Safety system for an industrial machine

Abstract

Safety system (100) for an industrial machine (50), comprising at least one access point (10) and including: - at least one camera (2) designed to capture an image of a predetermined underlying zone defining a monitored zone (3), wherein the monitored zone (3) comprises at least a first protection zone (3A) located further away from the access point (10) and a second protection zone (3B) located closer to the access point (10), and - a control unit (4) connected to this camera (2), wherein the system (100) is characterized in that the control unit (4) is designed to - to receive from the camera (2) at least one image of a zone below which is representative of a monitored zone (3) captured by the camera (2); - to determine the presence of a person inside the captured image; - to define at least one inclusion region (5) in the captured image that contains the identified person; - to check whether the containment region (5) is at least partially contained within the first protection zone (3A) and / or the second protection zone (3B); - to issue an alarm signal if the containment region (5) is at least partially contained within the second protection zone (3B).

Description

[0001] This invention relates to a safety system for an industrial machine. This invention is applicable in the field of safety and protection systems for industrial machines, particularly in the area of ​​safety systems that use machine vision to detect the entry of persons into the vicinity of the machine's working zone. A first type of available system for detecting the entry of persons into the vicinity of the machine's working zone uses cameras that, for example, employ the principle of interrupting the light reflected by a reflector.

[0002] In other words, such systems, which detect when a person enters the vicinity of the work zone, are equipped with a light source, e.g. an LED or a laser, that emits a beam of light towards the reflector, and a camera or optical sensor that detects the light returning to the system after reflection by the reflector.

[0003] The reflector reflects the light back in the direction from which it came, regardless of the angle of incidence.

[0004] When an object or person enters the area between the light source and the reflector, the light beam is blocked or disrupted. This creates a change that can be detected by the system, which can then identify the presence of the obstacle.

[0005] A second type of system for detecting people near the machine's work zone uses cameras that apply the principle of monitoring changes in background patterns.

[0006] This principle uses a camera to analyze the field of view and detect changes relative to a reference image of the background pattern. The system stores an image of the field of view when no objects or people are present, with this image representing the background pattern. A camera continuously monitors the area of ​​interest and captures new images in real time, and each image is compared to the reference image to identify any differences. If a change is detected, the system assumes that an object or person has entered the monitored area.

[0007] Despite the use of the aforementioned systems, accidents are known to occur involving people working near industrial machinery used to transport loads.

[0008] In the aforementioned sector, there is therefore a need for a system that is able to distinguish between the presence of people and that of loads, for example, near the unloading zone of an industrial machine.

[0009] The object of this invention is in particular to provide a safety system for an industrial machine that makes it possible to increase the reliability in distinguishing between persons and other objects that are in the vicinity of a working zone of the machine.

[0010] The main features of the invention are illustrated by the description of various embodiments, which are shown in the accompanying drawings as purely exemplary and without limitation. These drawings show: - Fig. 1 a schematic view of a component of the security system being described; - Fig. 2 a view of the system being described, which is used in a work zone; - Fig. 3 a top view of a working zone captured by the system being described; - Fig. 4 A schematic overview of a work zone in which the system consists of Fig. 1 is used; - Fig. 5A and Fig. 5B respective image sections received from above, captured by the system being described.

[0011] This description concerns a safety system for an industrial machine 50, which is referred to as system 100 for the sake of simplicity.

[0012] System 100 has at least one access point 10.

[0013] System 100 includes in particular a portal structure 20, which defines the access point 10.

[0014] System 100 includes at least one camera 2.

[0015] Camera 2 is located at the top of portal structure 20.

[0016] According to one aspect of this description, the system 100 includes a conveying device 8 designed to transport loads between a first and a second position.

[0017] The first and second positions occupied by the loads are successive positions arranged along the conveying device 8.

[0018] The conveying device 8 is designed in the form of a conveyor belt that moves the loads inside a work zone.

[0019] In Fig. Figure 4, for example, shows a conveyor belt that runs through a work zone in which an industrial machine 50 is operating. The industrial machine 50 can be any machine designed for handling loads.

[0020] Camera 2 is designed to capture an image of a predetermined underlying zone that defines a monitored zone 3.

[0021] The monitored zone 3 includes the work zone within which the conveying device 8 moves the loads, at least partially.

[0022] According to the example in Fig. In the embodiment shown in Figure 2, the monitored zone 3 comprises at least a first protection zone 3A and a second protection zone 3B.

[0023] The first zone 3A is further away from access point 10, and the second zone 3B is closer to access point 10.

[0024] Preferably, the second protection zone 3B is bounded on two sides by the posts of the portal structure 20.

[0025] In other words, the second zone 3B, being closer to access point 10, is the zone where, in order to avoid accidents, no persons may be present during the working phases of the machines.

