METHOD AND DEVICE FOR MULTI-DIMENSIONAL DIRECTIONAL MEASUREMENT OF GAMMA RADIATION IN THE DISTANT FIELD

Description

[0001] The invention relates to a method and a device for radiation detection, in particular a direction-measuring detector system for measuring the directional distribution of radiation intensity. The method supports both 3-dimensional and 2-dimensional direction measurement of one or more radiation sources in the far field.

[0002] In NBC and radiation protection, there is a demand for handy, compact radiation measuring devices that can provide directional information. Unlike medical imaging, where activity concentrations are displayed, directional measurement here means determining the directions of one or more radiation sources to assist in locating them.

[0003] Compact gamma spectrometers and nuclide identification devices are state-of-the-art. However, such devices cannot determine the angle of incidence of the radiation.

[0004] Currently, locating radiation sources in the field is only possible if emergency personnel walk or drive through the entire area with a radiation measuring device to establish an intensity distribution. This process is very complex and not always applicable, for example, when locating radiation sources in inaccessible terrain.

[0005] Complex devices composed of many individual detectors are known from the clinical setting and are used for medical imaging. These devices are complex and require complex algorithms for data analysis, which require considerable computing power. The algorithms calculate the activity concentrations of a radiopharmaceutical in a patient in the immediate vicinity of the radiation detectors (near field). In principle, these algorithms can also be used for the aforementioned compact, direction-resolving devices, provided they are adapted to the operating conditions in NBC and radiation protection.

[0006] For this purpose, the methods must be switched from the near field to the far field. The computational effort must be reduced to such an extent that real-time data processing is possible in parallel with data acquisition, so that the results are available live during the measurement. Furthermore, there are differences in radiation intensity. When detecting radioactive radiation sources, the radiation intensities recorded at the detection site are often far below those commonly used in medicine. This places increased demands on the algorithms, which must deliver reliable and reproducible results despite statistical limitations of the data sets. The directional measurement must still function reliably even when the radiation intensity of the sources to be measured is below the level of the natural radiation background. It is therefore necessary that the natural radiation background be taken into account in the evaluation.To locate radiation sources in the field, fast methods are needed that reconstruct a radiation far field, can be used even with low counter statistics and take into account the influence of natural radiation.

[0007] Several methods are known that have been developed for locating radiation sources in the field. US2012 / 0043467 describes a method for single-plane Compton cameras that is suitable for measuring the direction of a radiation source. However, this method has the disadvantage that it is limited to measuring the direction of one radiation source per radionuclide and cannot measure the directional distribution of multiple radiation sources. Furthermore, US 8,461,547 B2 describes a method that uses the directional information from Compton scattering processes to determine the direction of a radiation source. From GUEORGUIEV A ET AL: "A novel method to determine the directionality of radiation sources with two detectors based on coincidence measurements",NUCLEAR SCIENCE SYMPOSIUM CONFERENCE RECORD (NSS / MIC), 2010 IEEE, IEEE, October 30, 2010, pages 1525-1530, DOI: 10.1109 / NSSMIC.2010.5874031 ISBN: 978-1-4244-9106-3, is a method for determining the direction of radiation sources using two detectors and based on coincidence measurements. A set of coincidence events from each detector is recorded in different spectra. The energy distribution is statistically analyzed. The mean of each distribution is proportional to θ from the Compton scattering formula f(θ) = (EL - ER ) / (EL + ER ).

[0008] The state-of-the-art methods do not make use of image reconstruction techniques; instead, simple statistical evaluation techniques are used.

[0009] The task therefore aims to provide a method for compact, direction-resolving radiation measuring devices that can also be used for combined systems consisting of several radiation detectors and for segmented radiation detectors.

[0010] The invention therefore also aims to overcome the disadvantages of the prior art methods. In particular, it aims to provide reliable and efficient location of multiple radiation sources without any limitation on the number of radiation sources.

[0011] The objects are achieved by the method and device according to the independent claims. Advantageous further developments and preferred embodiments are specified in the subclaims.

[0012] The method according to the invention for multidimensional directional measurement of gamma radiation in the far field by means of a group of several energy-discriminating, mutually synchronized detectors for detecting radiation is characterized in that the method uses uni- and bidirectional Compton scattering processes and lookup tables LUT SK< , a defined function value f(E1,E2), a list of defined detector pairs with an identification numberi for defined detector pairs and one or more frequency distributions Y used to record measured values. The method according to the invention comprises the following steps: a) Setting up of the detector system for a measurement, comprising the following steps: Create a list of defined detector pairs, where the defined detector pairs include all pairs that are formed combinatorially from the set of detectors and each pair contains at least one detector made of a material with an atomic number of Z eff > 30, and these with an ID number i provides; Interconnection of all detectors in a coincidence circuit such that coincidence events in all defined detector pairs i = 1, ..., I be recorded; Mark of the two detectors of each defined detector pair iwith the numbers 1 and 2 respectively, whereby the detector with the lower atomic number receives the number 1 and the one with the higher atomic number receives the number 2; if both detectors are made of the same material, the designation as 1 or 2 is made arbitrarily; Define a function f(E1,E2) = (E2-E1) / (E1+E2) which is calculated from two energy values ​​E1 and E2; b) Acquiring measured values of coincidence events, if in two detectors each of all defined detector pairs i simultaneous interactions take place, whereby the measured values ​​originate from a radiation distribution in the far field and the measured values ​​are the interaction energies E1 and E2 of the radiation measured in the detectors; c) Associate of coincidence events with an identification number i ; d) Calculate the function value f(E1,E2) from two energy values ​​E1, E2 per coincidence event with the function f(E1,E2) defined in step a); e) Recordingof the coincidence events according to their identification number i and their function values ​​f(E1,E2) in one or more frequency distributions Y , where for each radionuclide a separate frequency distribution Y is present, f) Calculate one or more directional distributions X from the frequency distributions Y with a statistical image reconstruction method of emission tomography using lookup tables LUT SK< , where for each radionuclide a separate directional distribution X is present.

