System and method for monitoring railway facilities

Abstract

A method and system for monitoring a railway facility designed to identify rolling stock using posts (1, 2, 4) at entries / exits, signposted (6) and non-signposted (7) areas for identification at entries / exits and to create virtual blocks in non-signposted areas, facilitating the tracking of rolling stock. A local computer (28) from posts (1, 2, 4) processes signals and captured images to send them to a data center (400), wherein an image identification computer (33) detects the rolling stock, and another identification computer (37) identifies it by comparing signal information with historical data and extracting unique identifiers (13) associated with the rolling stock. An AEI computer (34) processes sensor signals to compare them with historical data and validate the identification. An image recognition computer (38) extracts information from signaling elements, differentiating signposted and non-signposted areas. Finally, this information is associated with the identified rolling stock and its geographic position by a tracking computer (39) at the data center (400).

Description

OBJECT OF THE INVENTION

[0001] The present invention falls within the technical field of monitoring systems in railway workshops and facilities.

[0002] Particularly, the present invention refers to a system and method for monitoring rolling stock through its identification, location and tracking in railway environments.BACKGROUND OF THE INVENTION

[0003] Railway workshops, depots, classification centers or complexes are characterized by having a large number of tracks to be able to carry out reception, parking, loading / unloading work, as well as rolling stock maintenance activities, among other types of operations. The identification and monitoring of rolling stock within the facility is highly valuable information when it comes to managing, planning and maintaining the facilities.

[0004] In this type of railway facilities it is common to find areas or tracks that have signaling facilities, that is, train presence detection systems (such as axle counters or track circuits), drives to automate the movement of turnouts or track devices and signs that control the entries, exits or operations carried out on those tracks. These specially signposted areas are on the entry and exit tracks, but given the large dimensions of this type of facilities and the features of the work carried out, it is also common to find areas and groups of tracks that lack signaling systems.

[0005] Other features of these railway environments are that the distance between the axes of two adjacent tracks is usually as small as possible to optimize space and can commonly be less than 5m, and on the other hand they usually have covered parts where the workshops are located, which are buildings or warehouses in which maintenance work is carried out. Additionally, many of these roads have electrification systems that generate electromagnetic fields and disturbances. All these features are factors that affect both the identification of rolling stock and its location and monitoring within the railway complex.

[0006] US20120056042A1 describes a system for tracking rolling stock in signposted facilities, commonly with axle counters where the identification of the rolling stock is carried out through the use of an automated equipment identification (AEI) reader or text recognition through optical character recognition (OCR) at the entries and exits of the facilities. This system has the limitation that it can only be used in signposted facilities. Not allowing rolling stock to be tracked on non- signposted tracks. Additionally, the data acquisition of data from the signaling system is carried out through the use of signals coming directly from axle counters.

[0007] US7826938B2 describes a system for tracking railcars through the use of an automated equipment identification (AEI) reader based on a radio frequency identification (RFID) reader. The problem with this system is that it requires a large number of readers located at the entries and exits of each track, which makes it a cost-inefficient solution. Furthermore, the system is poorly easy if the rail vehicles entering the facility lack RFID tags or if the message contained in the RFID tags is unknown.

[0008] US6637703B2 describes a system for tracking railcars through the use of an AEI reader that is also an RFID reader and the use of overhead cameras. Railcars on a track yard are tracked by recognizing patterns in video images acquired by cameras and signals acquired by readers. In general, pattern recognition using cameras is known to be less accurate and less reliable for moving objects under changing lighting and weather conditions. Furthermore, rolling stock belonging to different operators does not always have RFID tags installed, or the message it provides is encoded, or is repeated between vehicles, not allowing an association to be made between the vehicle and the RFID code being read.

[0009] US6511023B2 describes a system for tracking railcars using AEI readers and wheel counting stations. AEI readers identify a train running on a track. Wheel counting stations are located between AEI readers to increase identification locations. However, trains with the same number of wheels cannot be distinguished by that system. This is a particular problem at a track yard or train depot where most trains are relatively short, for example a single railcar, or the number of railcars in a train changes dynamically.

[0010] US6377877B1 describes a system for tracking railcars by comparing a location and a itinerary of a railcar. The location is obtained from a global positioning satellite system (GPS). Since the railcar is not specifically identified, incorrect information may be collected. Additionally, solutions based on the global navigation satellite system (GNSS) cannot provide adequate location accuracy indoors.

[0011] US20110017693A1 is a system to identify and track cargo containers that, by using OCR and RFID readers in a reading arc at the entries, tracks information on the position of turnouts and the loading and unloading of the containers. It is a system that focuses on the tracking of cargo containers, but does not contemplate or apply the tracking of rolling stock.

[0012] In view of the problems of the existing solutions, the objective technical problem presented is to improve, taking into account all the aspects and features of the railway environments mentioned above, the identification and tracking of rolling stock for monitoring and control within any railway environment.DESCRIPTION OF THE INVENTION

[0013] The present invention serves to solve the problem mentioned above, by providing a method and computer system for monitoring and controlling, in real time, in railway workshops and facilities, the use made by different vehicles or rolling stock (including locomotives, railcars and wagons) in these facilities.

[0014] The proposed method and system comprise carrying out, in a first step, the identification of rolling stock, through the combined use of sensors at the entries and exits of the railway facilities with which the electromagnetic, thermal, sound emissions and possible radio frequency identification tags ("RFID tags") that may be installed on the rolling stock are read and characterized. In the identification of rolling stock, positioning data (for example, global navigation satellite system or GNSS positions) of the rolling stock are also used, if available, and images from different electronic imaging devices associated with the railway facilities are used. The images and data collected are processed to identify the rolling stock, characterize the composition and read its unique identifier based on the unique numerical code given by the UIC or Railway International Union (UIC identifier) for each rolling stock.

[0015] The proposed method and system further comprise the monitoring of rolling stock within the railway facility that combines the data collected in the identification process, mentioned above, with data from non-intrusive equipment for the acquisition of electronic images, which comprises images captured by signaling devices and images from virtual blocks used to monitor rolling stock in non-signposted areas.

