WO2008142238A2

WO2008142238A2 - Method of determining the number of objects present in a given surface region by infrared thermography

Info

Publication number: WO2008142238A2
Application number: PCT/FR2008/000426
Authority: WO
Inventors: David Morio
Original assignee: Dgm Technologies
Priority date: 2007-03-29
Filing date: 2008-03-28
Publication date: 2008-11-27
Also published as: WO2008142238A3; FR2914461A1; FR2914462A1

Abstract

The invention relates to a method of determining the number of objects of a given time present in a given surface region using an infrared camera 3, comprising the following steps: a) acquisition of an image using an infrared camera 3; b) subdivision of the target region 1 into a plurality of grid units 6; c) calculation for each grid unit 6 of its average temperature TM; d) comparison, for each grid unit 6, of this average temperature TM with a prechosen reference temperature TS; e) determination, for each grid unit 6, of the difference between the temperatures TM and TS; f) determination of the number N of what are called “relevant” grid units having a difference between the temperatures TM and TS of greater than a reference value chosen beforehand according to the type of object; and g) multiplication of the number N by a predetermined coefficient, which expresses the average density of objects of a given type in a relevant grid unit, so as to obtain the number of objects of a given type present in the target region 1.

Description

Method for determining the number of objects present in a given surface area by infrared thermography

Technical field of the invention

The invention relates to the technical field of thermographic imaging and the exploitation of data thus generated. More particularly, it relates to the determination of the number of objects present in a given surface area, in particular the number of objects in a crowd, or in a flow of objects, from infrared thermography images.

State of the art

Despite the strong development of video surveillance (most often used as closed-circuit television (CCTV) in public and private spaces, there is now no simple - method to determine the number of objects present in a given polygon element. For example, the number of people who parade during a political event is generally considered to be of very different value by the organizers of the event and by the police. In controlled access sites, individual counting is easy, for example using turnstiles or automatic gates. But this approach is obviously not applicable to determine the affluence rate in front of the site or the number of people waiting, and more generally in a public space. Moreover, we do not always need to know with precision unitary the number of people, one can admit a certain margin of error.

Similarly, there is no simple automatic method for determining in real time the number of automobiles in a given area area, for example around a junction or a roundabout.

The exploitation of photographs or video images is possible using digital image analysis software, but comes up against the difficulty of finding an angle of view that allows individual counting of objects in a given surface area, knowing that this count can only be exact for images taken by a camera located exactly above the surface element to be analyzed, as described in the patent application WO 02/093487 (Koninklije Philips Electronics NV). In all other cases, there is a risk that large objects will hide small objects. In addition, image processing software must include a pattern recognition module, and are quite difficult to calibrate, especially when operating in real time. This software can therefore introduce a bias that is difficult to evaluate because a crowd is always random in its spatial distribution and, to a certain extent, in the individual nature of the objects of a given category that compose it. Thus, people can wear a hat or backpack, or can pull a wheeled suitcase or trolley, or a person can carry a child on his or her shoulders. Similarly, a truck may hide passenger vehicles, or an unusually shaped rolling machine may not be recognized by the software. According to the state of the art, it is mainly through sophisticated mathematical methods that the skilled person has tried to determine the number of people in a crowd from video images. These mathematical methods can, however, also be applied to video images from other spectral domains, such as the infrared domain, the sonar domain, and the radar domain (see patent application WO 2005/038717 (Intellivid Corp.).

Patent Application EP 0 967 583 A2 (Texas Instruments Inc.) discloses methods of analyzing statically acquired static video images for the purpose of detecting the presence of a human being in a given field of view.

US Patent 7,139,409 B2 (Siemens Corporate Research) describes a dynamic method of CCTV type image analysis, which is based on Markov type functions. US Patent 6,049,363 (Texas Instruments Inc.) discloses a statistical method for describing the evolution of a scene observed by a television or infrared-type camera.

In addition, it would be desirable to have additional information about the typology of the objects that make up the crowd or the flow. For example, it may be interesting to know if a stationary car is parked with the engine off, or if it is still part of the circulatory flow.

According to the state of knowledge of the inventor, in the state of the art there is no direct method based on optical measurements to determine the number objects of a given type that are present in a given area area.

Due to the absence of simple and automatic methods to determine the number of objects present in a surface area, including objects forming a crowd or a flow of objects, there are no exploitable databases either. to process these data, either for analytical purposes or for forecasting purposes.

