ACCESS CONTROL DEVICE

Abstract

Access control device (40), for preventing access by unauthorized individuals in an identification passage (P) of a premises, comprising: a device for receiving identifying data of an individual; a processing device (20) for continuous detection and counting, connected to the receiving device for identifying data, to obtain an access authorization signal (AA); and characterized in that it further comprises: a plurality of cameras connected to the processing device (20), each camera (11) of the plurality of cameras being configured to be oriented towards a lower part of the passage (P) to monitor footwear shapes (FW) in images of the lower part of the passage (P).

Description

ACCESS CONTROL DEVICE TECHNICAL SECTOR The technical field of the present invention is access control devices configured to prevent the authorized passage of more than one individual at a time. OBJECT OF THE INVENTION The present invention presents an access control device configured to prevent unauthorized individuals from entering a premises. STATE OF THE ART AND TECHNICAL PROBLEM TO BE SOLVED Many access control devices are known in the state of the art to prevent the authorized passage of more than one individual to a premises at a time. These control devices comprise an identification data receiver device, configured to receive identification data of an individual and to be placed in an identification passage, for entry and / or exit to the premises, the passage having a proximal end and a distal end. The individual's identification data may be read from an identification document, such as a national identity card, tax identification number, passport, authorization card, or from a fingerprint, or from facial recognition, among many others. These control devices also comprise a processing device connected to the identification data receiving device. The processing device is configured to send an output signal based on the received individual identification data and to obtain an access authorization signal, either granted or denied, from the individual based on the output signal. These control devices also include cameras to monitor the passage. Several factors contribute to the success of an access control device; for example, it is desirable that they be as unobtrusive as possible while maintaining security, with a view to shortening interaction times with the access control device. It would also be desirable for the access control device to be autonomous, that is, an access control solution without the need for permanent human assistance, where human action is relegated to those cases where an intrusion or unauthorized access is detected automatically. Secure automation of access without human control is very advantageous, especially in premises where relatively frequent entries and exits occur, for example, for eating, smoking, or for any other reason, where all this implies an extra workload and additional costs associated with these entries and exits of the premises. The technical problem to be solved is to reliably and robustly prevent malicious access by individuals who do not have authorized access to a premises, in an access control solution without the need for permanent human assistance, maintaining access security at all times. EXPLANATION OF THE INVENTION To overcome the drawbacks described in the previous section, the present invention relates to an access control device for preventing unauthorized individuals from entering a premises. The access control device is configured to prevent more than one authorized individual from entering the premises at a time. The access control device comprises an identification data receiver. The identification data receiver is configured to receive identification data from an individual and is installed in an identification passageway, either for entry and / or exit to the premises, the passageway having a proximal end and a distal end. In other words, the access control device may be located at an entrance and / or an entrance to the premises, configured to control access in an entry direction to the premises, and / or in an exit direction from the premises, according to needs. Thus, the terms "proximal" and "distal" throughout the present invention are labels. The control device is configured to be placed at an entrance and / or an exit of the enclosure, and to function as an access control device for the interior of the enclosure and / or the exterior of the enclosure, interchangeably. The individual's identification data may be read from an identification document, such as a national identity card, tax identification number, passport, authorization card, or from a fingerprint, or from facial recognition, among many others. The term "enclosure" should be understood in a generic way and will be understood throughout this document as a space contained within certain limits, generally closed, although it may also be open, and includes cases of an establishment, for example, a gaming hall or betting shop, a casino, a space, a room, a building, for example cultural or governmental, a sports venue, an educational venue, a theater, a cinema, facilities or premises of a company, a space for events, for example, cultural, ceremonial, musical events, among many other cases. Likewise, an access control device comprises a processing device connected to a device that receives identification data. The processing device is configured to send an output signal based on the received individual identification data and to obtain an access authorization signal, either granted or denied, from the individual based on that output signal. Throughout this document, the term "connected" includes connecting directly, or indirectly through one or more intermediary elements, two elements by means of a wireless connection (e.g., Wi-Fi or Bluetooth) or by means of control signal cables, which may be, for example, communication buses, or multi-wire or single-wire cables, fiber optic cables or Ethernet, among others. The access control device also