[0016] One aspect of the invention relates to a method for monitoring railway facilities by identifying rolling stock and tracking in railway facilities, both in the signposted areas and in the non- signposted areas of a railway facility, the method comprising the following steps: capturing images by one or more electronic imaging devices and capturing signals by at least one set of sensors, wherein the imagers and sensors are fixed on a fixing post installed within the railway facility at a known location that may be one of: i) an entry / exit area belonging to the signposted areas of the railway facility, wherein the fixing post is a rolling stock identification post, or ii) a passage area belonging to the non-signposted areas of the railway facility where the fixing post is a post for the generation of virtual blocks for tracking rolling stock, and iii) in a mixed post for the identification of rolling stock at entries / exits and the generation of virtual blocks for tracking rolling stock; locally processing at the railway facility, by a local computer, the signals captured by each set of sensors and the images captured by each electronic imaging device; sending the information processed by the local computer from the railway facility to a first data center that receives image information and signal information; recursively detecting in the received image information, by an image identification computer in the first data center, the presence of rolling stock at the known location of the electronic imaging device to which the received image information belongs; when the presence of rolling stock has been detected, identifying the detected rolling stock, by an identification computer in the first data center, using an identification module configured to perform a comparison of the received signal information with historical signal information previously analyzed by the identification computer, associated with a unique identifier of the rolling stock and stored in a database according to the unique identifier; extracting from the received image information, by an image recognition computer in the first data center, information of signaling elements, wherein: if the received image information comes from the signposted areas, the signaling element information comprises status information of rolling stock presence detectors, positions of turnouts and status of signals generated by a signaling monitoring module, if the received image information comes from the non-signposted areas, the signaling element information comprises status information of the virtual block determined by a virtual block tracking module; associating the information of the signaling elements extracted by the image recognition computer with the rolling stock identified by the identification computer and with a geographical position in the railway facility, for the tracking of the rolling stock identified by a tracking computer in the first data center.

[0017] Another aspect of the invention refers to a system for monitoring railway facilities configured to carry out the steps of the previously defined method.

[0018] Another aspect of the invention relates to a computer program product comprising instructions that, when the program is executed by a computer, cause it to carry out the method defined above.

[0019] Another aspect of the invention relates to a computer-readable medium comprising instructions that, when executed by the computer, cause it to execute the method defined above.

[0020] The scope of the present invention is defined by the independent and dependent claims that are attached hereto. For a more complete understanding of the invention, its objects and advantages, descriptive reference may be made to this text and the accompanying drawings.DESCRIPTION OF THE FIGURES

[0021] To complement the description that is being made and in order to help a better understanding of the features of the invention, in accordance with a preferred example of practical embodiment thereof, a set of drawings is attached as an integral part of said description wherein, for illustrative and non-limiting purposes, the following has been represented: Figure 1.- Shows a diagram of a railway facility to illustrate its different areas and show an example of the location of the equipment on the track. Figure 2.- Shows a train (rolling stock) at a railway facility with the different elements of the facility's monitoring system located on posts for the identification and monitoring of the rolling stock. Figure 3.- Shows a block diagram of a system for the identification and tracking of rolling stock at railway facilities, according to a possible embodiment of the invention. Figure 4.- Shows a flow chart of the rolling stock identification module used in the system according to a possible embodiment of the invention. Figure 5.- Shows a flow chart of the rolling stock tracking module used in the system according to a possible embodiment of the invention. Figure 6.- Shows a non-intrusive sub-system for the acquisition of signaling images that can be incorporated into the system of Figure 3, applicable to both operator stations and video walls, according to possible use cases of the invention. PREFERRED EMBODIMENT OF THE INVENTION

[0022] Below, with the help of the aforementioned figures, there is provided a detailed explanation of an example of a preferred embodiment of the object of the present invention.

[0023] A system is proposed for the identification of rolling stock at railway facilities such as the one shown in Figure 1. Typically, railway workshops and facilities consist of one or several entry / exit tracks, which contain signposted areas (6) and non-signposted areas (7), in which there are different types of built infrastructure (5), such as covered buildings or warehouses, wherein different rolling stock maintenance tasks are commonly carried out. The proposed system contemplates the use of at least one identification post (1) in the entry and exit areas, which is preferably low in height, and one or more posts for the generation of virtual blocks (2), which are used to monitor rolling stock in non-signposted areas (7). And depending on the track configuration of the railway facility, mixed posts (4) can also be used that combine identification at entries / exits and the generation of virtual blocks for tracking. For each of the posts (1, 2, 4) mentioned, one or more signal capture beams (3) for the rolling stock are defined, wherein each of the beams (3) can cover one or several tracks.

[0024] Figure 2 shows in more detail the different elements that make up each identification post (1), each virtual block generating post (2) or the mixed posts (4). A fixing post (12) about 3m high is fixed on a concrete base or anchored to a concrete slab, to which the rest of the components are anchored: at least one electronic imaging device (8), a cabinet (9) with a local computer (28) and communications equipment (29) comprising a radio frequency antenna (11), and a set of one or more sensors (10). The fixing post (12), being low in height, has the advantage that it can be easily installed between tracks of electrified tracks, avoiding the installation of gantry-type structures. Depending on the location on the track, the fixing post (12) includes one or multiple electronic imaging devices (8) and various types of sensors (10) for the characterization of the rolling stock, such as electromagnetic emissions detectors, thermal emissions detectors, sound emissions detectors and AEI readers. The set of sensors (10) are connected to the local computer (28) that performs local processing of data from sensors (10) and the images captured by each electronic imaging device (8). The local computer (28) of the railway facility or workshop (200) is connected to a secure communications equipment (29) configured to send the data to a first data center (400) that receives them through other communications equipment or data transport network (32) also being secured. Each rolling stock (for example, each railcar of the train drawn in Figure 2) carries a numerical code that contains a unique identifier (13) which is the UIC identifier, unique for each rolling stock. Additionally, the rolling stock may have RFID tags (14) for automated identification. Also optionally, the rolling stock may have satellite positioning devices (15), for example, GSNN equipment.

[0025] In the case of rolling stock with satellite positioning devices (15), the position is periodically reported to a computer (30) configured to process GSNN data located on the ground, at a second data center (300), shown in the Figure 3. The processed GSNN information is securely sent from the computer (30) of the second data center (300) to another computer (36) located at the first data center (400) through respective data transport networks (31, 32).

[0026] The components represented in Figure 3 are: 8 imaging device 10 sensors 19 operator station computer 20 display 21 video cable 22 video splitter 23 video capture computer 24 operator station data transport network 28 local computer 29 workshop data transport network 30 GNSS computer of the second data center 31 data transport network of the second data center 32 data transport network of the first data center 33 image identification computer 34 computer for Sensors and AEI 35 OCR computer 36 GNSS computer of the first data center 37 rolling stock identification computer 38 operator station image recognition computer 39 rolling stock tracking computer 40 computer for data storage 41 computer for human-machine interface (HMI) 42 data center data transport network with users 43 user data transport network with data center 44 mobile user terminals with HMI 45 fixed user terminals with HMI 200 railway workshop or facility 300 second data center 400 first data center 500 operator's station

[0027] The images received in the first data center (400) are processed by an image identification computer (33), which recursively detects the presence of rolling stock in all the images it receives from each electronic imaging device (8) whose location is known. And it performs a histogram comparison for the color distribution in each image using the Kullback-Leibler distance to measure their difference and, using a statistical classifier, obtain a confidence score of the probability of correctly identifying the presence of rolling stock and the direction of circulation thereof. In the images in which the presence of rolling stock has been detected, another statistical classifier based on neural networks, previously trained using samples of rolling stock images, is used to provide a security score of the probability of correctly identifying the composition, series, operator and manufacturer of the detected rolling stock.