[0013] The inventive method for measuring the direction of a radiation distribution in the far field applies equally to interconnected systems comprising multiple radiation detectors and to segmented radiation detectors. In one case, a group of multiple radiation detectors is interconnected; in the other case, the active detector medium is divided into several units. The group of radiation detectors or the segmented individual detector each form a temporally synchronized system. Each detector or each segment can measure the energy of the radiation deposited there through interaction and can be synchronized with other detectors or segments. For the purposes of the invention, a detector is understood to be a radiation detector or a detector segment that can detect radiation and be individually read out.

[0014] It is assumed that the measurement system is equipped with list-mode data acquisition and can process the data of each detector or segment.

[0015] The method according to the invention is particularly suitable for stationary or quasi-stationary direction measurement, with two embodiments being particularly advantageous, namely the 3-dimensional direction measurement and the 2-dimensional direction measurement.

[0016] Two-dimensional direction measurement is a special case of direction measurement when all radiation sources are located in one plane, for example, the horizontal plane. Arrays of planar radiation detectors are suitable device types for two-dimensional direction measurement. The device is oriented so that the detection plane is identical to the plane of the radiation sources.

[0017] The method for measuring direction according to the invention is also called direction-resolved coincidence spectroscopy. In coincidence spectroscopy, coincidences are investigated in which interactions occur simultaneously in two detectors (or segments) of the measuring device. Pairs of two detectors each are the basic imaging elements of the direction measurement. The energy inputs E1 and E2 recorded in a coincident detector pair are considered together with the spatial orientation of the detector pair.

[0018] Directionally resolved coincidence spectroscopy makes use of two interaction processes of gamma radiation: Compton scattering and photoelectric absorption.

[0019] The measuring device is considered a combination of detector pairs. First, some definitions of detector pairs will be given.

[0020] The detector pairs to be used for the direction measurement are selected. When making the selection, it is helpful to first consider the materials from which the detectors are constructed. It is sufficient to divide the materials into two classes. One group is formed by materials with a medium to high atomic number. For the purposes of the invention, a medium to high atomic number is defined as a number Z eff of greater than 30. If the detector materials are chemical compounds, then Z eff the effective atomic number, ie the average atomic number of all elements contained in the compound, taking into account the atomic masses of the elements and their stoichiometric composition.

[0021] For detector materials of medium to high atomic number Z eff> 30 include, for example, NaI, a frequently used material in scintillation detectors, and the rare earth halides such as CeBr3, LaBr3, LaCl3, and La(BrxCl1-x)3. Among the semiconductor materials with medium to high atomic numbers, Ge, GaAs, CdTe, and CdZnTe are particularly noteworthy. These materials are characterized by their high probability of photoelectric absorption in the energy range from a few tens of keV to 3 MeV.

[0022] The other group includes detectors made of low atomic number materials, whose atomic number Z eff is less than or equal to 30. These include plastic scintillators, organic crystals such as anthracene, stilbene, or p-terphenyl, pure and doped inorganic crystals such as CaF 2 :Eu, and the semiconductor material silicon. Materials with low atomic numbers Z eff ≤ 30 have a high probability of Compton scattering in the energy range from 100 keV to 3 MeV.

[0023] These two groups of detectors can be combined into two types of detector pairs, which are used to determine the direction of the incident radiation. A distinction is made between unidirectional and bidirectional detector pairs. Both types of detector pairs differ in terms of the direction of the scattered radiation.

[0024] A unidirectional detector pair contains both a detector from the low atomic number group and a detector from the medium to high atomic number group.

[0025] For example, if a plastic detector and a cerium bromide detector are combined, this is a unidirectional detector pair. In a unidirectional detector pair, each detector has a unique physical function. If a coincidence event with the radiation energy E1+E2 corresponding to the nuclide is observed, it can be determined beyond doubt that the incident radiation was first scattered in the plastic detector before being absorbed in the cerium bromide detector. The converse case, in which the radiation was first scattered in the cerium bromide detector before being absorbed in the plastic detector, has a probability close to zero.

[0026] In a unidirectional detector pair, the direction of the scattered Radiation is known. The radiation is scattered in the low-atomic-number detector toward the high-atomic-number detector.

[0027] Bidirectional detector pairs consist of a combination of two detectors from the medium and high atomic number group, for example, cerium bromide with cerium bromide or germanium with germanium. When using segmented semiconductor detectors, pairs of two segments are also bidirectional, since both segments are made of the same material.

[0028] In a bidirectional pair, unless further information is available, it is not possible to determine which of the two detectors scattered the incident radiation first. The direction of the scattered Radiation remains unknown. The radiation may have been scattered from detector 1 to detector 2, or from detector 2 to detector 1.

