Method for measuring atomic number by X-ray or Gamma rays

Abstract

X-ray or Gamma Ray Atomic Number Measurement Device: An innovative method for determining the effective atomic number (Zeff) using X-ray or gamma radiation, characterized by: A material discrimination technique employing two energy profiles of X-rays or gamma rays, enabling precise and direct measurement of the atomic number through a bijective function and eliminating the need for pre-recorded approximations. The method offers the advantage of being adaptable to various X-ray or gamma ray scanning methods, compatible with all existing sensors, and usable in uncontrolled environments. It also reduces X-ray or gamma ray exposure time. Main applications: baggage screening, medical imaging, industrial and security controls, offering a technological solution that overcomes the limitations of previous methods.The innovation lies in its ability to accurately analyze the atomic composition of materials, including those in living organisms, without compromising their integrity through high doses.

Description

Title of the invention: Method for measuring atomic number by X-ray or Gamma rays

[0001] The present invention relates to a method for measuring atomic number by X-ray or Gamma radiation.

[0002] The invention defines a method for discriminating materials based on two energy profiles of X-rays or Gamma rays, allowing for the precise determination of the effective atomic number (Zeff) of the scanned materials. This field is of crucial importance in numerous industrial, security, and medical applications, particularly for airport baggage screening and medical examinations using X-ray imaging.

[0003] Various methods have already been proposed, but none gives complete satisfaction. PREVIOUS STATE OF THE ART

[0004] According to known techniques, multi-energy X-ray or Gamma radiation material discrimination systems are based on three measurement configurations: Multi-emission / Single-sensor, Single-emission / Multi-sensor, and Single-emission / Single-sensor. Multi-emission / Multi-sensor configurations are also proposed; however, they are comparable to Multi-emission / Single-sensor methods.

[0005] According to a first technique, the Multi-emission / Single-sensor solution was introduced in 1976 by Alvarez and Macovski, and has been the subject of numerous algorithmic developments, although the measurement principle remains unchanged.

[0006] According to this method, several energy profiles of X-rays are emitted in order to allow the differentiation of the materials most sensitive to these variations.

[0007] This method has several significant drawbacks. It is not suitable for imaging living organisms because unwanted movement between two X-ray emissions limits image resolution, and multiple radiation doses are contrary to radiation protection principles. For example, a note from the French Nuclear Safety Authority (ASN) dated October 2024 indicates that airport scanners installed in 2024 expose scanned objects to doses exceeding 26 mSv per pass. Furthermore, this method generates many false positives due to its statistical calculation method.

[0008] According to a second technique, the Mono-emission / Multi-sensor solution was introduced after 1985 by researchers Bames and Hickey.

[0009] According to this method, a single energy profile is emitted by an X-ray tube, the voltage and shape of whose collimator remain constant. The X-rays are captured by a juxtaposition of detectors, called sandwich detectors. Each detector exhibits distinct energy response properties.

[0010] This method has several major drawbacks. Although it addresses the problems of the Multi-Emission / Single-Sensor method regarding the measurement of living organisms, the signal-to-noise ratio (SNR – the result of differentiating the energy profiles of the X-ray measurements obtained on each sensor) of the device does not allow for an accurate measurement of the effective atomic number (Zeff) of organic materials ([Fig. 0B] and [Fig. 7B]). Moreover, the inherent limitations of the energy response profiles of the sensors prevent any intrinsic improvement in detection.

[0011] According to a third technique, the single-emission / single-sensor solution emerged in the 2000s with the advent of room-temperature semiconductor spectrometers, such as those using CZT (cadmium zinc telluride) semiconductors. The semiconductor spectrometer makes it possible to differentiate interactions related to photoelectric phenomena from those related to Compton scattering.

[0012] This method has several significant drawbacks. Analyzing the materials traversed requires interpreting the results. Consequently, the measurement is not direct but comparative. It is therefore necessary to know the profiles of the materials traversed and to perform a comparative statistical study. Thus, the probability of encountering a material corresponds to a cloud in a vector space, and not a simple point, and the characteristic clouds of the materials overlap, resulting in high uncertainty. Finally, obtaining a measurement profile is incompatible with the principles of radiation protection in the study of living organisms.

[0013] In accordance with the state of the art, the algorithms initially developed for one of the three aforementioned configurations have been adapted for application to the other configurations, thus making it possible to optimize their performance, complete their functionalities and provide additional reference parameters for the calculation of the effective atomic number (Zeff).

