Device and method for counting people

Abstract

The invention relates to a device and a method for counting people intended to count people (5) passing through a passageway (1) to enter a room (6) of a building.According to the invention, the counting device (2a) comprises: - a time-of-flight distance sensor (2) including a transmitter capable of emitting a signal (3) forming a detection zone (9, 10) towards the passage zone (1) and a receiver capable of receiving a reflected signal corresponding to a part of the emitted signal (3) when it is reflected by a person (5) passing through the detection zone so as to measure the distance between the person (5) and the distance sensor (2) as a function of time, the detection zone (9, 10) having a constant position and orientation, - signal processing means analyzing a variation in the distance between the distance sensor (2) and the detected person (5) in order to determine the direction of movement of the person (5) with respect to the passage zone (1) and to count the number of people crossing the passage zone (1) in order to count the number of people (5) in the room (6).

Description

[0001] The present invention relates to a device and a method for counting people passing through a passage area and a ventilation system of a room of a building comprising such a counting device.

[0002] The invention finds its main application in counting people passing through a door of a building in order to quantify the occupancy of a room in real time.

[0003] There is currently a need to count the number of people in the different rooms of a building in order to control energy costs and optimize resource use. Knowing room occupancy allows, among other things, the adaptation of ventilation flow rates to the number of occupants in the rooms. It also allows building managers to track room occupancy rates in order to adapt the use of rooms or to estimate, for example, the need for cleaning.

[0004] We know of patent FR9415107 filed by the company AERECO in 1996 concerning a method and device for modulating the ventilation of premises. The person counting proposed in this patent is however not very precise because it is based on a measurement of agitation in a room.

[0005] Patents EP0718806 and WO2002011075 are also known, disclosing counting devices using "radar" or "ultrasonic" type distance measurements. These are generally very bulky and very expensive. No details are provided in these documents on the type of algorithm used for discrimination between input signals and output signals.

[0006] Patents US2003066875 and WO03091955 are also known, both of which disclose counting barriers consisting of lines of sensors generating several measurement points in order to process simultaneous passages in the detection zone. These solutions have an encroachment of the order of the width of the door and are not compact.

[0007] Patent FR2899003 is also known, which discloses a system for counting passers-by and their direction of passage, consisting of two pyroelectric cells. The device is optimized for outdoor use; it allows the direction of passage to be distinguished without depending on the sign of the temperature difference between the passer-by and the surrounding environment.

[0008] In patent WO2014179839, a laser scanner is used to count people or objects crossing a passageway. Laser scanner sensors contain a moving part that is not very robust and are expensive.

[0009] Patent WO2016096880 discloses a method for detecting moving elements in a building using a lidar, during which several scanning sequences are performed in order to assess the movements of people within a room. However, this detection system is complex and expensive. It is not suitable for an application for counting people in a room, nor for a room ventilation system.

[0010] Other solutions from the state of the art are disclosed, for example, in documents US2013016880 A1, WO2017114846 A1, WO2007138025 A1 and WO2017023202 A1.

[0011] The invention aims to resolve all or part of these drawbacks by proposing a robust, simple and compact people counting device and method so as not to harm the aesthetics of buildings.

[0012] The invention also aims to provide a ventilation system for a room in a building making it possible to adapt the ventilation to the number of people in the room.

[0013] The invention relates to a people counting device for counting people passing through a passage area to enter a room or part of a building comprising several areas or rooms.

[0014] The following examples describe an application for one room but the invention can be applied to a part of a building comprising several rooms.

[0015] According to the invention, the people counting device comprises: a time-of-flight type distance sensor comprising a transmitter capable of emitting a signal emitted at an angle α forming a detection zone towards the passage zone and a receiver capable of receiving a reflected signal corresponding to a part of the emitted signal when it is reflected by a person passing in the detection zone so as to measure the distance between the person and the distance sensor as a function of time, the detection zone having a constant position and orientation, and signal processing means analyzing a variation in the distance between the distance sensor and the detected person in order to determine the direction of movement of the person relative to the passage zone to carry out a count of the number of people in the room,the distance sensor being arranged relative to the passage area such that the signal reflected by a person entering the passage area in a first direction of passage is different from the signal reflected by a person entering the passage area in a second direction of passage which is opposite to the first direction of passage.

