CH715882B1 - Acoustic detection instrument and method for noise detection and identification of a vehicle. - Google Patents

Abstract

The present invention relates to a method of noise detection and identification of a moving vehicle and comprises: identifying the passage of a noisy vehicle; switching on a sound level meter of a portable acoustic detection instrument (1); aiming said sound level meter (20) in the direction of a noisy vehicle (10); measuring by the sound level meter (20) the sound volume of the vehicle (10) when it is in a first position P1; measuring the distance (D) between the sound level meter and one of the surfaces of the vehicle (10) which is in the first position P1; proceed to a first determination of the normalized sound intensity (11) related to a predetermined reference distance (Dr) relative to the vehicle (10); identifying the vehicle (10) if the corresponding normalized sound intensity (I1) emitted by the vehicle at the predetermined reference distance (Dr) is higher than a predetermined maximum value not to be exceeded. The invention also relates to an acoustic detection instrument (1) suitable for implementing said method.

Description

Technical area

The present invention relates to the field of noise pollution, as caused by noisy vehicles. The invention relates more particularly to methods and systems intended to measure the distance to a sound source and to determine its intensity with respect to a reference frame. The invention also relates to the recognition and identification of a noisy vehicle.

State of the art

[0002] Noise pollution is a recurring problem in modern times. A huge amount of motor vehicles, like cars, trucks, mopeds, trains and others are driving everywhere every day. Most of these vehicles are noisy because they use internal combustion engines. The pollution induced by the sound of vehicles can have serious nuisances on health. The range of frequencies emitted can vary greatly depending on the type of vehicle. From near ultrasound to infrasound, a wide range of wavelengths can be a source of stress or have pathological consequences, depending on the intensity, the duration of the noise and the sensitivity of people to see animals.

[0003] In some countries, vehicle noise is a major source of noise pollution. In France, for example, the human impact of transport noise pollution alone is estimated in 2016 at at least 50 billion euros per year. There are also special situations such as residential neighborhoods or areas close to hospitals or nursing homes where excessive noise produced by vehicles may be difficult to tolerate or even strictly prohibited.

[0004] To limit this, laws and standards have been put in place, in most countries, for all approved vehicles, in terms of noise emissions.

[0005] However, said standards are not necessarily respected. Some people change parts between noise checks on their vehicle so they can be louder than allowed, and they are therefore difficult to identify.

[0006]There is therefore a need to acquire devices making it possible to detect whether a vehicle is in violation of the standards, directly at the time of the facts of the offense and to identify it. This information would make it possible to carry out a check without giving the owner time to redo the modifications.

[0007] There are a number of instruments for measuring the noise of vehicles, but these instruments are fixed in a certain location and measure the noise of a passing vehicle at a range of relatively limited distances relative to the sound measuring instrument. . One of the disadvantages of such systems is that drivers can easily reduce their speed when passing near them, as is done when approaching fixed speed cameras to measure speed.

[0008] The situation becomes more complicated when the vehicles in question are not vehicles which move on a road or a rail but are aircraft such as, for example, helicopters or drones which will increasingly be no longer used for the transport of goods or people. Fixed or semi-mobile systems are not at all suitable for this type of vehicle.

Disclosure of Invention

The invention relates to a method and an acoustic detection instrument for measuring the sound volume of a vehicle which produces noise above a defined standard. The vehicle is understood in the broad sense and can include not only vehicles such as cars, motorcycles or trucks but any type of vehicle whether motorized or not.

In particular, the invention is embodied by a method for recognizing an acoustic signal from a moving vehicle. The process of the invention consists of: identify the passage of a noisy vehicle; engage a sound level meter of a portable acoustic detection instrument; aim the sound level meter in the direction of said noisy vehicle; measuring the sound volume of the noisy vehicle when it is in a first position; measuring the distance between the sound level meter and one of the surfaces of the vehicle which is in the first position; determining the sound intensity (11) of said noisy vehicle at a reference distance (Dr); identifying the vehicle (10) if the corresponding normalized sound intensity (I1) emitted by the vehicle at the predetermined reference distance (dR) is higher than a predetermined maximum value not to be exceeded.

[0011] In one implementation of the measurement of the sound level, the measurement of said distance, as well as the identification of the vehicle are done simultaneously.

[0012] In an advantageous implementation, after said sound measurements made in said first position, the sound level of the vehicle as well as the distance of the vehicle from the sound level meter are measured in a second position of the vehicle.

[0013] In a variant, a second value of the normalized sound intensity (12) is determined.

In a variant, the average of said first (11) and second (12) value of the normalized intensity is determined, said average being determined for a spectrum of predetermined frequencies.

[0015] In one embodiment, the identification of the vehicle is also made in said second position of the vehicle.

[0016] In a preferred mode, the vehicle is identified by identifying the vehicle registration number.

[0017] In a variant, the identification of the vehicle is done by recognizing the shape and/or the brand of the vehicle.

[0018] In a variant, the vehicle is identified by recognizing the vehicle registration number and the shape and/or brand of the vehicle.

[0019] In a variant, the vehicle is identified by recognizing an identification code integrated into the vehicle, this identification code being other than the vehicle registration number.

[0020] In an advantageous mode, an alarm signal is sent to the vehicle indicating that the vehicle is too noisy.

[0021] Alternatively, the alarm signal is displayed inside the vehicle in order to alert the driver of the noisy vehicle.

[0022] In an advantageous mode, said acoustic instrument comprises means for determining and recording the sound spectrum of the noisy vehicle.

The invention is also achieved by an acoustic instrument for recognizing an acoustic signal produced by a vehicle, the acoustic instrument includes a sound level meter and also includes: means for measuring the distance of the vehicle from the sound level meter; a unit for storing acoustic thresholds corresponding to maximum values not to be exceeded; a man-machine interface for displaying a representation of the corresponding standardized emitted sound level emitted by the vehicle at a predetermined reference distance (Dr), means for identifying the vehicle.

[0024] In one embodiment, said means for identifying the vehicle comprise a vision system making it possible to recognize the license plate of the vehicle.

In one embodiment, said means for identifying the vehicle comprise a vision system making it possible to recognize the make and/or the type of the vehicle.

[0026] In one embodiment, the acoustic instrument also includes a display of a representation of the non-normalized sound level.

In one embodiment, the acoustic instrument comprises at least two sound level meters and/or two means for measuring the distance from the vehicle relative to the sound level meter.

In one embodiment, the means for measuring the distance from the vehicle relative to the sound level meter are optical means.

[0029] In one embodiment, the optical means comprise a laser.

[0030] In one embodiment, the means for measuring the distance from the vehicle comprise a radar system.

[0031] In one embodiment, the means for measuring the distance of the vehicle comprise a system using triangulation techniques.

In one embodiment the sound of a vehicle can be detected up to a distance of 100m.

[0033] In one embodiment, the acoustic instrument is the size of a mobile phone, preferably having a length between 15cm and 20cm, a width between 7cm and 9cm and a thickness between 5mm and 15mm.

[0034] In one embodiment, the means for identifying the vehicle comprise lenses and the focal length of at least one of the lenses of the instrument has a focal length greater than the thickness of the instrument.

