Method and device for detecting passengers entering and / or exiting a public transport vehicle

Abstract

A method and apparatus (1) for detecting passengers getting on and / or off a means of transportation (T), comprising: a) an acquisition step (P1), in which at least one image having at least one portion displaying at least one portion of a door of the means of transportation (T) is acquired; b) a processing step (P2), in which, based on the at least one image, it is determined whether the door is in a closed state or an open state; and a detection step (P3), in which, if the door is in the open state, at least one passenger getting on or off the means of transportation (T) is detected.

Description

[Technical Field]

[0001] The present invention relates to a method and a device for detecting passengers entering and / or exiting a means of transport, in particular a means of public transport, where the presence of moving passengers must be detected for safety reasons and / or for the correct application of fares.

[0002] As is known, public transport vehicles such as trains, buses or tram cars are environments in which pathogens such as influenza viruses, SARS-CoV-2 viruses, etc. can easily spread as passengers frequently gather together, thereby facilitating the transmission of pathogens by airborne transmission.

[0003] To limit this phenomenon, it is necessary to employ personnel (e.g. ticket inspectors, drivers if the transport is not moving, etc.) to monitor the congestion of public transport to ensure passengers are spaced far enough apart.

[0004] However, this task can also be performed by an automatic system, such as a counting system, which processes and counts passengers getting on / off the means of transport when the means of transport opens or closes its doors, thereby providing information about how many passengers are currently traveling on the means of transport or staying in a certain part of it (e.g., a floor of a railway passenger car).

[0005] According to the state of the art, for such measurement systems to work, they must be connected to one or more of the vehicle's on-board safety-critical systems, so that information acquired and verified by one or more safety-critical systems, e.g., the open or closed state of a door, can be shared. This connection also requires that the measurement system be certified as a safety-critical system, so that the installation of measurement systems in a fleet of vehicles already in circulation becomes very costly (in terms of time and money); it should be pointed out that, in fact, these safety-critical systems to which the measurement system is connected would have to be re-certified. Summary of the Invention

[0006] The present invention aims to solve these and other problems by providing a method for detecting passengers entering and / or exiting a vehicle.

[0007] The present invention further aims to solve these and other problems by providing a device for detecting passengers entering and / or exiting a means of transportation.

[0008] The basic idea of ​​the present invention is to determine, via processing means, whether the door of the means of transport is in a closed or open state based on at least one image including at least one portion displaying at least one portion of the door of the means of transport, and to detect, via detection means (e.g., a people counter sensor, etc.), the passage of at least one passenger getting on or off the means of transport when the door is in an open state.

[0009] In this manner, passengers can be effectively counted without the need for a measurement system to be connected to safety-critical systems of the means of transport, such as, for example, rail carriages, tram cars, buses, etc. Thus, there is no need to certify the measurement system and / or re-certify the vehicle's on-board safety-critical systems.

[0010] Further advantageous features of the invention are set out in the accompanying claims. [Brief explanation of the drawings]

[0011] These features and further advantages of the present invention will become more apparent in the light of the following description of preferred embodiments thereof, as illustrated in the accompanying drawings, which are given herein merely as non-limiting examples. [Figure 1] 1 shows a block diagram representing a device for detecting passengers entering and / or exiting a means of transportation according to the invention. [Figure 2] 1 shows a flow chart illustrating a method for detecting passengers entering and / or exiting a means of transportation in accordance with the invention. [Figure 3] 1 shows a system including multiple devices according to the invention. DETAILED DESCRIPTION OF THE INVENTION

[0012] In this description, any reference to "one embodiment" will indicate that a particular configuration, structure, or feature is included in at least one embodiment of the present invention. Thus, phrases such as "in one embodiment" that may appear in different parts of this description do not necessarily refer to the same embodiment. Moreover, any particular configuration, structure, or feature may be combined as deemed appropriate in one or more embodiments. Therefore, the following references are used for brevity only and do not limit the scope or extension of various embodiments.