[0016] Preferably, the counting device is configured to detect only one person at a time.

[0017] Preferably, the distance sensor comprises a laser source emitting a conical-shaped light beam forming the detection zone whose apex angle is between 10° and 40°, and preferably equal to 24°.

[0018] Advantageously, the counting device is fixed at a position higher than the passage area. The range of the distance sensor is greater than 1.50 m.

[0019] Alternatively, the distance sensor comprises at least two sets of transmitters / receivers, including a first set of transmitters / receivers generating a first detection zone at the angle β relative to a horizontal direction B and a second set of transmitters / receivers generating a second detection zone at an angle β' relative to the horizontal direction B so as to cover more of the passage area.

[0020] The invention also relates to a method for counting people intended to count people crossing a passage area to enter a room of a building by means of a counting device as defined above.

[0021] According to the invention, the counting method comprises the following steps: transmitting a signal emitted at an angle α by a time-of-flight type distance sensor to the passage area, the emitted signal forming a detection area of constant orientation and position, receiving a reflected signal corresponding to a portion of the emitted signal when reflected by a person passing through the detection area allowing the distance sensor to generate a detection signal representative of the distance between the person and the distance sensor as a function of time, the signal reflected by the person entering the passage area in a first direction of passage being different from the signal reflected by the person entering the passage area in a second direction of passage which is opposite to the first direction of passage,analysis of the variation of the detection signal corresponding to a variation of the distance between the distance sensor and the person detected by signal processing means in order to determine the direction of movement of the person relative to the passage area to count the number of people entering the room and the number of people leaving the room, said analysis step comprising a sub-step of determining false detection of entry / exit of people in the case where the variations of the detection signal do not correspond to an entry or an exit, and calculation of the difference between the number of people entering the room and the number of people leaving the room to determine the number of people present in the room.

[0022] Preferably, the counting device is arranged on the side where the door opens. For example, the counting device may be positioned inside a room. The passage area comprises a door intended to close it and opening towards the inside of the room. The counting device is capable of detecting the opening or closing of the door so that this information is used by the signal processing means to participate in determining the direction of passage of the person.

[0023] Due to the angle α, the shape of the detection signal corresponding to a person entering a room is different from the shape of the detection signal corresponding to a person leaving the room.

[0024] Advantageously, the acquisition frequency of the distance sensor is between 10 ms and 30 ms and preferably 20 ms.

[0025] The step of analyzing the variation of the detection signal may include an analysis of the slopes of the detection signal.

[0026] Alternatively, the step of analyzing the variation of the detection signal comprises a comparison of the distances measured at the start and end of the detection signal.

[0027] According to another variant, the step of analyzing the variation of the detection signal uses learning algorithms of the decision tree or neural network type or predictive discriminant analysis or support vector machines. In this case, the learning algorithm comprises a step of constituting a learning base and a step of training a classification algorithm before the installation of the people counting device in a building and its use.

[0028] The invention thus provides a robust, simple and compact people counting device and method.

[0029] The metering device is easy to install, so installers do not require any special training. Furthermore, the metering device is discreet, so it does not detract from the aesthetics of the building.

[0030] The counting device makes it possible to count the passers-by under a door and their direction of passage in a simpler and more precise manner than those of the prior art.

[0031] The measurement point is unique (or reduced to a single point in the case where the signal is composite).

[0032] Some prior art devices include a movable part allowing scanning while the distance sensor according to the invention is entirely fixed.

[0033] The algorithm used according to document WO2016096880 performs evolutionary mapping and analyses it logically, while the algorithm according to an embodiment of the invention performs classification of temporal signals by learning methods ("Machine Learning").

[0034] A processing computer is also required for some of the prior art solutions, whereas the reflected signals according to the invention are processed on microprocessors embedded in the counting device.

[0035] The invention also relates to a system for ventilating a room of a building comprising at least one passage zone.

[0036] According to the invention, the ventilation system comprises at least one people counting device as defined above, associated with the passage area. The people counting device counts the number of people crossing the passage area to enter or exit the room to count the number of people inside the room. The people counting device transmits this counting information to a control device of the ventilation system to modulate the ventilation of the room according to the number of people in the room.