[0029] It should be noted that both uni- and bidirectional detector pairs are effective in determining the direction with respect to the incident radiation. The direction of the incidentRadiation cannot be unambiguously reconstructed from a single coincidence event. For the 2-dimensional case, the direction of incidence is fixed on a V-shaped line, while for the 3-dimensional case, it is fixed on a conical surface. The measurement data from a single event do not allow any conclusions about which arm of the V-shaped line or in which direction on the conical surface the radiation source lies.

[0030] Regarding the directional reconstruction of the incident Unidirectional and bidirectional detector pairs provide equivalent information for radiation. The measured data of many coincidence events possess the statistical properties required for directional measurement in both cases. The fact that the direction of the scattered radiation in a bidirectional detector pair is unknown, does not represent an experimental limitation on the direction determination of the incident radiation.

[0031] As will be explained in more detail below, the inventive technique of direction-resolved coincidence spectroscopy has as its essential distinguishing feature that it can be used for both uni- and bidirectional detector pairs.

[0032] The detectors of a device can be divided into different pairs. The number of possible combinations is determined by the detector materials. Detectors from the low atomic number group can be combined with everyone Detector from the other group with medium or high atomic number. Detectors from the group with medium or high atomic number can even be combined with everyone other detector in the device to form a pair.

[0033] The group of detectors with low atomic number, however, is not suitable for assembling intra-group pairs.

[0034] A measurement process can use a combination of unidirectional and bidirectional detector pairs or use exclusively unidirectional or exclusively bidirectional pairs for directional measurement.

[0035] After the uni- and bidirectional detector pairs have been selected and assigned an identification number i are provided, the two detectors of each pair are labeled 1 and 2 respectively. If a pair contains different detector materials, the detectors of the material with the lower atomic number are labeled 1, and the detectors of the material with the higher atomic number are labeled 2. If the pairs consist of detectors of the same material, the decision as to which detector in a pair is labeled 1 or 2 can be made arbitrarily, as long as it is consistently maintained in the data analysis.

[0036] In the following, it is assumed that all radiation detectors are calibrated for energy measurement. During the measurement process, coincidence events are acquired that form a triple based on their energy values ​​E1 and E2 in detectors 1 and 2 and the spatial orientation of the detector pair. The spatial orientation is determined by the identification number. iof the detector pair. The two energy values ​​E1 and E2 contain the information that, on the one hand, identifies the nuclide whose radiation is being observed, but, on the other hand, also measures the direction from which the radiation reaches the detector pair. Due to energy conservation, the sum of the recorded energy inputs E1+E2 is fixed at the radiation energy of the nuclide. Both measured values ​​E1 and E2 are strongly correlated with each other, and the directional information is essentially contained in the difference E2-E1. It is therefore possible to reduce the number of variables by defining a new measured value f(E1,E2) from the two energy values ​​E1 and E2, which approximately represents the directional information. The measured data, initially organized as a triplet ( i, E1, E2) can thus be combined to form a duplicate with the two characteristics i and f(E1,E2) are reduced.

[0037] Since stationary or quasi-stationary measurement conditions are assumed, the measurement data can be represented as a 2-dimensional frequency distribution Y Each event is grouped according to its two characteristics, i and f(E1,E2), into the frequency distribution Y To record the measurement data, for each detector pair i A histogram is created. For the histograms, the value range of f(E1,E2) must be divided into classes. If a coincidence event is registered, a suitable function value f(E1,E2) can be calculated from the two energy values ​​E1 and E2 in detectors 1 and 2, respectively. For the purposes of the invention, the function is defined as follows: f E 1 , E 2 = A = E 2 − E 1 E 1 + E 2

[0038] Substituting E2 with C-E1 in Equation (1) yields the function f(E1) = 1 - 2E1 / C, which is uniquely defined, continuous, and monotonic over the entire interval [0,C]. This function, with the properties mentioned, is suitable for both unidirectional and bidirectional detector pairs. In Compton cameras, however, the Compton angle ϑ is often used. ϑ = cos − 1 1 − m c 2 E 2 + m c 2 E 1 + E 2 with mc 2< = 511 keV, the rest energy of the electron.

[0039] However, equation (2) is not a permissible function f(E1,E2) in the sense of the invention, since ϑ( E 1) is not uniquely defined on the entire interval 0 ≤ E1 ≤ C.

[0040] A coincidence event is sorted according to its value f(E1,E2) into the histogram that corresponds to the detector pair i in which the coincidence occurred. The total histograms for all detector pairs i = 1, ..., I forms the 2-dimensional frequency distribution Y which is continuously filled during a measurement.

[0041] From the recorded frequency distribution Y the directional distribution is now X of the radiation sources. Depending on the experimental situation, two cases are distinguished. In the 3-dimensional directional measurement, the directional distribution is determined as a function dependent on two variables X ( ω,h ) on the celestial sphere. The celestial sphere is a sphere commonly used in astronomy and navigation to determine positions in the sky. It is an apparent sphere of any radius surrounding the measuring device on all sides, onto which the radiation sources are projected. The function value X ( ω, h ) describes the radiation intensity in a solid angle segment d Ω on the celestial sphere in the azimuth direction ω at the elevation angle h.

[0042] In 2-dimensional direction measurement, a function dependent on a variable X ( ω ) along the horizon circle. The function value X ( ω ) describes the radiation intensity in the direction ω . It is assumed that the sources are in the same plane as the detectors of the measuring device.

[0043] The task of the algorithm is to calculate the function X ( ω, h ) or X ( ω ) to determine.