Optical methods based on conventional photographic images, ie images acquired with visible light (corresponding to a wavelength of approximately 400 - 740 nm) emitted or reflected by the objects observed, do not appear therefore not be able to solve this problem satisfactorily. In addition, these techniques are sensitive to visibility conditions, lighting conditions and weather conditions: they work poorly or not at all at night, are hindered by sources of external lights (car headlights) and fog, snow, rain.

There are also image acquisition techniques with infrared light emitted by objects. This infrared light corresponds to the thermal radiation emitted by the objects observed. Infrared thermography is used in many technical fields, and especially for the detection of thermal leaks, hot spots or cold spots in buildings, industrial furnaces, reactors, industrial installations. It is also used in the medical field, as a medical diagnostic aid capable of detecting hot areas of the body; these hot areas may then be subject to a thorough examination involving other medical techniques, in order to diagnose diseases likely to lead to local warming of the tissue initially detected by infrared thermography. In all these uses, an infrared camera is used with a CCD detector (charge-coupled device). Such a camera gives a map of the temperature.

For temperatures close to ambient temperature, such as for example the temperature of a human body, infrared thermography measurement requires a very thorough knowledge of the operation of the camera and the software used, as well as a careful calibration of the parameters. camera and software.

Mikael Cronholm's article "Using IR Cameras to Detect SARS Fever", published in the journal InfraMatrix 2004, provides an example. In all applications that measure the temperature of a human body, it is essential that the person who is characterized by infrared thermography be individually, that is to say that it is individually at a certain distance and in a certain position relative to the camera. According to the state of the art, it is not possible to determine using an infrared camera a person with fever, and therefore having a temperature above normal, in a crowd.

It can be seen that according to the state of the art, infrared thermography is used to characterize individual objects, such as an oven, a building, a human, but not to characterize a collective of objects. However, in the context of the present invention, there is no interest at all in characterizing an individual object, but only a plurality of objects. US Patent 5,995,900 (Grumman Corp.) discloses a method of analyzing images acquired by an infrared camera to characterize, by comparison between images taken successively, the flow of motor vehicles present in a segment of a road. This analysis is done pixel by pixel, which makes the method quite unstable.

The problem that the present invention proposes to solve is to present a method for determining the number of objects present in a given surface area, which can be implemented automatically, permanently, in real time, and which can operate in all environmental conditions, and in particular in all visibility, lighting and climatic conditions likely to occur in a public or private place, inside or outside. Another problem is to present a method for determining the temporal evolution of these parameters, in particular their tendency in a given time interval.

DESCRIPTION OF THE FIGURES FIG. 1 schematically shows a target zone 1 of dimension 10 m × 5 m in front of a waypoint 2. FIG. 2 shows by way of illustration the geometrical parameters related to the position of the infrared camera 3 with respect to to the target zone 1. FIG. 3 shows an example of a grid of the target zone of FIG.

FIGS. 4, 5 and 6 show the diagram of three embodiments of a computer network enabling centralized operation of the data generated by the method according to the invention.

The following marks are used:

1 Target area

2 Waypoint

3 Infrared camera

4 Ground level

5 surface acic element within which the density of obj ects is considered significant

6 Grid unit

Description of the invention 1. Definitions We call here "objects" the material objects that make up the crowd or flow to be characterized, and in particular: humans, animals, land motor vehicles (such as automobiles), and others body emitting thermal radiation detectable by an infrared camera.

We mean by "crowd" a multitude of objects gathered in one place.

By "target area" we mean the surface element of the soil within which the number of objects present is determined.

The term "crossing point" designates in the case of a queue or flow of objects a point in front of which the first or the first of the objects constituting said queue or said stream are slowed down or stopped before being able to pass ; in the case of a continuous stream it can also be a virtual landmark that the objects constituting said flow must pass. In the first case, this crossing point can be for example, a ticket office, a cash register, a bus stop, a red light, a roundabout, a junction, a road accident site, a road construction site.

The term "transit time" t _p designates the average time that elapses between the arrival of the object at the point of passage and its crossing of said point of passage. This typically corresponds to the average processing time of an object at the waypoint.

The term "waiting time" t _a refers here to the average waiting time of an object in front of a defined waypoint.

