FR3065814A1 - METHOD AND SYSTEM FOR CHARACTERIZING A IONIZING RADIATION SOURCE TO BE CHARACTERIZED - Google Patents

Abstract

The invention relates to a method for characterizing an ionizing radiation source to be characterized. The method comprises the following steps: providing and normalizing N time signals of reference pulses corresponding to a first type of particle; choosing at least one discriminating magnitude; calculating from the N time signals of pulses and the discriminant magnitude of a reference matrix; acquisition and normalization of M time signals of pulses acquired from the source of ionizing radiation to be characterized; calculating from the N time signals of pulses and the discriminant magnitude of a characteristic matrix; calculating a first characteristic value of dissimilarity between the reference matrix and the characteristic matrix; characterizing the presence of neutrons in a radiation emitted by the source of ionizing radiation to be characterized as a function of a result of the comparison between the first calculated dissimilarity characteristic value and a first predetermined threshold.

Description

Holder (s): COMMISSIONER OF ATOMIC ENERGY AND ALTERNATIVE ENERGIES Public establishment.

Extension request (s)

Agent (s): BREVALEX Limited liability company.

FR 3 065 814 - A1

104) METHOD AND SYSTEM FOR CHARACTERIZING AN IONIZING RADIATION SOURCE TO BE CHARACTERIZED.

The invention relates to a method for characterizing a source of ionizing radiation to be characterized. The method comprises the following steps: supply and normalization of N time signals of reference pulses corresponding to a first type of particle; choice of at least one discriminating quantity; calculation from the N time signals of pulses and the discriminating quantity of a reference matrix; acquisition and normalization of M time signals of pulses acquired from the source of ionizing radiation to be characterized; calculation from the N time signals of pulses and the discriminating quantity of a characteristic matrix; calculating a first characteristic value of dissimilarity between the reference matrix and the characteristic matrix; characterization of the presence of neutrons in radiation emitted by the source of ionizing radiation to be characterized as a function of a result of the comparison between the first characteristic value of dissimilarity calculated and a first predetermined threshold.

METHOD AND SYSTEM FOR CHARACTERIZING A RADIATION SOURCE

IONIZING TO CHARACTERIZE

DESCRIPTION

TECHNICAL AREA

The present invention relates to a method and a system for characterizing a source of ionizing radiation to be characterized, as well as to a data medium readable by a computer on which is recorded a program including the instructions for executing steps of the method for characterizing d 'a source of ionizing radiation to characterize.

The invention applies to the detection and characterization of sources of ionizing radiation.

The terminology of “ionizing radiation” must be understood above and in the rest of this document in its broadest understanding, thus including both so-called “indirectly ionizing” radiation, such as gamma photons and neutrons, as so-called “ionizing” radiation as such, such as beta β particles (electrons and positrons) and alpha a particles (protons).

PRIOR STATE OF THE ART

With the context of international terrorism and the risks linked to the use of radioactive materials for the manufacture of so-called “dirty” bombs, the development of tools to detect and identify these radioactive materials has become a priority. These radioactive materials are generally identifiable by the fact that they constitute sources of ionizing radiation emitting neutrons.

The usual methods for detecting ionizing radiation comprising neutrons generally require the use of the isotope 3 of helium ³ He and of its capacities to absorb neutrons. However, with the increasing need for detection of radioactive material, there is a shortage of helium ³ He isotopes 3. It is therefore necessary to develop alternative methods for the detection of ionizing radiation not using the isotope 3 of helium ³ He.

Unlike the methods using the helium ³ He isotope 3, the alternative methods have the disadvantage of being as sensitive to neutrons as to gamma radiation. However, gamma photons are generally preponderant over neutrons and can be emitted by perfectly legitimate sources, such as, for example, bananas.

Among these alternative methods to the use of the helium isotope 3, those based on organic scintillators, as described by GHV Bertrand and his coauthors in their work published in 2015 in the scientific journal "Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment ”, volume 776 pages. 114 to 128, are particularly promising. Indeed, organic scintillators are relatively inexpensive and they make it possible to detect all of the neutrons, without being limited to thermal neutrons only.

However, with such scintillators, the discrimination between signals from neutrons and those from gamma photons remains particularly problematic. Indeed, this discrimination, if it is based on flight time, requires the use of two separate scintillators. Recent work, such as that of A.Buffler and his coauthors published in 2016 in the scientific journal “International Journal of Modem Physics Conférence Sériés” volume 44 pages 1660228, suggests, in order to eliminate this need to call upon two scintillators, d '' Use discrimination by computer-assisted pulse. Such discrimination requires calibration from signals produced by the interaction only of neutrons which are relatively difficult, if not impossible, to obtain, if not by pre-separation with a time-of-flight technique.

Thus the use of organic scintillator to characterize sources of ionizing radiation, in particular to discriminate the signals produced by the interaction of neutrons, requires a simple method of discrimination, that is to say not requiring the use of two scintiIlators in parallel or a calibration from signals produced by the interaction only of neutrons.

It should be noted that such a problem is not limited to organic scintillators only, this also applying to inorganic scintillators, and is also observed for the discrimination of other types of ionizing radiation, such as discrimination between sources of ionizing radiation, the ionizing radiation of which includes gamma y photons and beta β particles or alpha a particles.

STATEMENT OF THE INVENTION

The invention aims to at least partially resolve the above drawbacks and thus in particular aims to provide a method for characterizing a source of ionizing radiation to characterize and determine or not the presence of neutrons emitted by said source, the method not requiring only one scintillator and not requiring access to ionizing radiation comprising only neutrons, or to carry out a neutron separation beforehand.

To this end, the subject of the invention is a method for characterizing a source of ionizing radiation to be characterized, the ionizing radiation of which comprises a first type of particle and is capable of comprising a second type of particle, the method comprising the following steps :

supplying N time reference pulse signals, each time reference pulse signal corresponding to the detection of the first type of particle by a scintillator, N being a non-zero natural integer;

- choice of at least one discriminating quantity;

- For each reference pulse time signal, extracting the value of each discriminating quantity chosen to form a reference value corresponding to said discriminating quantity and to said reference pulse time signal;

- Acquisition of M temporal pulse signals acquired, each temporal pulse signal acquired corresponding to the detection, by the scintillator, of a photon originating from the source of ionizing radiation to be characterized, M being a non-zero natural integer;

- For each acquired time pulse signal, extracting the value of each discriminating quantity chosen to form a characteristic value corresponding to said discriminating quantity and said acquired time pulse signal;

- calculation of a first characteristic dissimilarity value from a first dissimilarity calculation between a reference matrix and a characteristic matrix, the reference matrix being formed by the reference values, the characteristic matrix being formed by the characteristic values ;

- comparison of the first characteristic dissimilarity value calculated with a first predetermined threshold;

characterization of the presence of the second type of particles s in a radiation emitted by the ionizing radiation source to be characterized as a function of the result of the comparison between the first characteristic value of dissimilarity calculated and the first predetermined threshold, in which the step of choice of at least one discriminating quantity comprises the choice of R discriminating quantities, R being a non-zero natural integer, and the first of dissimilarity between the reference matrix and the characteristic matrix comprises the following substeps:

- for each discriminating quantity identified by an index i, subdivision of an interval representative of the values taken by the discriminating quantity in k, sub-intervals, k, being a non-zero natural integer;

- calculation, for each of the reference matrix and the characteristic matrix, of the value of Πί ^ ι kj coefficients 0, each coefficient θ being associated with R sub-intervals, each of said R sub-intervals being associated with a respective discriminating quantity and being equal to a sub-interval of said respective discriminating quantity, a first coefficient θι and a second coefficient 0 ₂ distinct being such that there exists at least one characteristic quantity for which the corresponding sub-interval of the first coefficient θι is different from the sub- corresponding interval of the second coefficient 0 ₂ , the value of each of the coefficients θ being calculated as the number of time signals of reference pulses, respectively of time signals of pulses acquired, the discriminating magnitudes of which are in the P associated subintervals to said coefficient, the values of coefficients θ calculated for one of the reference matrix and the characteristic matrix that being multiplied by a corrective factor corresponding to the ration of the number of time pulse signals between the reference matrix and the characteristic matrix;

- calculation of the first characteristic dissimilarity value from a difference in absolute value implementing the coefficients θ calculated from the reference matrix and the characteristic matrix and from R successive sums, each of the R sums being associated with a respective discriminating quantity, and said difference being made directly between each coefficient θ calculated on the basis of the reference matrix and the coefficient θ calculated on the basis of the characteristic matrix corresponding to the same R sub-intervals, the first sum being made on based on the result of said difference and the following R1 sums based on the result of the sum directly preceding it.

In fact, the inventors have discovered that by using at least one discriminant quantity and calculating a characteristic value of dissimilarity between a matrix of discriminant quantity values calculated for a source of ionizing radiation consisting of gamma photons and a matrix of discriminant magnitude values calculated for the source of ionizing radiation to be characterized, it was possible to discriminate the presence of temporal signals of pulses coming from neutrons.

In addition, the use of a dissimilarity calculation based on the division into sub-intervals of each discriminating quantity followed by N sums makes it possible to identify all of the dissimilarities between the population of time signals of reference pulses and the population of pulse time signals acquired. The discrimination is therefore particularly improved with respect to the characterization methods according to the prior art.

It can be noted that thus, with such a method, within the framework of a preferred application of the invention in which the first type of particles is that of gamma photons y and the second type of particles is that of neutrons, it is possible characterize a source of ionizing radiation to be characterized in order to determine whether this source of ionizing radiation emits neutrons. The method according to the invention therefore requires only a single scintillator and is carried out on the basis of a calibration based on a reference ionizing source emitting only gamma photons.