[0037] Alternatively, it is also possible to manage the flow of air extracted from the room or the flow of air supplied and extracted.

[0038] The room may include multiple passageways. The ventilation system then includes multiple counting devices, each associated with a passageway. The counting devices are connected to each other to allow the number of people in the room to be counted.

[0039] The connection can be wired or wireless (radio or other communications).

[0040] The invention thus provides a ventilation system associated with a simple, compact and robust counting device.

[0041] The ventilation system is adapted to the number of people actually present in the room and is therefore more energy-efficient.

[0042] Other applications are also possible, notably to detect other moving elements such as animals or vehicles.

[0043] Other characteristics and advantages of the invention will emerge from reading the following description, given solely by way of example, with reference to the appended figures, which illustrate: figure 1 , a diagram of a people counting device fixed to the ceiling of a room of a building according to one embodiment of the invention; figure 2 , a profile diagram of this people counting device; figure 3 , a top view diagram of this people counting device; figure 4 , a top view diagram of a people counting device according to another embodiment of the invention; figure 5 , another diagram of a people counting device according to the invention fixed to the ceiling of a room and inside it; figure 6 , a graph representing the detection signal as a function of time when a person enters the room and the counting device is placed inside the room; figure 7 , a graph representing the detection signal as a function of time when a person leaves the room and the counting device is placed inside the room; figure 8 , a flowchart explaining how door openings and closings are taken into account for the classification of detection signals; figure 9 , a flowchart representing an example of a classification algorithm; figure 10 , a flowchart for implementing a learning algorithm for classifying detection signals, according to another embodiment of the invention.

[0044] THE figures 1 à 3 And 5represent a people counting device 2a intended to count people 5 crossing a passage area 1 to enter a room 6 of a building, according to one embodiment of the invention.

[0045] The building can be a building or a house for example.

[0046] The counting device 2a comprises a time-of-flight type distance sensor 2 comprising a transmitter for transmitting an emitted signal 3 forming a detection zone 9, 10 towards the passage zone 1 and for receiving a reflected signal corresponding to a part of the emitted signal 3 when it is reflected by a person 5 passing in the detection zone 9, 10 so as to measure the distance between the person 5 and the distance sensor 2 as a function of time.

[0047] Passage zone 1 corresponds to a door frame or an area around the door frame, for example.

[0048] The detection zone 9, 10 has a constant position and orientation. The counting device 2a is configured to detect only one person 5 at a time.

[0049] Preferably, the distance sensor 2 comprises a laser source emitting a conical-shaped light beam forming the detection zone 9, 10 whose apex angle is between 10° and 40°, and preferably equal to 24°.

[0050] The laser source is characterized by the generation of short light pulses.

[0051] When these pulses are emitted towards a person, the light is reflected by the person passing through the detection zone or light detection cone. The measurement distance is deduced from the flight time taken by the light pulses to reach the person and return to the receiver integrated in the distance sensor 2, so as to obtain a detection signal 4 representative of the distance between the person 5 and the distance sensor 2 as a function of time.

[0052] Thus, the reflected signal is received by the receiver which determines the distance between itself and the detected person from, on the one hand, the time elapsed between the emission of the laser pulse and the reception of the reflected signal and, on the other hand, the speed of light.

[0053] The emitted light or emitted signal 3 forms a cone-shaped detection zone 9, 10 of constant orientation and position. The angle α of the emitted signal 3 relative to a vertical direction A is constant.

[0054] The emitted signal 3 also forms this angle α with the passage zone 1, as shown in the figures 1 et 2 , so that an asymmetry exists between an entry passage and an exit passage. This asymmetry generates a difference between the reflected signals generated by the distance sensor 2 depending on the direction of passage of the person.

[0055] The shape of the detection signal 4 corresponding to a person entering a room 6 ( fig. 6 ) is different from the shape of the detection signal 4 corresponding to a person leaving room 6 ( fig.7 ).

[0056] Advantageously, the detection field of the distance sensor 2 crosses the passage zone 1 which is vertical.

[0057] The range of the distance sensor 2 is such that any person 5 passing through the passage zone 1 is detected. Advantageously, the range of the distance sensor 2 is greater than 1.50 m.

[0058] The counting device 2a and the distance sensors 2 are fixed at a position higher than the passage area 1.