[0044] The spatial directional distribution X ( ω, h ) of the radiation field (or X ( ω) in the 2-dimensional case) can now be calculated using any statistical image reconstruction method, as known from emission tomography. The statistical reconstruction methods allow uniform processing of the measurement data from the entire measuring device, regardless of whether a particular detector pair is unidirectional or bidirectional.

[0045] Common statistical image reconstruction methods include: Maximum Likelihood Expectation Maximization (MLEM) algorithm Ordered Subset Expectation Maximization (OSEM) algorithm List Mode - Maximum Likelihood Expectation Maximization (LM-MLEM) algorithm List Mode - Ordered Subset Expectation Maximization (LM-OSEM) algorithm

[0046] All of the methods mentioned are examples of algorithms whose application is not limited to medicine. Using the approach described above, according to this invention, they can also be readily used in NBC and radiation protection.

[0047] The method according to the invention is described in more detail below using exemplary embodiments and drawings. These explanations are merely exemplary and do not limit the general concept of the invention.

[0048] According to a further advantageous embodiment of the method according to the invention, the radiation sources emit a discrete and / or continuous distribution of radiation.

[0049] According to a further advantageous embodiment of the method according to the invention, the radiation sources emit not only gamma radiation but also particle radiation from a particle group of electrons, positrons, protons, ions and / or neutrons.

[0050] According to a further advantageous embodiment of the method according to the invention, the radiation originates from the radioactive decay of one or more radionuclides and / or the radiation is the prompt gamma radiation which arises during the absorption of proton or ion radiation in target materials.

[0051] According to a further advantageous embodiment of the method according to the invention, the radiation has a low intensity.

[0052] According to an advantageous embodiment of the method according to the invention, step a) additionally comprises, in part or in whole, the following sequence of steps: Calibrate the signals from all detectors as absorbed radiation energy E; Establish a suitable coordinate system; Capture the directions of the defined detector pairs, where in the 2-dimensional direction measurement the direction of detector pairi with the azimuth angle ϕ i is detected and in the 3-dimensional direction measurement the direction of the detector pair i with the azimuth angle ϕ i and the elevation angle β i is recorded; Classify of the measuring range for the function value f(E1,E2) into a number J equidistant measured value channels j = 1, ..., J ; Invest one or more 2-dimensional arrays with I · J Fields as data structures for storing frequency distributions Y , in which coincidence events according to their ID number i and its measured value channel j registered; for each radionuclide a separate 2-dimensional array Y created; Classify the detector pairs i in symmetry classes SK , where those detector pairs i, which are imaged onto other identical detector pairs when rotated or translated, to a symmetry class SK ( i ) are summarized; Invest one or more lookup tables LUT SK< for each symmetry class SK and each radionuclide, which has an angular range ϑ span from 0° to 180° and are divided into equidistant angular steps; Create the lookup tables LUT SK< by measurements with the detector system or by Monte Carlo simulations or by means of a theoretical model; Hand over all lookup tables LUT SK< for all symmetry classes and all radionuclides to an image reconstruction algorithm, which processes the measurement data and calculates the directional distributions X calculated.

[0053] Preferably, each radiation detector of a measuring device is calibrated for energy measurement. For radiation detectors in the medium to high atomic number group, the photoelectric effect can be used for energy calibration. For such detectors, energy calibration is performed using the photopeaks in the pulse height spectrum. This procedure is well known to experts in the field of radiation detection and will not be discussed further here.

[0054] However, energy measurement is less common for detectors in the low atomic number group, as these are often operated in pulse counting mode. Several calibration options are available for detectors with Z eff≤ 30 are available. The pulse height spectra of these detectors can be examined for the presence of photopeaks. If photopeaks are detectable, they can be used for energy calibration. If no photopeaks are detectable, the endpoint of the Compton energy spectrum represents a suitable alternative. The maximum energy observable in Compton scattering events EC,max , which is also called the Compton edge, results in E C , max = E γ 1 + m c 2 2 E γ from the radiation energy E γ of the respective radionuclide, with mc 2< = 511 keV, the rest energy of the electron. The position of the Compton edge can be determined experimentally by calculating the first derivative of the pulse height spectrum. The minimum point in the first derivative of the pulse height spectrum provides a good approximation to the Compton edge. The pulse height at the minimum point can then be calibrated as energy according to Equation (3).

[0055] Many common radiation detectors do not provide information about the interaction point of the radiation in the detector material. Therefore, it is necessary to make an assumption about the connecting line connecting the two interaction points in a detector pair. An obvious assumption is to choose the line connecting the two centers of the active detector media (the scintillator or the semiconductor). For each detector pair, there is then a connecting line passing through the two detector centers.

[0056] The connecting line of a detector pair is considered a directed quantity. A vector is assigned to the line. The vector originates at the center of detector 2 and points to the center of detector 1.

[0057] Now, the spatial directions of the detector pairs must be determined. This assumes parallel radiation incidence on the measuring device (far field). The radiation incidence is approximately parallel if the distances between all radiation sources are much greater than the dimensions of the measuring device.

[0058] First, a suitable coordinate system must be defined that applies to both the radiation sources and the radiation detectors. For 3-dimensional directional measurements, this is a spherical coordinate system, with the coordinate origin located at the spatial center of the detector array. In the 2-dimensional case, a polar coordinate system is used.

[0059] Imagine that the coordinate system can be shifted in space parallel to the Cartesian coordinate directions. For each detector pair, the coordinate system can thus be shifted to the center of detector 2.