The term "affluence rate" is used here in the context of a flow of objects that tends to a site having a defined maximum capacity C _max . For a number N _x of objects X, we determine the average passage time t _P of an object X at the point of passage P, as well as the average residence time t _s in the site. Then, the distribution of the number of inputs N _input (t) and the number of outputs N _out i _e (t) as a function of time is determined. (The affluence rate A (t) at a time t is defined as A (t)

= [Nen Tree _(t) - N _outgoing part (t)] / C max •

2. Detailed description of the invention

A first object of the present invention is a method for determining the number of objects present in a given surface area using an infrared camera, and in particular a method for determining the number of objects present in a given surface area. in front of a crossing point. This method also makes it possible to determine the average wait time in front of this crossing point, and the affluence rate.

We describe here an embodiment of the complete method of thermographic determination of the number of objects present in a surface area of a crowd.

Step 1: In a first step, select the target area. This means that depending on the purpose of the determination, and depending on the layout of the place, the position and size of the target area are selected. FIG. 1 shows by way of illustration such a target zone 1 with a dimension of 5 m × 10 m in front of a waypoint 2.

Step 2: In a second step, the position of the infrared camera 3 and its configuration, that is to say its optical and software parameterization, is determined. FIG. 2 shows by way of illustration the geometrical parameters related to the position of the infrared camera 3 with respect to the target zone 1 defined in step 1:

the distance to the ground L of the camera 3 relative to the center of the target zone,

the optical distance D of the camera 3 with respect to the center of the target,

- the height H of the camera in relation to the ground level 4.

The configuration of the camera and the data acquisition system must take into account the following parameters: the geometric parameters L, H and D as defined above, the acquisition frequency, the orientation of the camera, the objective used, the thermal resolution of the camera, its spectral response, the size of the pixel matrix (for example: 160 x 120 pixels). A spectral band with a wavelength between 8 μm and 12 μm is suitable for carrying out the present invention.

Step 3: In a third step, an image is acquired using the infrared camera 3 as configured in the second step. In practice, a plurality of successive images are acquired at regular intervals or not. Each pixel of this image corresponds to a temperature value.

For example, one can acquire an image every 30 seconds, knowing that the acquisition time of an image (of the order of a few milliseconds) is typically very short compared to this interval between two images.

This image is stored in digital form in a memory unit, for example a computer server for storing, saving and archiving the files of the data acquired.

The fourth and fifth steps described below refer to the processing of a single image. They are typically performed for each acquired image, in order to determine the evolution of the parameters deduced.

Step 4: In a fourth step, using a first algorithm, inside the target zone 1, the surface element 5 (or the surface elements) within which (or from which) the density of objects is considered significant. This first algorithm retrieves a file of the image acquired by the camera 3 which has been temporarily saved. This file contains the image data as well as all the temperature measurement points for all the pixels that form the image.

Since the position of the camera 3 with respect to the target zone 1 is known, the algorithm calculates the coordinates of the target zone 1. The following operations of the method relate solely to this target zone. The algorithm divides the target area into so-called "grid units" 6. We call this operation the "grid" of the target area and the units of the "grid units", regardless of the form of the units; in fact, the grid units 6 thus formed may have a square, rectangular or other shape. The algorithm then calculates for each grid unit 6 the average temperature T _m . This average temperature is compared with a reference temperature T ₅ previously defined according to a number of parameters specific to the conditions at time t.

Here we describe a procedure for determining T ₃ based on an iterative discrimination signal processing technique. This method is based on the assignment of each pixel (ie each temperature value) of the target area to one of the two basic classes of background and form. Such a problem is known in the field of digital image analysis under the name of binarization. If these two classes are sufficiently distant from each other, then we conclude that the cliché has two modalities. The separation threshold will then be calculated based on these two classes. Otherwise, it is considered impossible to determine a threshold and the threshold associated with the image corresponding to the previous capture is used. This decision is justified by the short time interval between the taking of two shots on a given site and the fact that the infrared camera is fixed and thus retains its point of view.

An advantageous method that meets the objective is the dynamic cloud method (also called the centers method). It is a method of iterative discrimination. The algorithm used for this embodiment of the invention integrates the discrimination of the thermographic data by dynamic clouds and a verification of their effective separability and takes into account the evolution of the surface temperature of the object as a function of the temperature. room. This function is adapted according to the object measured. For example, for human subjects this function is indicated in Annex 1. The algorithm for the automatic determination of the threshold temperature T ₅ is indicated in Annex 2.