comprises a plurality of cameras connected to the processing device. Each camera in the plurality is configured to face the lower part of the passageway to monitor images of the lower part of the passageway. Throughout the present invention, the term "camera" should be understood as synonymous with video camera or an electronic device capable of capturing video that includes monitoring sequential images at a certain rate, for example, between 30 and 60 frames per second, for example, of the charge-coupled device (CCD) type or of the complementary metal-oxide semiconductor (CMOS) type, among other options. Additionally, in the access control device, the processing device is configured to continuously detect shoe shapes of one or more individuals in the monitored images and to continuously count the detected shoe shapes. The access control device is configured to continuously detect and count such footwear shapes in real time, i.e., at a repetition rate on the order of fractions of a second, for example, the repetition rate or cycle or frequency of analysis of such detection and counting is between 30 and 60 times per second. Throughout this report, the term "footwear" should be understood in a generic sense and refers to garments to cover the foot, including cases of a shoe, a slipper, a boot, a clog, a sandal, a flip-flop, among many other options. Likewise, the processing device is configured to activate an alarm signal if the number of shoe shapes detected and counted by the processing device in the images monitored by any one camera of the plurality of cameras is equal to or greater than four. Additionally or alternatively to the preceding paragraph, the processing device is configured to activate an alarm signal if the number of shoe shapes detected and counted by the processing device in the images monitored by at least two cameras of the plurality of cameras is equal to or greater than three. In the present invention, the term "processor device" included in the access control device is a computer programmable unit, and includes cases of "controller", "microprocessor" or "control unit", "central processing unit" or its acronym "CPU", "computer", "microcomputer", "computer", "microcomputer", among others. In the present invention, the term "processing device" as understood in the access control device includes the case of a processing device with a single processing element; and also includes the case where the processing device is distributed across a plurality of processing elements, for example, where one processing element acts as the master of the other slave processors, among other options. Thus, for example, the present invention covers the case where, for example, the processing device comprises a first and a second processing element, wherein the first processing element is configured to send the output signal from the received individual identification data, and to obtain an access authorization signal granted or denied from the individual from the output signal; and wherein the second processing element is configured to continuously detect and count shoe shapes of one or more individuals in the monitored images, and to activate the alarm signal. Advantageously, the present invention proposes a differential solution to overcome the drawbacks of the technical problem posed, as well as others, through the plurality of cameras and the processing device with a differential functionality, to reliably and robustly undertake an autonomous access control that detects intrusions. Unlike the state of the art, the multiple cameras and the processing device are jointly configured to allow the activation of an alarm signal, through a specific function of monitoring, detection and counting of specifically footwear and intentional number, which makes the access control device effective and robust, even in a challenging case where two individuals try to bypass the access control device, a first individual with identification and another without identification located behind the first individual at all times. The plurality of cameras, conveniently oriented towards a central region of the passage, are arranged in such a way that the plurality of cameras leaves no blind spot for monitoring, and subsequent detection and counting of the footwear in the processing device. Advantageously, the specific number of detected shoe shapes that trigger the alarm - four or more in any one camera and / or three in at least two cameras - has been intentionally selected to avoid false alarm signal activations, resulting in increased reliability and autonomy of the solution proposed here. On the one hand, the activation of the alarm with four shoes in any one of the camera pluralities is an unequivocal indication that more than one individual at a time is present in the passage. On the other hand, the activation of the alarm signal with three shoe shapes counted in at least two cameras, and not just one, is due to the technical effect that, depending on its orientation relative to each of the cameras, the same shoe may be counted as two shoes by one camera when this camera sees two different portions of the same shoe; however, at least two cameras with three shoe shapes in each is already considered by the invention as a situation susceptible to intrusion and is therefore configured to activate the alarm signal. Optional features of the invention are described in a subsequent section detailing embodiments / examples, together with the advantages / technical effects they provide unless they are easily deducible from the whole of this document by a person skilled in the art. BRIEF DESCRIPTION OF THE DRAWINGS To complement the description of the object of the present invention and to aid in a better understanding of its distinguishing characteristics, this descriptive document includes, as an integral part thereof, a set of drawings, in which the following is represented for illustrative and non-limiting purposes: Fig. 1 shows a side perspective view of the access control device of the present invention