Brief description of the drawings

Other features of the present invention will appear more clearly on reading the following description, made with reference to the accompanying drawings, in which: Figure 1a is a schematic view of the use of an acoustic instrument according to the invention; FIG. 1b is a schematic view of the use of a miniaturized acoustic instrument according to the invention; FIG. 2 is a schematic view of an aircraft comprising an identification code and the implementation of sound and distance measurements made at two positions of the aircraft; Figure 3 shows a block diagram of the method of the invention; Figure 4 is a schematic view of a noise reference measurement setup; FIG. 5 a block diagram of the components and their communications for an audio measurement; Figure 6 is an electronic diagram used to amplify the audio signal; Figure 7 shows the directivity of a commercial microphone; Figure 8 is a schematic view of a signal aliasing effect during sampling; FIG. 9 is an electronic diagram used with the PIC microcontroller; Figure 10 is an electronic diagram used for reading/writing SRAMs; Figure 11 and Figure 12 show RAM write and read signals at SP; FIG. 13 is a block diagram of the components of the acoustic instrument as well as their communication bus; Figure 14 is a block diagram of the behavior of a PIC microcontroller during sampling; Figure 15 is a block diagram of the behavior of a Raspberry Pi depending on user action; Figure 16 shows experimental results obtained by a prototype of the invention; Figure 17 shows an optical parallax pattern for determining a distance from an object; Figure 18 illustrates the recognition of a vehicle registration; Figure 19 shows the relationship between the focal length of an optic and its angle of view; Fig. 20 shows a diagram of the relationship between an object's viewing angle, its dimension and its distance from an observation point; Figure 21 shows a miniaturized acoustic instrument; Figure 22 shows a diagram of the method of the sound measurement process and the identification of a noisy vehicle.

Embodiments of the Invention

As shown more specifically in Fig.1, the object of the invention relates to a new technique implemented using acoustic instrumentation to recognize an acoustic signal emitted by a vehicle. Vehicles such as cars, mopeds, trucks, even trains and planes must meet acoustic performance standards. There are different standards in different countries that allow the acoustic thresholds emitted by these vehicles to be set. These thresholds can be determined by several criteria and in general a threshold is determined by a sound level not to be exceeded at a certain distance, for example a distance of 10 m. The invention relates not only to the measurement and sound detection of a stationary vehicle but above all to the measurement of the sound intensity emitted by a moving vehicle. The invention relates in particular also to the identification of the license plate of the vehicle which would produce too much noise. The invention furthermore relates to a method for summoning the owner of the excessively noisy vehicle to present himself for an inspection of his vehicle.

[0037] It is further understood that the invention is not limited to motor vehicles such as, for example, an automobile, a motorcycle or a moped, a coach, a tram, a boat. The term „vehicle“ is here understood in its broadest sense and is here defined as any moving machine which produces noise, independent of its nature, its use or the source of the machine which is at the origin. noise to be measured by the method as well as the acoustic instrument 1 of the invention. The invention therefore also includes, for example, any other vehicle propelled by other means such as, for example, propulsion by the wind or gravity. The invention relates above all to moving vehicles but can also be applied, at least partially, to immobile or stationary vehicles, which includes noisy machines that are difficult to access, such as for example a drone which remains stationary at a great distance from the instrument. acoustic, for example a distance of more than 50m.

It is understood that the method and the measuring instrument of the invention are applicable to measurements of noise emitted by vehicles located at least 100m from the measuring instrument of the invention as described herein. document.

It is also understood that the sound level, also defined as sound intensity or measured noise intensity, can be expressed in decibels (dB) or in another measurement such as a pressure measurement expressed in Pascals for example.

Figure 1a illustrates an implementation configuration of the acoustic instrument 1 sound of the invention. In FIG. 1a, the noises represented by arrows B can come from different noise sources B3-B5. In general, the major contribution of noise generated by a vehicle 10 is produced by the engine of the vehicle 10. The noise of an engine can deteriorate with the misalignment of the vehicle 10 but also by the deterioration of the insulation under the protective cover of the engine. Other sources of noise can be added to the engine noise, such as those produced by worn tires or by the use of non-aerodynamic structures. The object of the invention is to determine the cumulative noise of all the noise sources of a noisy vehicle. Of course, the aim of the acoustic measurement is to reduce the noise of the vehicles 10, but also to reduce fuel consumption to protect the environment, since an engine that is too noisy will often be linked to excessive fuel consumption.

[0041] It is also included in the invention the detection of a vehicle 10 which includes any other source of noise which is not produced by the vehicle itself but produced by accessories such as sirens or loudspeakers, for cite only these 2 examples.

More specifically, the invention is embodied by a method for recognizing an acoustic signal B produced by a vehicle 10, in particular when this vehicle 10 is moving. This process comprises the following steps, illustrated in the block diagram of Figure 22: identifying a noisy vehicle 10; of course the passage includes the distance from the vehicle 10; of course it is understood that the noisy vehicle may be in motion or may be stationary; switching on a sound level meter 20 of a portable acoustic instrument 1 as described above; aim with the sound level meter 20 in the direction of said noisy vehicle 10; measuring the sound level emitted by the noisy vehicle 10 when it is in a first position P1; determining the distance D between the sound level meter 20 and one of the surfaces 10a, 10b, 10c of the vehicle 10 which is in the first position P1; determining a first normalized sound intensity I1 with respect to a determined reference distance Dr, illustrated in FIG. 1; such a normalized measurement can be determined by various means such as electronic and/or computer means as discussed later; identify the vehicle 10 if the normalized sound intensity I1 at the reference distance Dr is higher than an authorized predetermined standard, for example as described in https://publications.europa.eu/en/publication-detail/- /publication/fe98a835-e656-11e3-8cd4-01aa75ed71a1/language-fr.

In one embodiment, the measurement of the sound level, the measurement of said distance D, as well as the identification of the vehicle 10 are done simultaneously. It is understood that the term "simultaneously" is limited by what is possible to achieve taking into account the speeds of sound as well as the speed of the signal sent to the vehicle 10 and partially collected as in the case of a radar system using a laser. . It will be understood that the term "simultaneously" means that the measurement of the sound level, the measurement of the said distance D, as well as the identification of the vehicle are done in a period of time less than a second, preferably less than 500 ms or even more preferably less than 100ms (0.1 sec).

Figure 3 shows a block diagram of the method of the invention.

[0045] Figure 23 illustrates a diagram of the steps of the method of the invention.

When a user hears a noisy vehicle 10, stationary or moving, he turns on the instrument 1. This goes into stand-by. When the vehicle 10 passes, the instrument 1 is pointed at it and a trigger integrated preferably in a handle 1c of the instrument 1 is pressed. In a variant, the system 1 can be switched on automatically, for example according to the noise level corresponding to the sound volume detected. In a preferred mode, three data are taken simultaneously: the sound level coming from the direction pointed, the distance from the source pointed and an image of the vehicle 10, in particular of its license plate. It is understood that in variants the acoustic instrument 1 described below can comprise means to help the instrument 1 to be oriented towards a noisy vehicle 10. These means can comprise, for example, a heat source detector.

[0047] A healthy human ear can hear sound pressures ranging from 20 μPa to 20 Pa. This results in a ratio of 1:1,000,000, which is impractical in everyday use. A logarithmic scale of sound levels in decibels [dB] has therefore been implemented, which also better represents the perception of sound by humans. The sound propagating by wave having values sometimes positive, sometimes negative, it is necessary to take the effective value of this one over the duration of the measure to be able to have an average. The relation between the sound pressure, the effective pressure and a reference pressure is given by:

• p: sound pressure [Pa] • T: processed time interval [s] • prms: effective pressure [Pa] • pref: reference pressure [20 µPa] • Lp: sound level [dB]

When two sound sources emit at the same intensity, the sound level in decibels is not doubled, due to the logarithmic nature of decibels, the level will only increase by 3 decibels. This difference decreases if the additional noise is smaller than the base noise. Ltot = 10 log [(p1<2>+p2<2>+p3<2>+p4<2>)/p0<2>]; Ltot: Total sound level; pn: Sound pressure to be added. P0: Reference pressure (20 µPa)

For example, if a level of 10dB and 15dB are added together, the result will be 16.2 (the highest level (15) + 1.2). This particularity makes it possible not to have to isolate the measured sound source: the instrument being intended to probe the noisiest cars, they will cover the weaker surrounding noises.