[0028] The image identification computer (33) additionally performs the approximate detection of regions in the images that may contain alphanumeric characters, and more specifically of the numerical codes with the unique identifiers (13): UIC identifiers that are unique for each rolling stock. Once the detection of the numerical characters has been carried out, the image identification computer (33) cuts the images and proceeds to send the areas with the pixels that contain said characters to another OCR computer (35) for optical character recognition. The OCR computer (35) recursively makes changes to the scale, color softening, and focus of each cut image it receives before using a statistical classifier to identify characters that may be present in the images. This classifier uses a neural network, previously trained using virtual samples of characters, and obtains an assessment that must exceed a security score of the probability of correctly identifying each character. With the numerical sequence obtained that has exceeded the security thresholds, a validation method is carried out using the control (or verification) digits that are included in the UIC codes, identifiers (13), control digits that must be obtained through a mathematical operation with the rest of the figures of the UIC code, according to the numbering scheme established by the UIC. This validation method provides a system of unambiguous identification by the UIC code of the rolling stock. In another possible embodiment, part of the image processing carried out on the computers (33, 35) of the first data center (400) can be carried out on the local computer (28) installed on the post (12) of the facility or workshop (200) itself as edge computing to optimize response times and data sending bandwidth.

[0029] Data sent by the sensors (10) placed on the post (12) of the facility or workshop (200) are received in an AEI computer (34), also located at the first data center (400), which processes the signals captured as the rolling stock passes by. The AEI computer (34) analyzes the signals from each sensor (10) and, using a statistical classifier, makes a comparison of the signals received with the historical signals previously analyzed and stored in a database that have been previously related to the UIC identifier (13), obtaining a rating that shall exceed a security score of the probability of correctly identifying the received signal.

[0030] To carry out the identification of the rolling stock, the system comprises an identification module (4000). As indicated in the flow diagram of Figure 4, the identification module (4000) executes the following steps based on the reception of images and information of the train by AEI or other sensors and / or GNSS: If it receives valid images (4100) with the presence of rolling stock, it analyzes (4110) the images and cuts out texts and identifies series, operator, composition and track, sending said data extracted from the texts to a train identifier (4400) along with, after converting the image into text (4120) and checking if it corresponds to a valid train license plate (4130) - if not, the received image is discarded (4140) -, the obtained license plate. If it receives information (4200) by AEI or other sensors and it is a valid signal (i.e., it comes from rolling stock) and known (4210) according to the aforementioned classifier, it is passed directly to the train identifier (4400) of the computer (37); if not, it is checked if it is a new validated signal (4220) and: if it is valid, the signal is associated (4230) with a train identifier; if not, the received signal is discarded (4240). If it receives positioning information (4300) by GNSS satellite, it is passed directly to the train identifier (4400). If the train identifier (4400) provides a complete and unambiguous identification, it is checked if it is a new validated signal (4220) and: if it is valid, the signal is associated (4230) with a train identifier; if not, the received signal is discarded (4240). The result of the train identifier (4400), whether positive or not, is in any case sent as input information to the rolling stock tracking module shown in Figure 5.

[0031] In summary, if the signal is known and has exceeded the indicated security thresholds, the results of the statistical classifier that compares the received signals with the previously analyzed and stored historical signals are transmitted to the computer (37) for the identification of the rolling stock. But if the signals received from the sensors are valid, that is, they come from rolling stock but are not known since they do not resemble any previously analyzed and stored signal in the database, a process is carried out to try to associate the new signal received to a unique UIC code. This process of associating signals from sensors to UIC codes uses another statistical classifier that uses data from sensors of various inputs or outputs of the same railway vehicle that has been unequivocally identified as a result of the processes carried out in the computer (37). If said statistical classifier determines that the signals are unique and exceeds a previously established security threshold, the signals from the sensors are associated with the UIC code and this information is stored in the database used by the computer (34) for the signal identification.

[0032] The results provided by the computers (33, 34, 35 and 36) are sent to the computer (37) which uses a statistical classifier and provides a score with the probability of correctly identifying the composition, series, operator, manufacturer and UIC code, if this score exceeds a pre-established security threshold, the identification of the rolling stock is determined and the result is sent to a tracking computer (39).

[0033] Tracking of rolling stock inside the railway facility is carried out in the rolling stock tracking module, which is executed in the tracking computer (39), according to the flow diagram represented in Figure 5, wherein the monitoring module is divided into a signaling monitoring module (5010) and a virtual block monitoring module (5020) that use data coming respectively from i) the signaling equipment and ii) the virtual blocks that are generated in the non-signposted areas and strategically located at specific points of the facility to carry out correct tracking. As indicated above, railway facilities may have non-signposted areas (7), in which it is also necessary to track and monitor the activity carried out by the rolling stock in these areas (for example, tracks). Therefore, the proposed monitoring method uses the generation of virtual blocks in specific passage areas of the railway facility, such as the example of the location of virtual blocks (2) indicated in Figure 1, to carry out complete monitoring in the facility. These virtual blocks are known track areas in which the presence and direction of circulation of the rolling stock that has been identified is detected using the components installed on the track that capture the signals - from the imaging device (8), sensors (10 )- and that process and transmit them -local computer (28) and transport network (29)-, along with the processes carried out in the computers for image identification (33), sensors (34) and train identification ( 37), previously described for identification in Figure 3. Each virtual block provides the presence, direction and composition data of the rolling stock passing through the specific area where the virtual block is located.

[0034] The computer (39) provides at its output the information on the tracking, location and identification of the rolling stock in the monitored railway facility, using the information from the identification computer (37) and the image recognition computer (38). This tracking and tracking process in turn uses the date and order of entry and exit of the rolling stock on each of the tracks, in which its length is known, and the features and unique identifier of the rolling stock or composition. As a reference time in all processes, a GNSS clock source or an NTP (Network Time Protocol - NTP) time server is used to determine the timestamp of each record.

[0035] As illustrated in Figure 5, the signaling tracking module (5010) executes the following steps upon receiving images: If it receives valid images (5100), it analyzes (5110) the images and generates the status of the signaling system which, together with the result data (5120) of the identification module (4000), is sent as rolling stock tracking information at the workshop (5130) and the information is saved (5140).

[0036] As illustrated in Figure 5, the virtual block tracking module (5020) executes the following steps based on the reception of images and information of the train by AEI or other sensors: If it receives valid images (5200), it analyzes (5210) the images, detects the rolling stock and determines the direction of circulation, which is the information that is passed to determine the status of the virtual block (5220). If it receives information (5300) by AEI or other sensors and it is a valid signal (i.e., it comes from rolling stock) and known signal (5310), the information is passed directly to determine the status of the virtual block (5220) and if not, the received signal is discarded (5320).