[0014] All existing solutions have significant limitations in terms of accuracy and reliability, hindering their effective deployment in critical applications requiring accurate and secure material identification. Examples include medical analysis of soft tissues or the ability to leave a bottle of water in luggage without risk of it being mistakenly identified as an explosive. DETAILED DESCRIPTION OF THE CLAIMED INVENTION

[0015] An object of the present invention is to propose a new solution overcoming the limitations of existing technologies, thus meeting the critical needs of the security, industrial and medical sectors.

[0016] To this end, an advanced method is proposed that corrects the drawbacks of the three processes previously presented. Its main characteristics are: easy industrialization, precise measurement of the atomic number using only two energy configurations, no need for pre-recorded approximations, adaptability to disparities in X-ray energy profiles (SNR), compatibility with all X-ray or gamma-ray sensors offered in industry, adaptability to uncontrolled environments, and limited exposure of scanned bodies to X-rays, allowing the analysis of living beings without risk.

[0017] The present invention relates to a method that contributes to the improvement of single-emission / multi-sensor processes, said method being also applicable to single-emission / single-sensor or multi-emission / single-sensor processes and, by extension, to multi-emission / multi-sensor configurations. Said method comprises a measuring device characterized by: a. An X-ray or gamma ray emission system, comprising: • At least one X-ray or gamma radiation source, arranged, for example, as one of the configurations in [Fig. 3B], comprising at least one X-ray tube equipped with a target, said target being made of at least one material enabling the Bremsstrahlung effect, for example, Tungsten, Rhodium and Rhenium, • A primary collimator configured to: limit the propagation of X-rays in unwanted directions and attenuate low-energy X-rays not relevant to the measurement; b. A detection system, comprising: • X-ray or gamma ray detectors arranged in at least two superimposed rows, said detectors being chosen from known technologies for detecting X-rays or gamma rays; • A secondary collimator adapted to prevent measurement disturbances caused by scattered radiation, • At least one filter positioned between said sensors, said filter being; made of metals, composite materials or plastic and is characterized by a predefined and constant thickness; c. At least one of the following elements being advantageously implemented interchangeably.

[0018] According to the embodiment of the invention, said method is implemented at by means of an innovative physical modeling, designated as the "hyperbolic arc tangent model" (hereinafter referred to as the "Argth model"), specifically developed and optimized for said embodiment.

[0019] According to a general embodiment, a scanner performs measurements based on the contrast difference between a reference signal obtained without a target and a signal obtained in the presence of a target, this difference being characterized by a factor Ka determined according to Beer-Lambert's law on the attenuation of light radiation, according to the expression: Ka = e M or X and d are distances. In the case of nuclear radiation, X is generally denoted 1 with the The quantity q / p represents the interaction cross-section. In a configuration where radiation passes through a plurality of elements, the value of the mass interaction factor q / p corresponds to a resultant representing the sum of the atomic numbers of all the materials traversed. This value can be determined using various calculation methods, including arithmetic and geometric methods.

[0020] According to the state of the art, and more particularly according to the work of Taylor et al. published in 2012 in the journal "Medical Physics" under the title "Robust Calculation of Effective Atomic Numbers", the optimal approach for determining said value consists of using the calculation of the effective atomic number (hereinafter referred to as "Zeff").

[0021] According to the prior art, the mass interaction factor q / p evolves according to the energy of the gamma or X radiations and according to the predominance of the type of interaction as shown in [Fig.1B] and [Fig.1A].

[0022] According to the device of the invention, in the boundary zone between photoelectric predominance and Compton predominance, we model the value of q / p by the function Argth(Z / E) such that: with x and ô two characteristic constants in R* and, advantageously Z=Zeff [Math 1] q / p = JL- . Argth (ôg?)

[0023] According to the device of the invention, we measure two different energy values ​​or distributions, a low value and a high value. We denote A = - In (low energy Ka) and B = - In (high energy Ka) •

[0024] By referring to the so-called mass attenuation coefficient method, denoted R, explained in the article "Detection of explosive materials in dual-energy X-Ray security Systems" of October 2022 by researcher Ozan Yalçm of îskender Atilla Reyhancan University published in Volume 1040 of the journal Nuclear Instruments and Methods in Physics Research, we can demonstrate that the Argth model used in the present invention is true in the context where the energy distributions of the emissions used respect the following approximation: [Math 2] R= 5 _

[0025] According to the device of the invention and the assumption of the value of q / p of the Argth method, we can show that Z is proportional to: [Math 3] with co a ratio of high and low energies used such that co = JL and, advantageously Z=Zeff.