[0059] As shown in the figures 2, 3 And 5 , the counting device 2a is fixed on the ceiling 8 of the room 6, and therefore above the passage area 1.

[0060] The people counting device 2a comprises signal processing means analyzing the detection signal 4 in order to determine the direction of movement of the person 5 relative to the passage area 1 and to carry out a count of the number of people crossing the passage area 1 as well as their direction of passage to count the number of people in the room 6.

[0061] The signal processing means are preferably integrated into the counting device 2a.

[0062] According to one variant, the counting device 2a comprises an optical filter allowing the wavelength band used by the distance sensor to pass through. The filter is positioned in front of the distance sensor 2 in order to limit the measurement noise caused by ambient light.

[0063] According to a variant, the counting device 2a comprises a pyroelectric sensor whose detection field is wider than the detection zone 9, 10. The distance sensors are only activated when the pyroelectric sensor detects a presence in its detection field. This makes it possible to limit the consumption of the sensors and to increase the lifespan of the counting device 2a.

[0064] According to another variant, the counting device 2a comprises several distance sensors 2 to cover the entire passage area 1 or a large part of it.

[0065] The distance sensor 2 may comprise at least two sets of transmitters / receivers, a first set of transmitters / receivers generating a first detection zone 9 at the angle β relative to a horizontal direction B and a second set of transmitters / receivers generating a second detection zone 10 at an angle β' relative to the horizontal direction B, providing a compact configuration covering the entire passage area 1, as shown in the figure 4 .

[0066] Depending on the geometric configuration chosen, the cones or detection zones of each transmitter / receiver may or may not overlap.

[0067] The invention also relates to a method for counting people crossing a passage area 1 comprising a step of transmitting a signal emitted 3 by a distance sensor 2 of the time-of-flight type towards the passage area 1 at a constant angle α relative to a vertical direction A.

[0068] The counting method also comprises a step of receiving a reflected signal corresponding to a part of the emitted signal 3 when it is reflected by a person 5 passing in the detection zone 9, 10.

[0069] This makes it possible to measure a detection signal 4 representative of the distance between the person 5 and the distance sensor 2 as a function of time.

[0070] The counting method comprises a step of processing or analyzing a variation in the detection signal 4 by signal processing means corresponding to a variation in the distance between the distance sensor 2 and the detected person 5 in order to determine the direction of movement of the person 5 relative to the passage area 1 and to carry out a count of the number of people crossing the passage area 1 in order to count the number of people 5 in the room 6.

[0071] The acquisition frequency of the distance sensor 2 must be high so as to record a sufficient number of points even when a person passes quickly. Typically, an acquisition frequency of around 20 ms makes it possible to obtain around ten points during very rapid passages.

[0072] The acquisition frequency of the distance sensor 2 is between 10 ms and 30 ms and is preferably 20 ms.

[0073] When no one is located in the detection zone or cone 9, 10, the distance sensor 2 returns the information of absence of detection. A variation in the distance measurement occurs only when a person passes into the detection zone 9, 10.

[0074] The detection signals 4 generated by the distance sensor 2 are schematically represented on the graphs of the figures 6 et 7 .

[0075] As previously stated, the shape of the detection signal 4 corresponding to a person entering a room 6 ( fig. 6 ) is different from the shape of the detection signal 4 corresponding to a person leaving room 6 ( fig.7 ).

[0076] There figure 6 represents the detection signal 4, i.e. the distance d between the person 5 and the distance sensor 2 as a function of time when the person enters the room 6 and the distance sensor 2 is positioned in the room 6. The person 5 then moves towards the passage area 1 and the distance sensor 2.

[0077] The evolution of the detection signal 4 shows that the distance between person 5 and distance sensor 2 reduces as person 5 approaches the room. An entry from person 5 generates a decreasing signal represented by the decreasing slope 11.

[0078] When person 5 leaves the detection zone 9, 10, the detection signal 4 suddenly returns to its initial non-detection level.

[0079] As shown in the figure 7 , the distance d between person 5 and the distance sensor 2 increases abruptly when person 5 leaves room 6 because he suddenly arrives in the detection zone 9, 10.

[0080] The distance d gradually decreases as person 5 leaves room 6 and moves away from the passage area 1, as represented by the increasing slope 12.