As with any image processing algorithm, the following issues have been addressed: noise elimination, point reversal of classes, iterative discrimination methodology (in this case use of the previous threshold.

Here we describe an example for the grid procedure. The grid may use a grid unit 6 of dimension 1 m * 1 m for a crowd of humans, but other dimensions may be used depending on the nature of the objects and the characteristics of the crowd or the flow of objects. to analyze. For a crowd of humans, one can for example use a dimension grid unit as small as 0.25 m * 0.25 m, but according to the findings of the inventor, for a human crowd, it is not useful to choose a grid unit 6 smaller than that. For example, the average temperature may be 37 ° C for a human. In practice, depending on the environmental conditions, a correction coefficient is advantageously applied.

Advantageously, the ambient air temperature is measured at the same time as the acquisition of the thermography image, either (preferably) on the site itself, or in a central weather station in the city in question; in the latter case, the data can be retrieved instantaneously or in a delayed manner by a teletransmission means, such as an RSS feed. The best method is the use of a measurement probe located on or near the outer housing of the camera, provided that said probe is protected from direct sunlight. For example, the probe can be glued against the front window (which is typically germanium) of the camera, it is thus protected against direct sunlight by the protective cap of said window.

The threshold temperature T _{s is} then determined in real time, using the following parameters:

the temperature measured by the probe located at the camera at time t;

the temperature measured by the infrared camera 3 at the plane of focus at time t; - the values of the measuring range of the camera 3

(high value and low value) which are automatically generated by the camera before each shot; - the brightness levels of the camera plane

(indicated automatically by the camera);

- the historicity of the respective values of the parameters listed previously. T _s can be chosen at a value close to the highest temperature measured by the infrared camera; then T ₃ can be adjusted as follows: the actual site method is tested to determine the concordance between the number of objects actually present in the target zone and the number of persons determined by the method according to the invention with this value. T ₃ .

Then, for each grid unit 6, the difference between the average temperature on the surface unit and the threshold reference temperature is compared. It is considered that if this difference is greater than a value used as a decision criterion, which is empirically defined for each type of object (for example 20% for humans), then the density in the grid unit 6 is considered significant; otherwise, it is considered insignificant. Example 2 below illustrates this process.

Next, the number N of grid units having a significant density is counted, and this number of grid units is multiplied by a typical density of objects per relevant grid unit. This typical density is an empirical value that must be defined for each type of object and for each dimension of grid units. It can also depend on the site, ie the type of crowd or flow. For example, it can be considered for a human crowd that in a grid unit of 1 mx Im, there is an average of 4 people in a queue type configuration. This value may vary according to the type of crossing point: it will be larger for a queue in front of a cinema than for a queue in front of a supermarket checkout, where the presence of shopping carts keeps people away from each other. others. For a circulatory flow of motorized land vehicles, a grid unit of 1 mx 1 m is typically used, and a density of 0.175 for passenger cars, 0.1 for small urban vehicles, and 0.25 is typically used. for high-end vehicles.

The method used optionally to determine the typology of the objects will be explained later. This method aims at determining, for a given surface area, the nature of the objects present on this same surface, insofar as this nature can be deduced from thermal data. For example, we can distinguish between small passenger vehicles and highly motorized passenger vehicles.

The fourth step gives an average number of objects (N _mθ yen) present on the target zone.

Step 5: In a fifth step, the waiting time t _a and / or the affluence rate A is calculated using a second algorithm.

The waiting time t _a is calculated as follows: t _a = t _P (1 + N _mean in X t _P / G) where G is the number of wickets, or more generally the number of points by which the objects leave the target zone 1, serving the waypoint 2, and t _P is the transit time; this last parameter is specific to each waypoint and can be determined independently.

The affluence rate A (t) is calculated according to the formula A (t) = [N input (t) - N _output (t)] / Cma _x given above, where N input (t) is determined from t _a and N _SO rti _e is calculated taking into account the average residence time in the site.

Step 6: In an optional sixth step, at least one of the parameters N _moγ in / t _a and A is transmitted to at least one user and brought to his attention by at least one visual or acoustic means, and preferably a way that allows to deduce the temporal evolution of at least one of these parameters.