[0026] According to the result defined above, the JL function is independent of the thickness of the materials traversed for energy spectra of X or gamma emissions whose statistical variance of the distribution function is less than 16. In the case of variance greater than 16, it is advantageous to compensate the JL modeling to maintain high measurement accuracy.

[0027] According to the result defined above, the compensation in the event of a variance in the energy distribution of X-rays or gamma rays greater than 16 can be a function of the distance d such that: Z = f( rd). For example, for a distribution of the energy spectrum in Gaussian form with a variance of 150, Z oc. + 0.01 X d- This correction reduces the influence of the thickness to less than 0.5% of the calculation result. Thus, we model the atomic number Z in a polynomial, hereafter referred to as the compensated Argth model, such as: [Math4]Z = P(^.e-<^.axd), with the values ​​d, the thickness of the material and a a calibration value belonging to R* depending on an energy emission profile and, advantageously Z=Zeff.

[0028] According to the result defined above, correcting the influence of thickness on energy distributions with a variance greater than 16, without knowing the thickness of the materials or the energy spectrum of the emission system, is possible using the Argth method. To this end, it can be shown that a weighting factor, based on the Argth model, defined by [equation missing], must be subtracted from the function of the model. of the Argth, with the values ​​co, cp and o of the calibration eigenvalues ​​defining the invention belonging to R. Thus, we model the atomic number Z in a polynomial called the weighted Argth model below such that: [Math 5] Z = P(JL . e-^ . A^¥), with advantageously Z=Zeff.

[0029] The values ​​o, cp represent the fitting and correction values ​​of the dependence functions of the material thickness, material density and the temperature. These values ​​are obtained by correcting the slope between [Fig.5A] and [Fig.5B] according to an embodiment.

[0030] The polynomial used to calculate Z, advantageously Zeff, is obtained by linear extrapolation from the calculation of the optimal polynomial, for example, which can be calculated using the Lagrange method, between the result points of the compensated Argth function or the moderated Argth function at different known values ​​of Z, advantageously Zeff, as shown in [Fig. 0A] and [Fig. 7A]. This polynomial is advantageously defined separately along segments of thickness and according to the nature of the materials present, as defined by: a. Atomic numbers from 1 to 8 are organic materials, b. Atomic numbers from 8 to 10 are light inorganic materials, c. Atomic numbers from 10 to 12 are heavy inorganic materials, d. Atomic numbers from 12 to 17 are light metals, e. Atomic numbers from 17 to 29 are the intermediate metals, f. Atomic numbers greater than 29 are dense metals.

[0031] According to an advantageous embodiment, the determination of the polynomial along thickness segments and according to the nature of the materials present for the determination of Z, advantageously Zeff, is obtained from a first result characterized by the double-energy mass attenuation coefficient R, as shown in [Fig. 8]. Advantageously, the value of the coefficient R allows the nature of the materials to be identified, while the ratio between the measurement in the open air and the measurement with the body allows the corresponding thickness segment to be defined.

[0032] The value obtained by calculating the weighted Argth model or the compensated Argth model gives a very precise result using a simple bijective affine line ([Fig. 0A] and [Fig. 7A]). However, the value obtained is sensitive to the measurement error characteristic of ionizing particle detection. By way of comparison, for a measurement error corresponding to the standard deviation of the distribution law, the evolution of the result of the double-energy mass attenuation coefficient, R, will be less than 3%, whereas that of the weighted Argth model can reach 50% depending on the configuration, and that of the compensated Argth model will be less than 5%.

[0033] It is therefore necessary to correct the measurement error, according to one of the state-of-the-art methods on correlated quantities, advantageously by an NF ISO 5725 approach, before calculating the atomic number Z, advantageously Zeff.

[0034] Depending on the embodiment, a semiconductor sensor or scintillator will preferably require a type A correction (statistical error), such as the inverse Student's t-test or the least squares method, and a current or voltage measurement sensor will preferably require a type B correction (systematic error) of metrological adjustment of the components present. In general, sensors require both types of error correction.

[0035] According to the device of the invention, the properties of the Argth model also make it possible to help in adjusting the errors of the input values ​​by relying on the property. The ratio of this approximation remains between 1%. and 5% regardless of the energy distributions emitted, and maintains a propagation of measurement error equivalent to the standard deviation of the distribution law of each measured quantity. This property allows us to define a reference point for error correction.

[0036] According to the device of the invention, these corrections make it possible to reduce the measurement error on the error equivalent to the variation of the mass attenuation coefficient at double-energy R. This correction makes it possible to work with lower measurement values ​​and consequently with a lower exposure dose.