[0026] System 100 includes a control unit 4, which is connected to camera 2.

[0027] The at least one camera 2 and the control unit 4 are integrated inside the same enclosure body 1.

[0028] The control unit 4 is designed to receive at least one image from camera 2 of an underlying zone that is representative of a protection zone 3 captured by camera 2.

[0029] The control unit 4 includes an artificial intelligence module 6.

[0030] The artificial intelligence module 6 is defined using a pre-trained classifier.

[0031] In addition, the artificial intelligence module 6 is pre-trained using images that contain a depiction of a person, as well as images that do not contain a depiction of a person.

[0032] According to another aspect of this description, the Artificial Intelligence Module 6 is implemented using a neural network.

[0033] The neural network includes at least one input designed to receive information about an image captured by camera 2.

[0034] The control unit 4 is designed to detect the presence of a person inside the captured image.

[0035] Preferably, the artificial intelligence module 6 is designed to determine, based on the images received during training, whether a person is present in the captured image.

[0036] In other words, since the Artificial Intelligence Module 6 is pre-trained with images containing depictions of a person, as well as with images not containing depictions of a person, it is able to determine whether a figure that can be traced back to a person's body is present in any part of the image it captures.

[0037] The term "pre-trained" refers to the fact that the Artificial Intelligence Module 6 was trained on a set of images to perform a specific task: determining whether a person is present inside the input image. The Artificial Intelligence Module 6 is designed to calculate a numerical value that represents the probability of a person being present inside the captured image.

[0038] According to one aspect of this description, the numerical value calculated by the Artificial Intelligence Module 6 is a probability value between 0 and 1.

[0039] The output of the artificial intelligence module 6 is therefore a probability value between 0 and 1.

[0040] The Artificial Intelligence Module 6 is designed to compare the calculated numerical value (especially the percentage), which provides information about the probability of a person being present in the image, with a predefined threshold.

[0041] The predefined threshold provides information about a percentage value that corresponds to the presence of a person in the image.

[0042] The presence of a person in the image is therefore confirmed if the calculated numerical value exceeds the threshold.

[0043] Preferably, the presence of a person in the image is confirmed if the calculated percentage exceeds the threshold.

[0044] In other words, the presence of a person in the image is confirmed if the calculated percentage value is between the threshold and 1.

[0045] The control unit 4 is designed to define at least one inclusion region 5 in the captured image that contains the detected person.

[0046] The control unit 4 is designed to define at least one inclusion region 5 in the captured image, which contains the person detected by means of the artificial intelligence module 6.

[0047] The artificial intelligence module 6 is designed to define an inclusion region 5 of a predetermined shape in the captured image that contains the detected person.

[0048] It should be noted that this inclusion region is associated with the aforementioned probability that a person is present inside the captured image.

[0049] According to another, in Fig. The aspect of this description shown in 5A is that the predetermined shape of the inclusion region 5 is a rectangular shape.

[0050] In other words, inclusion region 5 surrounds the person from within.

[0051] The containment region 5 has variable dimensions depending on the position and arrangement of the person inside it.

[0052] If the person moves inside the monitored zone 3, in other words, the containment region 5 moves accordingly and also changes its dimensions according to the movements of the person it contains (i.e., according to the width occupied by the person in the image).

[0053] According to another aspect of this description, the inclusion region 5 is representative of the physical space that the person actually occupies in the room.

[0054] In other words, inclusion region 5 is located as close as possible to the boundary defined by the person in the image.

[0055] The inclusion region 5 preferably runs tangentially to the boundary defined by the person in the image.

[0056] Advantageously, the fact that the containment region 5 is accurate in capturing the physical boundary allows for the deployment of a system 100 with high reliability in detecting a person in a potentially hazardous work zone. According to another aspect of this description, the control unit 4 is designed to check whether the containment region 5 is at least partially contained within the first protection zone 3A and / or the second protection zone 3B.

[0057] In other words, the control unit 4 is designed to check the position of the person's containment region 5 in relation to the first protection zone 3A.

[0058] In other words, the control unit 4 is designed to check the position of the person's containment region 5 in relation to the second protection zone 3B.

[0059] According to the in Fig. In the embodiment shown in 5A, the inclusion region 5 is contained in the first protection zone 3A and is located outside the second protection zone 3B.

[0060] The control unit 4 is designed to issue an alarm signal if the containment region 5 is at least partially contained within the second protection zone 3B.

[0061] This alarm signal can be advantageously used, for example, to stop the machine.

[0062] More generally, the control unit 4 is designed to output an alarm signal when the containment region 5 is as shown in the diagram. Fig. 5B is contained in the second protection zone 3B.