[0037] The proposed system additionally includes a non-intrusive data extraction procedure from the system's signaling equipment, as shown in Figure 6, wherein on the left side there are presented the field equipment (16), the operator stations connected through an interlock (172) in a technical building (17) to operator stations in the centralized control center (18) wherein a video wall can also be available, while the right side of the Figure shows in more detail the blocks of the operator station equipment (610) and the "video wall" equipment (620). The technical building (17) and the centralized control center (18), which includes centralized traffic control computers or CTC (182), are connected by two communications networks (171, 181). The data extraction procedure from the system's signaling equipment obtains information on the status of train presence detectors, positions of the turnouts and status of the signals, among other data, and as it is non-intrusive, it does not affect the Safety Integrity Level established for signaling facilities. For this data extraction, the system uses a video cable splitter (22) that connects: a) to the video cables (21) that provide the images of the signaling system to displays (20) of the equipment of the operator stations (610) that can be both local and central, as well as control center operator stations; or, b) installed at the location of the video wall equipment (620), in the video cable that connects the CTC or Centralized Traffic Control computer (25) with the controller computer (26) of the video wall (27). This video splitter (22) provides the image that is represented on the displays (20) or on the video wall (27) to the local video capture computer (23) that collects the images and transforms them into a compressed video format to send them through a data transport network (24) for secured communications. A data center receives them through secure communications equipment (32) and the images are collected and analyzed by an image recognition computer (38) for signaling systems, which uses a statistical classifier based on neural networks, previously trained with images of representations of signaling command and control systems, and identifies all the elements of the transmitted image that make up the signaling system. All signaling elements contained in the images are identified and associated with the geographical position they occupy in the monitored railway facility. By using another statistical classifier, the image recognition computer (38) proceeds to analyze the color changes that occur in the pixels that make up each element of the signaling equipment represented in the image, to obtain a security score of the probability of correctly identifying the change in color and status of the signaling element. All signaling element statuses are stored in a database.

[0038] With data resulting from the extraction of information from the signaling equipment, including the position of turnouts, the status of the rolling stock presence detectors and signals from images and sensors, together with the information obtained in the virtual blocks of the specific areas of passage of the railway facility that are not signposted, the system is capable of positioning the rolling stock and generating tracking itineraries thereof.

[0039] Data provided by the proposed system represents an improvement in the monitoring of the railway facility because, based on them, it is possible to: identify and track the rolling stock used by the monitored railway facility, map the position of the rolling stock in real or quasi-real time, calculate the time that each railway vehicle has been in a position in the workshop, the number of operations carried out and provide this information through a Human-Machine Interface or HMI server to users who connect through a data transport network (43) with mobile (44) and / or fixed (45) user devices / terminals for viewing and using the application through a graphical interface. The data provided by the system can also be shared with other external systems through a machine-machine interface or M2M so that it can be reused.

Examples

Embodiment Construction

[0022]Below, with the help of the aforementioned figures, there is provided a detailed explanation of an example of a preferred embodiment of the object of the present invention.

[0023]A system is proposed for the identification of rolling stock at railway facilities such as the one shown in Figure 1. Typically, railway workshops and facilities consist of one or several entry / exit tracks, which contain signposted areas (6) and non-signposted areas (7), in which there are different types of built infrastructure (5), such as covered buildings or warehouses, wherein different rolling stock maintenance tasks are commonly carried out. The proposed system contemplates the use of at least one identification post (1) in the entry and exit areas, which is preferably low in height, and one or more posts for the generation of virtual blocks (2), which are used to monitor rolling stock in non-signposted areas (7). And depending on the track configuration of the railway facility, mixed posts (4) ...

OBJECT OF THE INVENTION

[0001] The present invention falls within the technical field of monitoring systems in railway workshops and facilities.

[0002] Particularly, the present invention refers to a system and method for monitoring rolling stock through its identification, location and tracking in railway environments.BACKGROUND OF THE INVENTION

[0003] Railway workshops, depots, classification centers or complexes are characterized by having a large number of tracks to be able to carry out reception, parking, loading / unloading work, as well as rolling stock maintenance activities, among other types of operations. The identification and monitoring of rolling stock within the facility is highly valuable information when it comes to managing, planning and maintaining the facilities.

[0004] In this type of railway facilities it is common to find areas or tracks that have signaling facilities, that is, train presence detection systems (such as axle counters or track circuits), drives to automate the movement of turnouts or track devices and signs that control the entries, exits or operations carried out on those tracks. These specially signposted areas are on the entry and exit tracks, but given the large dimensions of this type of facilities and the features of the work carried out, it is also common to find areas and groups of tracks that lack signaling systems.

[0005] Other features of these railway environments are that the distance between the axes of two adjacent tracks is usually as small as possible to optimize space and can commonly be less than 5m, and on the other hand they usually have covered parts where the workshops are located, which are buildings or warehouses in which maintenance work is carried out. Additionally, many of these roads have electrification systems that generate electromagnetic fields and disturbances. All these features are factors that affect both the identification of rolling stock and its location and monitoring within the railway complex.

[0006] US20120056042A1 describes a system for tracking rolling stock in signposted facilities, commonly with axle counters where the identification of the rolling stock is carried out through the use of an automated equipment identification (AEI) reader or text recognition through optical character recognition (OCR) at the entries and exits of the facilities. This system has the limitation that it can only be used in signposted facilities. Not allowing rolling stock to be tracked on non- signposted tracks. Additionally, the data acquisition of data from the signaling system is carried out through the use of signals coming directly from axle counters.

[0007] US7826938B2 describes a system for tracking railcars through the use of an automated equipment identification (AEI) reader based on a radio frequency identification (RFID) reader. The problem with this system is that it requires a large number of readers located at the entries and exits of each track, which makes it a cost-inefficient solution. Furthermore, the system is poorly easy if the rail vehicles entering the facility lack RFID tags or if the message contained in the RFID tags is unknown.

[0008] US6637703B2 describes a system for tracking railcars through the use of an AEI reader that is also an RFID reader and the use of overhead cameras. Railcars on a track yard are tracked by recognizing patterns in video images acquired by cameras and signals acquired by readers. In general, pattern recognition using cameras is known to be less accurate and less reliable for moving objects under changing lighting and weather conditions. Furthermore, rolling stock belonging to different operators does not always have RFID tags installed, or the message it provides is encoded, or is repeated between vehicles, not allowing an association to be made between the vehicle and the RFID code being read.

[0009] US6511023B2 describes a system for tracking railcars using AEI readers and wheel counting stations. AEI readers identify a train running on a track. Wheel counting stations are located between AEI readers to increase identification locations. However, trains with the same number of wheels cannot be distinguished by that system. This is a particular problem at a track yard or train depot where most trains are relatively short, for example a single railcar, or the number of railcars in a train changes dynamically.

[0010] US6377877B1 describes a system for tracking railcars by comparing a location and a itinerary of a railcar. The location is obtained from a global positioning satellite system (GPS). Since the railcar is not specifically identified, incorrect information may be collected. Additionally, solutions based on the global navigation satellite system (GNSS) cannot provide adequate location accuracy indoors.