according to a first embodiment; Fig. 2 is a front view of Fig. 1; Fig. 3 is the front view of Fig. 1, where some elements housed inside the frame of the access control device have been schematically represented for the sake of greater clarity; Fig. 4 shows a front view of the access control device of the present invention according to a second embodiment; Fig. 5 shows an example of an image monitored by a camera of the plurality of cameras, where two shapes of footwear of an individual have been detected and marked with delimiting rectangles with a first neural network and counted by the processing device; Fig. 6 shows an example of an image monitored by another camera of the plurality of cameras, where four shoe shapes of two individuals have been detected, indicated by delimiting rectangles, with the first neural network and counted by the processing device; Fig. 7a shows a schematic example of a second image monitored by the upper camera, where only one individual shape marked with a delimiting rectangle has been detected and counted, which is a difficult case of intrusion where two individuals try to bypass the access control device, where a first individual with identification tries to stay ahead of another individual without identification at all times; Fig. 7b shows a schematic example of a second image monitored by the intermediate camera, where only one head shape marked with a delimiting rectangle has been detected and counted, in the same difficult intrusion case as Fig. 7a; Fig. 8 shows a schematic plan view of Figs. 1 and 2, where portions of the proximal and distal ends of an access control device frame have been sectioned to show in detail the chambers of the plurality of chambers, and where a partition facing the access control device that partially delimits one side of the passage has been schematically represented; Fig. 9 shows an example of a second (empty) image monitored by an upper camera of the access control device of any one of Figs. 1 and 4, and where a plurality of linear position segments are shown, to virtually divide the passage and thus detect a direction of advance; Fig. 10 includes the elements to virtually divide the passage of Fig. 9, and an individual monitored by the upper camera and detected by the second neural network and a third neural network for detecting the shapes of individuals and their heads, respectively, has also been schematically represented; Figures 11 to 13 show different embodiments of a smart mobile access control device; Fig. 14 shows a flowchart according to a first embodiment; Fig. 15 shows a flowchart according to a second embodiment; Fig. 16 shows a flowchart according to a third embodiment; Figure 17 shows an implementation of detecting a direction of movement in relation to Figure 16; and Fig. 18 shows in detail a flowchart of detecting a direction of advance from Fig. 17. DETAILED EXPLANATION OF MODES OF REALIZATION / EXAMPLES Throughout Figs. 1 to 8 the numerical reference 40 generically designates the access control device (40) to prevent unauthorized individuals from entering a premises. Figures 1 to 4 show that the access control device (40) comprises an identification data receiver. The identification data receiver is configured to receive identification data from an individual. Figures 1 and 8 show that the access control device (40) is configured (provided) to be arranged in an identification passage (P) for entry and / or exit to the premises, the passage (P) having a proximal end (EP) and a distal end (ED). Figures 1 and 8 show that, preferably, the access control device (40) has no physical barrier at the proximal end (PE) nor at the distal end (DE) that prevents access by individuals, reducing access times. Preferably, the device receiving identification data comprises a device that reads an identification document (7), for example, a national identity card, tax identification number, passport and / or authorization card with a barcode or with a chip, or a combination of these. Optionally, the identification data receiving device comprises a fingerprint reader device (8) configured to receive identification data from one or more fingerprints. Optionally, the identification data receiving device may include an optional intermediate camera (14), which in addition to the functionalities described below, may also be configured to capture at least one facial image of the individual for identification purposes. Optionally, the access control device (40) includes a display screen (15) that indicates messages or alarms. Thus, the recipient of identifying data may include any one of the options described above, or a combination of these, among others. Likewise, the access control device (40) comprises a processing device (20), shown in Fig. 3, connected to the identification data receiving device. Optionally, Fig. 3 shows that the processing device (20) comprises a computer. The processing device (20) in Fig. 3 is configured to send an output signal based on the received individual identification data, and to obtain an electrical access authorization (AA) signal, either granted or denied, from the individual based on the output signal. The access authorization (AA) signal is shown in Figs. 14 to 16. Figures 1 to 4 show that the access control device (40) further comprises a plurality of cameras connected to the processing device (20). Each camera (11) of the plurality of cameras shown is configured to be oriented towards a lower part of the passage (P) to monitor images of the lower part of the passage (P), such as, for example, the images shown in Figures 5 and 6. Optionally, in Figs. 1 to 4 and 8, the plurality of chambers comprises two chambers (11). Optionally, in an option not shown, the plurality of cameras comprises the two cameras (11) mentioned above and also a third camera (11) arranged between the two cameras (11). The plurality of chambers (11) may comprise four, five, or a greater number of chambers (11). Additionally, in Figs.1 to 4 and 8, the plurality of chambers (11) comprises one chamber (11) configured to be disposed at the proximal end (EP), and another chamber (11) configured to be positioned at the distal end (ED). Additionally, following Figs. 1 to 4 and 8, each front face with an image sensor array of each camera (11) of the plurality of cameras (11) is oriented forming a first inclination between 30 and 60 degrees in a plan view, and preferably 45 degrees, with respect to a horizontal linear direction (L) of the passage (P). Additionally, as shown in Figs.1 to 4 and 8, each front face with an image sensor array of each camera (11) of the plurality of cameras (11) is oriented forming a second downward tilt, towards a floor plane of the passage (P). Continuing in Figs. 1 to 4 and 8, optionally, a front face with an array of image sensors for each camera (11) of the plurality of cameras is arranged at a height less than 1.20 meters from a horizontal plane of the passage floor (P). Figures 1 to 4 and 8 show that, optionally, the plurality of cameras (11) is arranged at a lower height than the identifying data receiving device. Advantageously, either or both of these height limitations contribute to detecting footwear shapes (FW) more reliably, since as the vertical position of the cameras (11) increases, footwear shapes (FW) are gradually deformed or viewed from angles that cause them to be detected sometimes and not at other times. Preferably, each camera (11) of the plurality of cameras, as well as each of the optional cameras (13, 14) that will be described later, is a video camera, for example, of the CCD or CMOS type, among other options. Optionally, in Fig. 3 the processing device (20) is connected to the plurality of cameras (11) and the optional cameras (13, 14) through an electronic board (28) which may be a controller interface. Additionally, Fig. 3 shows that the processing device (20) is connected to a router (29) to transmit the alarm signal and to optionally transmit the images monitored by the plurality of cameras (11) and / or the second images monitored by a top camera (13) which will be described later. In the access control device (40) of Figs. 1 to 4 and 8, the processing device (20) is configured to continuously detect and count footwear shapes (FW) of one or more individuals in the monitored images, as shown in Figs. 5 and 6. Likewise, the processing device (20) is configured to activate an alarm signal (AS), represented in Figs. 14 to 16, if the number of shoe shapes (FW) detected and counted by the processing device (20) in the images monitored by any camera (11) of the plurality of cameras is equal to or greater than four. Referring now to Figs. 5 and 6, in the image of a camera (11) in Fig. 5 the number of shoe shapes (FW) counted in the processing device (20) is equal to two, while in the image of a camera (11) in Fig. 6 the number of shoe shapes (FW) counted in the processing device (20) is equal to four, activating the processing device (20) the alarm signal (AS) which is electrical, and which, for example, is converted into an acoustic signal in a buzzer and / or is converted into a light signal in a screen or light pilot, for example the display screen (15). Additionally or alternatively to the preceding paragraph, the processing device (20) is configured to activate the alarm signal (AS), represented in Figs. 14 to 16, if the number of shoe shapes (FW) detected and counted by the processing device (20) in the images monitored by at least two cameras (11) of the camera plurality is equal to or greater than three. Referring now to an option not shown in Figs. 5 and 6, if in the image of a camera (11) in Fig. 5 the number of shoe shapes (FW) counted in the processing device (20) were equal to three, and in the image of a camera (11) in Fig. 6 the number of shoe shapes (FW) counted in the processing device (20) were equal to three, the processing device (20) would activate the alarm signal (AS). Preferably, the processing device (20) is configured to continuously detect and continuously count shoe shapes (SW) in the images using a first neural network. In short, a neural network is a model or algorithm, previously trained with positive training images (which do include an object to be detected) and negative training images (which do not include the object to be detected), to detect and position the object within images, in this case, from a video camera. The neural network is configured using deep learning techniques, such as convolutional neural networks, to detect and locate objects using bounding boxes. The squares or rectangles delimiting the shapes of footwear (FW) detected and counted in the images of respective cameras (11) of the plurality of cameras are represented in Figs. 5 and 6. The processing device (20) is configured to continuously detect and count such footwear shapes (FW) in real time, i.e., at a repetition rate on the order of fractions of a second, for example, the repetition rate or cycle or analysis frequency of such detection and counting is between 30 and 60 times every second. Preferably, the first neural network (previously trained with positive and negative images of footwear) is permanently stored in a memory area of ​​the controller device (20), installed locally in a memory of the controller device (20), which allows for more reliable detection and counting of a greater number of images per second. Alternatively, the controller device (20) is configured to receive data from a neural network (previously trained with positive and negative images of footwear) that is not permanently stored in any memory area of ​​the controller device (20). Preferably, Figs. 1 to 4 show that the access control device (40) further comprises an upper camera (13), connected to the processing device (20) and arranged above the plurality of cameras (11) and tilted downwards, to be oriented towards the passage (P) and monitor a second set of images of the passage (P). The processing device (20) is configured to continuously detect and count individual shapes (S) from the one or more individuals monitored in the second images. The