[0063] In other words, the control unit 4 is designed to issue an alarm signal when a person is located in the second protection zone 3B, i.e., near access point 10 (i.e., portal structure 20). Alternatively, the control unit 4 can be designed to issue an alarm signal when the containment region 5 is at least partially contained within the first protection zone 3A. Advantageously, the control unit 4 is designed to issue an alarm signal when the person's containment region 5 overlaps even partially with the first protection zone 3A and / or the second protection zone 3B (i.e., when the containment region 5 enters the first protection zone 3A and / or the second protection zone 3B).

[0064] It should be noted that the first protection zone 3A may correspond to the entire field of view of the camera or be defined by a section of the field of view of the camera (i.e., a section of the captured image).

[0065] Advantageously, the control unit 4 is designed to issue an alarm signal when the containment region 5 enters the potential danger zone (i.e., the second protection zone 3B).

[0066] The control unit 4 is operationally connected to a machine interface module 9.

[0067] The control unit 4 and the machine interface module 9 are preferably connected by means of a wired connection.

[0068] According to another aspect, the control unit 4 and the machine interface module 9 are preferably connected via a WLAN connection.

[0069] The control unit 4 is designed to send the alarm signal to the machine interface module 9.

[0070] When the machine interface module 9 receives the alarm signal, the conveyor device 8 is stopped by halting the movement of the loads arranged on it.

[0071] The conveying device 8 is therefore blocked by the control unit 4 of the system 100 from the moment at least one person is inside the second protection zone 3B.

[0072] It should be noted that the operating machine can therefore be programmed to stop the conveying device 8 when it receives the alarm signal from the machine interface 9.

[0073] According to another aspect of this description, the control unit 4 is operationally connected to a signaling device 11.

[0074] The control unit 4 is designed to send the alarm signal to the alarm device 11.

[0075] The control unit 4 is designed to send the generated alarm signal to the reporting device 11 when the containment region 5 of a person overlaps at least partially with the second protection zone 3B.

[0076] The signaling device 11 is designed, for example, to output an acoustic signal based on the alarm signal received by the control unit 4.

[0077] Advantageously, the System 100 enables the reduction of the risk of accidents occurring near the moving parts of industrial machines.

[0078] According to another aspect, the system 100 also includes a user interface 15.

[0079] The operator interface 15 is operationally connected to the control unit 4 and is designed to receive at least one piece of information about the state of the system 100 from it.

[0080] The operator interface 15 enables the system 100 to be diagnosed when an alarm signal is issued.

[0081] Advantageously, the user interface 15 enables a user to perform a diagnosis of the system 100.

[0082] Advantageously, the user of System 100 performs the diagnostics to check that System 100 is operating correctly.

[0083] In other words, the user of system 100 performs the diagnosis to prove that the control unit 4 sent the alarm signal due to the actual presence of a person's confinement region 5 inside the first protection zone and / or the second protection zone 3B.

[0084] According to another aspect of this description, the inclusion body 1 includes at least one supply connector 12.

[0085] The power supply connector 12 enables the enclosing body 1 to be connected to a power line that is necessary for its supply.

[0086] According to another aspect of this description, the enclosure body 1 includes a plug connection for transmitting at least the alarm signal.

[0087] The connector for transmission 22 is connected to the control unit 4 and transmits the alarm signal to the machine. According to another, in Fig. The aspect shown in section 1 of this description includes the containment body 1 of system 100, comprising a multitude of cameras 2.

[0088] Each camera 2 is operationally connected to the control unit 4. The cameras 2 are arranged opposite each other to capture areas that at least partially overlap.

[0089] Since each camera 2 does not capture an area region that is completely different from the neighboring camera, cameras 2 capture adjacent and partially overlapping areas.

[0090] The use of a large number of cameras 2 advantageously enables the provision of a redundant system 100, in which even the malfunction of one camera 2 does not affect the overall assessment carried out by the system 100.

[0091] The control unit 4 is designed to detect the presence of a person inside the image captured by the respective camera on the image captured by each camera 2 of the cameras.

[0092] According to another aspect, the control unit 4 is designed to define at least one inclusion region 5 in the captured image that contains the detected person.

[0093] The control unit 4 makes it possible to check whether the containment region 5 is at least partially contained within the first protection zone 3A and / or the second protection zone 3B.

[0094] Finally, the control unit 4 provides for the output of an alarm signal if the containment region 5 is at least partially contained within the second protection zone 3B.

[0095] The control unit 4 is designed to send an alarm signal to the reporting device 11 if the inclusion region 5 of the detected image, which contains the detected person, is at least partially contained within the second protection zone 3B.

[0096] Advantageously, the system 100 relating to the invention enables an increase in the reliability in distinguishing between persons and objects, in particular between the loads present in the working zone of the industrial machine 50.