Frequency weighting:

The ear can hear frequencies ranging from 20 Hz to 20 kHz; however, it has a different sensitivity depending on the frequency and this factor is taken into account in the measurement. A standardized weighting (A weighting of the IEC 61672-1 standard „Electroacoustics - Sound level meters“) has been established to have a better representation of the sound. This is the weighting used in most sound references, including legal automotive noise emission values

[0051] A sound level value evaluated in decibels according to A-weighting can be expressed in dBA, even if most devices and documents indicate it simply in dB. The equation corresponding to the transfer function is as follows: A(f) = 20log10(HA(f)) + 2.00 • HA(f): Gain as a function of frequency • A(f): Total gain [dB]

When the sound travels a distance, it becomes less and less audible. The loss of sound pressure over distance makes sound measurement unusable without the distance to the corresponding sound source. Thus, during sound volume measurements for controls and standards, a normalized distance is defined to be able to compare the results.

For an omnidirectional sound source, the sound intensity (expressed in W/m<2>) decreases according to the ratio 1/r<2>, where r is the distance traveled by the sound wave. The energy is distributed evenly over the sphere of the sound wave and the area of the latter increases according to the square of the distance. Since the sound intensity is proportional to the square of the sound pressure p. It follows that the sound pressure p, contrary to the intensity, decreases according to 1/r.

It thus makes it possible to establish the relationship p2/p1=r1/r2, or, in decibels:

In the method of the invention, the aim is to have the same sound measurement regardless of the distance between the acoustic instrument 1 and the vehicle 10. To do this, a reference distance dR must be defined. Several possibilities are of course possible. For example, it is possible to rely on sound measurement methods as carried out during automobile inspections according to the Swiss ordinance RS 741.41. (https://www.admin.ch/opc/fr/classifiedcompilation/19950165/index.html)

As illustrated in Figure 4 the vehicle 10 must accelerate to the maximum from line AA' to line BB'. The maximum intensity measured by the two microphones gives the result of the measurement. If we hypothesize that the car is a point sound source with omnidirectional emission, the maximum intensity will be measured at 7.5 m from the source. Measuring the distance with the vehicle 10 simultaneously with the sound recording and reducing the sound intensity to what it would be at 7.5 m would therefore make it possible to have a good approximation of whether or not the authorized values were respected.

There are different possibilities for determining a sound level at a certain reference distance, defined as a normalized level. For example, an electronic circuit can proceed to a correction of the sound level measured by the sound level meter 20 taking into account the measurement of the distance D between the sound level meter and one of the surfaces of the vehicle, knowing that the sound intensity decreases gradually by a factor which is inversely proportional to the square of said distance D. Said correction of the degression effect of the signal and the determination of the sound level at a reference distance Dr can also be done, or at least partially, by computer means.

It is well understood that the uncertainty of the measurement of the distance D between the sound level meter 20 and one of the surfaces of the car 10 will be determined by the section of the signal which is incident on said surface. In the case of a laser radar, for example, this section may be less than 500 mm<2>. In the case of a distance measurement of a vehicle such as a car or a motorcycle, the precision of the distance D measured is approximately 20cm over a typical distance of 30m.

Of course, means can be implemented to be able to measure the sound level at distances greater than 100 m, for example 200 m or 300 m. In this case, more sophisticated means must be deployed, such as filtering certain typical sound frequencies generated by the vehicle. In order to increase the sensitivity of sound detection, several sound level meters can be integrated into the acoustic instrument, these sound level meters 20 can have different directivity or sensitivity characteristics.

[0060] The standards for tolerated noise levels are of course different depending on the country. For example, according to Ordinance RS 741.41 of the Swiss law on emissions with regard to the sound volume of motor vehicles, the values authorized for cars are directly linked to the European standards in force. A sound volume measurement at a previously chosen reference distance is shown in figure 4, in which this distance Dr is 7.5m

In one embodiment, the sound level of the vehicle as well as the distance D of the vehicle with respect to the sound level meter 20 are also measured in a second position P2 of the vehicle 10, as illustrated in FIG. 2 where the vehicle represented is a drone 10.

[0062] In one embodiment, at least one second value of the normalized sound intensity I2 is determined. Of course, several values of the normalized intensity can be determined for one or more positions of the vehicle 10.

In a variant, the average of said first I1 and second I2 value of the normalized intensity is determined, said average being determined for a spectrum of predetermined frequencies. This makes it possible to obtain a more reliable value of the normalized sound noise.

In one embodiment, the identification of the vehicle 10 is done by identifying the registration number of the vehicle 10, which is typically done when the vehicle 10 is a car or truck or motorcycle or even mopeds. Flying machines can also include a registration number or a code which identifies them.

In a variant, the identification of the vehicle 10 is done by recognizing the shape and/or the brand of the vehicle 10. This can have several uses, for example during comparison tests of several types of vehicles, such as example cars or drones.

Of course, the invention is not limited to the recognition of a serial number 12. As illustrated, the identification of the vehicle 10 can be done by recognizing the serial number 12 of the vehicle 10 and the shape and/or of the vehicle brand 10.

In one embodiment the identification of the vehicle 10 is made by the recognition of an identification code C integrated into the vehicle 10. It is understood here that said identification code C is an apparent code on the vehicle or detectable by an instrument such as a bar code and may be other than the serial number 20 of the vehicle 10.

Said code C can be used to more easily recognize a vehicle. The C code may also be applied in a manner suitable to facilitate reading of the code such as in a fluorescent material. The C code can also be designed to reduce the risk that the vehicle owner will modify the vehicle to prevent the serial number or said code from being read, for example by using an anti-reflective layer or a layer that disperses the light coming from the code recognition system. . In the future, for example, vehicles 10 such as drones will probably also be provided with a registration number or a code C making it possible to identify their owner or additionally other properties of the vehicle 10. Said properties can be optical properties. Advantageously, said C code may comprise structures increasing the reflectivity. Said code C can be a bar code and can also comprise structures which reflect particularly well in a predefined spectral domain such as, for example, a range of wavelengths between 550 nm and 690 nm.

The distance measurement can be done by different methods. A first direct method consists in using a radar system as illustrated in FIG. 1. Such a radar can be a radar based on an optical system comprising a laser. Such radar systems are known and are not commented on in more detail here. The method of the invention can be implemented in various ways making it possible not only to determine the new sound but additionally to determine at the same time the speed or the distance traveled by a vehicle. For example, FIG. 2 shows a noise measurement situation of a drone 10 which is moving from a first position to a second position over a distance traveled P. By aiming at the vehicle in the two positions and by measuring the distances in the two positions P1, P2 it is possible to deduce the distance traveled by the vehicle 10 using an acoustic instrument 1 which comprises a GPS system or an accelerometer. In a similar way there are situations where the distance traveled P by the vehicle is known (for example a train at constant speed) it suffices to point the acoustic instrument 1 in two directions to be able to deduce the distances d1, d2 of the vehicle on the sound level meter. Of course, several distance measurement methods can advantageously be deployed by a single instrument 1, for example the combination of a triangulation technique and a radar measurement.

In an advantageous variant, an alarm signal can be generated and sent to the vehicle indicating that the vehicle 10 is too noisy.

In another variant, the alarm signal is displayed inside the vehicle 10 in order to alert the driver of the noisy vehicle 10.