In general, in the context of the present invention, the crossing point may be for example a wicket, a payment point, a junction, a roundabout, a meeting point, a point of sale, a checkpoint, control (including an identity checkpoint, a baggage checkpoint, an access authorization checkpoint or entry ticket), a distribution station, a door, a turnstile, a flow rate of said objects. For example, the method can be used to determine the number of humans present in a target area located in or in front of an amusement park, museum, stadium, concert hall, theater, theater, cinema hall, conference room, historical monument, tourist site, public or private administration, exhibition hall, airport, railway station, bus station, port area, or in the vicinity of any of these places. The method can also be used to determine the number of land vehicles in or in front of a parking lot, a road junction, a roundabout, a mandatory stopping point, a waiting area, a tunnel, a bridge, or in the vicinity of one of these places.

The method according to the invention can be used to determine a number of objects, for example a number of humans, animals or land motor vehicles, present in a given area area. This number is only an approximate number, but for most targeted applications, this accuracy is sufficient. This number of objects can be exploited in different ways, for example to estimate the waiting time on the tourist sites of a city or a district, to estimate the wait time to the attractions of a park of attractions, in front of the aperture of an administration or in front of a cinema, to estimate the rate of affluence of these sites, to characterize the intra-urban and other circulatory circulatory flow. The possibility of acquiring an infrared image sequence makes it possible to calculate these parameters in real time, and thus makes it possible to determine the trend of this waiting time or this affluence rate.

This information, especially the waiting time and its trend, that is to say its time evolution, are useful for the users of said crossing points and can be brought to their attention by any appropriate means, including visual or acoustic, such as a display device (especially a display screen) or an automatic acoustic message. Thus, the users can adjust their behavior so as to minimize their waiting time at this crossing point, by presenting themselves at the said crossing point at the moment when the waiting time displayed by a display device is low. The term "users of a crossing point" refers to the people who make up the crowd or flow in question, or who drive or operate the vehicles that make up the objects that make up the crowd or the flow.

This information, in particular the number of objects, the waiting time and the trend of these two parameters, are also useful for the operators of said crossing points, who may, for example, decide to limit access to said waypoint, diverting the flow of objects to another waypoint, or lengthening the path of objects forming said flow in front of the waypoint.

The method according to the invention also makes it possible to determine the typology of the number of these objects. More precisely, according to certain thermal characteristics detectable on an infrared image, it is possible to determine for each detectable type of objects its fraction. The thermal characteristics detectable and usable in the context of the present process are in particular: the thermal emissivity, the geometric dimensions, the measured temperature, the surface on which this temperature is measured.

For example, for humans, the thermal emissivity depends in particular on their dress and certain physiological characteristics (especially male / female, adult / child / elderly).

For motorized land vehicles, the thermal emissivity is difficult to exploit because it depends mainly external material. On the other hand, it is easy to distinguish between light passenger vehicles and highly motorized passenger vehicles: on the one hand, the first ones are generally smaller in size, on the other hand, the latter have a higher measured temperature. , and this temperature comes from a larger engine.

Similarly, it is possible to distinguish between a vehicle with the engine stopped and a vehicle with the engine stopped (which corresponds to the difference between a vehicle stopped in a traffic jam and a parked vehicle).

The data generated by the method according to the invention can be transmitted in real time or in deferred time to one or more users, and shared with one or more users simultaneously, sequentially or in any other way. Users can be on different sites or not. Users can be the operators of the crossing points, or the users of the crossing points, or any other person interested in this crossing, such as: the police, the professional actors, the general public, the tour operators. This transmission and sharing can be done through a web interface, a website available to users of the crossing points. The consultation of the data by the operators of the crossing points or by the users of the points of passage may be subject or not to a restriction of access and / or the payment of a fee. It can be done by any appropriate means, for example by an intelligent voice server, by sending text message or SMS messages, by displaying on screens and in particular on screens integrated in dynamic urban furniture. These Data can be archived and statistically processed to predict wait times and inflow rates.

Another object of the present invention is a method of exploiting data generated by the first method, in which:

(a) determining for at least one site S at least one parameter selected from the group consisting of the number of objects, the waiting time of said objects in front of the waypoint, the affluence rate, the trend of at least one of the three aforementioned parameters;

(b) at least one parameter is broadcast to at least one visual and / or auditory information broadcast station capable of bringing said at least one parameter to the knowledge of at least one user.

This visual and / or auditory information broadcast station is preferably selected from the group consisting of: a display panel, a display screen, a fixed or mobile phone, a fixed or portable computer, a speaker, a television set.