[0011] US20110017693A1 is a system to identify and track cargo containers that, by using OCR and RFID readers in a reading arc at the entries, tracks information on the position of turnouts and the loading and unloading of the containers. It is a system that focuses on the tracking of cargo containers, but does not contemplate or apply the tracking of rolling stock.

[0012] In view of the problems of the existing solutions, the objective technical problem presented is to improve, taking into account all the aspects and features of the railway environments mentioned above, the identification and tracking of rolling stock for monitoring and control within any railway environment.DESCRIPTION OF THE INVENTION

[0013] The present invention serves to solve the problem mentioned above, by providing a method and computer system for monitoring and controlling, in real time, in railway workshops and facilities, the use made by different vehicles or rolling stock (including locomotives, railcars and wagons) in these facilities.

[0014] The proposed method and system comprise carrying out, in a first step, the identification of rolling stock, through the combined use of sensors at the entries and exits of the railway facilities with which the electromagnetic, thermal, sound emissions and possible radio frequency identification tags ("RFID tags") that may be installed on the rolling stock are read and characterized. In the identification of rolling stock, positioning data (for example, global navigation satellite system or GNSS positions) of the rolling stock are also used, if available, and images from different electronic imaging devices associated with the railway facilities are used. The images and data collected are processed to identify the rolling stock, characterize the composition and read its unique identifier based on the unique numerical code given by the UIC or Railway International Union (UIC identifier) for each rolling stock.

[0015] The proposed method and system further comprise the monitoring of rolling stock within the railway facility that combines the data collected in the identification process, mentioned above, with data from non-intrusive equipment for the acquisition of electronic images, which comprises images captured by signaling devices and images from virtual blocks used to monitor rolling stock in non-signposted areas.

[0016] One aspect of the invention relates to a method for monitoring railway facilities by identifying rolling stock and tracking in railway facilities, both in the signposted areas and in the non- signposted areas of a railway facility, the method comprising the following steps: capturing images by one or more electronic imaging devices and capturing signals by at least one set of sensors, wherein the imagers and sensors are fixed on a fixing post installed within the railway facility at a known location that may be one of: i) an entry / exit area belonging to the signposted areas of the railway facility, wherein the fixing post is a rolling stock identification post, or ii) a passage area belonging to the non-signposted areas of the railway facility where the fixing post is a post for the generation of virtual blocks for tracking rolling stock, and iii) in a mixed post for the identification of rolling stock at entries / exits and the generation of virtual blocks for tracking rolling stock; locally processing at the railway facility, by a local computer, the signals captured by each set of sensors and the images captured by each electronic imaging device; sending the information processed by the local computer from the railway facility to a first data center that receives image information and signal information; recursively detecting in the received image information, by an image identification computer in the first data center, the presence of rolling stock at the known location of the electronic imaging device to which the received image information belongs; when the presence of rolling stock has been detected, identifying the detected rolling stock, by an identification computer in the first data center, using an identification module configured to perform a comparison of the received signal information with historical signal information previously analyzed by the identification computer, associated with a unique identifier of the rolling stock and stored in a database according to the unique identifier; extracting from the received image information, by an image recognition computer in the first data center, information of signaling elements, wherein: if the received image information comes from the signposted areas, the signaling element information comprises status information of rolling stock presence detectors, positions of turnouts and status of signals generated by a signaling monitoring module, if the received image information comes from the non-signposted areas, the signaling element information comprises status information of the virtual block determined by a virtual block tracking module; associating the information of the signaling elements extracted by the image recognition computer with the rolling stock identified by the identification computer and with a geographical position in the railway facility, for the tracking of the rolling stock identified by a tracking computer in the first data center.

[0017] Another aspect of the invention refers to a system for monitoring railway facilities configured to carry out the steps of the previously defined method.

[0018] Another aspect of the invention relates to a computer program product comprising instructions that, when the program is executed by a computer, cause it to carry out the method defined above.

[0019] Another aspect of the invention relates to a computer-readable medium comprising instructions that, when executed by the computer, cause it to execute the method defined above.

[0020] The scope of the present invention is defined by the independent and dependent claims that are attached hereto. For a more complete understanding of the invention, its objects and advantages, descriptive reference may be made to this text and the accompanying drawings.DESCRIPTION OF THE FIGURES

[0021] To complement the description that is being made and in order to help a better understanding of the features of the invention, in accordance with a preferred example of practical embodiment thereof, a set of drawings is attached as an integral part of said description wherein, for illustrative and non-limiting purposes, the following has been represented: Figure 1.- Shows a diagram of a railway facility to illustrate its different areas and show an example of the location of the equipment on the track. Figure 2.- Shows a train (rolling stock) at a railway facility with the different elements of the facility's monitoring system located on posts for the identification and monitoring of the rolling stock. Figure 3.- Shows a block diagram of a system for the identification and tracking of rolling stock at railway facilities, according to a possible embodiment of the invention. Figure 4.- Shows a flow chart of the rolling stock identification module used in the system according to a possible embodiment of the invention. Figure 5.- Shows a flow chart of the rolling stock tracking module used in the system according to a possible embodiment of the invention. Figure 6.- Shows a non-intrusive sub-system for the acquisition of signaling images that can be incorporated into the system of Figure 3, applicable to both operator stations and video walls, according to possible use cases of the invention. PREFERRED EMBODIMENT OF THE INVENTION

[0022] Below, with the help of the aforementioned figures, there is provided a detailed explanation of an example of a preferred embodiment of the object of the present invention.

[0023] A system is proposed for the identification of rolling stock at railway facilities such as the one shown in Figure 1. Typically, railway workshops and facilities consist of one or several entry / exit tracks, which contain signposted areas (6) and non-signposted areas (7), in which there are different types of built infrastructure (5), such as covered buildings or warehouses, wherein different rolling stock maintenance tasks are commonly carried out. The proposed system contemplates the use of at least one identification post (1) in the entry and exit areas, which is preferably low in height, and one or more posts for the generation of virtual blocks (2), which are used to monitor rolling stock in non-signposted areas (7). And depending on the track configuration of the railway facility, mixed posts (4) can also be used that combine identification at entries / exits and the generation of virtual blocks for tracking. For each of the posts (1, 2, 4) mentioned, one or more signal capture beams (3) for the rolling stock are defined, wherein each of the beams (3) can cover one or several tracks.