[0013] With reference to FIG. 1 , the following describes a device for detecting passengers entering and / or leaving a means of transport according to the invention, the means of transport being preferably public transport (e.g. railway carriages, tram cars, buses, etc.); the device 1 preferably comprises the following components: - processing means 11, e.g. one or more CPUs, GPUs, FPGAs etc., preferably via software and / or hardware means, implementing the trained state classifier by inputting thereto one or more training images showing one or more doors of the means of transport in either an open or closed state, and forcing the output of data defining the actual state of the doors represented in the training images; - a volatile memory means 12, for example a random access memory RAM, in signal communication with the processing means 11. When the device 1 is in an operational state, the volatile memory means 12 stores at least one image including at least one portion representing at least one portion of a door of the means of transport; - non-volatile memory means 13, preferably one or more magnetic disks (hard disks) or flash memories or other types of memory, in signal communication with the processing means 11 and the volatile memory means 12, wherein the non-volatile memory means 13 comprises at least one set of instructions implementing a method for detecting passengers entering and / or exiting a means of transport according to the invention. Furthermore, the non-volatile memory means 13 may also comprise information enabling the configuration and / or operation of the classifier (e.g. metadata, values ​​representing weights used by the classifier to detect the door status, etc.); - Input and / or Output (I / O) means 14 that can be connected to the acquisition means (e.g., a CMOS or CCD image sensor, etc.) and the detection means (e.g., a people counter sensor, etc.) and that are configured to acquire at least one image comprising at least one portion representing at least one portion of the door. Such input / output means 14 may include, for example, USB™, IEEE1394, RS232, RS485, IEEE1284 adapters, etc.; - transmitting means 15, preferably wired and / or wireless (for example a network interface such as 803.2 (also known as Ethernet) or 802.11 (also known as WiFi) or 802.16 (also known as WiMax)), or a data network interface such as GSM / GPRS / UMTS / LTE / 5G, TETRA, LoRa, XBee, ZigBee, etc., configured to transmit a signal in which at least one monitoring datum is coded to the monitoring computer, as further explained below; a communication bus 17 allowing information to be exchanged between the processing means 11, the volatile memory means 12, the non-volatile memory means 13, the input / output means 14 and the transmission means 15;

[0014] Referring also to FIG. 2, when the device 1 is in an operational state, said device 1 executes a set of instructions implementing a method according to the invention, said method comprising the following steps: - an acquisition stage P1, in which, via the input means 14, at least one image is acquired having at least one portion representing at least one portion of a door of the means of transport; a processing step P2, in which it is determined, via the processing means 11, based on said at least one image, whether said door is in a closed or open state; - a detection stage P3, in which, when the door is in said open state, at least one passenger getting on or off said means of transport is detected via said input means 14, for example by activating a measurement system which detects passengers getting on and / or off said means of transport and / or by acquiring a signal generated by said measurement system.

[0015] In this manner, passengers can be effectively counted without the need for the measurement system to be connected to safety-critical systems of the means of transport, such as rail carriages, tram cars, buses, etc. Furthermore, many false detections of passengers boarding and / or exiting the vehicle are avoided, as detection only occurs when the doors are open, i.e., when passengers are actually able to board and exit the vehicle. Thus, there is no need to certify the measurement system and / or re-certify the vehicle's on-board safety-critical systems.

[0016] However, it should be pointed out that steps P1, P2, and P3 of the method according to the invention may also be performed in separate devices (e.g., each one of steps P1-P3 may be performed by a single device capable of communicating over a communications network) without departing from the teachings of the present invention.

[0017] However, it should also be pointed out that the acquisition means and / or detection means may either be included in the device 1 according to the invention or may be provided as separate devices without departing from the teachings of the present invention.

[0018] An image classifier, such as a Support Vector Machine (SVM), a Convolutional Neural Network (CNN), or a Fully Connected Deep Neural Network (FC DNN), is input with a large number of points (pixels), including portions of the image that represent the portions of the vehicle doors whose open or closed state is to be detected.

[0019] According to the invention, the process preferably also comprises a training phase, and before executing processing phase P2, the classifier is trained by inputting several training images to it, so that the algorithm can identify the desired state according to the images received during the training phase. More particularly, the training step is preferably, but not necessarily, performed by using a workstation including a CPU having higher computing power than the processing means 11, the CPU preferably being configured to execute a set of instructions implementing the training algorithm.

[0020] More particularly, the method according to the invention preferably also comprises a training stage, in which, before carrying out the processing stage P2, the neural network is trained in such a way that it outputs the desired (second) state data, when at least one training image is input to the neural network. This allows the device 1 to be flexibly adapted in the course of its service life, for example so that the images show the doors of the means of transport from different perspectives should it be moved from one carriage to another.

[0021] In this manner, passengers can be effectively counted without having to connect device 1 to the vehicle's safety-critical systems, thus eliminating the need to certify the measurement system and / or re-certify the vehicle's on-board safety-critical systems.

[0022] With reference to FIG. 3, the following describes a system S for detecting passengers getting on and / or off a means of transport, for example a train T.

[0023] System S consists of the following elements: - a plurality of devices 1 according to the invention, each one being arranged in the vicinity of at least one door of the means of transport; - a monitoring unit 2 (for example a server, an industrial PC, etc.) configured to receive from said device 1 data relating to the open or closed state of a door D included in the means of transport T; at least one network device (e.g., an Ethernet switch, router, etc.) that allows the device 1 and the monitoring unit 2 to communicate with each other; Equipped with.