In one embodiment, said acoustic instrument 1 comprises means for determining and recording the sound spectrum of the noisy vehicle 10 . This makes it easier to identify the origin of the noise produced by the vehicle, which can be advantageously used to identify an improvement in the vehicle to be communicated to its driver. In luxury vehicles, for example, information such as the spectral content of the noise detected by the acoustic instrument 1 can be transmitted to its driver or to the autonomous piloting system of the vehicle.

[0073] In an advantageous variant, the acoustic instrument 1 can comprise means for displaying the distribution of noise around a vehicle, similar to a thermographic analysis of heat leaks from a building. In this case, for example, red zones can visualize local overruns of at least 50% of the standard, an orange color when the sound level is between +30% and 50% of the authorized normalized sound level, and green zones would indicate areas where the noise level is well below the standards, for example less than 30% of the authorized standard.

The measurable noise level corresponding to the sound volume emitted by a source depends on the characteristics of the sound level meter, i.e. its directivity, and the measurable distance depends, i.e. on the technical means used to measure this distance. . With the development of commercial microphones and lasers it can reasonably be expected that a noise measurement at distances greater than 100 m can be determined.

The invention is also embodied by an acoustic instrument 1 for recognizing an acoustic signal produced by a vehicle 10. The instrument 1 of the invention corresponds to several desired criteria: the instrument of the invention must be light and quickly usable when the situation arises; the measurement must make it possible to have an estimate of the level of noise emitted to know whether the maximum emission standard is respected or not; the noise value must correspond to the noise perceptible by the human ear within a radius of 7 to 10 meters around the source; the acoustic instrument 1 does not necessarily have to be used as a reference for a fine. Above all, he must determine whether a vehicle must pass a noise emission test according to legal standards; the instrument must be usable up to a range of approximately 100 meters; the system must operate without being recharged for at least 8h to 10h, corresponding to one working day for an operator.

More specifically the acoustic detection instrument 1 of the invention, illustrated in Figure 1a, 1b, 2 comprises at least: means for measuring the distance D of the vehicle 10 relative to the sound level meter 20; a unit for storing sound level detection thresholds; correction means for determining the normalized sound intensity at a previously determined reference distance Dr; a man-machine interface 1a for displaying the normalized sound level emitted relative to a previously chosen reference distance Dr; means for identifying the vehicle 10.

In one embodiment, said correction means are electronic means or computer means. In a variant, said correction means are a combination of an electronic circuit and computer means.

Said correction means make it possible to take into account the noise attenuation effects (also defined as a degression effect). This correction is carried out by a processor just after a conversion into decibels with a digital filter. The formula used is the simple logarithmic addition mentioned above, which is none other than 1/ (r<2>) converted into a logarithm. It is performed in the „Data processing“ step illustrated in fig.3, or more precisely in „Calculating dBA + Display“ in fig.15. It is a physical calculation performed by the calculation unit (the processor) with the measured data. Unlike previous methods in this kind of calculation a programmed digital filter is used rather than an analog filter, as explained later.

In one embodiment, said means for identifying the vehicle 10 comprise a vision system 40 making it possible to recognize the license plate 12 of the vehicle 10 and/or to recognize the brand and/or the type of the vehicle.

In one embodiment, the acoustic instrument 1 also includes a display 1a of the non-standard sound level. The instrument 1 can comprise an interface 1b comprising means for programming parameters of the instrument such as a keyboard 1e. The instrument 1 can also include warning elements 1d which can be sound or optical elements to indicate the level of sound detected.

In one embodiment, the acoustic instrument 1 comprises at least two sound level meters and/or two means for measuring the distance D from the vehicle relative to the sound level meter 20. Said sound level meters and/or the two means for measuring the distance D can be different

In one embodiment, the means for measuring the distance D from the vehicle 10 relative to the sound level meter 20 are optical means.

[0083] In one embodiment, the optical means 4 comprise at least one laser.

In one embodiment, the means for measuring the distance D of the vehicle 10 comprise a radar system.

In one embodiment, the means for measuring the distance D of the vehicle 10 comprise a system using triangulation techniques, as illustrated in Figures 2, 17.

In an advantageous variant, the acoustic instrument 1 of the invention may comprise at least two sound level meters 40, which may be different. In another variant, the acoustic instrument 1 of the invention may comprise at least two distance measurement systems and/or at least two imaging systems. For example, one of the vision systems can have a wide viewing angle and the second can have a high magnification. Said two shooting systems can be integrated into a single one and include zoom optics. In a variant, this zoom lens can be triggered according to a measured noise level and according to a predetermined level programmed in the acoustic instrument 1.

It is understood that the acoustic instrument 1 of the invention can be implemented for any parked vehicle. In a mode of use where the vehicle is stationary, the distance measurement may be inactive.

In one embodiment, the acoustic instrument 1 is arranged so as to be able to carry out noise measurements at distances greater than 150m or even more than 200m.

In one embodiment, the acoustic instrument is the size of a mobile phone, preferably having a length between 10cm and 25cm, a width between 5cm and 15cm and a thickness between 3 and 20 mm.

Example of an embodiment of an acoustic instrument prototype 1 according to the invention:

Sound measurement implementation:

[0091] Figure 5 shows an example of the implementation of a sound acquisition

The prototype is based on a single-board nanocomputer, in particular a Raspberry Pi 2 (ref: https://fr.wikipedia.org/wiki/Raspberry_Pi). The following steps are executed.

When the user wants to take a measurement, the Raspberry Pi sends a signal to a PIC microcontroller. The amplified signal from a microphone is connected to an analog-digital converter AD of the latter, which samples it and stores it in a random access memory of the RAM type. When the PIC indicates that it has finished sampling, the Raspberry Pi retrieves the values from the memories and keeps them internally. The user can then choose to display the converted value in dBA or to create a .wav file to listen to or process the recording.

Microphone implementation

[0094] The first step in recording sound is to measure the output of a microphone. The prototype microphone is a model CMI-4537-SN69 from CUI inc. for its low price, unidirectional directivity and ease of use. Its signal is amplified using the Texas Instruments LM386 audio amplifier, as shown in Figure 6.

[0095] Resistor R1 and capacitor C1 constitute the classic diagram for supplying and withdrawing the DC component of the microphone. Potentiometer R2 allows to adjust the gain of the input, capacitor C2 allows to have the maximum gain of the amplifier (200) while capacitors C3 and C4 as well as resistor R3 are manufacturer's recommendations. Capacitor C5 is used to filter the input to cut out disturbances from communication between other components (see sampling). It forms an RC filter with the potentiometer with a cutoff frequency of around 26.5 kHz. This value varies with the value of R2, but these variations are not large enough in our operating range to have an influence on the output. A second RC filter at 25.5 kHz (R4 and C6) is placed at the output to eliminate the remains of noise which are amplified by the circuit.

[0096] The series capacitor proposed at the output by the manufacturer has been removed. The purpose of this is to remove the DC DC component from the signal. However, to simplify powering the system to a single DC DC power supply, this component should be retained. It should be noted that the output value of the LM386 is not rail-to-rail. For a 3.3 V supply, the output range extends from 600 mV to 2.6 V, with a central value of 1.6 V acting as an offset.

The microphone used has a unidirectional directivity as illustrated in FIG. 7: it picks up the sound well in front of it, slightly on the sides and almost not behind. High noise will drown out lower sounds. The use of a unidirectional microphone also makes it possible to attenuate sounds coming from directions other than the one pointed, which further improves the precision of the measurement.

Sampling

In order to be able to record analog values such as sound, sampling must be carried out. According to the Nyquist-Shannon theorem, the minimum sampling frequency should be 40 kHz (twice the maximum measured frequency, in our case the maximum audible frequency 20 kHz). However, there is a aliasing phenomenon, illustrated in figure 8, which can occur if signals exist above 20 kHz: they would be mirror-aliased on Fs/2 to end up on the audible band and generate noise.