[0060] First, the 3-dimensional case is considered: The spherical coordinates of detector 1 in this shifted coordinate system, whose origin is in detector 2, define a characteristic azimuth and elevation angle for this detector pair. Each detector pair i thus has such a characteristic azimuth angle ϕ i and elevation angle β i . Both angles associated with the detector pair ϕ i and β i are stored in the software. They are used in the evaluation procedures.

[0061] In the 2-dimensional case, it is a polar coordinate system that for each detector pair i parallel to the Cartesian coordinate directions to the center of detector 2. The azimuth angle ϕ i of detector 1 in this shifted coordinate system is calculated for each detector pair i stored in a software.

[0062] As a further component, the method according to the invention uses a system matrix that describes the response of the device to a point radiation source. Since the computational effort for data processing scales with the size of the system matrix, it is helpful to utilize the symmetry properties of the measuring device to keep the system matrix as small as possible. For this purpose, the detector pairs of a measuring device are divided into symmetry classes.

[0063] Those detector pairs that map onto other identical detector pairs upon rotation or translation belong to a symmetry class. They are grouped together and all treated identically. The system matrix is ​​defined as a set of lookup tables. Each lookup table describes, for each symmetry class, the response behavior of the detector pairs of this symmetry class to an infinitely distant, point-like radiation source. The lookup tables can be derived from a simulation calculation, a theoretical model, or from measurements with point-like reference sources.

[0064] There is a lookup table for each symmetry class and each radionuclide. The totality of all lookup tables forms the system matrix of the measuring device. The lookup tables represent the mathematical relationships between the radiation sources and the measured data.

[0065] The elements of a lookup table are the mapping functions LUT j [ϑ]. They represent the probabilities that a radiation source at an angle ϑ to the detector pair axis will produce coincidence events in bin j of the measured value distribution f(E1,E2). The index j indexes the j-th class of the measured value distribution f(E1,E2).

[0066] According to a further advantageous embodiment of the method according to the invention, in step a) the lookup tables LUT SK< generated by measurements with the detector system by performing the following sequence of steps: Create a rule for generating and validating the lookup tables LUT SK< from reference measurements, which selects the reference sources, takes into account the nuclide type, and defines the measurement conditions under which the measurements are to be carried out; Carry outof reference measurements for all steps previously defined in the regulation; Create and Validate the lookup tables LUT SK< according to the previously defined rule for evaluating the measurement data from the reference measurements; Capture the natural radiation background b SK ( i )< for all symmetry classes SK; Exclude the natural radiation background b SK< from the lookup tables LUT SK< ;

[0067] The following describes an embodiment of the invention that experimentally generates the lookup tables using point-like reference sources. All reference sources should be far enough away from the measuring device that the radiation falls on the device in a quasi-parallel manner (far field).

[0068] It is advantageous to perform measurements with several reference sources of different nuclides. The sources should be selected to represent the energy range in which the measuring device is used. The distance between the sources should be large enough to ensure parallel radiation incidence. Ideally, the distance should be as large as possible, given the strength of the source, in order to obtain measurement data in reasonable acquisition times.

[0069] For each source, a series of measurements is performed at different angular positions. Care must be taken to ensure that the source is always at the same radial distance from the device center in all measurements. All measurements are performed with the same measurement time t. Data acquisition should be carried out under the same conditions as those intended for operational use.

[0070] If the operating conditions of the measuring instrument require that not only relative intensities but also absolute radiation intensities are provided in units of measurement such as pGy sr -1< s -1< or nSv sr -1< h -1<, it must be ensured that only calibrated reference sources are used.

[0071] In all other cases, it is sufficient if the lookup table data are acquired at a constant relative irradiance; absolute calibration of the irradiance is then unnecessary. Sources of any intensity can be used, provided their activity is sufficient to produce good counter statistics.

[0072] After completing the measurements with the reference sources, the natural radiation background should be recorded in a further measurement without the source. The background measurement should be performed with the same measurement duration t as the previous measurements with the reference sources.

[0073] The procedure for obtaining the lookup tables from the measurement data is the same for all nuclides and is described here as an example for one nuclide.

[0074] It is assumed that all detector pairs are divided into symmetry classes. Each symmetry class has an identification number Ski), which for each detector pair i is to be defined.

[0075] For each symmetry class SK a separate lookup table LUT SK< which spans the angular range from 0° to 180° and is divided into equidistant angular steps.

[0076] The data recorded at a specific angular position must be evaluated for each detector pair. According to the current position of the reference source at an azimuth angle ω and an elevation angle h an angular distance ϑ i for each detector pair i calculated: ϑ i = arccos cosβ i cos h cos ω − φ i + sinβ i sin h

[0077] In the above expression, ϕ i the azimuth angle and β i the elevation angle of detector pair i In the 2-dimensional case, the angular distance ϑ i with ϑ i = arccos cos ω − φ i be calculated.

[0078] The process is repeated for all angular positions of the reference source.

[0079] The measurement data are then sorted according to the symmetry classes of the detector pairs to generate the lookup tables. For all pairs i a symmetry class Ski) the measured values ​​are at the angular distances ϑ i The goal is to find for each symmetry class SK the corresponding lookup table LUT SK< in the entire range of values ​​from ϑ = 0 ° to fill with data up to 180 °.

[0080] For each value ϑ in the range of ϑ = 0 ° up to 180 ° the data set is now ϑ i a symmetry class SKselected that is closest to the current value ϑ. This data set is copied to position ϑ in the lookup table. This process is repeated until all positions ϑ from 0° to 180° in a lookup table are filled with data. This process is repeated for all symmetry classes. The lookup tables of a measuring instrument are generated according to this procedure.