[0024] Figure 2 shows in more detail the different elements that make up each identification post (1), each virtual block generating post (2) or the mixed posts (4). A fixing post (12) about 3m high is fixed on a concrete base or anchored to a concrete slab, to which the rest of the components are anchored: at least one electronic imaging device (8), a cabinet (9) with a local computer (28) and communications equipment (29) comprising a radio frequency antenna (11), and a set of one or more sensors (10). The fixing post (12), being low in height, has the advantage that it can be easily installed between tracks of electrified tracks, avoiding the installation of gantry-type structures. Depending on the location on the track, the fixing post (12) includes one or multiple electronic imaging devices (8) and various types of sensors (10) for the characterization of the rolling stock, such as electromagnetic emissions detectors, thermal emissions detectors, sound emissions detectors and AEI readers. The set of sensors (10) are connected to the local computer (28) that performs local processing of data from sensors (10) and the images captured by each electronic imaging device (8). The local computer (28) of the railway facility or workshop (200) is connected to a secure communications equipment (29) configured to send the data to a first data center (400) that receives them through other communications equipment or data transport network (32) also being secured. Each rolling stock (for example, each railcar of the train drawn in Figure 2) carries a numerical code that contains a unique identifier (13) which is the UIC identifier, unique for each rolling stock. Additionally, the rolling stock may have RFID tags (14) for automated identification. Also optionally, the rolling stock may have satellite positioning devices (15), for example, GSNN equipment.

[0025] In the case of rolling stock with satellite positioning devices (15), the position is periodically reported to a computer (30) configured to process GSNN data located on the ground, at a second data center (300), shown in the Figure 3. The processed GSNN information is securely sent from the computer (30) of the second data center (300) to another computer (36) located at the first data center (400) through respective data transport networks (31, 32).

[0026] The components represented in Figure 3 are: 8 imaging device 10 sensors 19 operator station computer 20 display 21 video cable 22 video splitter 23 video capture computer 24 operator station data transport network 28 local computer 29 workshop data transport network 30 GNSS computer of the second data center 31 data transport network of the second data center 32 data transport network of the first data center 33 image identification computer 34 computer for Sensors and AEI 35 OCR computer 36 GNSS computer of the first data center 37 rolling stock identification computer 38 operator station image recognition computer 39 rolling stock tracking computer 40 computer for data storage 41 computer for human-machine interface (HMI) 42 data center data transport network with users 43 user data transport network with data center 44 mobile user terminals with HMI 45 fixed user terminals with HMI 200 railway workshop or facility 300 second data center 400 first data center 500 operator's station

[0027] The images received in the first data center (400) are processed by an image identification computer (33), which recursively detects the presence of rolling stock in all the images it receives from each electronic imaging device (8) whose location is known. And it performs a histogram comparison for the color distribution in each image using the Kullback-Leibler distance to measure their difference and, using a statistical classifier, obtain a confidence score of the probability of correctly identifying the presence of rolling stock and the direction of circulation thereof. In the images in which the presence of rolling stock has been detected, another statistical classifier based on neural networks, previously trained using samples of rolling stock images, is used to provide a security score of the probability of correctly identifying the composition, series, operator and manufacturer of the detected rolling stock.

[0028] The image identification computer (33) additionally performs the approximate detection of regions in the images that may contain alphanumeric characters, and more specifically of the numerical codes with the unique identifiers (13): UIC identifiers that are unique for each rolling stock. Once the detection of the numerical characters has been carried out, the image identification computer (33) cuts the images and proceeds to send the areas with the pixels that contain said characters to another OCR computer (35) for optical character recognition. The OCR computer (35) recursively makes changes to the scale, color softening, and focus of each cut image it receives before using a statistical classifier to identify characters that may be present in the images. This classifier uses a neural network, previously trained using virtual samples of characters, and obtains an assessment that must exceed a security score of the probability of correctly identifying each character. With the numerical sequence obtained that has exceeded the security thresholds, a validation method is carried out using the control (or verification) digits that are included in the UIC codes, identifiers (13), control digits that must be obtained through a mathematical operation with the rest of the figures of the UIC code, according to the numbering scheme established by the UIC. This validation method provides a system of unambiguous identification by the UIC code of the rolling stock. In another possible embodiment, part of the image processing carried out on the computers (33, 35) of the first data center (400) can be carried out on the local computer (28) installed on the post (12) of the facility or workshop (200) itself as edge computing to optimize response times and data sending bandwidth.

[0029] Data sent by the sensors (10) placed on the post (12) of the facility or workshop (200) are received in an AEI computer (34), also located at the first data center (400), which processes the signals captured as the rolling stock passes by. The AEI computer (34) analyzes the signals from each sensor (10) and, using a statistical classifier, makes a comparison of the signals received with the historical signals previously analyzed and stored in a database that have been previously related to the UIC identifier (13), obtaining a rating that shall exceed a security score of the probability of correctly identifying the received signal.

[0030] To carry out the identification of the rolling stock, the system comprises an identification module (4000). As indicated in the flow diagram of Figure 4, the identification module (4000) executes the following steps based on the reception of images and information of the train by AEI or other sensors and / or GNSS: If it receives valid images (4100) with the presence of rolling stock, it analyzes (4110) the images and cuts out texts and identifies series, operator, composition and track, sending said data extracted from the texts to a train identifier (4400) along with, after converting the image into text (4120) and checking if it corresponds to a valid train license plate (4130) - if not, the received image is discarded (4140) -, the obtained license plate. If it receives information (4200) by AEI or other sensors and it is a valid signal (i.e., it comes from rolling stock) and known (4210) according to the aforementioned classifier, it is passed directly to the train identifier (4400) of the computer (37); if not, it is checked if it is a new validated signal (4220) and: if it is valid, the signal is associated (4230) with a train identifier; if not, the received signal is discarded (4240). If it receives positioning information (4300) by GNSS satellite, it is passed directly to the train identifier (4400). If the train identifier (4400) provides a complete and unambiguous identification, it is checked if it is a new validated signal (4220) and: if it is valid, the signal is associated (4230) with a train identifier; if not, the received signal is discarded (4240). The result of the train identifier (4400), whether positive or not, is in any case sent as input information to the rolling stock tracking module shown in Figure 5.

[0031] In summary, if the signal is known and has exceeded the indicated security thresholds, the results of the statistical classifier that compares the received signals with the previously analyzed and stored historical signals are transmitted to the computer (37) for the identification of the rolling stock. But if the signals received from the sensors are valid, that is, they come from rolling stock but are not known since they do not resemble any previously analyzed and stored signal in the database, a process is carried out to try to associate the new signal received to a unique UIC code. This process of associating signals from sensors to UIC codes uses another statistical classifier that uses data from sensors of various inputs or outputs of the same railway vehicle that has been unequivocally identified as a result of the processes carried out in the computer (37). If said statistical classifier determines that the signals are unique and exceeds a previously established security threshold, the signals from the sensors are associated with the UIC code and this information is stored in the database used by the computer (34) for the signal identification.

[0032] The results provided by the computers (33, 34, 35 and 36) are sent to the computer (37) which uses a statistical classifier and provides a score with the probability of correctly identifying the composition, series, operator, manufacturer and UIC code, if this score exceeds a pre-established security threshold, the identification of the rolling stock is determined and the result is sent to a tracking computer (39).