To limit this, the sampling frequency is increased to 100 kHz. Frequencies likely to fold into the audible band would start at 80 kHz, frequencies at which signals are much better cut by filters. The sampling time is set to one second, a compromise between capture speed and data consistency: there will therefore be 100,000 samples per measurement.

PIC microcontroller implementation

[0100] A Raspberry Pi does not have a real-time clock and its operating systems do not allow this kind of application. This task is therefore entrusted to a second processor. Due to its relative ease of use and access to the necessary programming hardware, a PIC32 MX170F256B microcontroller from Microchip was used. This one has an internal 10-bit AD converter, an SPI port and the ability to use an external oscillator to accelerate its modules.

The PIC32 shown in Figure 9 contains more modules than it has legs; it offers the possibility of choosing the legs on which the modules are used. This includes the programming module (RB0 and RB1), the AD converter (RB3) and the SPI (RB11, RB13, RB14). The „Busy“ and „Start“ signals, used to communicate with the Raspberry Pi, as well as the RAM selection are placed on conventional inputs-outputs.

RAM memory implementation

The microcontroller does not have enough RAM memory to store all the samples. The sampling being done on 10 bits, the data would be stored on 2 bytes. This amounts to a total amount of data of 200,000 bytes for one second at 100 kHz. To store these values during sampling, it is thus necessary to have recourse to very fast external memories. Two 128 KB 23LC1024 SRAM memories communicating in SPI (SPI: a serial communication interface, in English Serial Peripheral Interface), illustrated in figure 10, were chosen for the prototype there was no memory in DIL box (for prototyping) having enough room to store all the data in just one of them.

The two memories are connected in parallel on the SPI bus, they are each selected in turn by the processor.

The measurement of a write to RAM in SPI is illustrated in figure 11.

[0105] The MOSI and SCLK symbols in figures 11 and 12 are defined by: • MOSI: Master Out Slave In: The SPI uses a master-slave system, which means that a device makes the requests and gives the orders (master) and that the other can only respond (slave). The MOSI pin represents the output pin for the master and thus the input pin for the slave (line where the orders are sent); • MISO: Master In Slave Out: Input pin for the master and output pin for the slave (line where the responses are sent); • SCLK: Serial Clock: wire on which rotates a „clock“ (ie a clock, periodic square signal of a defined frequency) used to synchronize the communication.

In the above example, the instruction arrives first (0x02 for a write), then the 24-bit address concerned, and finally two data bytes. Memory is configured to operate sequentially; when an additional byte is written without reordering, it will be written to the next address. In this example, 0x01 will be written to address 0x00D772 and 0xFF to 0x00D773. The data is on 16 bits (10 bits from the AD converter rounded up), so it is stored on two bytes. The address must therefore be incremented two by two.

Reading happens in a similar way to writing and is illustrated in FIG. 12: the instruction is sent (0x03 for reading), then the address. Then the clock sends two groups of eight clock ticks to collect the two bytes on a MISO line not shown here. The first will be that of the requested address, and the second that of the following address.

Raspberry Pi Implementation

As shown in the block diagram, the Raspberry Pi must communicate with the PIC microcontroller and with the RAMs as shown in Figure 13. This has an integrated SPI module as well as various inputs-outputs. The Raspberry Pi also has a UART port which will be used when implementing the distance sensor. This part is discussed in the chapter dedicated to it.

For better accessibility and flexibility during testing, software has been developed on a Linux operating system (Raspbian). This one was programmed using QT Creator to be able to have a user-friendly graphical interface. This allows testing and modifying the code, as well as processing the results directly on the operating system. The Raspberry Pi of the prototype requires a keyboard and a mouse connected to the USB ports as well as a screen on the HDMI port, which will no longer be necessary during the industrialization of the system.

PIC Programming

[0110] When the system is switched on, the main function waits until the Raspberry Pi asks it to perform a sampling (FIG. 14). As soon as this sets the „Start“ signal to 1, the PIC sets the „Busy“ signal to the high level. This is followed by the initialization and start-up of the SPI, ADC and timer modules, as well as the launch of the AD conversion and the initialization of the RAMs. The main function then goes back to waiting until the end of the sampling. During this time, the timer will call its interrupt every 10 µs to recover the values. During the interruption, the value of the AD converter is first read and the PIC is asked to restart a conversion. Starting the conversion at this time allows it to be done in parallel with sending data to RAM, which saves time. Then follows the selection of the RAM with which you want to communicate in SPI, then sends the value read from the converter. The RAM address is then incremented by two (see section „RAM memory implementation“). Half of the 16-bit values (50,000) are sent to the first RAM. When it's done (address = 100'000), the second RAM is selected, the address is reset and the process starts again. When 50,000 new values have been sent to the second RAM, the timer and the SPI communication of the PIC have stopped and it indicates that the sampling is finished. The PIC can therefore reset the „Busy“ signal to the low level to allow the Raspberry Pi to recover the values and continue the work. When the hand on the SPI is released, all the legs relating to the SPI are configured as input to avoid having an influence.

Programming the Raspberry Pi

The operating diagram of the essential functions of a Raspberry Pi of prototype 1 is illustrated in FIG. 15, namely an acquisition function, producing a sound.wav file, playing a sound.wav file, converting to dBA. The user interface for managing audio capture consists of four buttons. The acquisition button sends a signal to the PIC to perform the sampling, then retrieves the data for internal storage. Two buttons are for reading and writing to a file*. wave. The last one applies the A-weighting filter to the data and outputs the value in dBA. ;Function „Acquisition“;[0112] Pressing the acquisition button starts the measurement process. First, a "Start" line connected to the PIC is pulled high to indicate that it should start sampling. This responds by setting the „Busy“ line to the high level for the duration of the acquisition. The Raspberry Pi is waiting for „Busy“ to come back low so that the data is all stored in the RAMs. Then, it activates the SPI, reads 100,000 bytes from each RAM in groups of two and stores them in an array of 100,000 16-bit signed values. The data is 10 bits and has an offset (see section „Microphone implementation“). The offset is removed, then the values multiplied to cover the 16 bits (-32'768 to 32'767). ;[0113] In the measurement above, the low active selection of the RAM memories is displayed during acquisition. During a read or write, the RAMs are quickly selected, then deselected. We can see in the first part the use of the RAMs by the PIC for the sampling of one second (500 ms per RAMs), then in the second part the taking of the relay by the Raspberry Pi to read the two RAMs. Only the first part should be in real time, the second being only data recovery. After sampling, the "Convert into dBA" function is called automatically. ;Function „Make sound.wav“;[0114] For experimentation and control reasons, a file *.wav (Waveform Audio File Format), a simple playable audio file is created from the computer. This format is composed according to the following table: 0 1 2 3 4 5 6 7 8 9 10 11 ChunklD ChunkSize Format 12 13 14 15 16 17 18 19 20 21 22 23 Subchunk1ID Subchunk1Size AudioFormat NumChannels 24 25 26 27 28 29 30 31 3 2 33 34 35 SampleRate ByteRate BlockAlign BitsPerSample 36 37 38 39 40 41 42 43 44 45 46 47 Subchunk2ID Subchunk2Size data 0 data 1 48 49 50 51 52 53 54 55 56 57 58 59 data 2 data 3 data 4 data 5 data 6 ...

Each number represents a byte. The file is separated into three sections (chunk), each containing its name (ID), size and content. The first is the RIFF (Resource Interchange File Format) header, common to the file corresponding to this model. This is followed by the format section, indicating the characteristics of the file. Finally, the data section contains all the audio values of the file.