It may be noted that by characteristic value of dissimilarity, a value characterizing the difference between the N reference time pulse signals and the M time pulse acquired signals comprising the abovementioned calculation steps must be understood, this characteristic value being able to be a characteristic value of dissimilarity as such, such as a statistical difference between the reference values and the characteristic values, or a likelihood value, such as the inverse of a statistical difference between the reference values and the values feature. Of course, a person skilled in the art is able to use both a characteristic value of dissimilarity and a likelihood value, these values being generally linked to each other by a simple inversion calculation.

In the case where the characteristic value of dissimilarity is a likelihood value, the calculation of the characteristic value of dissimilarity may also include an operation of inversion of the result obtained after the N ^th sum.

It will be noted that the term reference matrix and characteristic matrix must be understood broadly and as an array of numerical values regardless of the representation actually used for the processing of the data to implement the method. according to the invention. Thus, it is perfectly conceivable, without departing from the scope of the invention, that the reference values, respectively the characteristic values, be manipulated in the form of N vectors, respectively M vectors, each vector comprising the set reference values, respectively characteristic values, extracted for a given pulse time signal.

The steps of extracting the value of each discriminating quantity, whether for the reference time pulse signals and for the acquired pulse time signals, can comprise at least one sub-step chosen from the following sub-steps :

normalization of each time pulse signal;

reduction of electronic noise of each time pulse signal.

Such sub-steps make it possible to obtain an easy comparison between the time pulse signals and therefore a calculation of a characteristic dissimilarity value which is perfectly representative of the dissimilarity between the reference time pulse signals and those acquired.

The at least one discriminating quantity can be selected from:

- an integral of at least a portion of the reference time pulse signal or acquired time pulse signal;

- a time characteristic of the reference time pulse signal or acquired time pulse signal;

- a value, in at least one instant, of the reference time pulse signal or time pulse acquired signal;

- a function of the variance of at least a portion of the reference time pulse signal or acquired time pulse signal;

- a value derived from the reference time pulse signal or acquired time pulse signal.

Such discriminating quantities can make it possible to highlight the distinguishing characteristics between the time signature pulse signals of an interaction of a particle of the first type and those of an interaction of a particle of the second type. Indeed, the integral of at least one portion can for example be representative of a secondary peak present only in the pulse time signals obtained for the particles of the second type while the characteristic time can be representative of a time of more important interaction for these same particles of the second type.

The terminology of "derivative value of a time pulse signal" should be understood broadly, thus encompassing, for example, the average derivative values over a portion of the time pulse signal, the maximum derivative values of the time pulse signal, and the derivative values of the time pulse signal at a given time.

The step of choosing at least one discriminating quantity may be subsequent to the step of supplying the N time signals of reference pulses and to the step of normalizing each time signal of reference pulses, the step of choosing at least one discriminating quantity comprising the following sub-steps:

supply of P time signals of comparison pulses, each time signal of comparison pulses corresponding to the detection, by the scintillator, of a particle of the first and or second type of particles emitted by a source of ionizing radiation emitting a ionizing radiation comprising particles of the first type and particles of the second type, P being a non-zero natural integer;

- construction of a plurality of sets of discriminating quantities;

for each reference time pulse signal, and for each set of discriminating quantities, extraction of the value of each discriminating quantity from said set of discriminating quantities to form a first comparison value corresponding to said discriminating quantity and to said time signal reference pulse;

- For each comparison pulse time signal, and for each set of discriminating quantities, calculation of the value of each discriminating quantity of said set of discriminating quantities to form a second comparison value;

- for each set of discriminating quantities, calculation of a second characteristic dissimilarity value from a second dissimilarity calculation between a first matrix consisting of the first comparison values extracted for said set of discriminating quantities and a second matrix consisting of the second comparison values extracted for said set of discriminating quantities;

- choice of at least one discriminating quantity corresponding to one of the sets of discriminating quantities, the choice being made on the basis of the second characteristic values of dissimilarity calculated for each of the sets of discriminating quantities.

During the sub-step of choosing at least one discriminating quantity corresponding to one of the sets of discriminating quantities, the at least one corresponding discriminating quantity chosen can be that which corresponds to the set of discriminating quantities for which the second characteristic value of dissimilarity presents an extremum characteristic of maximum dissimilarity.

With such sub-step of choosing at least one discriminating quantity, it is possible to optimize the discrimination, since it is possible to base the choice of this or these discriminating quantities on a maximization of the discriminating quantity.

The step of choosing at least one discriminating quantity may further include the following sub-steps:

supply of Q time signals of test pulses, each time signal of test pulses corresponding to the detection, by the scintillator, of a particle of the first type emitted by a source of ionizing radiation emitting an ionizing radiation comprising particles of the first type, Q being a non-zero natural integer;

for each test pulse time signal, and for each set of discriminating quantities, extracting the value of each discriminating quantity from said set of discriminating quantities to form a third comparison value corresponding to said discriminating quantity and to said time signal test pulse;

- for each set of discriminating quantities, calculation of a third characteristic dissimilarity value from the second dissimilarity calculation between a first matrix consisting of the first comparison values extracted for said set of discriminating quantities and a second matrix consisting of the third values of comparison extracted for said set of discriminating quantities;

- comparison of the third characteristic dissimilarity value calculated for each of the sets of discriminating quantities with a second predetermined threshold;

- selection from among the sets of discriminating quantities of those having a minimum dissimilarity from said comparison between the third characteristic value of dissimilarity calculated for each of the sets of discriminating quantities at a second predetermined threshold;

- choice of at least one discriminating quantity corresponding to one of the sets of discriminating quantities, the at least one corresponding discriminating quantity chosen being that corresponding to the set of discriminating quantities from the sets of selected discriminating quantities whose second value characteristic of dissimilarity presents an extremum characteristic of maximum dissimilarity.

With such sub-step of choosing at least one discriminating quantity, it is possible to optimize the discrimination without risk of erroneous detection, since it is possible to base the choice of this or these discriminating quantities on a maximization of the quantity discriminating while limiting the risks of false positive thanks to the calculation of the third characteristic value of dissimilarity.

At least one discriminating quantity can be noted Q _tot and calculated according to:

where s is the reference time pulse signal or the time signal d ', respectively the acquired pulse time signal;

t _max is the time coordinate of the maximum of the reference time pulse signal, respectively of the acquired time pulse signal; and ti is a time coordinate calculated according to:

tl i-max + PT-HM where β is a strictly positive real; and t _H M is the time coordinate, subsequent to t _max , for which the value of the reference pulse time signal, respectively of the acquired pulse time signal, is 0.5.

Such a value Q _tot makes it possible to discriminate the temporal pulse signals having a stronger, or weaker interaction, vis-à-vis those of the temporal pulse signals.

At least one discriminating quantity can be noted Q _t aii and calculated according to:

tl s (t) dt ^ 0 where s is the reference time pulse signal, respectively the acquired time pulse signal;

t ₀ is a time coordinate calculated according to:

to t _max + OC.t ^ jv [where a is a strictly positive real;

t _max is the time coordinate of the maximum of the reference time pulse signal, respectively of the acquired time pulse signal; and t _H M is the time coordinate, subsequent to t _max , for which the value of the reference pulse time signal, respectively of the acquired pulse time signal, is 0.5;

and ti is a time coordinate calculated according to:

tl t _max + p.tfijv [where β is a strictly positive real.

Such a value Q _t aii makes it possible to discriminate the temporal pulse signals having a delayed interaction with respect to those of the temporal pulse signals.

Each set of discriminating quantities comprises at least one of Q _to t and Q _tai |, the sets being distinguished two by two by at least one of the values a and β.

In this way, it is possible to define in an optimized way the limits of the sums used for the calculation of the value of the discriminating quantities and thus the characteristic value of dissimilarity which will be calculated.

The at least one discriminating quantity can be noted D and is the variance of the portion of the reference time pulse signal, respectively the acquired time pulse signal, between t ₀ and ti, where:

t ₀ is a temporal coordinate calculated according to: f-0 f-maxf ^ -f-HM where a is a strictly positive real;

t _max is the time coordinate of the maximum of the reference time pulse signal, respectively of the acquired time pulse signal; and t _H M is the time coordinate, subsequent to t _max , for which the value of the reference pulse time signal, respectively of the acquired pulse time signal, is 0.5; ti is a time coordinate calculated according to:

f-l f-max + P-f-HM where β is a strictly positive real.

Such a discriminant quantity makes it possible to characterize the presence of temporal pulse signals differentiating from the temporal reference pulse signals. This discriminating quantity is therefore particularly suitable for demonstrating a population of two types.

Each set of discriminating quantities can include D, the sets being distinguished in pairs by at least one of the values a and β.

Thus, it is possible to select the discriminating quantity D making it possible to provide a quantity of dissimilarity perfectly characteristic of the presence of particles of the second type.