processing device (20) is configured to activate an alarm signal (AS) if the number of individual forms (S) detected and counted in the second images monitored by the upper camera (13) is equal to or greater than two, as shown in Figs. 15 and 16. Preferably, the forms of individuals (S) are human bodies. Preferably, the processing device (20) is configured to continuously detect and continuously count the shapes of individuals (S) in the second images by means of a second neural network as shown in Figs. 7a and 10. Figure 7a shows a schematic example of a second image monitored by the upper camera (13), which is a challenging case where two individuals attempt to circumvent the access control device (40). A first individual with identification attempts to stay ahead of another individual without identification at all times. The processing device (20) has only detected and counted one individual shape (S) with delimiting rectangles, so in this and other cases it is necessary to resort to the monitoring, detection, and counting of shoe shapes (FW) jointly performed by the plurality of cameras and the processing device (20) of the present invention. In this preferred option, Figs. 6 and 7a represent that the processing device (20) of Fig. 3 is configured to continuously detect and continuously count shoe shapes (FW) and individual shapes (S) at the same time (in parallel), i.e., simultaneously, as also represented by Fig. 15. Preferably, Figures 1 to 4 show that the access control device (40) further comprises an intermediate camera (14), positioned at a lower height than the upper camera (13) and at a higher height than the plurality of cameras. The intermediate camera (14) is configured to face the passage (P) and to monitor third images of the passage (P) not shown in the figures. The intermediate camera (14) is connected to the processing device (20). The processing device (20) is configured to detect and count head shapes (H) of the one or more individuals monitored in the third images monitored by the intermediate camera (14). The processing device (20) of Figs. 1 to 4 is configured to activate an alarm signal (AS) if the number of head shapes (H) detected and counted in the third images is equal to or greater than two, as shown in Figs. 15 and 16. Figure 7b shows a schematic example of a third image monitored by the intermediate camera (14), which presents the same challenging scenario described for Figure 7a, where two individuals attempt to circumvent the access control device (40). A first individual with identification attempts to stay ahead of another individual without identification at all times. In Figure 7b, the processing device (20) only detected and counted one head shape (H) with a bounding rectangle. Therefore, in this and other cases, it is necessary to implement the monitoring, detection, and counting of shoe shapes (FW) jointly performed by the plurality of cameras and the processing device (20) of the present invention. In this preferred option, Figs. 6, 7a and 7b represent that the processing device (20) of Fig. 3 is configured to continuously detect and continuously count shoe shapes (FW), individual shapes (S), and head shapes (H) all at once, i.e., simultaneously (in parallel), as also represented by Fig. 16. Alternatively, in an option not shown in the figures, the head shapes (H) of the one or more individuals in Figs. 7b and 10 may be detected and counted by the processing device (20) of Figs. 1 to 4 from the second images monitored by the upper camera (13). Preferably, the access control device (40) further comprises a proximal detector device (1, 2, 3) and a distal detector device (4, 5, 6), one on each side of the identifying data receiving device. The proximal detector device (1, 2, 3) and a distal detector device (4, 5, 6) are connected to the processing device (20), optionally via an electronic board (28), configured to detect an individual at the proximal (EP) and distal (ED) ends, respectively. Each proximal and distal detection device comprises a plurality of detection elements, which act as a barrier. Each of the detection elements in the figures is a laser detector, although alternatively it may be a photocell. Alternatively, in an option not shown in the figures, each proximal and distal sensing device comprises only one sensing element, for example, a laser detector or a photocell. In the example in the figures, the proximal detector device (1, 2, 3) comprises a plurality of laser detection elements, specifically three: a lower laser detector (1), an intermediate laser detector (2), and an upper laser detector (3). In the example in the figures, the distal detector device (4, 5, 6) comprises a plurality of laser detection elements, specifically three: a second lower laser detector (4), a second intermediate laser detector (5), and a second upper laser detector (6). Preferably, the access control device (40) is configured for a passage width (P), here defined as the transverse linear distance to the linear direction (L) of the passage (P), between 1.20 and 1.50 meters, by the range of the proximal detector device (1, 2, 3) and the distal detector device (4, 5, 6). In one case, the proximal detector device (1, 2, 3) may be arranged at the proximal end (PE) of the passage (P) associated with an entrance side of the enclosure, and the distal detector device (4, 5, 6) may be arranged at the distal end (DE) of the passage (P) associated with an exit side of the enclosure. In another case, the proximal detector device (1, 2, 3) may be arranged at the proximal end (PE) of the passage (P) associated with an exit side of the enclosure, and the distal detector device (4, 5, 6) may be arranged at the distal end (DE) of the passage (P) associated with an entrance side of the enclosure. The processing device (20) is configured to, upon detection of the individual, initiate monitoring with the plurality of cameras (11) and the continuous detection and counting of footwear shapes (FW). The processing device (20) is configured to send the alarm signal (AS) activated upon detection of an individual in the distal detector device (4, 5, 6) who has not obtained an access authorization (AA) signal granted, i.e., an individual who has obtained an access authorization (AA) signal denied or who has not been identified by the identification data receiving device. Advantageously, the proximal detector device (1, 2, 3) and the distal detector device (4, 5, 6), in combination with the plurality of cameras, provide a unique user identification solution that is minimally intrusive to the individual and requires minimal supervision and / or human intervention, since the alarm signal is only sent (notified) if the individual passes the distal detector device (4, 5, 6) and thereby passes the distal end (ED) of the passage (P), associated with an unauthorized entry to or exit from the premises. Preferably, the access control device (40) further comprises a smart mobile device (34, 35, 36) configured to receive the alarm signal (AS). Advantageously, in the event of an intrusion, a person responsible for access security can perform other tasks elsewhere, without being at the access control device (40), and is reliably and robustly informed of intrusions. Figures 11 to 13 show that, optionally, the smart mobile device (34, 35, 36) may be a smart watch (34), a smart mobile phone (35), or a tablet (36), respectively. On a screen (37) of any of the smart mobile devices of Figs. From 11 to 13, data associated with the alarm signal (AS) sent may be displayed, for example, such data includes a message or characteristic indicative of intrusion and / or a video of the moment the intrusion occurred. Optionally, the alarm signal (AS) is sent by the processing device (20), for example, through the router module (29) among many other options. Optionally, Figs. 1 and 2 show that the access control device (40) further comprises two auxiliary detector devices (16), one arranged on each side of the identification data receiver device, configured to detect improper handling of the identification data receiver device. Advantageously, the two auxiliary detection devices (16) prevent an individual located outside the passage (P) from inserting their hand into the identification data receiving device to identify themselves by fingerprint, while the cameras (11) monitor another unauthorized individual. Additionally, in Figs. 1 and 2, each auxiliary detection device (16) comprises a plurality of detection elements (9), which act as a barrier. Each of the detection elements (9) in the figures is a laser detector, although alternatively it may be a photocell. Alternatively, in an option not shown in the figures, each auxiliary detection device (16) comprises only one detection element (9), for example, a laser detector or a photocell. In Figs. 1 and 2, the distal detector device (4, 5, 6) comprises three detection elements (9), arranged around the identification data receiver device, delimiting it on both sides or flanks of the same. Preferably, in the first embodiment of Figures 1 to 3 of the access control device (40), it further comprises a frame (30). The frame (30) houses the identification data receiving device, the processing device (20), and the plurality of cameras. At one end of the frame (30), associated with the proximal end (PE) of the passage (P), one camera (11) from the plurality of cameras is mounted. At the other end of the frame (30), associated with the distal end (DE) of the passage (P), another camera (11) from the plurality of cameras is mounted. Optionally, continuing in the first embodiment of Figs.1 to 3, the proximal detector device (1, 2, 3), the distal detector device (4, 5, 6) and the upper chamber (13) are mounted on the frame (30). In the second embodiment of the access control device (40) of Fig. 4, it is shown that it comprises a central frame (33), a proximal frame (31) and a distal frame (32). The central frame (33) houses the identification data receiving device, the processing device (20) and the plurality of cameras (11). A chamber (11) from the plurality of chambers is mounted on the proximal frame (31). The proximal frame (31) is configured to be positioned at the proximal end (PE). Another chamber (11) from the plurality of chambers (11) is mounted on the distal frame (32). The distal frame (32) is configured to be positioned at the distal end (ED). Advantageously, this alternative option of a central frame (33), a proximal frame (31) and a distal frame (32), and intentional arrangement of the chambers (11), allows the installation of the access control device (40) in passages (P) that are not merely linear, for example, curved passages, T-shaped or L-shaped passages, among other options. Preferably, in Fig. 4, the camera (11) mounted on the proximal frame (31) and the camera (11) mounted on the distal frame (32) are connected to the processing device (20) either physically by means of a cable or by means of a wireless connection. Referring now to Figs. 14, 15 and 16, they show a flowchart. The access control device (40) of Figs. 1 to 13 comprises software with instructions, preferably stored in a memory of the processor device (20) or alternatively in an external memory such as a USB memory, for, in conjunction with adapted devices included in the access control device (40): (a) monitor images of the lower part of the passage (P) with each camera (11) of the plurality of cameras, monitoring images in one camera (11) at stage (a.1) and monitoring other images in another camera (11) at stage (a.2); (b) continuously detecting and counting with the processing device (20) footwear shapes (FW) of one or more individuals in the images monitored in stage (a); detecting and counting in stage (b.1) footwear shapes (FW) monitored in stage (a.1) and detecting and counting in stage (b.2) footwear shapes (FW) monitored in stage (a.2); and (c) receive identification data of an individual on the identification data receiving device; (d) send an output signal, using the processing device (20), based on the individual identification data received in