The data usable by this method can also come from another source capable of providing these data, such as the individual counting using a tourniquet or any other means, or the analysis of video images based on the visible light.

In this method, which is illustrated schematically in FIG. 4, the data obtained via different visual and / or auditory information broadcasting stations, such as the Internet, local networks, the voice server, text messages or SMS messages, display screens, display screens for dynamic urban furniture.

In this method is displayed on at least one visual information broadcast station for each of a plurality of (M) sites (Si, ..., S _M ), or for a subset of said (M) sites, simultaneously or successively at least one parameter of the same type, selected from the group selected from the group consisting of the number of objects, the waiting time of said objects in front of the waypoint, the affluence rate, the trend of at least one of the three aforementioned parameters, so as to allow the user to compare said parameters.

For each of the sites and its sub-sites (for example: a counter, a cash desk, a traffic light), a reference surface is determined, based on the average passage time, which defines the target area of measurement for each camera. This reference surface as determined corresponds to a precise waiting time.

Yet another object of the present invention is a computer network comprising a plurality of infrared cameras and a plurality of computer machines, to allow the centralized exploitation of the data generated by this plurality of cameras installed on a single site or on a plurality of sites. . This networking of the data generated by the methods according to the invention can be done according to several embodiments. In a first embodiment, there is at least one site S, and possibly several (M) sites Si, .. S _M. Each site is observed by at least one infrared camera C and possibly several (N) cameras Ci, ... C _N. Thus, the camera Ci, ₂ is the first camera of the second site.

For each site S, this camera, or this plurality of cameras, is connected by a wireless network or a wired network to a first computer machine R _s of the router type. This router R ₃ is advantageously a router type VPN (Virtual Private Network), to ensure a high level of security. This embodiment is shown schematically in FIG. 4.

In an advantageous variant of this embodiment, the plurality of cameras is hosted on a local area network (for example a virtual private network (VPN) type network) in order to guarantee a high level of security, and the router R ₃ can provide interface between the VPN and the outside internet world so that the bilateral connections between the local network and the operating software, advantageously installed on the computer machine P mentioned below, are completely secure.

To establish the connection between the camera and said network

(Preferably a network type LAN) can be used for example a RJ45 type cable or a wireless connection type WiFi or WiFi / GPRS. Said network may advantageously apply a known protocol such as that standardized by the IEEE 802.3 standard, known under the trademark Ethernet. The computer machine R, or the computer machines Rs = i, • », Rs = M are connected by wireless and / or wired teletransmission means, such as the Internet, to a second computer machine P, which centralizes the processing site-specific data.

This second computer machine P can then be connected, by means of wireless and / or wired teletransmission, such as a telephone network, to stations known as user stations U to which it distributes, spontaneously or on request of said users, messages. This embodiment is shown schematically in FIG. 5. This communication between the computer machine P and the user stations U can possibly be done via computer machines V which ensure the distribution of said messages. By way of example, V may be a server dedicated to telephone messages, or messages of the SMS type, or an internet server, and U may be a fixed or portable phone, a computer, a portable message receiver of the type email (e-mail).

In another embodiment, which provides a better level of computer security than the previous one, is inserted in the local network for each site S between the cameras C and R routers a computer machine E server type. This server E processes all the data from the cameras C and returns to the computer machine P only the final parameters calculated. This embodiment is shown schematically in FIG. 6. Then, in both embodiments, the final data generated (in particular: the number of objects, the time waiting for said objects in front of the waypoint, the affluence rate, the trend of at least one of the three aforementioned parameters) can be broadcast to users, for example to the operator of the waypoint and / or the user of the crossing point, and / or to other persons interested in the crossing point. This diffusion can be done in real time or delayed, by any appropriate means, as indicated above, for example by means of a visual and / or auditory information broadcast station, such as a billboard panel. display, a display screen, a landline or mobile phone, a desktop or laptop computer, a speaker, a television set.

Advantages of the invention

The invention has many advantages. The measurement method according to the invention can be implemented automatically, and the data can be transmitted automatically to an operating center. The data generated by the method according to the invention improves the predictability of waiting times and affluence rates in front of sites. Thus, it is possible to reduce waiting times and affluence rates significantly. Tourists and tour operators can optimize their visits according to waiting times and day-to-day attendance rates, and can plan their journeys (and therefore their costs) in the short and long term thanks to archived data from previous seasons. The tourist and circulatory fluidity in the city and in particular near the tourist sites is improved, because the communication of the waiting times and the affluence rate of the different sites in real time allows actors to adapt their behavior; this leads to natural self-regulation of circulatory flows. This adaptation of the behavior of the actors to the circulatory flows presents numerous advantages for certain individuals, in particular for the people with reduced mobility. This improvement in circulatory flow also has ecological benefits.