During the step of constructing the plurality of sets of discriminating quantities, a given set of discriminating quantities can comprise S discriminating quantities, S being a non-zero natural integer, and the second calculation of dissimilarity between a first matrix and a second matrix, for the given set of S discriminating quantities, can include the following sub-steps:

- for each discriminating quantity identified by an index j, subdivision of an interval representative of the values taken by the discriminating quantity in kj sub-intervals, kj being a non-zero natural integer;

- calculation, for each of the first matrix and of the second matrix, of the value of Πρ = ι kj coefficients δ, each coefficient δ being associated with S subintervals, each of said S subintervals being associated with a respective discriminating quantity and being equal to a sub-interval of said respective discriminating quantity, a first coefficient δι and a second coefficient δ ₂ distinct being such that there exists at least one characteristic quantity for which the corresponding sub-interval of the first coefficient δι is different from the sub- corresponding interval of the second coefficient δ ₂ , the value of each of the coefficients δ being calculated as the number of time signals of reference pulses, respectively of time signals of acquired pulses, the discriminating magnitudes of which are in the S associated subintervals to said coefficient, the values of coefficients δ calculated for one of the first matrix and of the second matrix being multiplied es by a corrective factor corresponding to the ration of the number of time pulse signals between the first matrix and the second matrix;

- calculation of a characteristic value of dissimilarity from a difference in absolute value using the coefficients δ calculated from the first matrix and the second matrix and from S successive sums, each of the S sums being associated with a respective discriminant quantity, and said difference being made directly between each coefficient δ calculated on the basis of the first matrix and the coefficient δ calculated on the basis of the second matrix corresponding to the same R sub-intervals, the first sum being made on the basis of the result of said difference and the following S s based on the result of the sum directly preceding it.

Such a second calculation of a characteristic value of dissimilarity makes it possible to obtain a sum of differences corresponding to each sub-interval of the discriminating quantities. It is therefore easy with such a calculation to locate the presence of a second population of type of particles, such as the temporal pulse signals of particles of the second type, with respect to the first homogeneous population that constitutes the signals. reference pulse times which correspond to only one type of particles, namely the particles of the first type.

There may also be provided a step for estimating the number and / or a pulse rate corresponding to the second type of particle.

The first type of particles and the second type of particles can be selected from the group comprising:

- gamma photons as the first type of particles and neutrons as the second type of particles,

- gamma photons as the first type of particles and beta particles as the second type of particles,

- gamma photons as the first type of particles and alpha particles as the second type of particles,

- beta particles as the first type of particles and alpha particles as the second type of particles.

The invention further relates to a system for characterizing a source of ionizing radiation to be characterized, comprising:

a scintillator, a photomultiplier coupled to the scintillator, and a processing unit adapted to process the time pulse signals supplied by the photomultiplier and obtained during the absorption of ionizing radiation by the scintillator, the characterization system being in that the processing unit is configured to implement a characterization method according to the invention.

Such a characterization system has the advantages linked to the implementation of the method according to the invention.

The invention further relates to a computer-readable data medium on which a program including instructions for executing steps of the method according to the invention is recorded.

Such a data carrier has the advantages linked to the implementation of the method according to the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention will be better understood with the aid of the description which follows, given solely by way of nonlimiting example and made with reference to the appended drawings in which:

FIG. 1 illustrates a schematic view of a characterization system according to the invention,

- Figure 2 is a graph of two time averaged pulse signals obtained with an organic scintillator, the first time averaged pulse signal having been obtained with a source of ionizing radiation emitting only gamma photons, the second time signal average pulse having been obtained with a source of ionizing radiation emitting gamma photons and neutrons,

FIG. 3 is a graph of a time pulse signal obtained for a scintillator, the graph appearing of the values used for the calculation of discriminating quantities according to the invention,

FIGS. 4A and 4B are two graphs showing the distribution according to two discriminating quantities of respectively a set of pulses coming from a source of ionizing radiation comprising only gamma photons and a set of pulses coming from a source of ionizing radiation containing neutrons and gamma photons,

FIGS. 5A and 5B illustrate graphically the differences on which the intermediate calculation of a characteristic dissimilarity value according to the invention is based, this for a discriminating quantity Q _tot and a discriminating quantity Q _t ai, respectively,

- Figures 6A and 6B are graphs illustrating the variation of a characteristic value of dissimilarity for two discriminating quantities between a matrix of time signals of reference pulses produced by the interaction of gamma photons and respectively a matrix of time signals d comparison pulses produced by the interaction of neutrons and gamma photons and a matrix of time signals of comparison pulses produced by the interaction only of gamma photons,

- Figure 7 illustrates the result of an intermediate calculation for the calculation of a characteristic value of discrimination from a discriminating quantity of variance.

The different possibilities (variants and embodiments) must be understood as not being mutually exclusive and can be combined with one another.

DETAILED PRESENTATION OF PARTICULAR EMBODIMENTS

FIG. 1 illustrates a system for characterizing 10 a source of ionizing radiation 20 to be characterized, the ionizing radiation of which comprises a first type of particle and is capable of comprising a second type of particle according to the invention.

If in the context of this detailed description only the case of a first type of particle which is the gamma photons and of a second type of particle which is the neutrons is described, the invention is of course not limited to this alone possibility. Indeed, a person skilled in the art is perfectly capable, from the present disclosure, of applying the invention to the discrimination of other types of particles present in ionizing radiation. The invention thus also more particularly targets discrimination of:

- sources of ionizing radiation comprising gamma photons y and beta β particles vis-à-vis a reference ionizing radiation source whose radiation contains only gamma γ photons, the first type of particles then being gamma photons γ and the second type of particle being beta β particles;

- sources of ionizing radiation comprising gamma γ photons and alpha particles vis-à-vis a reference ionizing radiation source whose radiation contains only gamma γ photons, the first type of particles then being gamma photons γ and the second type of particle being alpha a particles;

- sources of ionizing radiation comprising beta β particles and alpha particles vis-à-vis a reference ionizing radiation source whose radiation comprises only beta β particles; , the first type of particles then being beta β particles and the second type of particles being alpha a particles.

Such a characterization system 10 comprises: an organic scintillator 11, a photomultiplier 12 coupled to the organic scintillator 11, and a processing unit 13 adapted to process the temporal pulse signals supplied by the photomultiplier 12 and obtained during the absorption of ionizing radiation by the organic scintillator 11.

It may be noted that if the invention relates more particularly to characterization systems 1 with organic scintillator, it is not limited to these characterization systems only and also relates to characterization systems with inorganic scintillator such as semiconductor scintillators. Thus, according to such a possibility, not illustrated, the characterization system may not include a photomultiplier and may include a system for amplifying the signal from the semiconductor scintillator other than a photomultiplier.

Thus, in general, the characterization system comprises a scintillator, such as the organic scintillator 11, an amplification system, such as the photomultiplier 12, and a processing unit adapted to process the temporal pulse signals supplied by the amplification system and obtained during the absorption of ionizing radiation by the scintillator

The processing unit 13 is in particular adapted to implement a method for characterizing the source of ionizing radiation 20 to be characterized.

The processing unit 13 is selected from:

a dedicated processing circuit, the configuration of which, both in terms of components and programming, is suitable for implementing the method for characterizing the source of ionizing radiation 20 to be characterized,

a computer comprising an acquisition card by which the computer is connected to the photomultiplier 12 and a data medium readable by a computer on which a program is recorded including the instructions for executing steps of the method for characterizing the radiation source ionizing 20 to be characterized,

a hybrid configuration comprising a processing circuit and a computer configured for the implementation of the method for characterizing the source of ionizing radiation 20 to be characterized.

Such a characterization process comprises the following steps:

supplying N time reference pulse signals, each time reference pulse signal corresponding to the detection of a photon y by a scintillator, N being a non-zero natural integer;

- choice of at least one discriminating quantity;

- For each reference pulse time signal, extracting the value of each discriminating quantity chosen to form a reference value corresponding to said discriminating quantity and to said reference pulse time signal;

- Acquisition of M temporal pulse signals acquired, each temporal pulse signal acquired corresponding to the detection, by the scintillator, of a photon originating from the source of ionizing radiation to be characterized, M being a non-zero natural integer;

- For each acquired time pulse signal, extracting the value of each discriminating quantity chosen to form a characteristic value corresponding to said discriminating quantity and said acquired time pulse signal;

- calculation of a first characteristic dissimilarity value from a first dissimilarity calculation between a reference matrix and a characteristic matrix, the reference matrix being formed by the reference values, the characteristic matrix being formed by the characteristic values ;

- comparison of the first characteristic dissimilarity value calculated with a first predetermined threshold;

characterization of the presence of neutrons in radiation emitted by the source of ionizing radiation to be characterized as a function of the result of the comparison between the first characteristic value of dissimilarity calculated and the first predetermined threshold.

With such a method, it is possible to determine whether the characteristic radiation source 20 to be characterized has an emission comprising neutrons.

The step of supplying the N reference pulse time signals, can consist of:

either by characterizing a reference ionizing radiation source whose emission of ionizing radiation is known to include only gamma y photons, this characterization preferably being carried out with the characterization system 10, or by characterizing several sources of reference ionizing radiation, the emissions of ionizing radiation of which are known to contain only gamma y photons, these characterizations being preferably carried out with the characterization system 10, ie the recovery of temporal pulse signals representative of simulated from previous characterizations of reference ionizing radiation source whose ionizing radiation emissions are known to contain only gamma photons y.

In order to illustrate the choice of at least one discriminating quantity, FIG. 2 compares these time signals of averaged pulses 301, 302 obtained respectively with a source of ionizing radiation emitting only gamma photons y and a source of radiation ionizing emitting photons y and neutrons. It can thus be seen that if these time signals of averaged pulses 301, 302 are relatively similar, these are distinguished mainly by two characteristics which are:

the time signal averaged from the ionizing radiation source emitting only gamma photons has a faster decrease therein, and therefore a width at half height smaller than the time signal averaged from the ionizing radiation source emitting y photons and neutrons , the time signal averaged from the source of ionizing radiation emitting photons y and neutrons has a preponderant secondary peak compared to that of the time signal averaged from the source of ionizing radiation emitting only gamma photons y.