[0037] According to the device of the invention, each measurement can represent a pixel and constitute an image, comparable to the sinograms of density measurements from a CT scanner or a baggage scanner. This sinogram of the atomic number measurement can, after Radon transformation and Fourier filtering, provide an image of the contrast of the atomic number variations or an image allowing the densitometric image to be colorized.

[0038] According to the device of the invention, the colors of the image can correspond to intuitive representation criteria, characterized by color gradients specific to each type of material, for example: a gradient from red to orange for organic materials, a gradient from blue to black for metals, and a gradient from yellow to green for inorganic materials.

[0039] The features of the invention mentioned above, as well as others, will become clearer upon reading the following description of an exemplary embodiment, said description being made in relation to the accompanying drawings. Brief description of the drawings

[0040] [Fig.1A] represents the relative importance of the three main types of Gamma-X ray interaction, taken from The Atomic Nucleus by RD Evans, 1955;

[0041] [Fig.1B] represents a resultant curve of the p / p value with respect to the energy of the photons, sum of the different energy-matter interactions, characterized by the LANL laboratory of Los Alamos;

[0042] [Fig.2A] represents a table of values ​​of the Zeff and the mass attenuation coefficient of dual-energy R taken from the publication "Detection of explosive materials in dual-energy X-Ray security Systems" by Ozan Yalçm;

[0043] [Fig.2B] represents the relationship between the double-energy mass attenuation coefficient R and the effective atomic number Zeff, when the X-ray emissions are mono-energetic;

[0044] [Fig.3A] represents the schematic diagram of the device configuration in Single-emission / Multi-sensor with sandwich sensors separated by a copper layer in the context of a customs or airport baggage scanner;

[0045] [Fig.3B] represents three CT scanner configurations in Multi-emission / Single-sensor with one X-ray source (A), Multi-emission / Single-sensor with several X-ray sources (B) and Multi-emission / Multi-sensors (C);

[0046] [Fig.4A] represents the distribution of the energy spectrum emitted by the X-ray tube and received by photodiode-type semiconductor sensors, without any intervening body between the tube and the sensors, in a Single-Emission / Multi-Sensor configuration with an intervening 0.2mm copper filter;

[0047] [Fig.4B] represents the distribution of the energy spectrum received by the photodiode-type semiconductor sensors, without any interposed body between the tube and the sensors, in a Multi-emission / Single-sensor configuration with two voltage values ​​of the X-ray tube and two collimators configured differently;

[0048] [Fig.5A] represents the result of the variation of the Argth method with respect to thickness and density with uncalibrated values ​​q> = 0 and o = 1.

[0049] [Fig.5B] represents the result of the variation of the Argth method with respect to thickness and density with the calibrated values ​​q> = 1.39 and o = 7.

[0050] [Fig.6A] represents the curve obtained from the relationship between the Argth method divided by 3 and Zeff for organic materials with the distribution of the energy spectrum of [Fig.4A] and the calibration of [Fig.5B].

[0051] [Fig.6B] represents the curve obtained from the relationship between the value of the mass attenuation coefficient at double-energy R and Zeff with a distribution of the energy spectrum of [Fig.4A].

[0052] [Fig.7A] represents the curve obtained from the relationship between the Argth method divided by 3 and Zeff for organic materials with the distribution of the energy spectrum of [Fig.4B] and the calibration of [Fig.5B].

[0053] [Fig.7B] represents the curve obtained from the relationship between the value of the mass attenuation coefficient at double-energy R and Zeff with a distribution of the energy spectrum of [Fig.4B].

[0054] [Fig.8] represents the algorithm for choosing the polynomial according to the nature of the material. DETAILED DESCRIPTION OF IMPLEMENTATION METHODS

[0055] The detailed description sets out two distinct embodiments of a measurement chain, namely: a first mode characterized by the determination of the weighted Argth polynomial characteristic of organic materials in a Single Emission / Multi-sensor configuration using a Tungsten X-ray tube, and a second mode defines the weighted Argth polynomial characteristic of organic materials in a Multi-Emission / Single-sensor configuration with a Tungsten X-ray tube.

[0056] According to the device of the invention, in a single-emission / multi-sensor measurement configuration, collimated X-rays (450) are emitted by a tungsten X-ray tube (310) powered by a voltage of 150kV. These X-rays are emitted to pass through an object or body, such as a suitcase (330) positioned on a conveyor belt (340) or a living subject (390), depending on the intended application, whether in the context of an airport scanner (301) or a computed tomography scanner (302).