[0081] The measurement without a reference source is used to measure the natural radiation background. The natural radiation background can be assumed to be isotropic. All background data sets belonging to the same symmetry class can be averaged. The averaged data set is b j SK and applies to all detector pairs of a symmetry class SK.

[0082] The radiation background contained in the lookup tables can be removed by subtracting from each element of the lookup table LUT j SK ϑ the underground b j SK in the binj the measured value distribution f(E1,E2) is subtracted.

[0083] A central component of the method according to the invention are the recorded lookup tables. Lookup tables are energy-specific and apply to the characteristic radiation energy of the radiation source. Various options are possible for configuring a measuring device for specific operating conditions. If the intended use is linked to a manageable number of nuclides, it may be useful to create a set of lookup tables for each nuclide. If, however, the measuring device is to be used in a broad energy range with very different nuclides, it is advantageous to perform an additional data processing stage. The lookup tables can be fitted with functions. The fit parameters are recorded for different gamma energies and stored as energy-dependent functions. During a measurement, lookup tables for any gamma energies can be generated from the stored energy-dependent functions.The device can thus be dynamically adapted to specific situations. Radiation sources can be imaged across a wide energy range.

[0084] According to a further advantageous embodiment of the method according to the invention, in step e), a selection condition is applied with respect to the energy sum E1+E2 of the energies detected in both detectors of a pair. In embodiments where several radionuclides are present, the selection conditions are used to determine a separate frequency distribution for each radionuclide in step e). Y and / or in step f) a separate directional distribution for each radionuclide X to calculate.

[0085] When recording the 2-dimensional frequency distribution Y It is recommended to select the coincidence events with a selection condition for the energy sum E1+E2. Only those values ​​f(E1,E2) should be included in the distribution Y whose energy sum E1+E2 lies within a predefined range around the energy value corresponding to the characteristic nuclide energy whose radiation is detected. Using such a selection condition, it is possible to evaluate separate frequency tables for each nuclide.

[0086] According to a further advantageous embodiment of the method according to the invention, in step e) several selection conditions with respect to the energy sum E1+E2 are applied for radionuclides with several gamma energies and / or for such radionuclides with several gamma energies one or more frequency distributions Y created.

[0087] When measuring the directions of sources of a nuclide that emits radiation along multiple lines, it is up to the user to decide whether the coincidence events should be recorded in one or more frequency distributions. For example, Co-60 has two closely spaced lines at 1173 keV and 1332 keV. Coincidence events of both emissions can easily be classified into the same frequency distribution. Because the two emission lines are so close to each other, the 2-dimensional frequency distributions are also very similar and can be evaluated together without difficulty. However, if a nuclide has several widely separated gamma lines, it is recommended to record them separately. The frequency distributions belonging to a nuclide are later combined in the evaluation process.

[0088] According to a further advantageous embodiment of the method according to the invention, the statistical image reconstruction method in step f) is part of or entirely the Maximum Likelihood Expectation Maximization (MLEM) method, the Ordered Subset Expectation Maximization (OSEM) method, the List Mode - Maximum Likelihood Expectation Maximization (LM-MLEM) method and / or the List Mode - Ordered Subset Expectation Maximization (LM-OSEM) method.

[0089] In the following, two embodiments of the invention are described, such as a spatial directional distribution X ( ω, h ) or a flat directional distribution X ( ω ) can be calculated using the Maximum Likelihood Expectation Maximization (MLEM) algorithm. The statements apply to devices in which the sum ∑ i,j Y ij all entries in the frequency distribution Y has no directional dependence. In the presence of many detector pairs, the directional dependence of individual detector pairs averages out; the total count rate ∑ i,j Y ij is then approximately direction independent.

[0090] According to a further advantageous embodiment of the method according to the invention, in step f) a 3-dimensional directional distribution is determined for each detected radionuclide X small = X ( ω k , hl ) according to X kl n + 1 = X kl n KL ∑ i , j Y ij ∑ i , j LU T j SK i ϑ ω k h l φ i β i Y i , j ∑ k ′ = 1 K ∑ l ′ = 1 L LU T j SK i ϑ ω k ′ h l ′ φ i β i X k ′ , l ′ n from an initial distribution ∀ k = 1 , … , K , l = 1 , … , L X kl 0 = 1 using lookup tables LUT SK< and an angular distance function according to ϑ ω k h l φ i β i = arccos cosβ i cos h l cos ω k − φ i + sinβ i sin h l where K and L are user-defined numbers of pixels for the azimuth angle ω k and the elevation angle h l are.

[0091] The following describes an algorithm for reconstructing a 3-dimensional directional distribution. The surface of the celestial sphere is first divided into latitude and longitude. Each of the equal-sized surface elements on the celestial sphere corresponds to an image pixel. (k,l) for which the irradiance X small = X ( ω k , hl ) at the azimuth angle ω k and elevation angle h l should be calculated. Make sure that the classes for the elevation angle h l be chosen so that the sine of the elevation angle is uniformly distributed in the interval (-1,1). There are a total of K·L pixels X small for K classes ω k and L classes h l .