Thus, when choosing the at least one discriminating quantity, the discriminating quantity can be selected from:

- an integral of at least a portion of the reference time pulse signal or acquired time pulse signal, such as for example the integral of the portion corresponding to the preponderant secondary peak for the source of ionizing radiation emitting photons y and neutrons;

- a time characteristic of the reference time pulse signal or acquired time pulse signal, such as a fall time;

a value, in at least one instant, of the reference time pulse signal or acquired time pulse signal, such as an amplitude corresponding to the secondary peak;

- a function of the variance of at least a portion of the reference time pulse signal or acquired time pulse signal, such a value making it possible to illustrate the preponderance or not of the secondary peak;

a value of derivative of the reference time pulse signal or time pulse signal acquired, such a value possibly being representative of the faster decay observed for the time signal averaged from the source of ionizing radiation emitting only gamma photons y .

In a conventional configuration of the invention, each of the steps of extracting the value of each discriminating variable, whether for the reference time pulse signals or for the acquired pulse time signals, comprises the sub-steps following:

normalization of each reference or acquired time pulse signal;

reduction of the electronic noise of each time signal of reference or acquired pulse n, for each time signal of reference or acquired pulse calculation of the value of each discriminating quantity chosen to form the corresponding characteristic value.

Of course, depending on the configuration of the characterization system 1, each of the steps of extracting the value of each discriminating quantity may not include one or both sub-steps from the normalization sub-step and the electronic noise reduction step.

It may be noted that in the context of the invention, during the step of choosing the at least one discriminating quantity, it is preferentially chosen at least two discriminating quantities, these discriminating quantities being preferably chosen from an integral of at least at least a portion of the reference time pulse signal or acquired time pulse signal and a function of the variance of at least a portion of the reference time pulse signal or acquired time pulse signal.

During this step of choosing the at least one discriminating quantity, it is therefore chosen R discriminating quantities, R being a non-zero natural integer.

During the step of calculating the first characteristic value of dissimilarity, the reference matrix can be a matrix comprising R columns and N rows, each column corresponding to a respective discriminating quantity and containing the reference values extracted for said discriminating quantity. Similarly, during this same step, the characteristic matrix can be a matrix comprising R columns and M rows, each column corresponding to a respective discriminating quantity and containing the characteristic values extracted for said discriminating quantity.

Of course, it is also possible, without departing from the scope of the invention, for the reference matrix to be a matrix comprising N columns and R rows or even a matrix comprising a single column and NxR rows, even NxR columns and a single line, as long as said matrix is formed by the set of extracted reference values. In order to facilitate the calculation of the characteristic value of dissimilarity, the characteristic matrix has, of course, a dimensioning similar to that of the reference matrix.

The step of calculating the characteristic value of dissimilarity comprises the following substeps:

- for each discriminating quantity identified by an index i, subdivision of an interval representative of the values taken by the discriminating quantity in k, sub-intervals, k, being a non-zero natural integer;

- calculation, for each of the reference matrix and the characteristic matrix, of the value of Πί ^ ι kj coefficients Θ, each coefficient θ being associated with R sub-intervals, each of said R sub-intervals being associated with a respective discriminating quantity and being equal to a sub-interval of said respective discriminating quantity, a first coefficient θι and a second coefficient θ ₂ distinct being such that there is at least one characteristic quantity for which the corresponding sub-interval of the first coefficient θι is different from the sub- corresponding interval of the second coefficient θ ₂ , the value of each of the coefficients θ being calculated as the number of time signals of reference pulses, respectively of time signals of pulses acquired, of which the values of discriminating quantities extracted are in the P sub -intervals associated with said coefficient, the values of coefficients θ calculated for one of the reference matrix and the characteristic matrix being multiplied by a corrective factor corresponding to the ration of the number of time pulse signals between the reference matrix and the characteristic matrix;

- calculation of the first characteristic value of dissimilarity from a difference in absolute value using the coefficients θ calculated from the reference matrix and the characteristic matrix, and from R successive sums, each of the R s being associated to a respective discriminating quantity, and said difference being made directly between the coefficients θ calculated on the basis of the reference matrix and the coefficients θ calculated on the basis of the characteristic matrix corresponding to the same R sub-intervals, the first sum being on the basis of the result of said difference and the following R1 sums on the basis of the result of the sum directly preceding it.

It will of course be noted that for the sake of brevity, the terminology "a difference between each coefficient θ calculated on the basis of the first matrix and the coefficient θ calculated on the basis of the second matrix corresponding to the same R sub-intervals", s' means fliti kj differences, each corresponding to a respective coefficient calculé calculated on the basis of the first matrix.

In the same way, a sum given among the R sums, apart from the R ^th sum, corresponds to a plurality of sums. Thus the T ^th sum, T being between 1 and R, R excludes, corresponds to Π _{ί = 1} ^ sums, ki corresponding to the sub-intervals of the quantities for which the corresponding sum has not yet been calculated.

In this way, the characteristic value of dissimilarity corresponds to a sum of the differences in absolute value between the distribution of the values of discriminating quantities obtained for the time signals of reference pulses and the time signals of characteristic pulses. It is also possible to highlight the differences in the distribution of the temporal pulse signals linked to the presence of possible neutrons in the ionizing radiation 21 emitted by the ionizing radiation source 20 to be characterized.

It may be noted that it is also possible in the context of the invention, as will be illustrated in connection with FIGS. 6A and 6B, to make one or more sums prior to the difference, this or these sums being each associated with a respective discriminating magnitude. Indeed, such sums made before the difference is perfectly equivalent to having chosen a number of discriminating quantities reduced by the number of sums made before said difference. Thus, if such a possibility is clearly not preferable, it is perfectly compatible with the invention, since it just induces, in addition to the steps claimed in the invention, additional calculations reducing the number of discriminating quantities actually implemented artwork.

In order to optimize the choice of the at least one discriminating quantity, the latter being chosen after the supply of N time signals of reference pulses, the step of choosing said at least one discriminating quantity may include the following steps :

supply of P time signals of comparison pulses, each time signal of comparison pulses corresponding to the detection, by the scintillator, of a photon y and / or of a neutron emitted by a source emitting ionizing radiation comprising photons y and neutrons, P being a non-zero natural integer;

- construction of a plurality of sets of discriminating quantities;

for each reference time pulse signal, and for each set of discriminating quantities, extraction of the value of each discriminating quantity from said set of discriminating quantities to form a first comparison value corresponding to said discriminating quantity and to said time signal reference pulse;

for each time comparison pulse signal, and for each set of discriminating quantities, extraction of the value of each discriminating quantity from said set of discriminating quantities to form a second comparison value corresponding to said discriminating quantity and to said time signal comparison pulse;

- for each set of discriminating quantities, calculation of a second characteristic dissimilarity value from a second dissimilarity calculation between a first matrix consisting of the first comparison values extracted for said set of discriminating quantities and a second matrix consisting of the second comparison values extracted for said set of discriminating quantities;

- choice of at least one discriminating quantity corresponding to one of the sets of discriminating quantities, the choice being made on the basis of the second characteristic values of dissimilarity calculated for each of the sets of discriminating quantities.

In a conventional configuration of the invention, each of the steps of extracting the value of each discriminating quantity, whether for the reference pulse time signals or for the comparison pulse time signals, comprises the sub- following steps :

normalization of each time reference or comparison pulse signal;

reduction of the electronic noise of each time reference or comparison pulse signal, for each time reference or comparison pulse signal calculation of the value of each discriminating quantity chosen to form the corresponding characteristic value.

Of course, depending on the configuration of the characterization system 1, each of the steps of extracting the value of each discriminating quantity may not include one or both sub-steps from the normalization sub-step and the electronic noise reduction step.

The sub-step of constructing a plurality of sets of discriminating quantities, may consist of a selection of several sets of discriminating quantities considered to be representative of the difference between the time pulse signal of a gamma photon γ and the signal. temporal pulse of a neutron after these two signals have been. Thus during the construction of each of the sets of discriminating quantities, each of the discriminating quantities can be selected from:

- an integral of at least a portion of the reference time pulse signal or acquired time pulse signal, such as for example the integral of the portion corresponding to the preponderant secondary peak for the source of ionizing radiation emitting photons y and neutrons;

- a time characteristic of the reference time pulse signal or acquired time pulse signal, such as a fall time;

a value, in at least one instant, of the reference time pulse signal or acquired time pulse signal, such as an amplitude corresponding to the secondary peak;

- a function of the variance of at least a portion of the reference time pulse signal or acquired time pulse signal, such a value making it possible to illustrate the preponderance or not of the secondary peak;

a value of derivative of the reference time pulse signal or time pulse signal acquired, such a value possibly being representative of the faster decay observed for the time signal averaged from the source of ionizing radiation emitting only gamma photons y .

It will be noted that this same sub-step of constructing a plurality of sets of discriminating magnitudes can consist of a choice of at least one function of discriminating magnitude which is a function of one variable or several variables, each of the sets of discriminating magnitudes. corresponding to a value of the given variable or variables and comprising a discriminating quantity resulting from the discriminating quantity function for said one or said values of the given variable or variables.

Thus, such a discrimination magnitude function can be, for example, an integral of a portion of the reference time pulse signal or acquired time pulse signal, one or more limits of said portion being defined by the one or more given variables, or a value derived from the reference time pulse signal or time pulse signal acquired at a position of said time pulse signal defined from said one or more given variables.

According to an advantageous possibility of the invention, at least one function of discriminating magnitude of each of the sets of discriminating magnitudes are selected from the group comprising:

a discriminating quantity denoted Q _tot corresponding to an integral of a portion of the reference time pulse signal or temporal pulse signal acquired, a discriminating quantity denoted Q _tai i, corresponding to an integral of a portion of the time signal d As a reference pulse or acquired time pulse signal, a discriminating quantity is denoted D corresponding to a function of the variance of at least a portion of the reference time pulse signal or acquired time pulse signal.