[0057] The detection system is based on a sandwich-type sensor array (320), which consists of two distinct sensor layers dedicated to energy measurement. The first layer (321) is optimized for detecting low energies, while the second layer (323) preferentially detects high energies. These two layers are separated by an interlayer metallic layer; in this example, it is made of 0.2 mm thick copper foil, ensuring energy discrimination of the incident X-rays.

[0058] According to the device of the invention, when there is no body to be scanned interposed between the X-ray tube and the Sandwich detector (320), the low energy sensor (321) receives the energy profile (410) of the graph (401) of [Fig.4A] and the high energy sensor (323) receives the energy profile (420) which is the result of (410) after passing through the detector (321) and the filter (322).

[0059] When the body (330) or (390) is intercalated, the X-rays emitted by the tube will interact with the matter of the body according to its energy, as shown in (102) on the [Fig.1B] of the variations of the cross section, and according to its atomic number (101) as represented on the [Fig.1A] illustrating the relative importance of the interactions of the X-rays.

[0060] The variation in counting on the low energy detector (321) and high energy detector (323), compared to the no-load measurement respectively (410) and (420), will allow us to define the nature of the elements traversed which strongly interacted on the energies in the interaction elbow from 40keV to 200keV of the curve (120) of the graph (102) and of atomic number from 1 to 30, where the Compton and photoelectric effect overlap as shown on the curve (110) of the graph (101).

[0061] According to the device of the invention, the measurement value for high energy (812), denoted C, and for low energy (811), denoted D, which can be in the form of a Counting or measuring current or voltage is corrected by adding a correction value to reduce the impact of the statistical measurement error, such as Ce = C + c (821) and De = D + d (822). An algorithm based on the NF ISO 5725 approach, for example, least squares using measurements of adjacent pixels on a 3x3 pixel square (9 values), and the defining property of the Argth method, namely ~g-^, allows us to reduce the intrinsic measurement error and define the c and d probable to less than 1% error. For example, the error correction algorithm following; for all circumferential measurements of C and D such that, Ci - {c <c<Ci+ / c and Di - < D < Di + and for each measurement of the Ci and Di of the 8 surrounding pixels, plus the pixel to be corrected, have a ratio of 1 < ) x 1.000 < 3 [Math 2]then, V A. I / c = + c) e <D= 2'Jd + d) 2. This helps to reduce the error The measurement error is at least a factor of 10. This algorithm is possible provided the measurement is taken on a sensor bar, such as those found in medical or baggage scanners. With simple sensors, it is possible to repeat the same measurement multiple times.

[0062] The corrected values ​​Ce and De are divided by the reference measurements at no load, without the bodies (330) or (390), of the sensors (321) and (323) to give the Ka baSSe coefficients energy and high-energy Ka allow us to obtain the values ​​of A = -In (low-energy Ka) and B = -In (high-energy Ka). From these values, the double-energy mass attenuation coefficient R (830) is defined, allowing the appropriate choice of the polynomial (840) for calculating the Zeff as a function of the segment thickness and the nature of the materials analyzed. The values ​​of A and B are directly used by this polynomial to define the value of the Zeff and by the same token, identify the nature of the materials using the Zeff table (201).

[0063] Prior to determining Zeff, the device of the invention requires calibration of the values ​​co, q> and o for each segment of thickness and nature of the materials analyzed. The value co is determined by estimating the energy difference of the expected values ​​of the two low and high energy profiles such that co = JL. The coefficients q> and o are correction factors for the actual energy profiles exiting X-ray tubes.

[0064] The variation of the dependence function between the calculated value and Zeff is strongly dependent on the energy profile present. This evolution is particularly visible in the difference in shape between curve (210) in [Fig. 2B], representing the dependence function between the double-energy mass attenuation coefficient R, theoretically defined in the literature with single-energy spectra energy and the function (620) of the same coefficient plotted in [Fig.6B] with spectra from X-ray tube.

[0065] During calibration, and according to the energy profile of the emitted X-ray tube, the procedure consists of taking reference materials (501), defining the slope of dependence between these materials with respect to the variation in thickness, for example, from 1 to 15 cm of water (521) and a second material, with a sufficiently different Zeff, such as 1 to 15 cm of cannabis resin (511), and adjusting the factor q> so that this slope becomes zero for water (512) on the correction graph (502) [Fig. 5B] and the factor o to homogenize the correction so that it is as close to 0 as possible for cannabis (522). In the case of our example energy profile, the correction value q> is and o is 7 for a co = 3.