step (c); (e) obtaining an access authorization (AA) granted or denied signal from the individual at the processing device (20) from the output signal sent in step (d); and (f) activate an alarm signal (AS) with the processing device (20): if, in step (b), the number of footwear shapes (FW) detected and counted by the processing device (20) in the images monitored by any one camera (11) of the plurality of cameras (11) is equal to or greater than four, in any one of the detect and count sub-steps (b.1) and (b.2); and / or If, in step (b), the number of footwear shapes (FW) detected and counted by the processing device (20) in the images monitored by at least two cameras (11) of the camera plurality is equal to or greater than three, i.e., if in both detect and count sub-steps (b.1) and (b.2), the footwear shapes (FW) are equal to or greater than three. Preferably, Fig. 16 illustrates that the software comprises instructions for, in the access control device (40): (g) monitor a second set of images of the passage (P) with the upper camera (13); (h) continuously detect and count with the processing device (20) forms of individuals (S) from the one or more individuals monitored in the second images of stage (g); and (i) activate an alarm signal (AS), with the processing device (20), if in step (h) the number of forms of individuals (S) detected and counted is equal to or greater than two. In this preferred option, Fig. 15 represents that the continuous detection and continuous counting of shoe shapes (FW) and individual shapes (S) is carried out simultaneously (in parallel), i.e., at the same time. Preferably, Figs. 15 to 18 illustrate that the software comprises additional instructions to the above for (j) detecting a direction of advance (SA) of the forms of individuals (S), towards the proximal end (AEP) or towards the distal end (AED), with the processing device (20). In one option, the software comprises instructions to, in the access control device (40), detect a forward direction (SA) of the stage (j) by means of the stages of: (k) virtually divide the passage (P) , with the processing device (20) , defining a plurality of linear segments (T) of invariable position, optionally parallel to each other, and optionally perpendicular to a linear direction (L) of the passage (P) , in the second monitored images of the passage (P); (l) positioning the shapes of individuals (S) in the second images continuously with the processing device (20), through an associated virtual geometric shape, for example, a virtual bounding rectangle and / or a virtual point associated with object detection of a second neural network; and (m) compare the position of the associated geometric virtual shape with respect to the position of a plurality of linear segments to detect the direction of advance of stage (j). In Fig. 10, when a point (G) associated with the box or rectangle delimiting the shape of the individual (S) (or equivalently the shape of the head (H)), is positioned from a perpendicular linear segment (T) to a perpendicular linear segment (T) adjacent to the previous one, according to the linear direction (L) of the passage (P), the direction of advance (SA) can be determined. In one option, Fig. 16 illustrates that the software comprises instructions for, on the access control device (40): (o) monitor a third set of images of the passage (P) with the intermediate camera (14); (p) detect and count head shapes (H) with the processing device (20), of the one or more individuals monitored in the third stage images (o); and (q) activate an alarm signal (AS) with the processing device (20) if in step (o) the number of head shapes (H) detected and counted in the third images is equal to or greater than two. In this option, Fig.16 represents that the continuous detection and continuous counting of head shapes (H), shoe shapes (FW), and individual shapes (S) is carried out simultaneously (in parallel), that is, at the same time. Preferably, the software comprises instructions for the access control device (40) to execute step (b) of continuously detecting and counting shoe shapes (FW) with a first neural network. Preferably, the software comprises instructions for the access control device (40) to perform step (h) continuously detecting and counting shapes of individuals (S) with a second neural network. Preferably, the software comprises instructions for the access control device (40) to execute step (h) continuously detect and count shapes of individuals (S) using a third neural network. Preferably, the first neural network, the second neural network, and the third neural network are permanently stored in respective memory areas of the processor device (20). Preferably, the software comprises instructions for the access control device (40) to further execute the following step: (r) detect an individual in the proximal detector device (1, 2, 3); (s) detect an individual in the distal detector device (4, 5, 6); and instructions to, upon detection of an individual in the proximal detector device (1, 2, 3) of stage (r), initiate stage (a) of monitoring with the plurality of cameras and stage (b) of continuously detecting and counting footwear shapes (FW); and instructions to (u) send the alarm signal (AS) activated by the processing device (20), upon detection of an individual in the distal detector device (4, 5, 6) of stage(s), who has not obtained an access authorization signal (AA) granted in stage (e) of obtaining. Alternatively, in an option not shown in the figures, the detections of stages (r) and (s) may be performed with the upper camera (13) and / or the middle camera (14). Preferably, the software comprises instructions for the access control device (40) to further execute the step of (v) sending the alarm signal (AS) to a smart mobile device (34, 35, 36). Figures 11 to 13 show that, optionally, the smart mobile device (34, 35, 36) may be a smart watch (34), a smart mobile phone (35), or a tablet (36), respectively. The scope of the present invention is defined by the following claims.