[0081] When the counting device 2a is located outside the room 6, the passage zone 1 being located between the distance sensor 2 and the room 6, the detection signal 4 shown on the Figure 6 corresponds to an output and the detection signal 4 represented on the Figure 7 corresponds to an entry

[0082] The signal processing is preferably carried out by a microprocessor embedded in the counting device 2a.

[0083] The processing means are capable of differentiating the detection signals 4 of people entering the room from the detection signals 4 of people leaving the room.

[0084] The processing means calculates the number of people entering the room and the number of people leaving the room. The difference between these two numbers gives the number of people present in the room.

[0085] The step of analyzing the variation of the detection signal 4 includes a sub-step of determining false detection of entry / exit of people 5.

[0086] During the analysis of the variation of the detection signal 4, the variations of the detection signal 4 not corresponding to an entry or an exit are classified as false detections to remove from the count the people passing near the passage zone 1 but not crossing it. The algorithm of the processing means discriminates between entries, exits and false detections. The false detections are linked to the detection of a person in the field of vision of the counting device 2a without them entering or leaving the room 6.

[0087] The reasons for variations in the shapes of the detection signals 4 are varied. The measurement noise of the distance sensors 2 is a primary cause of signal variability.

[0088] User behavior is also sometimes very erratic. For example, a round trip causes a signal to increase then decrease, or vice versa.

[0089] And a person standing in the doorway of passage zone 1 generates a plateau signal. For these different reasons, the signal processing algorithm must be particularly robust in discriminating between different cases.

[0090] According to a possible variant, the distance sensor 2 is positioned inside the room 6 and the passage area 1 comprises a door 7 intended to close it and opening towards the inside of the room 6.

[0091] The distance sensor 2 detects the opening or closing of the door 7 so that this information can be used by the signal processing means.

[0092] This solution provides the ability to detect door openings and closings 7 by simply comparing the measured distance to a defined distance detection threshold. This information is very useful for building management and ventilation.

[0093] It is possible to integrate door openings and closings into the classification algorithm.

[0094] According to the algorithm shown on the figure 8 , the people counting method 5 comprises: a step of acquiring the signal 100, i.e. emitting the emitted signal 3 towards the passage area 1 and receiving a part of the emitted signal 3 when it is reflected by a person 5 heading towards or moving away from the passage area 1 to obtain the detection signal 4, a step of analyzing 200 the variation of the detection signal 4 during which the opening or closing of the door 7 is analyzed by the signal processing means.

[0095] A door closed at the last points before the variation of the detection signal 4 (symbol 210) corresponds to a person entry 5 (symbol 220).

[0096] Otherwise, a door closed at the first points following the variation of the detection signal 4 (symbol 230) corresponds to a person exit 5 (symbol 240).

[0097] Otherwise, it is analyzed during a classification step 300 if there is entry, exit or false detection of an entry / exit of person 5 according to the symbol 400.

[0098] In case door 7 opens outside room 6 (towards the corridor for example), distance sensor 2 is placed outside room 6.

[0099] On the diagram of the figure 9 , the labels “input” and “output” are then reversed.

[0100] According to a possible variant, the step of analyzing the variation of the reflected signal comprises an analysis of the slopes 11, 12 of the detection signal 4.

[0101] According to another possible variant, the step of processing or analyzing the variation of the detection signal 4 comprises a step of comparing 300 the distances measured at the start and end of the detection signal 4, as illustrated in the diagram of the figure 9 .

[0102] There figure 9 is an example of a classification algorithm used for example in step 300 of the algorithm of the figure 8 .

[0103] The analysis of the variation of the detection signal 4 includes: a step 310 of calculating values at the start a and end b of detection signal 4, and a step 320 of calculating the difference c=ab.

[0104] If c is greater than an entry threshold (symbol 330), an entry of person 5 is detected (symbol 410).

[0105] Otherwise, the classification algorithm includes another step during which it is analyzed whether c is less than an exit threshold (symbol 340). If so, an exit of person 5 is detected (symbol 420).

[0106] Otherwise, the algorithm deduces a false detection of entry / exit of people 5 (symbol 430).