[0066] The identification of the Zeff is obtained by determining the linear extrapolation polynomial of the curve (610) between the reference values ​​of the calculation of JL q-3*-^ _ ^_J__ e-7x^ and 'c Zeff representing the reference material. The Graph (601) of [Fig.6A] shows the result in the proposed embodiment for organic materials with atomic numbers from 4 to 8. A first-degree polynomial y = a X + b with, following the example of the embodiment, a = -1.72 and b = 18.7. This polynomial allows for a precision of less than 1% (theoretical, not considering the propagation of measurement error).

[0067] The nature of the materials according to the determined Zeff is done by searching for the closest value in a list table of Zeff with respect to the materials (201).

[0068] According to the device of the invention, 2D and 3D images can be produced by a multitude of Zeff measurements organized in matrix form of sinogram and processed by calculation with the Radon transform and Fourier filter.

[0069] According to the device of the invention, in a multi-emission / single-sensor (360) or, by extension, multi-emission / multi-sensor (370) measurement configuration, collimated X-rays (450) are emitted by at least two tungsten X-ray tubes (382) and (383) firing pulsed X-rays one after the other or simultaneously (384) and (385) and being powered with two separate voltages of 120 kV and 150 kV. These X-rays are emitted to pass through an object or body, such as suitcases (330) or a living subject (390), depending on the intended application, whether in the context of a dual scanner (370) or a medical computed tomography scanner (360). The detection system is based on a single (387) or dual (388) and (389) sensor line.

[0070] According to the device of the invention, when there is no body to be scanned interposed between the X-ray tube and the detector (387) or detectors (388) and (389), the X-ray tube (382) or (384) emits the low-energy profile (430) of the graph (401) of the [Fig.4A] and the X-ray tube (383) or (385) emits the high energy profile (440).

[0071] The position of the body (330) or (390) between the X-ray tubes modifies the interaction of the emitted X-rays. These rays, coming from the tubes (382), (383), (384) or (385), will interact differently with the matter according to two main parameters: the energy of the X-rays, the impact of which is illustrated by the variation of the cross section shown in [Fig.1B] (102), and the atomic number of the body, the relative importance of whose interactions is visualized in [Fig.1A] (101).

[0072] The difference in counting observed on the detector(s) following shots from the high-energy tube (383) or (385) generating the X-ray energy profile (440) and shots from the low-energy tube (382) or (385) generating the X-ray energy profile (430), compared to no-load measurements, makes it possible to identify the nature of the traversed elements that significantly interacted. This analysis focuses on the energy range between 40 and 200 keV, as represented on curve (120) of graph (102), and for elements with atomic numbers from 1 to 30. In this area, a characteristic overlap of the Compton and photoelectric effects is observed, as illustrated on curve (110) of graph (101).

[0073] According to the device of the invention, the measurement value for high energy, denoted C, and for low energy, denoted D, which can be in the form of a count or a current or voltage measurement, are corrected by adding a correction value to reduce the impact of the statistical measurement error, such that Ce = C + c and De = D + d. An algorithm based on the NF ISO 5725 approach, for example least squares using measurements of adjacent pixels, on a 3-pixel by 3-pixel square, i.e., 9 values, and the defining property of the Argth method, namely __ q-^, allows us to reduce the intrinsic error of the measurement and to define the probable c and d at less than 1% error. For example, the following error correction algorithm; for all circumferential measurements of C and D such that, C, and D i - <D <D i + , et pour chaque mesure des Ci et D; des 8 pixels entourant, plus the pixel to be corrected, have the ratio of 1 < I ) x 1000 < 3 [Math 2] then, v jA / / C = + c^'D = ^(ü + d) 2. This helps to reduce the error The measurement error is at least a factor of 10. This algorithm is possible provided the measurement is taken on a sensor bar, such as those found in medical or baggage scanners. When measuring on a sensor, it is possible to repeat the same measurement multiple times.

[0074] The corrected values ​​Ce and De are divided by the reference measurements at no load, without the bodies (330) or (390), of the sensors (321) and (323) to give the Ka baSSe coefficients High-energy Ka and energy Ka allow us to obtain the values ​​of A = -In (low-energy Ka) and B = -In (high-energy Ka). From these values, the double-energy mass attenuation coefficient R (830) is defined, enabling the appropriate selection of the polynomial (840) for calculating the Zeff based on the thickness of the segments and the nature of the materials analyzed. The values ​​of A and B are directly used by this polynomial to define the Zeff value and thereby identify the nature of the materials using the Zeff table (201).