[0092] Maximum Likelihood Expectation Maximization (MLEM) is an iterative procedure that determines initial values ​​for the directional distribution X needed: ∀ k = 1 , … , K , l = 1 , … , L X kl 0 = 1

[0093] The MLEM iteration rule for calculating a new representation X kl n + 1 from the previous values X kl n is: X kl n + 1 = X kl n KL ∑ i , j Y ij ∑ i , j LU T j SK i ϑ ω k h l φ i β i Y ij ∑ k ′ = 1 K ∑ l ′ = 1 L LU T j SK i ϑ ω k ′ h l ′ φ i β i X k ′ l ′ n

[0094] Herein refers Y ij in the frequency table Y registered counters (the measured values) for a detector pair i in the bin j the measured value distribution f(E1,E2). When calling the lookup tables Lur SK ( i )< is the angular distance ϑ between a pixel (k,l) on the celestial sphere and a pair of detectors i In the 3-dimensional case, the angular distance ϑ can be calculated according to ϑ ω k h l φ i β i = arccos cosβ i cos h l cos ω k − φ i + sinβ i sin h l be calculated, where ϕ i the azimuth angle and β i the elevation angle of detector pair i are.

[0095] The MLEM algorithm is a method that uses a frequency distribution, namely the measured frequency distribution Y , which indicates how often coincidences have occurred in certain detector pairs for certain measured values, and from this a new frequency distribution is calculated, namely the desired directional distribution X , which indicates the frequency with which radiation from certain angles of incidence ω k and h l arrives at the detection location.

[0096] Equation (7) is formulated in such a way that the iteration rule can be executed in parallel with the data acquisition, while still continuously calculating coincidence counters Y ij in the frequency distribution Y be accumulated.

[0097] According to a further advantageous embodiment of the method according to the invention, in step f) a 2-dimensional directional distribution is determined for each detected radionuclide X k = X ( ω k ) according to X k n + 1 = X k n K ∑ i , j Y ij ∑ i , j LU T j SK i ϑ ω k φ i Y ij ∑ k ′ = 1 K LU T j SK i ϑ ω k ′ φ i X k ′ n from an initial distribution ∀ k = 1 , … , K X k 0 = 1 using lookup tables LUT SK<and an angular distance function according to ϑ ω k φ i = arccos cos ω k − φ i where K is a user-defined number of pixels for the azimuth angle ω k is.

[0098] In a further embodiment of the invention, a flat directional distribution X k = X ( ω k ) are calculated using the MLEM algorithm. An initial distribution is initialized ∀ k = 1 , … , K X k 0 = 1 and an iteration loop with the iteration rule: X k n + 1 = X k n K ∑ i , j Y ij ∑ i , j LU T j SK i ϑ ω k φ i Y ij LU T j SK i ϑ ω k ′ φ i X k ′ n The angular distance ϑ in the 2-dimensional case is ϑ ω k φ i = arccos cos ω k − φ i calculated.

[0099] The directional distribution X of the radiation field is initially unitless. The values X small represent the relative radiation intensities in the different directions of incidence ω k and h l However, it is also possible to specify the radiation intensity using the unit of dose rate per solid angle. This requires first setting up the detector system so that nuclide-specific dose rates can be measured.

[0100] To measure nuclide-specific dose rates, count rates of interaction events are used, whose total energy corresponds to the respective nuclide energy. A variety of options are available. On the one hand, the nuclide-specific count rates of all medium- to high-atomic number detectors included in the device can be used. On the other hand, count rates of coincidence events whose total energy E1+E2 lies within a specific interval around the nuclide energy can also be used. These count rates can be calibrated in units of the equivalent dose rate or in units of the absorbed dose rate (in air). The procedure is briefly outlined here.

[0101] The calibration is performed with a reference source of calibrated activity Λ. The dose rate Ḋ or the equivalent dose rate Ḣ *(10) at the center of the device can be calculated from the activity Λ of the source, its distance r and the nuclide-specific dose rate constants Γ and Γ H The (physical) dose rate Ḋ results from: D ˙ = dD dt = Λ Γ r 2 and the equivalent dose rate Ḣ *(10) from: H ˙ * 10 = dH * 10 dt = Λ Γ H r 2

[0102] The values ​​of the dose rate constant Γ and the equivalent dose rate constant Γ H for the various nuclides can be found in tables in the relevant specialist literature.

[0103] From the available nuclide-specific count rates of the device, a count rate or a combination of count rates must now be selected that is suitable for representing the dose rate while avoiding systematic errors. One possible cause of systematic errors in dose rate measurements is an uneven response of the device with respect to the direction of the incident radiation. The count rates can be affected, for example, by shadowing effects. Coincidence count rates, in particular, are characterized by their strong directional dependence. The directional sensitivity of coincidence events, which is advantageous for directional measurement, can be a significant source of error in dose rate measurements.

[0104] In the following, two embodiments of the invention are presented, showing how a measurement of nuclide-specific dose rates with small systematic errors is possible.

[0105] For devices where shadowing effects are negligible, it is useful to use the detectors m = 1, ..., M high atomic number detected nuclide-specific counter N m in the photo peak. The count rate Ż , defined as: Z ˙ = d Max N 1 , … , N M dt can be a good measure of the dose rate if there is at least one high atomic number detector in each direction that is unobstructedly exposed to the incident radiation.

[0106] An alternative approach is available for devices consisting of many detectors. Here, the sum of all coincidence counters in the frequency distribution Y serve as a measure of the absorbed dose, provided that the total count rate Z ˙ = d ∑ i , j Y ij dt is independent of direction.