The discriminating quantity Q _tot is a discriminating quantity calculated according to:

where s is the reference pulse time signal, respectively the acquired pulse time signal, t _max is the time coordinate of the maximum of the reference pulse time signal, respectively of the acquired pulse time signal, and ti is a time coordinate calculated according to:

tl fmax + PJ-HM where β is a strictly positive real, and t _H M is the time coordinate, subsequent to t _max , for which the value of the reference pulse time signal, respectively of the acquired pulse time signal, is worth 0.5.

The discriminating quantity Q _t aii is a discriminating quantity calculated according to:

where s is the reference time pulse signal, respectively the acquired time pulse signal, t ₀ is a time coordinate calculated according to:

t-0 t-maxTæt-HM where a is a strictly positive real, t _max is the time coordinate of the maximum of the reference time pulse signal, respectively of the acquired time pulse signal, and t _H M is the time coordinate , subsequent to t _max , for which the value of the reference time pulse signal, respectively of the acquired time pulse signal, is 0, and ti is a time coordinate calculated according to:

tl t-max + P-t-HM where β is a strictly positive real.

The discriminating quantity D is the variance of the integral of the portion of the reference time pulse signal, respectively the acquired time pulse signal, between t ₀ and ti, where t ₀ is a time coordinate calculated according to:

t-0 t-maxTæt-HM where a is a strictly positive real, t _max is the time coordinate of the maximum of the reference time pulse signal, respectively of the acquired time pulse signal, and t _H M is the time coordinate , subsequent to t _max , for which the value of the reference time pulse signal, respectively of the acquired time pulse signal, is 0.5, ti is a time coordinate calculated according to:

t-l t-max + P-t-HM where β is a strictly positive real.

In other words, D is the variance of the discriminating magnitude Q _ta îi of the reference time pulse signal, respectively the acquired time pulse signal.

FIG. 3 illustrates the calculation of the values Q _tot , Qtaii and D by appearing on an example of a time pulse signal, in addition to the portions of the time pulse signal corresponding to Qtot and Qtaii, the times corresponding to t ₀ , ti, t _max and T _H m the parameters a and β are real values which are used to take into account the response time of the scintillator 11 and of the photomultiplier 12. If it is possible to determine them from the characteristic data of the set scintillator 11 and photomultiplier 12, it is nevertheless also possible, as illustrated below this document, to determine them by the implementation of a step of choosing the at least one discriminating quantity.

Thus, with such functions of discriminating quantities, it is possible to form a plurality of sets of discriminating quantities based which are distinguished in pairs by at least one of the parameters a and β.

During the sub-step of choosing at least one discriminating quantity corresponding to one of the sets of discriminating quantities, the at least one corresponding discriminating quantity chosen can be that which corresponds to the set of discriminating quantities for which the second characteristic value of dissimilarity presents an extremum characteristic of maximum dissimilarity.

Thus, in the case where the characteristic value of dissimilarity is a characteristic value of dissimilarity as such which makes it possible to characterize the differences between the time signals of reference pulses and the time signals of comparison pulses, the extremum characteristic of a dissimilarity is a maximum. In the case where the characteristic value of dissimilarity is a likelihood value which makes it possible to characterize the similarities between the time signals of reference pulses and the time signals of comparison pulses the characteristic extremum of a maximum dissimilarity is a minimum .

As a variant to such a sub-step of choosing at least one discriminating quantity corresponding to one of the sets of discriminating quantities and this in order to limit the risk, during the characterization of the source of ionizing radiation 20 to be characterized, the step of choosing at least one discriminating quantity further comprises the following sub-steps:

supply of Q time signals of test pulses, each time signal of test pulses corresponding to the detection, by the scintillator, of a photon y emitted by a source emitting ionizing radiation comprising γ photons, Q being a non-zero natural number,

for each test pulse time signal, and for each set of discriminating quantities, extracting the value of each discriminating quantity from said set of discriminating quantities to form a third comparison value corresponding to said discriminating quantity and to said time signal test pulse;

- for each set of discriminating quantities, calculation of a third characteristic dissimilarity value from the second dissimilarity calculation between a first matrix consisting of the first comparison values extracted for said set of discriminating quantities and a second matrix consisting of the third values of comparison extracted for said set of discriminating quantities;

- comparison of the third characteristic dissimilarity value calculated for each of the sets of discriminating quantities with a second predetermined threshold;

selection from among the sets of discriminating magnitudes those which have a minimum dissimilarity from said comparison between the third characteristic value of dissimilarity calculated for each of the sets of discriminating magnitudes at a second predetermined threshold;

- choice of at least one discriminating quantity corresponding to one of the sets of discriminating quantities, the at least one corresponding discriminating quantity chosen is that corresponding to the set of discriminating quantities among the sets of selected discriminating quantity whose second value characteristic of dissimilarity presents an extremum characteristic of maximum dissimilarity.

As indicated above, the step of extracting the value of each discriminating quantity comprises the following sub-steps:

normalization of each test pulse time signal; reduction of the electronic noise of each test pulse time signal, for each test pulse time signal calculation of the value of each discriminating quantity chosen to form the corresponding characteristic value.

Of course, depending on the configuration of the characterization system 1, each of the steps of extracting the value of each discriminating quantity may not include one or both sub-steps from the normalization sub-step and the electronic noise reduction step.

During this step of choosing the at least one discriminating quantity, a given set of discriminating quantities comprises S discriminating quantities, S being a non-zero natural integer. The second and third dissimilarity characteristic values can be calculated from a dissimilarity calculation similar to that of the first dissimilarity characteristic value. Thus the dissimilarity calculations of the second and third characteristic dissimilarity values can include the following sub-steps:

- for each discriminating quantity identified by an index j, subdivision of an interval representative of the values taken by the discriminating quantity in kj sub-intervals, kj being a non-zero natural integer;

- calculation, for each of the first matrix and of the second matrix, of the value of Πρ = ι kj coefficients δ, each coefficient δ being associated with S subintervals, each of said S subintervals being associated with a respective discriminating quantity and being equal to a sub-interval of said respective discriminating quantity, a first coefficient δι and a second coefficient δ ₂ distinct being such that there exists at least one characteristic quantity for which the corresponding sub-interval of the first coefficient δι is different from the sub- corresponding interval of the second coefficient δ ₂ , the value of each of the coefficients δ being calculated as the number of time pulse signals of the first matrix, respectively of time pulse signals of the second matrix, from which the values of discriminating quantities extracted are in the S subintervals associated with said coefficient, the values of coefficients δ calculated for one of the first ma the second matrix being multiplied by a corrective factor corresponding to the ration of the number of time pulse signals between the first matrix and the second matrix;

- calculation of a characteristic value of dissimilarity from a difference in absolute value using the coefficients δ calculated from the first matrix and the second matrix and from S successive sums, each of the S sums being associated with a respective discriminant quantity, and said difference being made directly between each coefficient δ calculated on the basis of the first matrix and the coefficient δ calculated on the basis of the second matrix corresponding to the same R sub-intervals, the first sum being made on the basis of the result of said difference and the following S s based on the result of the sum directly preceding it.

It will of course be noted that for the sake of brevity, the terminology "a difference between each coefficient δ calculated on the basis of the first matrix and the coefficient δ calculated on the basis of the second matrix corresponding to the same S subintervals", s' means Πρ = ι kj differences, each corresponding to a respective coefficient δ calculated on the basis of the first matrix.

In the same way, a sum given among the S sums, apart from the S ^th sum, corresponds to a plurality of sums. Thus the U ^th sum, U being., [11 between 1 and S, S excludes, corresponds to Ilj _{= 1} kj sums, kj corresponding to the subintervals of the discriminating quantities for which the corresponding sum has not yet been calculated .

Thus, with such a calculation, the calculations of the second and of the third characteristic value of dissimilarity are distinguished only in that the second matrix is made up of the second comparison values calculated for said set of discriminating quantities for the calculation of the second value. characteristic of dissimilarity while the second matrix consists of the third comparison values calculated for said set of discriminating quantities for the calculation of the second characteristic value of dissimilarity.

It may be noted that, according to an advantageous possibility of the invention, the first predetermined threshold can be determined from the result of the step of choosing the at least one discriminating quantity. Thus in the case where the choice of the at least one discriminating quantity involves the calculations of the second and the third characteristic value of dissimilarity, the first predetermined threshold can be selected in an interval delimited by the second and third characteristic values of dissimilarity for the set of selected discriminating quantities.

According to another possibility of the invention, the characterization method can also include an additional step of estimating the number and / or a rate of time pulse signals corresponding to a neutron. Such an estimate can in particular be calculated from the first characteristic value of dissimilarity since this characteristic value is, when the latter is a characteristic value of dissimilarity as such, directly proportional to the rates of time signals of pulses corresponding to a neutron.

FIGS. 4A to 6B illustrate a first example of implementation of such a method in which the choice of the at least one discriminating quantity is carried out from a set of discriminating quantities comprising Q _tot and Q _t aii and for which the choice of the at least one discriminating quantity calls for a calculation of the third characteristic values of dissimilarity.

In the context of this first example of implementation, four sources of ionizing radiation have been selected, three sources of ionizing radiation whose emission of ionizing radiation consists only of gamma photons y and one source of which the emission of ionizing radiation is made up of gamma y photons and neutrons. The three sources of ionizing radiation whose emission of ionizing radiation consists only of gamma photons are sources consisting respectively of the isotope 22 of sodium ²² Na, the isotope 137 of cesium ¹³⁷ Cs and the isotope 60 of cobalt ⁶⁰ Co. The source, the emission of ionizing radiation of which consists of gamma y photons and neutrons, is a source of isotopes 252 of californium ²⁵² Cf.