[0116] The RIFF Header includes: • ChunkID: Contains the letters [RIFF] in ASCII (0x52494646) • ChunkSize: Total file size excluding the first two fields in bytes. The size of the first two sections being constant in our case, the value will be 36 + SubChunk2Size for an uncompressed file (PCM), which gives for our file 200'036 (0x30D64). • Format: Contains the letters [WAVE] in ASCII (0x57415645)

[0117] Section format: • Subchunk1ID: Contains the [fmt] characters in ASCII (0x666d7420) • Subchunk1Size: Size of the section except the first two fields in bytes. For a classic PCM file, the size is 16 (0x10). • AudioFormat: Encoding format. The value for a typical uncompressed sampled file is 1. • NumChannels: Specifies the number of channels. For mono as in our case, the value is 1. • SampleRate: Sampling frequency, here 100'000 (Ox186AO) • ByteRate: Number of bytes per second. 16-bit values on a channel at 100,000 Hz gives 200,000 (0x30D40). • BlockAlign: Number of bytes per block. 16 bits on 1 channel gives 2. • BitsPerSample: Number of bits per sample, here 16 (0x10).

[0118] Data section: • Subchunk2ID: Contains the letters [data] in ASCII (0x64617461) • Subchunk2Size: Size of the section except the first two fields in bytes. For 100,000 values on 16 bits, the size is 200,000 (0x30D40) • Data: The sampled values are then written one after the other here. In 16 bits, the values must be signed.

[0119] When the button is pressed, the program creates a sound.wav file (or modifies it if it already exists) in the compilation folder, writes the header according to predefined values, then inserts the stored data internally. The file can then be played by an audio playback program or opened in Matlab.

"Read sound.way" function

[0120] Given that we only work with a single file format in the prototype (1 second at 100,000 16-bit samples), reading the file is very simple: the program looks for a sound file. wav in the compilation folder, skips the header and reads the data. These then become usable as if they had just been measured with the PIC.

"Convert into dBA" function

This button makes it possible to apply the weighting filter A to the table of values, to recover the effective value of the latter and to scale it to have a value in dBA. This filter is necessary to have a sound value representative of what the human ear hears. There were two possibilities to implement the filter: • Analogically, at the microphone output; • Digitally, from the Raspberry Pi.

[0122] The digital method was preferred. It allows the filter to be tested on the computer before implementation to ensure accuracy and saves components. This also makes it possible to generate .wav files, the original unfiltered values being available on the program.

[0123] To avoid having to make too large a correction when applying the corrective offset, the values are first scaled by unit to go from -1 to 1. application of the filter: this has been resized on Matlab to become a digital filter with a sampling frequency of 100 kHz. After applying the filter, the rms value is calculated and converted to dBA. Finally, an offset representing the gain of the microphone is defined at the time of calibration so that the results correspond to reality.

Digital filter

The digital filter was first generated on Matlab using the base filter cited in theory. The sample rate has been set to 100 kHz and the filter looks like this:

The base filter being of the 6th order, it was separated using Matlab into 3 chained filters of 2nd order. The reason is that the higher the order of the filter, the more precise its coefficients must be. For comparison, the graph below compares the analog filter (blue), the 6<th>order digital filter (red) and the three 2<th>order digital filters put in series.

The coefficients of the digital filters are all 32 bits; the Raspberry Pi processor being on 32 bits, it allows to be faster for the calculation. The three 2<th> order filters prove to be sufficiently precise in the range used (20 Hz to 20 kHz), so it is not necessary to use coefficients on more bits. The three filters used are:

The whole must also be multiplied by a gain of 0.1753793 to be similar to the analog filter. To be able to program these filters, they must first be transformed into difference equations. The purpose of the operation is to be able to calculate each of the output values from the current and previous values. For example, to obtain the 10<th> output value, we apply coefficients on the input values 10, 9 and 8, as well as the output values 9 and 8, then we add the results. The filters being of order 2, we use values up to k - 2, k being the index of the value we want to calculate. The exponent of z in the equations represents the index shift. Since it is impossible to use future values, the first step is to negative the exponents by dividing the numerator and denominator by z<2>:

[0128] It is then necessary to isolate the output of interest, that is to say Y(z): Y1(z) = U(z<-1>) + 0.5884756 U(z<-2>) + 0.9295781 Y (z<-1>) – 0.2160289 Y(z<-2>) Y2(z) = U(z) – 2.0002375 U(z-<1>) + 1.0002375 U(z<-2>) + 1.9479506 Y(z<-1>) – 0.9482561 Y(z<-2>) Y3(z) = U(z) – 1.9997631 U(z<-1>) + 0.9997632 U(z<-2>) + 1.9974139 Y(z<-1>) – 0.9974147 Y(z<-2>)

Finally, we can replace z by the index k to obtain the final equations: Y1[k] = U[k–1] + 0.5884756 U[k–2] + 0.9295781 Y[k–1] – 0.2160289 Y[k–2] Y2[k] = U[k] – 2.0002375 U[k–1] + 1.0002375 U[k–2] + 1.9479506 Y[k–1] – 0.9482561 Y[k–2] Y3[k] = U[k] – 1.9997631 U[k–1] + 0.9997632 U[k–2] + 1.9974139 Y[k–1] – 0.9974147 Y[k–2]

Each of the functions therefore represents the way of calculating a specific element of the array. To program these functions, it is now sufficient to calculate each of the 100,000 values of the first output array with the values to be filtered as input, then of the second by taking the first as input, and finally of the third with the second as input. To finalize the application of the A-weighting filter, the previously mentioned gain must be applied to each of the values in the table.

Test of the prototype of the acoustic instrument 1 of the invention, outdoors

The prototype of the acoustic detection instrument 1 was tested in a real situation and aimed above all at testing the application of the attenuation calculation. To avoid disturbing, the test is carried out in the forest rather than in the city, the test requiring that a source of noise must be strong enough to drown out the others.

Material:

[0132] 1 Bushnell Tour v3 Laser Distance Sensor 1 Sound Source (Lawn Mower) 1 Roline RO-1350 sound level meter

Measurement setup:

The vehicle 10 in question for the test was a lawn mower placed on the side of the road, turned on. The person carrying out the measurement stands at certain distances defined using the laser sensor, then measures the sound volume captured. Measurements are taken further and further away.

Measurement result:

The measurements were taken every 10 m up to 70 m. The distance measurement with this laser has an accuracy of ±1 m, while the sound level meter remains within ±1 dB of fluctuation. The measurements are shown in the table below as well as in Figure 17.

Table 1

[0135] Distance [m] 1 10 20 30 40 50 60 70 Sound [dBA] 81 66 60 55 52 50 48 47

It can be seen that the attenuation in dBA resembles the 1/r type in the shape of the curve, which suggests that the attenuation in Pascal is of the 1/r type in accordance with the theory. This also means that the closer you are to the source, the more the tolerance on the distance will have an impact and the further you are from it, the more the tolerance on the sound volume will have an impact.

To better understand this phenomenon, here is a table in which each line represents a measured value reported over all the other distances (the first line is the value measured at 1 m reported over the other distances, the second is the value measured at 10 m, etc.). The following table shows outdoor measurements over all test distances.