[0075] Prior to determining Zeff, the device of the invention requires calibration of the values ​​co, q, and o for each segment of thickness and the nature of the materials analyzed. The value co is determined by estimating the energy difference between the expected values ​​of the two low- and high-energy profiles such that co = JL. The coefficients q and o are correction factors for the actual energy profiles exiting the X-ray tubes.

[0076] The dependence function between the calculated value and Zeff varies significantly depending on the energy profile encountered. This evolution is particularly manifested by the difference in shape between the curve (210) of [Fig.2B], which represents the dependence function of the dual-energy mass attenuation coefficient R defined theoretically in the literature with mono-energetic spectra, and the function (720) of the same coefficient plotted in [Fig.7B] which uses spectra from X-ray tubes (430) and (440).

[0077] During calibration, and according to the energy profile of the emitted X-ray tube, the procedure consists of taking reference materials (501), defining the slope of dependence between these materials with respect to the variation in thickness, for example, from 1 to 15 cm of water (521) and a second material, with a sufficiently different Zeff, such as 1 to 15 cm of cannabis resin (511), and adjusting the factor q> so that this slope becomes zero for water (512) on the correction graph (502) [Fig. 5B] and the factor o to homogenize the correction so that it is as close to 0 as possible for cannabis (522). In the case of our example energy profile, the correction value q> is JL and o is 9 for a co = 1.

[0078] The identification of the Zeff is obtained by determining the linear extrapolation polynomial of the curve (610) between the reference values ​​of the calculation of .JL . _ ZL x 1,0 3 X ct 'c Zeff representing the reference material. Graph (601) in [Fig. 6A] shows the result in the proposed embodiment for organic materials with atomic numbers from 4 to 8. A first-degree polynomial y = aX + b + b with, following the example of the embodiment, a = -0.414 and b = 9.29. This polynomial allows for a precision of less than 1% (theoretical, not considering the propagation of measurement error).

[0079] The nature of the materials according to the determined Zeff is done by searching for the closest value in a list table of Zeff with respect to the materials (201).

[0080] According to the device of the invention, 2D and 3D images can be produced by a multitude of Zeff measurements organized in matrix form of sinogram and processed by calculation with the Radon transform and Fourier filter.

[0081] POTENTIAL INDUSTRIAL APPLICATIONS OF THE INVENTION

[0082] The techniques of the invention allow for the precise identification of the materials present. This innovation offers a solution for significantly speeding up passage through high-traffic areas such as airports, as it eliminates the need to remove items from luggage. It also simplifies medical analyses by identifying soft tissues or highlighting fluids without the addition of allergenic contrast agents.

[0083] The developed solution offers a high capacity for discriminating between different liquids and gels such as water, oil, creams, and explosive or illicit products like drugs. This precision simplifies security procedures: identifying drugs or explosives in baggage without the need for smears for chemical testing.

[0084] This innovative technique has promising medical applications, notably the improved visualization of human soft tissues. It allows for the precise identification of Zeff variations over an area of ​​the same density; this makes it possible, for example, to differentiate lymph from blood, whose densities are very similar (1.1 kg / dm³ and 1.06 kg / dm³).

[0085] Moreover, this method should make it possible to revolutionize the early diagnosis of cancer by enabling the identification of cancerous nodules that have absorbed a specific sugar, thus offering an alternative to traditional PET SCAN techniques which lack precision and cannot be used without strong motivation due to the cost of imaging and especially due to the high radioactive dose received by the patient.

Claims

Demands

1. A method for measuring the atomic number Z of a material using X-ray or gamma radiation, characterized in that it implements a physical model based on a hyperbolic arctangent function relating the mass absorption coefficient p / p to the atomic number Z, the incident photon energy E and constants specific to the material traversed, said model being applied from attenuation measurements obtained for two distinct X or gamma energy profiles, said measurements being corrected to compensate for statistical errors of type A, systemic errors of type B, or both, the method being applicable to imaging or measurement devices in the following configurations: multi-emission / single-sensor, single-emission / multi-sensor, single-emission / single-sensor, or multi-emission / multi-sensor.

2. Method for measuring the atomic number Z by X or gamma radiation according to claim 1, characterized in that the physical model uses a hyperbolic arctangent function to determine the mass absorption coefficient p / p according to the formula p / p = Arffth (ô.^)' °where 'c is the atomic number, E is the energy of the incident photon, and ô and / are constants characteristic of the materials traversed.

