that is outside a safety distance from the platform edge). These will then not be checked.

[0081] In a sixth step, a warning signal is generated and output, or further processed in a computer-aided process (WRN). This warning signal does not interrupt train traffic because there is no immediate risk of collision. However, the warning signal could be issued to the driver of an approaching train in the cab so that they can prepare to brake. Therefore, a recursion to step 2 occurs, meaning that further images of track section STA are generated.

[0082] In a seventh step (7), a further check is performed to determine whether the detected obstacle is a train (abbreviated: TRN?). A train is not usually interpreted as an obstacle because it is part of normal railway operations for trains to pull up to platform BS. In this case, the process continues to step 11. To ensure that the train is permitted to pull up to platform BS as scheduled, a [missing information] can be [missing information]. Figure 6 A comparison with a timetable, in particular a current actual timetable, for example in the control center LZ via the first interface S1, is not shown. If the train is not expected at platform BS as scheduled, an obstacle signal can be generated because the unscheduled train could pose an obstacle to a scheduled train.

[0083] If no train could be detected, this means that another obstacle is present. There is an imminent risk of collision with the obstacle. In this case, the process continues to step 8. In this eighth step, a warning signal is generated, which can be output or processed computationally (abbreviated as ALT).

[0084] In a ninth step, a safety measure (MSR) is implemented. This could, for example, consist of the train applying the emergency brakes. Following trains may also need to be stopped. The derivation of these measures is known per se and therefore requires no further explanation within the scope of this invention description.

[0085] In a tenth step (10), the process is terminated (abbreviated: STOP). When train service resumes, the process can be restarted with step 1.

[0086] In the eleventh step, a test routine (TST) is executed. This test routine serves to detect camera failures and thus advantageously increases the safety level during the process. Errors in the process can be detected early in this way and used, for example, to initiate maintenance measures, while operation can continue due to the redundancy of the sensors used. The test routine can be started in parallel with the process, which is why a recursion to step two also occurs.

[0087] In a twelfth step, a check is performed to determine whether the total number m of obstacles detected by all sensors, divided by the redundancy n (n = 1 for simple overlap, etc.), results in a natural number N (in short: m / n = N?). If this is the case, it indicates that all sensors have detected the obstacle. Of course, two sensors could have failed simultaneously, which would go unnoticed. However, this scenario is far less likely than a single sensor failing, which is why this check significantly increases the reliability of the process. No further action is required.

[0088] However, if the result is not a natural number, this indicates that a sensor has produced faulty results or has failed. Therefore, in step 13, an error signal (ERR) is output. This error signal could, for example, be used to initiate maintenance of the system. If a higher level of safety is required, the error signal can also be used as a reason to stop the process and train operations (not shown). Reference symbol list

[0089] ARS Far-facing edge segment ATA Antennas B1 First image B2 Second image B3 Third image B4 Fourth image B5 Fifth image B6 Sixth image B7 Seventh image BRS Platform-side edge segment BS Platform CP1 First computer CP2 Second computer CP3 Third computer FZ Vehicle GL Track HD1 Live obstacle HD2 Inanimate obstacle LZ Control center RB Edge area PR1 First processor PR2 Second processor PR3 Third processor PR5 Fifth processor PR6 Sixth processor PR7 Seventh processor PR8 Eighth processor PRJ Projection RUR Computing environment S1 First interface S11 Eleventh interface S12 Twelfth interface S13 13th interface S15 15th interface S16 16th interface S17 17th interface S18 18th interfaceInterface S1 first interface S2 second interface S3 third interface S4 fourth interface S5 fifth interface S6 sixth interface S7 seventh interface SE1 first storage unit SE2 second storage unit SE3 third storage unit SE5 fifth storage unit SE6 sixth storage unit SE7 seventh storage unit SE8 eighth storage unit SN1 first sensor SN2 second sensor SN3 third sensor SN4 fourth sensor SN5 fifth sensor SN6 sixth sensor SN7 seventh sensor SN8 eighth sensor SN9 ninth sensor STAG track section UKP enclosing bodies URS transition-side edge segments ZB central area.

Claims

1. Method for monitoring a track section (STA) located at a platform (BS) in which the track section (STA) is recorded by several imaging sensors for monitoring purposes, characterized by the fact that a) the sensors are aligned with the track section (STA) in such a way that images with an overlap can be generated, whereby as a result of the overlap every location of the track section (STA) is present in at least two images, b) the generated images are analyzed by computer for obstacles in the track section (STA), c) a signal indicating the presence of an obstacle is generated if an obstacle has been detected by the analysis, d) a signal indicating a fault is generated if the obstacle has only been detected once by the analysis.

2. Method according to claim 1, characterized by the fact thatEach of the sensors belongs to either a first group of sensors or a second group of sensors, with the generated images of the first group each bordering on adjacent images in a boundary area, and the generated images of the second group also each bordering on adjacent images in a boundary area.

3. Method according to claim 1 or 2, characterized by the fact that Each image is divided into a central area (ZB) and a peripheral area (RB) that at least partially surrounds this central area (ZB), wherein e) a warning signal is generated as an obstacle signal if an obstacle has been detected in the peripheral area (RB), and f) an alarm signal is generated as an obstacle signal if an obstacle has been detected in the central area (ZB).

4. Method according to claim 3, characterized by the fact thatthe sensors are aligned with the track section (STA) in such a way that the respective edge areas (RB) of the images extend between the respective central areas (ZB) of the images and the platform edge.

5. Method according to any one of the preceding claims, characterized by the fact that For each boundary area (RB), at least one vector is defined that specifies a direction in which an obstacle can move into the part of the track section (STA) covered by the image.

6. Method according to any one of the preceding claims, characterized by the fact that Found obstacles are supplemented by a computer-aided processing of the image data by an enclosing body (UKP), and g) the volume of the body is calculated and h) the position of a projection (PRJ) of the body onto the track section (STA) is calculated.

7. Method according to any one of the preceding claims, characterized by the fact thatA test routine is performed when a vehicle (FZ) is located on the track section (STA).

8. Method according to claim 7, characterized by the fact that During the execution of the test routine, it is checked whether the vehicle (FZ) was recognized as the object.

9. Method according to claim 8, characterized by the fact that During the execution of the test routine i) the number of times all sensors have detected the object as a vehicle (FZ) is counted, where m is the number of detections of the vehicle (FZ), j) an error signal is only generated if the number m is not a multiple of a number n, where, as a result of the overlap, each location of the track section (STA) is present in the number of n images.

10. Railway system comprising a platform (BS) and a track section (STA) located thereon, wherein a large number of imaging sensors for monitoring the track section (STA) are installed in the railway system, characterized by the fact thatthe railway system has a computing environment (RU) that is set up to carry out a method according to one of the preceding claims.

11. Computer program product containing program instructions executable by a computing environment (RU) such that at least steps b), c) and d) of the method according to any one of claims 1 - 9 are executed.

12. Computer-readable storage medium containing data which are stored as data records on the storage medium, such that the data records make the computer program product according to the last preceding claim executable.

Images