All of these ionizing radiation sources were characterized using a standard liquid scintillator, the BC-501-A® scintillator marketed by

Saint-Gobain® to define the N reference time pulse signals and the M acquired time pulse signals. Thus, with N equal to 16000, the time signals of reference pulses consist of time signals of reference of each of the three sources of ionizing radiation whose emission of ionizing radiation consists only of gamma photons. Likewise, M is also equal to 16000 and the pulse time signals acquired consist of 16000 time signals acquired from the source of the isotope 252 of californium ²⁵² CF.

In accordance with the characterization method according to the invention, each of the reference and acquired time pulse signals was this by dividing them by the maximum value of their amplitude. FIG. 2 thus illustrates the respective averages of the reference time pulse signals and of the acquired time pulse signals.

Once the pulse time signals, the steps for calculating the reference values and the characteristic values were implemented. To do this, the discriminating quantities Qtot and Qtail were calculated for each of the reference and acquired pulse time signals. To do this, and according to the result of the choice of the at least one discriminating quantity presented below, the parameters a and β were fixed at 4 and 20 respectively.

The 32,000 reference values and these 32,000 extracted characteristic values then made it possible to constitute the reference matrix and the characteristic matrix. The reference matrix is thus a matrix comprising two columns, one corresponding to Q _tot and the other to Q _ta ii and 16,000 rows corresponding to the time pulse signals.

In order to allow a calculation of the characteristic value of dissimilarity, the interval representative of the values taken by Q _tot has been subdivided into 300 sub-intervals while the interval representative of the values taken by Q _tai | was subdivided into 150 subintervals.

For each of the reference matrix and the characteristic matrix, 45,000 coefficients θ each corresponding to a given subinterval of

Qtot and a given sub-interval of Q _tai | have been calculated. Each of the coefficients θ is equal to the number of time pulse signals of said matrix whose reference values, respectively characteristic values, whose values of discriminating quantities are in the sub-intervals of Q _tot and Q _tai | corresponding. The 45,000 coefficients θ calculated from the reference matrix thus make it possible to obtain a matrix of 300 by 150. Similarly, the 45,000 coefficients θ calculated from the characteristic matrix thus make it possible to obtain a matrix of 300 by 150 .

FIG. 4A graphically illustrates the values of coefficients θ obtained for the reference matrix, with the values of the sub-intervals of Q _to x on the abscissa and the values of the sub-intervals of Q _tai on the ordinate, the values of the coefficients θ being figured using a color scale. It can thus be seen in this figure that in the case of the reference values, the discriminating quantities have a relatively small variance since they are distributed in a relatively dense set 321. Thus, the values of the discriminating quantity Q _tot are between 60 and 90 and those of the discriminating quantity Q _tai | are between 0 and 13.

Concerning the values of coefficients θ obtained for the characteristic matrix, which are represented on figure 4B according to the same representation as figure 4A, one notes a much stronger variance of the discriminating quantities. Indeed, it is noted that these discriminating quantities are distributed into a first relatively dense set 322 and a second, more extensive set 323. Thus, the values of the discriminating quantity Q _tot are between 55 and 140 and those of the discriminating quantity Q _tai | are between 2 and 47.

Once the coefficients θ are calculated for both the reference matrix and the characteristic matrix, the calculation of the first characteristic value of dissimilarity is carried out from two sums and a difference.

The calculation of the first characteristic value of dissimilarity can consist in realizing for each of the intervals of the magnitudes Q _tot and Q _tai |, a first difference in absolute value between the coefficients θ of the reference matrix and the characteristic matrix, followed by two are successive sums of the result of this difference on respectively one of the discriminating quantities among Q _tot and Q _tai | and the other of the discriminating quantities among Q _tot and Q _tai |.

In order to illustrate the result of such a calculation of the first characteristic dissimilarity value, an intermediate calculation was carried out to illustrate the observable dissimilarities from the discriminating quantity Q _tot . This calculation consists in successively:

- a first sum of the coefficients θ of the reference matrix, respectively of the matrix characterizing, by summing, for each of the sub-intervals of the discriminating quantity Q _tot , the number of time pulse signals of each sub-intervals of the discriminating quantity Q _t aii corresponding to said interval, the result of each of the first sums for the reference matrix 331 and for the characteristic matrix 332 being shown in FIG. 5A- for each of the sub-intervals of the discriminating quantity Q _tot , a difference in absolute value between the result of the first sum obtained for the reference matrix and the result of the first sum obtained for the characteristic matrix, the result of this difference being shown in the form of hatched portions 333 in FIG. 5A.

In the same way, a similar intermediate calculation was carried out to illustrate the dissimilarities observable from the discriminating quantity Q _t aii · This calculation consists in successively:

- a first sum of the coefficients θ of the reference matrix, respectively of the matrix characterizing, by summing, for each of the sub-intervals of the discriminating quantity Q _tai i, the number of time pulse signals of each sub-intervals of the discriminating quantity Q _tot corresponding to said interval, the result of each of the first sums for the reference matrix 341 and for the characteristic matrix 342 being shown in FIG. 5B,

- for each of the sub-intervals of the discriminating quantity Q _tai i, a difference in absolute value between the result of the first sum obtained for the reference matrix and the result of the first sum obtained for the characteristic matrix, the result of this difference being shown as hatched portions 343 in Figure 5B.

Thus with such a calculation of a characteristic value of dissimilarity, it is possible to add up the differences observable in FIG. 5A and represented by the hatched portions 333 with those observable in FIG. 5A and represented by the hatched portions 343.

From the calculation of the characteristic value of dissimilarity, it is possible to estimate the rate of time pulse signals produced by the neutron interaction. Indeed, the rate of time pulse signals produced by the interaction of neutrons is directly proportional to the characteristic value of dissimilarity calculated. In the configuration used for this implementation, that is to say with a source of ionizing radiation whose ionizing radiation comprises gamma photons and neutrons which is a source of ionizing radiation comprising the isotope 252 of Californium ²⁵² Cf, it was estimated that 29.9% of the 16,000 time pulse signals were time pulse signals produced by the interaction of neutrons.

Such a process is not limited, of course, not only to liquid organic scintillators and is applicable to any organic scintillator whatever its type. In order to illustrate this possibility, the method was implemented according to this same first example of implementation on three other non-liquid organic scintillators, an E32® scintillator designed and produced by the CEA and EJ-299-33® scintillators and EJ-200® both marketed by Ejen Technology®. The results are summarized in the following table 1 in which is indicated for each of the scintillators the estimation of the rate of time signals of pulses produced by the neutron interaction obtained for the source of ionizing radiation comprising the isotope 252 of californium ²⁵² Cf in parallel with a rate of temporal pulse signals produced by the interaction of neutrons for a source of ionizing radiation whose ionizing radiation contains only gamma photons y, this source having been simulated from a random subsampling to 8000 pulse time signals from the 16000 reference pulse time signals.

Scintillator Estimated rate of time pulse signals produced by the interaction of neutrons ²⁵² Cf Source y

BC-501-A® 29.9 4 E32® 26.0 4 EJ-299-33® 26.3 4 EJ-200® 23.2 4

Table 1: Estimation of the rate of temporal neutron pulse signals for respectively a source of ionizing radiation comprising the isotope 252 of Californium and for a subsample of the temporal reference pulse signals, this for four organic scintillators.

Of course, for each of these scintillators the parameters a and β have been adapted in order to take into account the temporal characteristics of each of the scintillators. The procedure implemented for such an adaptation is specified below in connection with the BC-501-A scintillator.

It can thus be noted that if the rates of temporal pulse signals exhibit a relatively large variation, the estimated rate is each time significant with respect to that, representative of the error rate, obtained for temporal signals of pulses obtained for a source of ionizing radiation, the ionizing radiation of which comprises only photons y.

Within the framework of this same first example of implementation, it has also been possible to show that a number of 2000 time pulse signals being sufficient to estimate a rate of time pulse signals produced by the interaction of neutrons with a margin of error of 4% and that this margin of error becomes less than 1% for a number of 8000 time signals of pulses. In addition, even with a reduced number of time pulse signals, namely 100 time pulse signals, it was possible to obtain a characteristic value of significant dissimilarity demonstrating the presence of neutrons in the ionizing radiation emitted for the light source. ionizing radiation comprising isotope 252 of californium ²⁵² Cf.

It can also be noted that the results of temporal pulse rates obtained for the ionizing radiation source comprising the isotope 252 of californium ²⁵² Cf, and therefore the corresponding characteristic value of dissimilarity, and that obtained for the sub-sampling of the time signals reference pulse, and therefore the corresponding characteristic value of dissimilarity, can be used to define the first predetermined threshold. Thus in the context of this first example of implementation, the first predetermined threshold can be chosen to be between the characteristic value of dissimilarity obtained for the source of ionizing radiation comprising the isotope 252 of californium ²⁵² Cf and that obtained for the sub-sampling of reference pulse time signals.

FIGS. 6A and 6B illustrate the step of choosing the set of characteristic quantities within the framework of this first example of implementation with in FIG. 6A the calculation of the second characteristic value of dissimilarity and in FIG. 6B the calculation of the third characteristic value of dissimilarity.