[0033] Tracking of rolling stock inside the railway facility is carried out in the rolling stock tracking module, which is executed in the tracking computer (39), according to the flow diagram represented in Figure 5, wherein the monitoring module is divided into a signaling monitoring module (5010) and a virtual block monitoring module (5020) that use data coming respectively from i) the signaling equipment and ii) the virtual blocks that are generated in the non-signposted areas and strategically located at specific points of the facility to carry out correct tracking. As indicated above, railway facilities may have non-signposted areas (7), in which it is also necessary to track and monitor the activity carried out by the rolling stock in these areas (for example, tracks). Therefore, the proposed monitoring method uses the generation of virtual blocks in specific passage areas of the railway facility, such as the example of the location of virtual blocks (2) indicated in Figure 1, to carry out complete monitoring in the facility. These virtual blocks are known track areas in which the presence and direction of circulation of the rolling stock that has been identified is detected using the components installed on the track that capture the signals - from the imaging device (8), sensors (10 )- and that process and transmit them -local computer (28) and transport network (29)-, along with the processes carried out in the computers for image identification (33), sensors (34) and train identification ( 37), previously described for identification in Figure 3. Each virtual block provides the presence, direction and composition data of the rolling stock passing through the specific area where the virtual block is located.

[0034] The computer (39) provides at its output the information on the tracking, location and identification of the rolling stock in the monitored railway facility, using the information from the identification computer (37) and the image recognition computer (38). This tracking and tracking process in turn uses the date and order of entry and exit of the rolling stock on each of the tracks, in which its length is known, and the features and unique identifier of the rolling stock or composition. As a reference time in all processes, a GNSS clock source or an NTP (Network Time Protocol - NTP) time server is used to determine the timestamp of each record.

[0035] As illustrated in Figure 5, the signaling tracking module (5010) executes the following steps upon receiving images: If it receives valid images (5100), it analyzes (5110) the images and generates the status of the signaling system which, together with the result data (5120) of the identification module (4000), is sent as rolling stock tracking information at the workshop (5130) and the information is saved (5140).

[0036] As illustrated in Figure 5, the virtual block tracking module (5020) executes the following steps based on the reception of images and information of the train by AEI or other sensors: If it receives valid images (5200), it analyzes (5210) the images, detects the rolling stock and determines the direction of circulation, which is the information that is passed to determine the status of the virtual block (5220). If it receives information (5300) by AEI or other sensors and it is a valid signal (i.e., it comes from rolling stock) and known signal (5310), the information is passed directly to determine the status of the virtual block (5220) and if not, the received signal is discarded (5320).

[0037] The proposed system additionally includes a non-intrusive data extraction procedure from the system's signaling equipment, as shown in Figure 6, wherein on the left side there are presented the field equipment (16), the operator stations connected through an interlock (172) in a technical building (17) to operator stations in the centralized control center (18) wherein a video wall can also be available, while the right side of the Figure shows in more detail the blocks of the operator station equipment (610) and the "video wall" equipment (620). The technical building (17) and the centralized control center (18), which includes centralized traffic control computers or CTC (182), are connected by two communications networks (171, 181). The data extraction procedure from the system's signaling equipment obtains information on the status of train presence detectors, positions of the turnouts and status of the signals, among other data, and as it is non-intrusive, it does not affect the Safety Integrity Level established for signaling facilities. For this data extraction, the system uses a video cable splitter (22) that connects: a) to the video cables (21) that provide the images of the signaling system to displays (20) of the equipment of the operator stations (610) that can be both local and central, as well as control center operator stations; or, b) installed at the location of the video wall equipment (620), in the video cable that connects the CTC or Centralized Traffic Control computer (25) with the controller computer (26) of the video wall (27). This video splitter (22) provides the image that is represented on the displays (20) or on the video wall (27) to the local video capture computer (23) that collects the images and transforms them into a compressed video format to send them through a data transport network (24) for secured communications. A data center receives them through secure communications equipment (32) and the images are collected and analyzed by an image recognition computer (38) for signaling systems, which uses a statistical classifier based on neural networks, previously trained with images of representations of signaling command and control systems, and identifies all the elements of the transmitted image that make up the signaling system. All signaling elements contained in the images are identified and associated with the geographical position they occupy in the monitored railway facility. By using another statistical classifier, the image recognition computer (38) proceeds to analyze the color changes that occur in the pixels that make up each element of the signaling equipment represented in the image, to obtain a security score of the probability of correctly identifying the change in color and status of the signaling element. All signaling element statuses are stored in a database.

[0038] With data resulting from the extraction of information from the signaling equipment, including the position of turnouts, the status of the rolling stock presence detectors and signals from images and sensors, together with the information obtained in the virtual blocks of the specific areas of passage of the railway facility that are not signposted, the system is capable of positioning the rolling stock and generating tracking itineraries thereof.

[0039] Data provided by the proposed system represents an improvement in the monitoring of the railway facility because, based on them, it is possible to: identify and track the rolling stock used by the monitored railway facility, map the position of the rolling stock in real or quasi-real time, calculate the time that each railway vehicle has been in a position in the workshop, the number of operations carried out and provide this information through a Human-Machine Interface or HMI server to users who connect through a data transport network (43) with mobile (44) and / or fixed (45) user devices / terminals for viewing and using the application through a graphical interface. The data provided by the system can also be shared with other external systems through a machine-machine interface or M2M so that it can be reused.

Examples

Embodiment Construction

[0022]Below, with the help of the aforementioned figures, there is provided a detailed explanation of an example of a preferred embodiment of the object of the present invention.

[0023]A system is proposed for the identification of rolling stock at railway facilities such as the one shown in Figure 1. Typically, railway workshops and facilities consist of one or several entry / exit tracks, which contain signposted areas (6) and non-signposted areas (7), in which there are different types of built infrastructure (5), such as covered buildings or warehouses, wherein different rolling stock maintenance tasks are commonly carried out. The proposed system contemplates the use of at least one identification post (1) in the entry and exit areas, which is preferably low in height, and one or more posts for the generation of virtual blocks (2), which are used to monitor rolling stock in non-signposted areas (7). And depending on the track configuration of the railway facility, mixed posts (4) ...