[0097] By increasing the reliability in distinguishing between persons and objects in the work zone, the system 100 advantageously enables the reduction of the number of alarm signals erroneously emitted by the alarm device 11.

[0098] The fact that the containment region 5 is accurate in detecting the physical boundary of the person advantageously allows for the reduction of the number of alarm signals erroneously issued by the reporting device 11.

[0099] Since the emission of an alarm signal generally involves the interruption of the conveying device 8, reducing the number of alarm signals erroneously emitted by the signaling device 11 (so-called false alarms) enables the reduction of the number of interruptions of the conveying device 8 of the system 100.

[0100] Reducing the number of interruptions of the conveyor device 8 of the system 100 essentially enables an improvement in the productivity and operation of the system 100.

[0101] In particular, the features described above advantageously enable the provision of a System 100 that meets high safety and productivity standards.

Claims

[1] Safety system (100) for an industrial machine (50), comprising at least one access point (10) and including: - at least one camera (2) designed to capture an image of a predetermined underlying zone defining a monitored zone (3), wherein the monitored zone (3) comprises at least a first protection zone (3A) located further away from the access point (10) and a second protection zone (3B) located closer to the access point (10), and - a control unit (4) connected to this camera (2), wherein the system (100) characterized by is that the control unit (4) is designed to - to receive from the camera (2) at least one image of a zone below which is representative of a monitored zone (3) captured by the camera (2); - to determine the presence of a person inside the captured image; - to define at least one inclusion region (5) in the captured image that contains the identified person; - to check whether the containment region (5) is at least partially contained within the first protection zone (3A) and / or the second protection zone (3B); - to issue an alarm signal if the containment region (5) is at least partially contained within the second protection zone (3B). [2] System (100) according to the preceding claim, wherein the control unit (4) comprises an artificial intelligence module (6) defined by a pre-trained classifier. [3] System (100) according to the preceding claim, wherein the artificial intelligence module (6) is implemented by means of a neural network. [4] System (100) according to the preceding claim, wherein the neural network comprises at least one input designed to receive information relating to an image captured by the camera (2). [5] System (100) according to one or more of claims 2 to 4, wherein the artificial intelligence module (6) is pre-trained using images containing a representation of a person, as well as using images not containing a representation of a person. [6] System (100) according to one or more of the preceding claims, wherein the artificial intelligence module (6) is designed to determine, on the basis of the images received during training, whether a person is present in the captured image. [7] System (100) according to one or more of the preceding claims, wherein the artificial intelligence module (6) is designed to calculate a numerical value that is representative of a probability that a person is present inside the captured image. [8] System (100) according to the preceding claim, wherein the artificial intelligence module (6) is designed to compare the calculated numerical value, which provides information about the probability of the presence of a person in the image, with a predefined threshold, wherein the presence of a person in the image is confirmed if the calculated percentage value exceeds the threshold. [9] System (100) according to one or more of claims 2 to 8, wherein the artificial intelligence module (6) is designed to define in the captured image an inclusion region of a predetermined shape containing the detected person. [10] System (100) according to the preceding claim, wherein the predetermined shape is a rectangular shape. [11] System (100) according to one or more of the preceding claims, wherein the control unit (4) is operationally connected to a communication module (9) and is designed to send the alarm signal to the communication module (9). [12] System (100) according to one or more of the preceding claims, wherein the control unit (4) is operationally connected to a signaling device (11) and is designed to send the alarm signal to this signaling device (11). [13] System (100) according to the preceding claim, wherein the signaling device (11) is designed to emit an acoustic signal based on the alarm signal. [14] System (100) according to one or more of the preceding claims, comprising a portal structure (20) defining said access point (10), wherein the camera (2) is arranged on the top of the portal structure (20). [15] System (100) according to one or more of the preceding claims, comprising a conveying device (8) designed to convey loads. [16] System (100) according to one or more of the preceding claims, comprising a plurality of cameras (2) operationally connected to the control unit (4) and arranged in relation to each other to capture at least partially overlapping regions, wherein the control unit (4) is designed to perform the following operations on the image captured by each camera (2) of the cameras: - Determining the presence of a person inside the image captured by the respective camera (2); - Define in the captured image at least one inclusion region (5) that contains the identified person; - Check whether this containment region (5) is at least partially contained within the first (3A) and / or the second (3B) protection zone; - Issuing an alarm signal if the containment region (5) is at least partially contained within the second protection zone (3B). [17] System (100) according to one or more of the preceding claims, wherein the camera (2) and the control unit (4) are integrated inside the same enclosure body (1). [18] System (100) according to the preceding claim, wherein the enclosing body (1) comprises at least one supply connector (12) and a connector for transmitting (22) at least the alarm signal.

Images