[0107] Comparing the distances measured at the start and end of the signal is a relevant approach, since these points a and b reflect the positioning of person 5 when they enter and leave the detection zone.

[0108] According to another possible variant illustrated on the figure 10 , the step of analyzing the variation of the detection signal 4 uses learning algorithms of the decision tree or neural network type or predictive discriminant analysis or support vector machines.

[0109] The following describes a possible method of obtaining 300' and using 900 a learning algorithm. The obtaining method comprises a step of constituting a learning base 500 and a step of training the classification algorithm 600.

[0110] The method comprises a step of using the classification algorithm 900 in-situ, i.e. after the installation of the people counting device 2a 5 in a building.

[0111] Thus, the method for obtaining a learning algorithm comprises a step of constituting a learning base 500 comprising: a step of acquiring the signal 100, i.e. transmitting the transmitted signal 3 to the passage zone 1 and receiving a part of the transmitted signal 3 when it is reflected by a person 5 passing through the detection zone of a room 6 to generate the detection signal 4, a step of affixing the entry, exit or false detection label 510, determined for example by a reference sensor (position of the sensor relative to the room, for example), a step of adding this labeled data to a learning base 520.

[0112] These steps are repeated until the database size is sufficient.

[0113] The method for obtaining a learning algorithm then comprises a step of training the classification algorithm 600 comprising: a step of choosing the type of algorithm and selecting the discriminating characteristics of the detection signals 610, a step of optimizing the algorithm on the learning basis 620, a test step, in order to determine whether the success rate is sufficient 630.

[0114] If not, the steps of choosing and optimizing algorithm 610 and 620 are repeated.

[0115] If the success rate is sufficient, the classification algorithm is obtained in step 640.

[0116] The method of using the learning algorithm which then follows comprises a step 700 of implementing the classification algorithm in a people counting device 2a during the manufacturing step of the latter.

[0117] The people counting device 2a is then installed in a building at step 800.

[0118] The method of use then comprises a step of using the in-situ classification algorithm 900 comprising: a step 100 of acquiring the detection signal as described previously, a step 300 of analyzing the variation of the detection signal 4 by the classification algorithm obtained during the step of training the classification algorithm 600 to determine whether the person is entering or leaving the room or whether it is a false detection (symbol 400).

[0119] The invention also relates to a ventilation system for a room 6 of a building comprising at least one passage zone 1 and a person counting device 2a as defined previously.

[0120] Preferably, a single counting device 2a is associated with the passage zone 1.

[0121] The people counting device 2a determines the number of people 5 present in the room 6 and transmits the information to a ventilation system control device to manage or adjust the air flow to be blown into the room 6 according to the number of people 5 present inside the room 6.

[0122] Alternatively, it is also possible to manage the flow of air extracted from the room or the flow of air supplied and extracted.

[0123] Alternatively, the room 6 comprises several passage zones 1 each associated with a counting device 2a.

[0124] The processing means is capable of centralizing the information from several counting devices 2a to integrate the entries and exits from the different passage zones 1.

[0125] In the context of specific uses in which rooms 6 or halls of a building are not used at certain times (meeting rooms in the tertiary sector, for example), the processing means can integrate a reset of the number of people in the room in order to compensate for drifts due to possible detection errors.

Claims

1. A people counting device (2a) intended to count people (5) crossing a passage area (1) to enter a room (6) of a building, the counting device (2a) comprising: - a time-of-flight type distance sensor (2) comprising a transmitter capable of emitting an emitted signal (3) at an angle α) forming a detection area (9, 10) towards the passage area (1) and a receiver capable of receiving a reflected signal corresponding to a portion of the emitted signal (3) when it is reflected by a person (5) passing within the detection area (9, 10) so as to measure the distance between the person (5) and the distance sensor (2) as a function of time, the detection area (9, 10) having a constant position and orientation, - signal processing means analyzing a variation of the distance between the distance sensor (2) and the detected person (5) in order to determine the direction of movement of the person (5) relative to the passage area (1) to perform a counting of the number of people (5) in the room (6), characterized in that the distance sensor (2) is disposed relative to the passage area (1) such that the signal reflected by a person entering the passage area (1) in a first direction of passage is different from the signal reflected by a person entering the passage area (1) in a second direction of passage which is opposite to the first direction of passage.