In the context of this first example of implementation, each set of discriminating quantities consists of the discriminating quantity Q _tot and the discriminating quantity Q _t aii and the distinction between the sets of discriminating quantities is provided by a variation of the parameters a and β. Thus, as shown in FIGS. 6A and 6B, the parameter aa was varied between 0.5 and 12 and the parameter β was varied between 4 and 100.

Once the sets of discriminating quantities thus constructed, the choice of at least one discriminating quantity among these sets was made by means of 16000 time pulse signals obtained from the ionizing radiation source comprising the isotope 252 of the californium ²⁵² Cf, and 8000 time pulse signals obtained by a random subsampling of the reference signals. Thus, said 16000 pulse time signals form the comparison pulse time signals and said 8000 pulse time signals form test pulse time signals.

From these time reference pulse signals, in application of the method according to the invention and similarly to the calculation of the characteristic value of dissimilarity explained with reference to FIGS. 4A to 5B, first comparison values are calculated for each. discriminating quantities and for each of the sets of discriminating quantities. In the same way, from the time signals of comparison pulse and from respectively the time signals of test pulse, are calculated of the second values of comparison, respectively of the third values of comparison, for each of the discriminating quantities and for each sets of discriminating quantities.

In order to obtain the second characteristic dissimilarity values, a dissimilarity calculation, according to the principle already described, was carried out from the first and second comparison values, this for each of the sets of discriminating quantities. These characteristic values of dissimilarity make it possible to estimate a rate of temporal pulse signals produced by the interaction of neutrons among the temporal signals of comparison pulses for each of the sets of discriminating magnitudes. The rates of time pulse signals produced by the interaction of neutrons thus estimated are shown graphically in FIG. 6A, this as a function of the values of the parameters a and β for said sets of corresponding discriminating quantities. Of course, it is preferable to obtain a maximum rate, this indicating that a maximum number of time pulse signals produced by the interaction of neutrons could be detected and allowing optimized discrimination of the sources of ionizing radiation, ionizing radiation includes neutrons and y photons.

In an identical manner and in order to obtain the third characteristic values of dissimilarity, a dissimilarity calculation was carried out from the first and third comparison values this for each of the sets of discriminating quantities, while taking into account the ratio of time signals pulse between the reference pulse time signals and the test pulse time signals. These characteristic values of dissimilarity make it possible to estimate a rate of temporal pulse signals produced by the interaction of neutrons among the temporal signals of test pulses for each of the sets of discriminating magnitudes. The rates of temporal pulse signals produced by the interaction of neutrons thus estimated are represented graphically in FIG. 6B this as a function of the values of the parameters a and β for said sets of corresponding discriminating quantities. Here, since the time signals of test pulses contain only γ photons, it is necessary to minimize the rate of time signals of pulses produced by the interaction of neutrons.

Thus, the choice of the set of discriminating quantities therefore consists in choosing the parameters a and β making it possible to maximize the rate of temporal signals of pulses produced by the interaction of neutrons for the temporal signals of comparison pulses while ensuring a rate of temporal pulse signals produced by the interaction of neutrons for the temporal test pulse signals does not exceed a second predetermined threshold, namely 3%. In the example illustrated in FIGS. 6A and 6B, the pair of parameters a and β (4, 20) makes it possible in particular to meet these constraints. Note that this graph shows that other pairs of values are also possible.

Thus, with such a choice of at least one discriminating quantity consisting of the discriminating quantities Q _tot and Qtaü for the pair of parameters a and β (4.20), a characterization system is obtained offering an optimized discrimination between the sources of ionizing radiation emitting ionizing radiation comprising y photons and neutrons vis-à-vis those whose ionizing radiation contains only y photons.

FIG. 7 illustrates the result of an intermediate calculation according to a second example of implementation in which the discriminating quantity is the discriminating quantity D, the parameters a and β being presenting the values 4, 20 determined for the first example of implementation artwork.

According to the same principle described in the context of the first example of implementation and from the same time reference and acquired pulse signals, it is possible to calculate the reference values and the characteristic values of the discriminating quantity D on the basis of the calculation already described. These 16,000 reference values and these 16,000 characteristic values are then used respectively to constitute a reference matrix and a characteristic matrix. The reference matrix is thus a matrix comprising a column corresponding to D and 16,000 rows corresponding to the time pulse signals.

In order to allow a calculation of the characteristic value of dissimilarity, the interval representative of the values taken by D has been subdivided into 200 sub-intervals.

For each of the reference matrix and the characteristic matrix, 200 coefficients θ each corresponding to a given subinterval of D were calculated. Each of the coefficients θ is equal to the number of time pulse signals of said matrix are in the corresponding sub-interval of D. The 200 coefficients θ calculated from the reference matrix thus make it possible to obtain a matrix of 200 by 1 In an identical manner, the 200 coefficients θ calculated from the characteristic matrix thus make it possible to obtain a matrix of 200 by 1.

FIG. 7 graphically illustrates the values of coefficients θ obtained for the reference matrix 362 and for the characteristic matrix 361, with the values of the sub-intervals of D on the abscissa and the values of the coefficients Θ on the ordinate. We can thus see in this figure that in the case of the reference values, the discriminating quantities appear in the form of a single peak with a strong amplitude of 4000. For the characteristic matrix, we note a stronger variance of the discriminating quantities . Indeed, we note that these discriminating quantities are distributed in the form of a first peak of relatively large amplitude of 2700, and a secondary peak of a lower amplitude reaching 250.

Once the coefficients θ are calculated for both the reference matrix and the characteristic matrix, the calculation of the first characteristic value of dissimilarity is carried out from a difference and a sum.

This calculation of the first characteristic value, the result of an intermediate step is shown in Figure 7, consists of successively:

- for each of the sub-intervals of the discriminating quantity D, a difference in absolute value between the coefficients θ calculated for the reference matrix and for the characteristic matrix, the result of this difference being represented in the form of hatched portion 363 on the figure 7,

- a sum of the result of the differences obtained for each of the intervals of the discriminating quantity D.

From this calculation of the characteristic value of dissimilarity, in the same way as for the first example of implementation, it is possible to estimate the rate of time signals of pulses produced by the interaction of neutron. Indeed, the rate of time pulse signals produced by the interaction of neutrons is directly proportional to the characteristic value of dissimilarity calculated. In the configuration used for this implementation, that is to say with a source of ionizing radiation whose ionizing radiation comprises gamma photons and neutrons which is a source of ionizing radiation comprising the isotope 252 of Californium ²⁵² Cf, it was estimated that 27.2% of the 16,000 time pulse signals were time pulse signals produced by the interaction of neutrons. This result is therefore very close to that obtained in the context of the first of the first example of implementation whereas this second example of implementation is based on only one discriminating quantity.

Claims (15)