Table 2

[0138] Measurement at 1 m [dBA] 81 61 54.9794 51.45757 48.9588 47.0206 45.43697 44.09804 Measurement at 10 m [dBA] 86 66 59.9794 56.45757 53.9588 52, 0206 50.43697 49.09804 Measurement at 20 m [dBA] 86.0206 66.0206 60 56.47817 53.9794 52.0412 50.45757 49.11864 Measurement at 30 m [dBA] 84.54243 66.0206 58 .52183 55 52.50123 50.56303 48.9794 47.64046 Measurement at 40 m [dBA 84.0412 64.54243 58.0206 54.49877 52 50.0618 48.47817 47.139241 Measurement at 5 0m [dBA] 83.9794 63.9794 57.9588 54.43697 51.9382 50 48.41638 47.07744 Measurement at 60 m [dBA] 83.56303 63.56303 57.54243 54.0206 51.52183 49.58 362 48 46, 66106 Measurement at 70 m [dBA] 83.90196 63.90196 57.88136 54.35954 51.86076 49.92256 48.33894 47

It can be seen that the further one moves away, the more the calculations correspond to the measurements. This indicates that the distance tolerance has more influence than the noise tolerance and that with a more precise distance measurement, one could take measurements closer to the source.

Fixed vehicle movement

It is necessary to take into account the movement of the noise source being measured. The sound measurement is done over one second, the vehicle will thus be farther from the measuring instrument at the end of the sampling than at the beginning. At long distances, there will be little difference in sound volume, but this becomes significant the closer you get to the source. The error can however be partially corrected by taking into account the temporality of the distance measurement. For the duration of the measurement, the sound attenuation can be approximated as a straight line. The average calculated for the sound would thus correspond approximately to the central value of the sampling if the sound volume emitted by the car is constant, and the average output by the instrument would therefore represent the sound received from the vehicle when the sampling is halfway. accomplished. It is therefore at this moment that the distance measurement must be taken to limit the error.

Limitation of tolerances

The tolerance on the distance measurement has less influence the further away. At 10m ±1m can be a few dB, while at 60-70m the difference is barely perceptible. If we go back to the attenuation calculation:

The value r2 is the distance at which the sound volume is to be related and r1 is the measured distance. The tolerance of L1 is fixed and corresponds to the precision of the instrument's loudness measurement. Since r2 is a constant (7.5 m), it is therefore necessary to have a large value of r1 (and thus a small tolerance value in %) to improve the accuracy of the measurement. It is therefore preferable to measure at medium or long distance. The disadvantage of long distance measurement is the increased influence of noise. As the measured sound volume decreases, ambient noise will have more influence on the measurement. This phenomenon is however limited by the fact that the element measured is defined as being the noisiest in the surroundings. However, the wind or too great a distance can still distort the results.

Conclusion

The test described above gives an idea of the performance of an acoustic instrument 1 of the invention. The results showed that despite the potential disturbances from the outside, the theoretical calculations remain applicable. The limit of sound level detection depends on the intensity of the sound source and the implementation of optical and electronic improvements to the instrument.

In the following chapter, various aspects of an embodiment of a distance measurement module are commented on.

Distance measurement

The acoustic instrument must be able to indicate, from a distance of up to 100 m, the sound produced by a car at 7.5 m from it. For this, it is necessary to know the distance between the vehicle and the person measuring.

[0146] Distance measurement up to 100 m has a limited number of possibilities and two of them (methods A, B) have been analyzed in more detail.

Comparison between two images (method A), illustrated in FIG. 17: in this technique, two photo instruments are separated by about ten centimeters taking an image simultaneously. The difference between these two images makes it possible to calculate the distance separating us from the object in question. This method has proven to be complicated with state-of-the-art techniques and is not the best choice due to the complexity of the image analysis compared to the second solution (B).

[0148] Laser sensor (method B: this method consists of using an industrial laser sensor for distance measurement. These have a high range, high precision and standardized interfaces. This solution seems preferable for the acoustic instrument 1 The details of this system are described below.

Laser distance measurement (method B)

To confirm the viability of the laser solution, tests were carried out with a Bushnell Tour v3 laser sensor, initially designed for golf. It has been confirmed that it is possible with this technology to measure the distance of even distant moving vehicles. On this finding, a laser from the same firm was sought. Looking for a compact laser that can be used with the Raspberry Pi, the choice fell on the Truyens S210. The S210 is a long distance laser that can measure from 46 cm up to 750 m for an accuracy of ±4 cm. It has a UART interface to be programmed and used with the Raspberry Pi. It measures 104.4 × 81.7 × 41.6 mm, which makes it integrable into a portable instrument. It also has a laser pointer that allows you to target the vehicle you want to measure. It is powered from 12 to 24 V, so a battery in this voltage range will need to be used.

[0150] For lack of a better alternative currently on the market, the laser is overqualified for the acoustic instrument 1. It would be preferable in the future to use a less expensive laser with a shorter measurement range (100-150 m ) as far as possible.

Example of capturing an image: case of a license plate

To be able to have a follow-up to the detection of an excessively noisy motorist, it is necessary to be able to find him. It was discussed to insert a camera to capture the license plate of the vehicle at the time of measurement.

In the case of the integration of a camera, it is necessary to check the resolution that the latter must have so that the license plate is readable. First of all, it is necessary to define the minimal resolution that the plate in itself can have. For this, a license plate image was shrunk on the computer until it was no longer readable.

[0153] The plate in the image illustrated in FIG. 18 above has a width of 55 pixels. A photo taken in action will be of lower quality, so an increase to a minimum of 100 pixels is best as an approximation. A camera is defined by its resolution and the focal length of its lens. A standard lens has a focal length of 50mm, which represents the human field of view. The more the focal length increases, the more the field of vision decreases, and vice versa. For the same camera, photographing a plate at 100 meters will have greater resolution with greater focal length. Figure 20 illustrates the viewing angles according to the focal length of the imaging optics.

[0154] By setting the maximum distance for capturing the image at 100 meters and knowing that a license plate has a size of 50 cm x 11 cm at the back, we can calculate the minimum resolution according to the focal length what must a camera have so that the plate represents 100 pixels at 100 meters away. This is illustrated in figure 21. Using the following formulas, we obtain the results shown in the table above for different focal lengths:

Table 3

[0155] Focal length (mm) 400 (6°) 200 (12°) 100(24°) 80 (30°) 50(47°) Min width (px) 2096.311171 4204.169411 8502.262467 10717.9677 17392.4951

It may be noted that with a conventional focal length (50 mm), a very high resolution is required to be able to read a license plate. With a focal length of 200mm or higher, the numbers become more reasonable. A 4k resolution for the camera would theoretically be enough to read a plate at 100m. However, keep in mind that the higher the focal length, the more precise you have to be when capturing the image. It would be necessary to carry out tests to verify that this factor is not too disabling during use in a real situation. To be able to use the device in any light, it may be considered to take infrared photographs if color is not necessary to recognize the offending vehicle.

[0157] In conclusion, the production of the prototype of the acoustic instrument 1 and the tests carried out have shown that the improvements must relate to the elimination of the use of a screen and to replacing the Raspberry Pi with a more suitable processor. to an embedded system.

Miniaturization of the acoustic instrument 1 of the invention

In one embodiment, the acoustic instrument 1 of the invention has a miniaturized size in order to have a typical dimension of a mobile telephone.