Claims

1. A method to monitor railway facilities with signposted areas (6) and non-signposted areas (7) where rolling stock circulates, characterized by comprising: - capturing images by at least one electronic imaging device (8) and signals by at least one set of sensors (10), the electronic imaging device (8) and the set of sensors (10) fixed on a fixing post (12) installed within a railway facility in a known location selected from: i) an entry / exit area belonging to the signposted areas (6) of the railway facility wherein the fixing post (12) is a post for the identification (1) of rolling stock, or ii) a passage area belonging to the non-signposted areas (7) of the railway facility wherein the fixing post (12) is a post for the generation of virtual blocks (2) for the tracking of rolling stock, and iii) in a mixed post (4) for the identification of rolling stock at entries / exits and the generation of virtual blocks for tracking rolling stock; - locally processing at the railway facility, by a local computer (28), the signals captured by each set of sensors (10) and the images captured by each electronic imaging device (8); - sending the information processed by the local computer (28) from the railway facility to a first data center (400) that receives image information and signal information; - recursively detecting in the received image information, by an image identification computer (33) in the first data center (400), the presence of rolling stock at the known location of the electronic imaging device (8) to which the received image information belongs; - when the presence of rolling stock has been detected, identifying the detected rolling stock, by an identification computer (37) at the first data center (400), through an identification module (4000) configured to perform a comparison of the received signal information with historical signals information previously analyzed by the identification computer (37), associated with a unique identifier (13) of the rolling stock and stored in a database according to the unique identifier (13); - extracting from the received image information, by an image recognition computer (38) at the first data center (400), information of signaling elements, wherein: - if the received image information comes from signposted areas (6), the signaling element information includes status information of rolling stock presence detectors, positions of the turnouts and status of the signals generated by a signaling monitoring module (5010), - if the received image information comes from non-signposted areas (7), the signaling element information comprises state information of the virtual block determined by a virtual block tracking module (5020); - associating the information of the signaling elements extracted by the image recognition computer (38) with the rolling stock identified by the identification computer (37) and with a geographical position at the railway facility, for tracking the rolling stock identified by a monitoring computer (39) in the first data center (400).

2. The method according to claim 1, characterized in that recursively detecting the presence of rolling stock by the image identification computer (33) comprises performing, for each image of the received image information, a comparison of distribution histograms of color to measure a color difference and, using a first statistical classifier, obtain a security score of the probability of identifying the presence of rolling stock and the direction of circulation of the rolling stock.

3. The method according to claim 2, characterized in that recursively detecting the presence of rolling stock by the image identification computer (33) further comprises obtaining, by means of a second statistical classifier based on neural networks previously trained with image samples of rolling stock, a security score of the probability of identifying the composition, series, operator and manufacturer of the rolling stock identified by the first statistical classifier.

4. The method according to any of the preceding claims, characterized by further comprising performing by the image identification computer (33) a detection of regions with alphanumeric characters in each image of the received image information, wherein the alphanumeric characters contain a numerical code of a unique identifier (13) of the rolling stock and are sent to an OCR computer (35) for recognition of the alphanumeric characters using OCR.

5. The method according to claim 4, characterized by further comprising validating the alphanumeric characters recognized by the OCR computer (35) using the control digits of the UIC codes assigned for the unambiguous identification of all rolling stock and verifying that the identifier unique (13) of the rolling stock corresponds to one of the valid UIC codes.

6. The method according to any of the preceding claims, characterized by further comprising analyzing, by an AEI computer (34) at the first data center (400), the signal information received from each set of sensors (10) comparing the received signal information with historical signals information previously analyzed by the AEI computer (34) and stored in a database, and obtaining, through a third statistical classifier, a value that is compared with a security score of the probability of identifying the received signal.

7. A computer program product comprising instructions that, when the program is executed by a computer, cause the computer to carry out the method of claims 1-6.

8. A computer-readable medium comprising instructions that, when executed by a computer, cause the computer to carry out the method of claims 1-6.

9. A system to monitor railway facilities with signposted areas (6) and non-signposted areas (7) through which rolling stock circulates, characterized by comprising: - at least one fixing post (12) installed within a railway facility at a known location that is selected from: i) an entry / exit area belonging to the signposted areas (6) of the railway facility wherein the fixing post (12) is a post for the identification (1) of rolling stock, or ii) a passage area belonging to the non-signposted areas (7) of the railway facility wherein the fixing post (12) is a post for the generation of virtual blocks (2) for the tracking of rolling stock, and iii) in a mixed post (4) for the identification of rolling stock at entries / exits and the generation of virtual blocks for tracking rolling stock; - at least one electronic imaging device (8) configured to capture images and at least one set of sensors (10) configured to capture signals, the electronic imaging device (8) and the set of sensors (10) fixed on the fixing post (12); - the fixing post (12) comprising a local computer (28) configured to locally process the signals captured by each set of sensors (10) and the images captured by each electronic imaging device (8); - a first data center (400) that receives image information and signal information sent by the local computer (28) from the railway facility, the first data center (400) comprising: - an image identification computer (33) configured to recursively detect in the received image information the presence of rolling stock at the known location of the electronic imaging device (8) to which the received image information belongs; - an identification computer (37) configured to identify the detected rolling stock by means of an identification module (4000) configured to perform, when the presence of rolling stock has been detected, a comparison of the received signal information with historical signal information previously analyzed by the identification computer (37), associated with a unique identifier (13) of the rolling stock and stored in a database according to the unique identifier (13); - an image recognition computer (38) configured to extract information about signaling elements from the received image information, wherein: - if the received image information comes from signposted areas (6), the signaling element information includes status information of rolling stock presence detectors, positions of the turnouts and status of the signals generated by a signaling monitoring module (5010), - if the received image information comes from non- signposted areas (7), the signaling element information comprises status information of the virtual block determined by a virtual block tracking module (5020); - a tracking computer (39) configured to track the identified rolling stock, associating the signaling element information extracted by the image recognition computer (38) with the rolling stock identified by the identification computer (37) and with a geographical position in the railway facility.

10. The system according to claim 9, characterized in that the image identification computer (33) is further configured to perform, for each image of the received image information, a comparison of color distribution histograms to measure a difference of color and, using a first statistical classifier, obtain a security score of the probability of identifying the presence of rolling stock and the direction of circulation of the rolling stock.

11. The system according to claim 10, characterized in that the image identification computer (33) is further configured to obtain, by means of a second statistical classifier based on neural networks previously trained with samples of images of rolling stock, a security score of the probability of identifying the composition, series, operator and manufacturer of the rolling stock identified by the first statistical classifier.

12. The system according to any of claims 9-11, characterized by further comprising an OCR computer (35) at the first data center (400), and wherein the image identification computer (33) is further configured to perform a detection of regions with alphanumeric characters in each image of the received image information, wherein the alphanumeric characters contain a numerical code of a unique identifier (13) of the rolling stock, and send them to the OCR computer (35) for the recognition of alphanumeric characters using OCR.

13. The system according to claim 12, characterized in that the identification computer (37) is configured to further validate the alphanumeric characters recognized by the OCR computer (35) using the check digits of the UIC codes assigned for the unambiguous identification of all rolling stock and verify that the unique identifier (13) of the rolling stock corresponds to one of the valid UIC codes.

14. The system according to any of claims 9-13, characterized by further comprising an AEI computer (34) at the first data center (400), configured to analyze the signal information received from each set of sensors (10 ) comparing the received signal information with historical signals information previously analyzed by the AEI computer (34) and stored in a database, and obtain, by means of a third statistical classifier, a value that is compared with a security score of the probability of identifying the received signal.

Images