[0024] As mentioned above, the device 1 preferably comprises acquisition means and has a shape that allows it to be placed above the door of the means of transport.

[0025] This advantageously allows the acquisition means to acquire images representing the door over its length, and preferably also its striker, to ensure accurate detection of the open or closed state of the door; in particular, it advantageously allows to detect if the door is not fully closed (e.g. because it has been tilted by a broken or loose component such as a screw, nut, etc.), thereby reducing the probability that a false closed or false open state may be detected.

[0026] In this manner, passengers can be effectively counted without having to connect device 1 to the vehicle's safety-critical systems, thus eliminating the need to certify the measurement system and / or re-certify the vehicle's on-board safety-critical systems.

[0027] More particularly, the device 1 has a shape that allows it to be arranged in the housing of a lamp, preferably a lamp adapted to illuminate the passageway while passengers are getting on and off the vehicle.

[0028] In this way, device 1 can advantageously be installed in a means of transport T that was not originally pre-arranged for said device 1, and passengers can be effectively counted without device 1 needing to be connected to the means of transport's safety-critical systems, thereby avoiding the need to certify the measurement system.

[0029] In the above combination, the acquisition means preferably includes a wide-angle lens.

[0030] This type of lens, together with the fact that device 1 is positioned above the door, allows the acquisition means to acquire images surrounding a larger portion of the door, thereby further reducing the probability of detecting a false close or false open state, and effectively counts occupants without device 1 needing to be connected to the vehicle's safety-critical systems, thus eliminating the need to certify the measurement system and / or re-certify the vehicle's on-board safety-critical systems.

[0031] Preferably, the monitoring unit 2 is also configured to capture the passage of at least one passenger getting on or off the means of transport T through a particular door of the means of transport T, the passage being detected by the detection means when the door monitoring device 1 detects that the door is in an open state. This allows the monitoring unit 2 to determine the total number of people on the means of transport T on which it is installed and, optionally, to tell how many people are present in a particular section of the means of transport T, for example a particular carriage and / or class of carriage, thereby preventing passengers from crowding together and / or to detect the presence of any ticketless passengers by comparing the total number of people determined by the monitoring unit 2 with the booking data.

[0032] To detect such passage, the system S may comprise detection means, such as strain gauges, capacitance sensors, radar sensors, etc., which are preferably placed near the door or at a point that a user must cross when passing through the door to board or disembark from a means of transport, making it possible to detect the passage of one or more persons.

[0033] Alternatively, or in combination with the above, the device 1 preferably comprises detection means, for example an FPGA or the like, configured to detect the passage of at least one passenger (during a detection phase P3) based on at least one image acquired by the acquisition means. It should be pointed out that the detection means may be included in the processing means 11 or may be identical to the processing means 11.

[0034] This makes it even easier to install the device 1 in vehicles that were not originally pre-configured for it, thereby achieving effective counting of passengers without the need to also install (dedicated) sensors that allow for detecting the passage of people and without the need to connect the device 1 to the safety-critical systems of the vehicle. Thus, there is no need to certify the measurement system and / or re-certify the vehicle's on-board safety-critical systems.

[0035] It should be emphasized that this feature and the combined use of a wide-angle lens means that the image acquired by the acquisition means shows, in addition to the door and its striker, also one or more passengers about to board / exit the means of transport T, thereby allowing the detection of at least one passenger boarding / exiting the means of transport; in fact, the wide-angle lens is particularly advantageous because it keeps the passengers in the image for a longer period of time, thereby increasing the probability that at least one passenger will be detected.

[0036] In addition to the above, the detection means may also be configured to detect (during detection phase P3) the direction of movement of at least one passenger based on at least one image acquired by the acquisition means, i.e. to detect whether at least one passenger is getting on or off the means of transport (T) (i.e. whether the detected passenger is entering or exiting the means of transport).

[0037] In addition to facilitating the installation of the device 1 in the means of transport, this also results in passengers being counted more effectively and in a similar manner, without the need to install two or more (dedicated) sensors that make it possible to detect the passage of a person, and without the need to connect the device 1 to the safety-critical systems of the means of transport. Thus, there is no need to certify the measurement system and / or re-certify the vehicle's on-board safety-critical systems.

[0038] Of course, the examples described so far are subject to numerous variations.