3. A method for measuring the atomic number Z by X-ray or gamma radiation according to claim 1 or 2, characterized in that the atomic number Z is determined as being proportional to an expression derived from the mass absorption coefficient p / p modeled by the hyperbolic function, of the form: Z oc . q-co^, in which A = -ln(low energy Ka) and B = -ln(high energy Ka), with low energy Ka and high energy Ka being coefficients obtained by the ratio between an attenuation measurement, or an attenuation measurement corrected for systematic errors of type A or B, and a reference measurement without entities, all carried out under radiation, respectively of low and high energies, and co is a ratio between the high and low energies used.

4. A method for measuring the atomic number Z by X-ray or gamma radiation according to claim 3, characterized in that the atomic number Z is determined using a compensated polynomial function, derived from the physical model based on the arc- function hyperbolic tangent, according to the formula Z = P(^. - a X where P is a polynomial function, .JL is the ratio defined in claim 3, d is the thickness of the material traversed, and a is a calibration value dependent on the emission energy profile.

5. Method of measuring the atomic number Z by X or gamma radiation according to claim 3, characterized in that the atomic number Z is determined using a weighted polynomial function, derived from the physical model based on the hyperbolic arctangent function, according to the formula Z = P(.JL - A(p.e'a^^ °where P is a polynomial function, .JL is the ratio defined in claim 3, q> and o are calibration values ​​depending on the emission energy profile, and co is the ratio between the high and low energies defined in claim 3.

6. Method for measuring the atomic number Z by X-ray or gamma radiation according to any one of claims 1 to 5, characterized in that calibration values ​​a, cp and a, defined in claims 4 and 5, are determined as a function of a first estimate of the atomic number Z obtained using the mass attenuation coefficient R = B / A.These calibration values ​​are adjusted using coefficients of a polynomial specific to each range of atomic numbers representative of a type of material traversed, so that: for atomic numbers from 1 to 8, the coefficients of a polynomial specific to organic materials are applied; for atomic numbers from 8 to 10, the coefficients of a polynomial specific to light inorganic materials are applied; for atomic numbers from 10 to 12, the coefficients of a polynomial specific to heavy inorganic materials are applied; for atomic numbers from 12 to 17, the coefficients of a polynomial specific to light metals are applied; for atomic numbers from 17 to 29, the coefficients of a polynomial specific to intermediate metals are applied. #for atomic numbers greater than or equal to#29, the coefficients of a polynomial specific to dense metals are applied, said. polynomials being derived from a physical model based on a hyperbolic arctangent function.

7. A method for measuring the atomic number Z by X-ray or gamma radiation according to any one of claims 1 to 6, characterized in that attenuation measurement values ​​A and B of the Lambert Béer law, obtained from the passage of X-ray or gamma radiation through a traversed material, are corrected for errors according to a method conforming to standard NF ISO 5725, or according to a correction of type A errors, type B errors, or both, so that attenuation measurements satisfy a ratio condition 1 < X 1000 < 3, where R = is the mass attenuation coefficient and D = a-Ar, said condition being derived from a proportionality property between R and D in the physical model based on the hyperbolic arctangent function.

8. Method for measuring the atomic number Z by X or gamma radiation according to any one of claims 1 to 7, characterized in that a calibration of the calibration value a, defined in claim 4, is carried out to minimize the influence of the thickness of the materials traversed on the function Z = P(JL - a X d^ said calibration being carried out using attenuation measurements carried out on at least two calibration materials, within the framework of the physical model based on the hyperbolic arctangent function.

9. Method for measuring the atomic number Z by X or gamma radiation according to any one of claims 1 to 7, characterized in that a calibration of the calibration values ​​q and o, defined in claim 5, is carried out to minimize the influence of the thickness of the materials traversed and of an energy distribution of the emitted X or gamma radiation on the function Z = P(g^ - A(p) said calibration being carried out using attenuation measurements carried out on at least two calibration materials, within the framework of the physical model based on the hyperbolic arctangent function.

10. A method for measuring the atomic number #Z by X-ray or gamma radiation according to any one of claims #1 to #9, characterized in that each value of the atomic number #Z, calculated #:# using a model based on a hyperbolic arctangent function according to claim #2, or a compensated polynomial according to claim #4, or a weighted polynomial according to claim #5, is attributed to an image element, constituting: a pixel of a reconstructed two-dimensional image, or a voxel of a tomographic volume or a sinogram, generated by an X-ray or gamma-ray imaging device.

Images