1. A method for characterizing a source of ionizing radiation to be characterized, the ionizing radiation of which comprises a first type of particle and is capable of comprising a second type of particle, the method comprising the following steps: supplying N time reference pulse signals, each time reference pulse signal corresponding to the detection of the first type of particle by a scintillator, N being a non-zero natural integer; - choice of at least one discriminating quantity; - For each reference pulse time signal, extracting the value of each discriminating quantity chosen to form a reference value corresponding to said discriminating quantity and to said reference pulse time signal; - Acquisition of M temporal pulse signals acquired, each temporal pulse signal acquired corresponding to the detection, by the scintillator, of a photon originating from the source of ionizing radiation to be characterized, M being a non-zero natural integer; - For each acquired time pulse signal, extracting the value of each discriminating quantity chosen to form a characteristic value corresponding to said discriminating quantity and to said acquired time pulse signal; - calculation of a first characteristic dissimilarity value from a first dissimilarity calculation between a reference matrix and a characteristic matrix, the reference matrix being formed by the reference values, the characteristic matrix being formed by the characteristic values ; - comparison of the first characteristic dissimilarity value calculated with a first predetermined threshold; characterization of the presence of the second type of particle in radiation emitted by the source of ionizing radiation to be characterized as a function of the result of the comparison between the first characteristic value of dissimilarity calculated and the first predetermined threshold, in which the choice step at least one discriminating quantity comprises the choice of R discriminating quantities, R being a non-zero natural integer, and in which the first of dissimilarity between the reference matrix and the characteristic matrix comprises the following substeps: - for each discriminating quantity identified by an index i, subdivision of an interval representative of the values taken by the discriminating quantity in k, sub-intervals, k, being a non-zero natural integer; - calculation, for each of the reference matrix and the characteristic matrix, of the value of Πί ^ ι kj coefficients 0, each coefficient θ being associated with R sub-intervals, each of said R sub-intervals being associated with a respective discriminating quantity and being equal to a sub-interval of said respective discriminating quantity, a first coefficient θι and a second coefficient 0 ₂ distinct being such that there exists at least one characteristic quantity for which the corresponding sub-interval of the first coefficient θι is different from the sub- corresponding interval of the second coefficient 0 ₂ , the value of each of the coefficients θ being calculated as the number of time signals of reference pulses, respectively of time signals of pulses acquired, of which the values of discriminating quantities extracted are in the P sub -intervals associated with said coefficient, the values of coefficients θ calculated for one of the reference matrix and the characteristic matrix being multiplied by a corrective factor corresponding to the ration of the number of time pulse signals between the reference matrix and the characteristic matrix; - calculation of the first characteristic dissimilarity value from a difference in absolute value implementing the coefficients θ calculated from the reference matrix and the characteristic matrix and from R successive sums, each of the R sums being associated with a respective discriminating quantity, and said difference being made directly between each coefficient θ calculated on the basis of the reference matrix and the coefficient θ calculated on the basis of the characteristic matrix corresponding to the same R sub-intervals, the first sum being made on based on the result of said difference and the following R1 sums based on the result of the sum directly preceding it. 2. Method for characterizing an ionizing radiation source to be characterized according to claim 1, in which at least one discriminating quantity is selected from: - an integral of at least a portion of the reference time pulse signal or acquired time pulse signal; - a characteristic time of the reference pulse time signal or acquired pulse time signal; - a value, in at least one instant, of the reference time pulse signal or time pulse acquired signal; - a function of the variance of at least a portion of the reference time pulse signal or acquired time pulse signal; - a value derived from the reference time pulse signal or acquired time pulse signal. 3. A method of characterizing an ionizing radiation source to be characterized according to claim 1 or 2, wherein the step of choosing at least one discriminating quantity is subsequent to the step of supplying the N time signals of pulses. of reference and in the step of normalization of each time reference pulse signal, the step of choosing at least one discriminating quantity comprising the following sub-steps: supply of P time signals of comparison pulses, each time signal of comparison pulses corresponding to the detection, by the scintillator, of a particle of the first type or of the second type emitted by a source of ionizing radiation emitting a radiation ionizing agent comprising particles of the first type and particles of the second type, P being a non-zero natural integer; - construction of a plurality of sets of discriminating quantities; for each reference time pulse signal, and for each set of discriminating quantities, extraction of the value of each discriminating quantity from said set of discriminating quantities to form a first comparison value corresponding to said discriminating quantity and to said time signal reference pulse; for each time comparison pulse signal, and for each set of discriminating quantities, extraction of the value of each discriminating quantity from said set of discriminating quantities to form a second comparison value corresponding to said discriminating quantity and to said time signal comparison pulse; - for each set of discriminating quantities, calculation of a second characteristic dissimilarity value from a second dissimilarity calculation between a first matrix consisting of the first comparison values extracted for said set of discriminating quantities and a second matrix consisting of the second comparison values extracted for said set of discriminating quantities; - choice of at least one discriminating quantity corresponding to one of the sets of discriminating quantities, the choice being made on the basis of the second characteristic values of dissimilarity calculated for each of the sets of discriminating quantities. 4. A method of characterizing an ionizing radiation source to be characterized according to claim 3, wherein during the sub-step of choosing at least one discriminating quantity corresponding to one of the sets of discriminating quantities, the at least a corresponding discriminating quantity chosen is that which corresponds to the set of discriminating quantities for which the second characteristic dissimilarity value has an extremum characteristic of maximum dissimilarity. 5. A method of characterizing an ionizing radiation source to be characterized according to claim 3, in which the step of choosing at least one discriminating quantity further comprises the following substeps: supply of Q time signals of test pulses, each time signal of test pulses corresponding to the detection, by the scintillator, of a particle of the first type emitted by a source of ionizing radiation emitting an ionizing radiation comprising particles of the first type, Q being a non-zero natural integer; for each test pulse time signal, and for each set of discriminating quantities, extracting the value of each discriminating quantity from said set of discriminating quantities to form a third comparison value corresponding to said discriminating quantity and to said time signal test pulse; - for each set of discriminating quantities, calculation of a third characteristic dissimilarity value from the second dissimilarity calculation between a first matrix consisting of the first comparison values extracted for said set of discriminating quantities and a second matrix consisting of the third values of comparison extracted for said set of discriminating quantities; - comparison of the third characteristic dissimilarity value calculated for each of the sets of discriminating quantities with a second predetermined threshold; - selection from among the sets of discriminating quantities of those having a minimum dissimilarity from said comparison between the third characteristic value of dissimilarity calculated for each of the sets of discriminating quantities at a second predetermined threshold; - choice of at least one discriminating quantity corresponding to one of the sets of discriminating quantities, the at least one corresponding discriminating quantity chosen being that corresponding to the set of discriminating quantities from the sets of selected discriminating quantities whose 5 second characteristic value of dissimilarity presents an extremum characteristic of a maximum dissimilarity. 6. Method for characterizing an ionizing radiation source to be characterized according to any one of claims 1 to 5, in which at least one 10 discriminating quantity is noted Q _tot and is calculated according to: where s is the reference time pulse signal or the time signal d ', respectively the acquired pulse time signal; t _max is the time coordinate of the maximum of the reference time pulse signal, respectively of the acquired time pulse signal; and 15 ti is a time coordinate calculated according to: ti t _max + pt _R M where β is a strictly positive real; and t _H M is the time coordinate, subsequent to t _max , for which the value of the reference pulse time signal, respectively of the acquired pulse time signal, is 0.5. 7. A method of characterizing an ionizing radiation source to be characterized according to any one of claims 1 to 6, in which at least one discriminating quantity is denoted Q _t aii and calculated according to: tl s (t) dt where s is the reference time pulse signal, respectively the acquired time pulse signal; t ₀ is a time coordinate calculated according to: t () t _max + CX.t ^] v [where a is a strictly positive real; t _max is the time coordinate of the maximum of the reference time pulse signal, respectively of the acquired time pulse signal; and t _H M is the time coordinate, subsequent to t _max , for which the value of the reference pulse time signal, respectively of the acquired pulse time signal, is 0.5; and ti is a time coordinate calculated according to: tl t _max + p.tpiM where β is a strictly positive real. 8. Method according to claims 3, 6 and 7 alone or in combination with claim 4 or 5, wherein each set of discriminating quantities comprises at least one of Q _tot and Q _tai |, the sets being distinguished in pairs by at least one of the values a and β. 9. Method according to any one of claims 1 to 8, in which at least one discriminating quantity is denoted D and is the variance of the portion of the reference time pulse signal, respectively the time pulse signal acquired, included between t ₀ and ti, where: t ₀ is a time coordinate calculated according to: t ₀ t _max + CX.t ^] v [where a is a strictly positive real; t _max is the time coordinate of the maximum of the reference time pulse signal, respectively of the acquired time pulse signal; and t _H M is the time coordinate, subsequent to t _max , for which the value of the reference pulse time signal, respectively of the acquired pulse time signal, is 0.5; ti is a time coordinate calculated according to: f-l f-max + PJ-HM where β is a strictly positive real. 10. Method according to claims 3 and 9 alone or in combination with any one of claims 4 to 8, wherein each set of discriminating quantities comprises D, the sets being distinguished in pairs by at least one of the values a and β. 11. The method according to claim 3 alone or taken in combination with any one of claims 4 to 10, in which, during the step of constructing the plurality of sets of discriminating quantities, a given set of discriminating quantities comprises S discriminating quantities, S being a non-zero natural integer, and in which the second dissimilarity calculation between a first matrix and a second matrix, for the given set of S discriminating quantities, comprises the following substeps: - for each discriminating quantity identified by an index j, subdivision of an interval representative of the values taken by the discriminating quantity in kj sub-intervals, kj being a non-zero natural integer; - calculation, for each of the first matrix and of the second matrix, of the value of Πρ = ι kj coefficients δ, each coefficient δ being associated with S subintervals, each of said S subintervals being associated with a respective discriminating quantity and being equal to a sub-interval of said respective discriminating quantity, a first coefficient δι and a second coefficient δ ₂ distinct being such that there exists at least one characteristic quantity for which the corresponding sub-interval of the first coefficient δι is different from the sub- corresponding interval of the second coefficient δ ₂ , the value of each of the coefficients δ being calculated as the number of time pulse signals of the first matrix, respectively of time pulse signals of the second matrix, from which the values of discriminating quantities extracted are in the S subintervals associated with said coefficient, the values of coefficients δ calculated for one of the first ma the second matrix being multiplied by a corrective factor corresponding to the ration of the number of time pulse signals between the first matrix and the second matrix; - calculation of a characteristic value of dissimilarity from S sums and of a difference in absolute value implementing the coefficients δ calculated from the first matrix and the second matrix, each of the S sums being associated with a quantity respective discriminant, and said difference being made from the set of results of one of said S sums, between the results obtained from the coefficients δ calculated on the basis of the first matrix and the results obtained from the coefficients δ calculated on the basis of the second matrix, the rest of the S sums made after the difference has been calculated on the basis of the result of said difference. - calculation of a characteristic value of dissimilarity from a difference in absolute value implementing the coefficients δ calculated from the first matrix and the second matrix and from S successive sums, each of the S sums being associated with a respective discriminant quantity, and said difference being made directly between each coefficient δ calculated on the basis of the first matrix and the coefficient δ calculated on the basis of the second matrix corresponding to the same R sub-intervals, the first sum being made on the basis of the result of said difference and the following S s on the basis of the result of the sum directly preceding it. 12. A method of characterizing an ionizing radiation source to be characterized according to any one of claims 1 to 11, in which there is further provided a step of estimating the number and / or of a corresponding pulse rate to the second type of particles. 13. method for characterizing an ionizing radiation source to be characterized according to any one of claims 1 to 12, in which the first type of particles and the second type of particles are selected from the group comprising: - gamma photons as the first type of particles and neutrons as the second type of particles, - gamma photons as the first type of particles and beta particles as the second type of particles, - gamma photons as the first type of particles and alpha particles as the second type of particles, - beta particles as the first type of particles and alpha particles as the second type of particles. 14. Characterization system (10) of a source of ionizing radiation to be characterized comprising: a scintillator (11), a photomultiplier (12) coupled to the scintillator (11), and a processing unit (13) adapted to process the temporal pulse signals supplied by the photomultiplier and obtained during the absorption of ionizing radiation by the scintillator (11), the characterization system (10) being in that the processing unit (13) is configured to implement a characterization method according to any one of claims 1 to 13. 15. Data medium readable by a computer on which is recorded a program including the instructions for executing steps of the method according to any one of claims 1 to 13. S.62199 1/5 40 60 80 t (ns)