Figure 21 shows a non-exhaustive embodiment of such an embodiment. Dimensions and components shown in Figure 22 are examples only and do not correspond to actual dimensions. The different elements can have different aspect ratios according to variants. Also, the relative location of the various elements of the exemplary embodiment may vary and other elements may be added in the miniaturized acoustic instrument of Figure 21. Specifically, in order to reduce the dimensions of the instrument various paths optics can be deviated using mirrors or beam splitters. In the embodiment of FIG. 21, two vision systems are integrated into the acoustic detection instrument 1. A first system is based on a short focal length lens 40a and a second system is based on a lens having a long focal length. The integration of at least two systems of different focal lengths makes it possible in particular to be able to capture an image with a wide field of view and an enlarged image of a detail of the vehicle-

FIG. 21 shows a section along a line A-A' of the optical system having a long focal length. The light picked up by the lens 48b is deflected by 45° by an at least semi-reflecting mirror and is transmitted to an image detector 404 by a second mirror 403. Such a configuration makes it possible to have a large focal length in a acoustic detection instrument 1 which has a thickness e which can be 10% to 50% smaller than the focal length of at least one of the lenses 34, 40b, 32 of the system. In one embodiment one of the optical paths, for example the paths 402, 406 in FIG. 21, comprises a beam which is deflected by 90° in the acoustic detection instrument 1. In a variant, using several mirrors , at least one of the optical paths could be designed to achieve a length of 2 x L + y x w, where L is the length of the instrument and w is its width, y being a predefined number. By choosing for example y = 1.5 in the case of the optical system comprising the lens 40b, the detector 404 would be close to the lens 40b and 3 deflection mirrors would be positioned in the corners inside the acoustic detection instrument 1. To optimize the geometric design, zigzag-shaped optical paths can be implemented. The dimensions of a portable instrument are preferably: a length between 10cm and 25cm, a width between 5cm and 15cm and a thickness between 3mm and 20mm. Other dimensions are of course also possible. It is understood that certain detectors such as that of the receiver 34a of the backscattered light of the optical radar as well as the optical transmitter, for example a laser 32a can be arranged in the plane of the instrument. or in thickness as shown in Figure 21.

In a variant, the optical instrument comprises an electronic layer which is placed on the back of the instrument as shown in the example of FIG. 21. Of course, the electronics and the necessary power supplies can be placed in the spaces gaps between the different elements. The acoustic detection instrument 1 may include a viewfinder 60 and may also include two viewfinders being arranged to allow binocular vision to the observer. Of course, certain elements such as the viewfinder or at least one of the lenses of the instrument can be arranged to allow movement of these elements. For example the viewfinder can be retractable.

In a variant embodiment, other optical systems can be integrated into the instrument, such as an infrared flash, for example. Of course, the instrument can include an external battery 5000 making it possible to increase the autonomy of the instrument.

The acoustic detection instrument 1 of the invention as well as the method for detecting noise and determining a normalized sound intensity with respect to a reference distance Dr of a vehicle can be applied in the field industrial. For example, the instrument and the method of the invention can be used for acoustic measurements of an installation of machinery producing noise. In particular, the acoustic detection instrument 1 of the invention is useful for measuring, controlling or monitoring the noise produced by noisy installations which are difficult to access. These installations can be static or include parts or elements that are in motion.

Claims (23)

1. A method of noise detection and identification of a moving vehicle (10), characterized in that it consists of: – identify the passage of a noisy vehicle (10); – activate a sound level meter of a portable acoustic detection instrument (1); – aim with said sound level meter in the direction of said vehicle (10); – measuring, by said sound level meter, the sound volume of said vehicle (10) when it is in a first position (P1); – measure the distance (D) between the sound level meter and one of the surfaces of the vehicle which is in the first position (P1); – determining the normalized sound intensity (11) of said vehicle (10) at a predetermined reference distance (Dr); – identify the vehicle (10) if the corresponding standardized sound intensity (I1) emitted by the vehicle at the predetermined reference distance (Dr) is higher than a predetermined maximum value not to be exceeded. 2. Method according to claim 1, characterized in that the measurement of the sound volume, the measurement of said distance (D), as well as the identification of the vehicle are done simultaneously. 3. Method according to claim 1 or 2, characterized in that, after said sound measurements made in said first position (P1), the sound volume of the vehicle as well as the distance (D) of the vehicle relative to the sound level meter (20) are measured in a second position (P2) of the vehicle (10). 4. Method according to claim 3, characterized in that at least a second normalized sound intensity (12) of the vehicle (10) is determined at said reference distance (Dr) relative to said vehicle (10). 5. Method according to claim 4, characterized in that an average of said first (11) and second (12) value of the sound intensity, normalized with respect to said reference distance (Dr), is determined, the said average being determined for a spectrum of predetermined frequencies. 6. Method according to one of claims 3 to 5, characterized in that the identification of the vehicle (10) is also made in said second position (P2) of the vehicle (10). 7. Method according to one of claims 1 to 6, characterized in that the identification of the vehicle (10) is done by identifying the registration number of the vehicle (10). 8. Method according to one of claims 1 to 6, characterized in that the identification of the vehicle (10) is done by recognizing the shape and / or the brand of the vehicle (10). 9. Method according to one of claims 1 to 8, characterized in that the identification of the vehicle (10) is made by the recognition of an optical identification code (C) integrated into the vehicle (10), this code optical identification (C) being other than the serial number (20) of the vehicle (10). 10. Method according to one of claims 1 to 9, characterized in that an alarm signal is sent to the vehicle indicating that the vehicle (10) is too noisy. 11. Method according to claim 10, characterized in that the alarm signal is displayed inside the vehicle (10) in order to alert the driver of the noisy vehicle (10). 12. Method according to one of claims 1 to 11, characterized in that said acoustic detection instrument (1) comprises means for determining and recording the sound spectrum of the noisy vehicle (10). 13. Acoustic detection instrument suitable for implementing the method for noise detection and identification of a moving vehicle according to one of claims 1 to 12, said instrument being an acoustic detection instrument (1) for measuring a sound volume produced by a vehicle (10), the acoustic detection instrument (1) comprising a sound level meter (20), characterized in that the acoustic detection instrument (1) comprises: - Means for measuring the distance (D) of the vehicle (10) relative to the sound level meter (2); – a unit for storing acoustic thresholds corresponding to predetermined maximum values not to be exceeded; – a man-machine interface for displaying the corresponding standardized sound intensity (I1) emitted by the vehicle (10) at a predetermined reference distance (Dr); - Means for identifying the vehicle (10). 14. Acoustic detection instrument (1) according to claim 13, characterized in that said means for identifying the vehicle (2) comprise a vision system (6) making it possible to recognize the license plate (20) of the vehicle. 15. Acoustic detection instrument (1) according to claim 13 or 14, characterized in that said means for identifying the vehicle comprise a vision system (6) making it possible to recognize the make and/or the type of the vehicle (10) . 16. Acoustic detection instrument (1) according to one of claims 13 to 15, characterized in that the acoustic detection instrument (1) also comprises a display of the sound volume of the vehicle (10) at said distance (D) . 17. Acoustic detection instrument (1) according to one of claims 13 to 16, characterized in that the acoustic detection instrument (1) comprises at least two sound level meters and/or two means for measuring the distance (D) from the vehicle relative to the sound level meter (20). 18. Acoustic detection instrument (1) according to one of claims 13 to 17, characterized in that the means for measuring the distance (D) of the vehicle (10) relative to the sound level meter (2) are optical means. 19. Acoustic detection instrument (1) according to claim 18, characterized in that the optical means (4) comprise a laser. 20. Acoustic detection instrument (1) according to claim 18 or 19, characterized in that the means for measuring the distance (D) from the vehicle (10) is a radar system. 21. Acoustic detection instrument (1) according to one of claims 17 to 19, characterized in that the means for measuring the distance (D) from the vehicle (10) comprise a system using triangulation techniques. 22. Acoustic detection instrument (1) according to one of claims 13 to 21, characterized in that its dimension has the size of a mobile phone, that is to say having a length (I) between 10 cm and 25 cm, a width (w) between 5 cm and 15 cm and a thickness (e) between 3 mm and 20 mm. 23. Acoustic detection instrument (1) according to claim 22, characterized in that the detection means for identifying the vehicle comprise several lenses (34, 40b, 32) and that the focal length of at least one of the lenses of the instrument (34, 40b, 32) is greater than the thickness (e) of the instrument.