Claims

1. An access control device (40) for preventing unauthorized access to an identification passage (P) of a premises, comprising: a device for receiving identifying data of an individual; a processing device (20) for continuous detection and counting, connected to the receiving device for receiving identifying data, to obtain an access authorization signal (AA); and characterized in that it further comprises: a plurality of cameras connected to the processing device (20), each camera (11) of the plurality of cameras being configured to be oriented towards a lower part of the passage (P) to monitor footwear shapes (FW) in images of the lower part of the passage (P). 2.Access control device (40) according to claim 1, further comprising an upper camera (13), connected to the processing device (20) and arranged above the plurality of cameras (11) and inclined downwards, to monitor second images of the passage (P), such that the processing device (20) has the functionality of activating an alarm signal (A) upon detecting and counting two or more individual shapes (S) in the second images. 3.An access control device (40) according to claim 2, further comprising an intermediate camera (14) connected to the processing device (20), arranged at a lower height than the upper camera (13) and a higher height than the plurality of cameras, for monitoring third images of the passage (P); such that the processing device (20) has the functionality of activating an alarm signal (A) upon detecting and counting two or more head shapes (C) in the third images.

4. An access control device (40) according to claim 2 or 3 further comprising: a proximal detector device (1, 2, 3) for detecting an individual at a proximal end (EP) of the passage (P) and a distal detector device (4, 5, 6) for detecting an individual at a distal end (ED) of the passage (P), one on each side of an identifying data receiving device. 5.An access control device (40) according to any one of the preceding claims, further comprising an intelligent mobile device (34, 35, 36) configured to receive the alarm signal (AS).

6. An access control device (40) according to any one of the preceding claims, further comprising two auxiliary detector devices (16), one arranged on each side of the identification data receiver device, for detecting improper handling of the identification data receiver device.

7. An access control device (40) according to any one of the preceding claims, wherein the plurality of cameras (11) is arranged at a lower height than the identification data receiver device. 8.Access control device (40) according to claim 7, wherein a front face with an array of image sensors for each camera (11) of the plurality of cameras (11) is arranged at a height of less than 1.20 meters from a horizontal plane of the passage floor (P).

9. Access control device (40) according to any one of the preceding claims, wherein the plurality of cameras (11) comprises one camera (11) configured to be arranged at the proximal end (EP), and another camera (11) configured to be positioned at the distal end (ED). 10.An access control device (40) according to any one of the preceding claims, further comprising a frame (30) on which the identification data receiving device, the processing device (20), and the plurality of cameras are mounted; and wherein one camera (11) of the plurality of cameras is mounted at one end of the frame (30), and wherein another camera (11) of the plurality of cameras is mounted at the other end of the frame (30).

11. An access control device (40) according to claims 2 and 10, wherein the proximal detector device (1, 2, 3), the distal detector device (4, 5, 6), and the upper camera (13) are mounted on the frame (30). 12.An access control device (40) according to any one of claims 1 to 9, comprising: a central frame (33) on which the identification data receiving device, the processing device (20), and the plurality of cameras (11) are mounted; a proximal frame (31) on which one camera (11) of the plurality of cameras is mounted, configured to be disposed at the proximal end (EP); and a distal frame (32) on which another camera (11) of the plurality of cameras (11) is mounted, configured to be disposed at the distal end (ED).

13. An access control device (40) according to claims 2 and 12, wherein the proximal detector device (1, 2, 3) is mounted on the proximal frame (31), and the distal detector device (4, 5, 6) is mounted on the distal frame (32).