2. The people counting device (2a) according to claim 1, characterized in that the distance sensor (2) comprises a laser source emitting a cone-shaped light beam forming the detection area (9, 10) whose point angle is comprised between 10° and 40°, and preferably equal to 24°.

3. The people counting device (2a) according to any one of claims 1 or 2, characterized in that the counting device (2a) is fixed at a position higher than the passage area (1), the range of the distance sensor (2) being greater than 1.50 m.

4. The people counting device (2a) according to any one of claims 1 to 3, characterized in that the distance sensor (2) comprises at least two sets of transmitters / receivers, including a first set of transmitters / receivers generating a first detection area (9) at the angle β) relative to a horizontal direction (B) and a second set of transmitters / receivers generating a second detection area (10) at an angle β') relative to the horizontal direction (B) so as to cover more of the passage area (1).

5. A ventilation system for a room (6) of a building comprising at least one passage area (1), characterized in that it comprises at least one people counting device (2a) as defined according to any one of claims 1 to 4, associated with the passage area (1), the people counting device (2a) counting the number of people (5) crossing the passage area (1) to enter or exit the room (6) and recording the number of people (5) inside the room (6), the counting information being transmitted to a control device of the ventilation system to modulate the ventilation of the room (6) according to the number of people (5) in the room (6).

6. The ventilation system for a room of a building according to claim 5, characterized in that the room (6) comprises several passage areas (1), the ventilation system comprising several counting devices (2a), each associated with a passage area (1), the counting devices (2a) being connected to each other to enable the counting of the number of people in the room (6).

7. A method for counting people (5) intended to count people (5) crossing a passage area (1) to enter a room (6) of a building, by means of a counting device (2a) as defined according to any one of claims 1 to 4, characterized in that it comprises the following steps: - emission of an emitted signal (3) at an angle α) by a time-of-flight type distance sensor (2) towards the passage area (1), the emitted signal (3) forming a detection area (9, 10) having constant orientation and position, - reception of a reflected signal corresponding to a portion of the emitted signal (3) when it is reflected by a person (5) passing within the detection area (9, 10) allowing the distance sensor (2) to generate a detection signal (4) representative of the distance between the person (5) and the distance sensor (2) as a function of time, the signal reflected by the person (5) entering the passage area (1) in a first direction of passage being different from the signal reflected by the person (5) entering the passage area (1) in a second direction of passage which is opposite to the first direction of passage, - analysis of the variation of the detection signal (4) corresponding to a variation of the distance between the distance sensor (2) and the person (5) detected by signal processing means in order to determine the direction of movement of the person (5) relative to the passage area (1) to count the number of people (5) entering the room (6) and the number of people (5) exiting the room (6), said step of analyzing the variation of the detection signal (4) comprising determining false detections of entry / exit of people (5) in the case where the variations of the detection signal (4) do not correspond to an entry or an exit, and - calculation of the difference between the number of people (5) entering the room (6) and the number of people (5) exiting the room (6) to determine the number of people (5) present in the room (6).

8. The method for counting people according to claim 7, characterized in that the acquisition frequency of the distance sensor (2) is comprised between 10 ms and 30 ms and preferably 20 ms.

9. The method for counting people according to any one of claims 7 or 8, characterized in that the passage area (1) comprises a door (7) intended to close it and opening towards the inside of the room (6), the counting device (2a) being capable of detecting the opening or closing of the door (7) so that this information is used by the signal processing means to participate in determining the direction of passage of the person (5).

10. The method for counting people according to any one of claims 7 to 9, characterized in that the step of analyzing the variation of the detection signal (4) comprises an analysis of the slopes (11, 12) of the detection signal (4).

11. The method for counting people according to any one of claims 7 to 9, characterized in that the step of analyzing the variation of the detection signal (4) comprises a comparison of the distances measured at the start and end of the detection signal (4).

12. The method for counting people according to any one of claims 7 to 9, characterized in that the step of analyzing the variation of the detection signal (4) uses learning algorithms of the decision tree or neural network or predictive discriminant analysis or support vector machine type, obtaining the learning algorithm (300') comprising a step of constituting a learning base (500) and a step of training a classification algorithm (600) carried out prior to the installation of the people (5) counting device (2a) in a building and its use.

Images