The following examples illustrate embodiments of the invention, but do not limit its scope. Examples;

Example 1 Description of a Typical Infrared Camera A Thermovision ™ A20-M type camera provided by the FLIR Systems company (spectral band: 8-12 μm, Peltier effect stabilized focal plane matrix, consisting of 19,200 test points was used. measurement, thermal sensitivity better than 0.1 ⁰ C at 30 ⁰ C, capable of measuring a temperature between -2O ⁰ C and + 250 ⁰ C, with a calibration ensuring accuracy and repeatability of the measurement of ± 0.5 ⁰ C at 35 ° C), a ThermaCAM ™ type lens supplied by FLIR Systems with a 19 ° opening, and a HEB type case. With this equipment and the parameters chosen, at an optical distance D of 100 m from the target area, one pixel corresponds to a square of 6 cm side.

Example 2

This measurement was performed at night outside, at an ambient temperature of 8 ⁰ C, in front of a crossing point in front of which waited up to 20 people. The target area was a rectangle of dimension 2m * 3.7m. The camera was positioned at a distance to the ground L of the target area of 12.7 m at an optical distance D of the target area of 24.4 m and a height H of the target area of 20.8 m. The emissivity (defined as the ratio of radiation emitted by a real body on radiation emitted by a black body) was 0.96. In this simulation, the average passage time at the waypoint was 60 seconds. The people were arranged in two rows. We recorded with the infrared camera (objective of 12 °, 160 * 120 pixels, temperature range + 5 ° C to +30 ⁰ C) two series of 14 shots. The first set consisted of winter dress conditions (warm clothing, from the head (hat or hat) to the feet (closed shoes or boots)). The second set represented summer dress conditions. Tables 1 and 2 summarize the results obtained. The first column (I) shows the number of people actually in the queue. The second column (II) indicates the number of the grid unit, which was 1 mx 1 m in size. The third column (III) indicates for each grid unit the average temperature (in ⁰ C) measured using the infrared camera. The ambient temperature Ti measured by infrared thermography in the focusing plane was 9.50 ^° C. The reference temperature T ₃ was 27.0 ^° C. In this case, the decision threshold beyond which we consider that a grid unit is populated was 1.2 x Ti. The fourth column (IV) shows the number of people calculated from these data. It can be seen that the agreement between the number of people calculated and the number of people actually present is good but not perfect. In practice, one usually does not need a perfect match. The concordance can be improved by choosing a unit Grid of smaller dimension, for example 0.25 mx 0.25. It can also be seen that this method can be implemented under unfavorable environmental conditions, and especially during the night.

Appendix 1: Function used for human subjects

Inputs: tair // room temperature Outputs: tmin // minimum temperature of subjects tmin = 0.71 * tair + 8

Annex 2: Algorithm for the automatic determination of the threshold temperature T ₃ (called "threshold" in the algorithm):

Inputs: temp [] // the temperature table thresholdprec // the previous threshold tfondprec / / previous background temperature Variables: background [] // the background pixels form [] // the pixels of the form nfond // number of individuals in the background number of individuals in the deep form // the prototype of the protoform background // the prototype of the form pnfond // previous number of individuals in the background pnforme // previous number of individuals in the finite form // should we stop? nbiter // number of iterations Outputs: threshold // the threshold to be calculated

nbiter = 0 finite = false nfond = 0 nform = 0 // initial prototypes. We guide research by not choosing at random. protofond = MIN (temp) protoform = MAX (temp) while (ifini) && (nbiter <10) background. empty () 5 bottom = nfond; nfond = 0 form. empty () pnforme = nform; nform = 0 10 for each indiv in temp // assignment of individuals to a class if indiv <40 ⁰ C if norm (indiv - protofond) <norm (indiv

- protoform)

15 bottom.add (indiv) nfond + = 1 otherwise form. add (indiv) form + = 1

20 so finsi finpour

// calculation of new prototype protofond = sum (background) / nfond; Protoform = sum (form) / form; if (nfond == pnfond) // it works also with nforme and pnforme fini = true; Repeat if protoform <skin_threshold (protome) threshold = protome * thresholdprec / tfondprec otherwise threshold = (protoform + 2 * protome) / 3 35 finsi