[0039] Although some of the possible variations of the invention have been described above, it will be apparent to those skilled in the art that other embodiments may in fact be implemented and that some elements may be replaced by other technically equivalent elements. The present invention is therefore not limited to the illustrative examples set forth above, but may be subject to various modifications, improvements or substitutions of equivalent parts and elements without departing from the basic inventive idea, as set forth in the following claims. (Other possible items) (Item 1) 1. A method for detecting passengers entering and / or exiting a means of transportation (T), comprising: - an acquisition stage (P1), in which, via the input means (14), at least one image is acquired having at least one portion representing at least one portion of a door of said means of transport (T); - a processing stage (P2), in which it is determined, via processing means (11), based on said at least one image whether said door is in a closed or open state; a detection step (P3), in which, when the door is in the open state, at least one passenger entering or exiting the means of transport (T) is detected via the input means (14); A method for providing (Item 2) 2. The method according to claim 1, wherein during the detection step (P3), the passage of the at least one passenger is detected based on the at least one image acquired during the acquisition step (P1). (Item 3) 3. The method according to claim 2, wherein during the detection step (P3), a direction of movement of the at least one occupant is also detected based on the at least one image acquired during the acquisition step (P1). (Item 4) A device (1) for detecting passengers entering and / or leaving a means of transport (T), - input means (14), acquisition means for acquiring at least one image having at least one portion representing at least one portion of a door of said means of transportation (T); detection means for detecting passengers getting on and / or off said means of transport (T); input means (14) that can be connected to - processing means (11), obtaining said at least one image via said input means (14); determining whether the door is in a closed or open state based on the at least one image; detecting, via the input means (14), at least one passenger getting on or off the means of transport (T) when the door is in the open state; Processing means (11) configured to A device (1) comprising: (Item 5) 5. The device (1) according to item 4, wherein the device (1) comprises the acquisition means. (Item 6) 6. The device (1) according to item 5, wherein the device (1) has a shape that allows it to be placed above the door of the means of transportation (T). (Item 7) 7. The device (1) according to item 6, wherein the device (1) has a shape that allows it to be positioned within a housing of a lamp. (Item 8) 8. The device (1) according to any one of items 4 to 7, wherein the acquisition means comprises a wide-angle lens. (Item 9) 9. The device (1) according to any one of items 4 to 8, comprising a detection means configured to detect the at least one passenger getting on or off the means of transportation (T) based on the at least one image acquired by the acquisition means. (Item 10) 10. The device (1) according to item 9, wherein the detection means is also configured to sense whether the at least one detected passenger is getting on or off the means of transportation (T). (Item 11) 4. A computer program product that can be loaded into the memory of an electronic computer and that comprises portions of software code that cause said electronic computer to carry out the steps of the method according to any one of items 1 to 3.

Claims

1. 1. A method for detecting passengers entering and / or exiting a means of transportation, comprising: an acquisition step, in which, via input means, at least one image is acquired having at least one portion representing at least one portion of a door of said vehicle; a processing step, in which via processing means it is determined based on said at least one image whether said door is in a closed or open state; a detection step, in which, when the door is in the open state, at least one passenger entering or exiting the means of transport is detected via the input means; A method comprising:

2. The method of claim 1 , wherein during the detecting step, the passage of the at least one occupant is detected based on the at least one image acquired during the acquiring step.

3. 3. The method of claim 2, wherein during the detecting step, a direction of movement of the at least one occupant is also detected based on the at least one image acquired during the acquiring step.

4. 1. A device for detecting passengers entering and / or exiting a means of transportation, comprising: - an input means, - acquisition means for acquiring at least one image having at least one portion representing at least one portion of a door of said vehicle; - detection means for detecting passengers getting on and / or off said means of transport; input means which can be connected to processing means, acquiring said at least one image via said input means; determining whether the door is in a closed or open state based on the at least one image; when the door is in the open position, detecting via the input means at least one passenger getting on or off the vehicle; a processing means configured to A device comprising:

5. The device of claim 4 , wherein the device comprises the obtaining means.

6. 6. The device of claim 5, wherein the device has a shape that allows it to be placed over a door of the vehicle.

7. The device of claim 6 , wherein the device has a shape that allows it to be positioned within a lamp housing.

8. 8. The device of claim 4, wherein the acquisition means comprises a wide-angle lens.

9. 8. The device of claim 4, further comprising a detection means configured to detect the at least one passenger getting on or off the means of transportation based on the at least one image acquired by the acquisition means.

10. 10. The device of claim 9, wherein the detection means is also configured to sense whether the at least one detected occupant is boarding or exiting the vehicle.

11. A computer program product that can be loaded into the memory of an electronic computer and that comprises portions of software code that cause said electronic computer to carry out the steps of the method according to any one of claims 1 to 3.

Images