Claims

A method of determining the number of objects of a given type present in a given area area (called target area 1) using an infrared camera 3, comprising the following steps: a) acquiring an image at With the aid of an infrared camera 3, b) the target zone 1 is subdivided into a plurality of units called grid units 6, c) the average temperature T _M is calculated for each grid unit 6, d) is compared for each grid unit 6, this average temperature T _M at a predetermined reference temperature T _s , e) for each grid unit 6, the difference between the temperatures T _M and T _s is determined, f) the number N is determined. of grid units having a difference between the temperatures T _M and T ₃ greater than a reference value previously chosen according to the type of object, said grid units being called "relevant grid units", g) multiplying the number N by a previously determined coefficient, which expresses the average density of objects of a given type in a so-called relevant grid unit, to obtain the number of objects of a given type present in the target zone 1.

2. Method according to claim 1, characterized in that (i) said objects are part of a plurality of objects, in the form of a crowd of objects or a flow of objects, and in that

(ii) said target zone 1 is a waiting or slowing zone in front of a crossing point 2 that said objects must cross.

3. The method of claim 1, further comprising a step of calculating the waiting time of said objects in front of said waypoint 2, and / or a step of calculating the inflow rate of said objects in said target zone 1.

4. Method according to claim 3, characterized in that said waypoint is selected from the group consisting of: a wicket, a payment point, a crossroads, a roundabout, a meeting point, a point of sale, a checkpoint (including an identity checkpoint, a baggage checkpoint, an access clearance checkpoint or entry ticket), a distribution station, a door, a turnstile, a flow rate limiter of said objects.

5. Method according to any one of claims 1 to 4, characterized in that said objects are humans, that said target zone 1 is located in or in front of a place selected from the group consisting of: amusement park, museum, stadium, concert hall, theater room, auditorium, cinema room, conference room, historical monument, tourist site, public or private administration, exhibition hall, airport, railway station, bus station, port area , or around any of these places.

6. Method according to any one of claims 1 to 4, characterized in that

(i) the said objects are motor land vehicles, and in that

(ii) said target zone 1 is located in or in front of a location selected from the group consisting of: a car park, a road junction, a roundabout, a mandatory stopping point, a waiting area, a tunnel, a bridge , or around one of these places.

A computer network for acquiring and operating the data generated by the method according to any one of claims 1 to 6, comprising: (a) a plurality of N infrared cameras (Ci, ..., C _N ) arranged in at least one site S and possibly several (M) sites (Si, ..., S _M ),

(b) a first router-type computer machine R ₃ per site S, connected to the plurality of N infrared cameras by a data transmission means, and preferably an instant data transmission means,

(c) a second computer machine P connected to each of the first computer machines R ₃ by a data transmission means, and preferably an instantaneous data transmission means, said machine P centralizing the processing of the data acquired on each site S,

(d) a plurality of user stations U connected to the second computer machine P by means of remote data transmission, and which receive messages from said computer machine P.

8. Computer network according to claim 7, further comprising for each site S _x server computer machine E _x , and wherein said server type computer machine processes the data acquired by the cameras Ci, _x , ... C _N , _{X of} said site, and transmits to the computer machine P the result of said data processing.

The method of exploiting data generated by the method of any one of claims 1 to 6, wherein:

(a) at least one site S determines at least one parameter selected from the group consisting of: - the number of objects, the waiting time of said objects in front of the waypoint, the affluence rate, the trend at least one of the three aforementioned parameters;

(b) the at least one parameter is broadcast to at least one visual and / or auditory information broadcasting station capable of carrying the said at least one parameter to the knowledge of at least one user, the said information broadcasting station visual and / or auditory being preferentially selected from the group consisting of: a display panel, a display screen, a fixed or mobile phone, a fixed or portable computer, a speaker, a television set.

10. The method of claim 9 wherein is displayed on at least one visual information broadcast station for each of a plurality of (M) sites (Si, ..., S _M ), or for a subset said (M) sites, simultaneously or successively at least one parameter of the same type, selected from the group selected from the group consisting of: the number of objects, the waiting time of said objects in front of the waypoint, the rate of affluence, the trend of at least one of the three aforementioned parameters, so as to allow the user to compare said parameters.