[0107] The suitability of a particular variable ŻThe calibration of the dose rate can be checked by moving a reference source around the device at a constant distance r. If the count rate remains Ż constant while the source is moved around the device, the condition is met. The variable Ż can then be used to calibrate the dose rate.

[0108] For the calibration of the dose rate with the count rate Ż A background measurement is recorded without a source. The count rate thus determined is called Ż b Device-specific calibration constants can now be C and CH be defined with which the dose rate Ḋ D ˙ = C Z ˙ − Z ˙ b and the equivalent dose rate Ḣ *(10) H ˙ * 10 = C H Z ˙ − Z ˙ b from the background-corrected count rate Ż - Ż b can be calculated.

[0109] Using the device-specific calibration constants C and CHDirection-dependent dose rates per solid angle can now be calculated. For this purpose, the intensity values X small multiplied by a scaling factor. In the following it is assumed that the count rate Ż according to Eq. (15) and the calibration constants C and CH according to Eq. (16) or (17).

[0110] The dose rate Ḋ per solid angle d Ω is dD ω k h l dt d Ω = C Z ˙ − Z ˙ b 4 π X kl the equivalent dose rate Ḣ *(10) per solid angle d Ω with: dH * 10 ω k h l dt d Ω = C H Z ˙ − Z ˙ b 4 π X kl calculated.

[0111] In the 2-dimensional case: dD ω k dt d ω = C Z ˙ − Z ˙ b 2 π X k and: dH * 10 ω k dt d ω = C H Z ˙ − Z ˙ b 2 π X k

[0112] For the dose rate range up to 1 µSv / h, the (physical) dose rate per solid angle can be specified, for example, using the unit pGy sr -1< s -1<. The equivalent dose rate per solid angle can be measured, for example, in units of nSv sr -1< h -1<.

[0113] If the dose rate per solid angle is averaged over all directions of incidence, the dose rate is Ḋ , which is received from all directions, divided by 4π. The same applies to the equivalent dose rate: The equivalent dose rate per solid angle averaged over all directions of incidence is equal to the equivalent dose rate Ḣ *(10), divided by 4π.

[0114] The invention further relates to a device for carrying out the method according to the invention. The device according to the invention comprises a group of several synchronized detectors for detecting radiation, wherein at least one detector material has an atomic number of Z eff > 30 and all detectors measure the energies E that occur in interactions of the radiation with the detector materials.

[0115] Furthermore, the device according to the invention has system electronics which register coincidence events when in each two detectors from a list of defined detector pairs i simultaneous interactions take place, where the list of defined detector pairs includes all pairs that can be formed combinatorially from the set of all detectors, and the defined detector pairs contain at least one detector made of a material with an atomic number of Z eff > 30 included.

[0116] Furthermore, the device according to the invention comprises a data acquisition system which establishes a ranking for both detectors involved in a coincidence event, which defines a first and a second detector, and which orders the energies (E1, E2) measured in coincidence events according to their designation 1, 2 and stores them in a chronological list with the attributes { i, E1, E2} and the detection time tstores; where in each detector pair i The detector with the lower atomic number is assigned the number 1 and the one with the higher atomic number is assigned the number 2; if both detectors in a pair have the same atomic number, the designation 1 or 2 is chosen arbitrarily.

[0117] In addition, the device according to the invention has an analysis unit which calculates a function value f(E1,E2) = (E2-E1) / (E1+E2) for each coincidence event from the energies (E1,E2) stored by the data acquisition system and classifies the coincidence events according to their detector pair number i and their function values ​​f(E1,E2) in one or more frequency distributions Y recorded and with a statistical image reconstruction method of emission tomography using lookup tables LUT SK< from each frequency distribution Y one directional distribution each X of the radiation field, whereby for each radionuclide a frequency distribution Y and one directional distribution each X present.

[0118] According to an advantageous embodiment of the device according to the invention, it is designed partially or completely as a Compton camera, a Compton telescope, a single plane Compton camera, a neutron camera and / or a dual gamma / neutron camera.

[0119] For the purposes of the invention, a device is also understood to mean a measuring device, a device for short, or a detector system.

[0120] According to a further advantageous embodiment of the device according to the invention, the entirety of all detectors of the device has an annular shape, in which the detectors are arranged in a ring and / or in which none, one, or more central detectors are present inside the ring. Furthermore, the ring preferably contains four or five, and particularly preferably six, plastic scintillation detectors. Furthermore, one or two scintillation detectors made of NaI, CSI, CeBr3, and / or LaBr3 are preferably located inside the ring.

[0121] According to a further advantageous embodiment of the device according to the invention, it is preferably constructed from scintillation detectors of the size 1"x1", 1.5"x1.5", 2"x2" and / or 3"x3".

[0122] According to a further advantageous embodiment of the device according to the invention, a scintillation detector is used as the detector and / or the scintillator is designed as a monolithic block or as a pixelated scintillator module and / or which is formed from pure or doped materials from the group of PVT, anthracene, stilbene, p-terphenyl, CaF 2 , BaF 2 , Nal, CeBr 3 , LaBr 3 , LaCl 3 , La(Br x Cl 1-x ) 3 , Csl, SrI 2 , CLYC, CLBC, CLCB, CLLB, BGO, LSO, LYSO, GAGG, YAP and / or YAG.

[0123] According to a further advantageous embodiment of the device according to the invention, a semiconductor detector is used as the detector and / or the semiconductor is designed as a segmented or unsegmented semiconductor and / or which has a planar or coaxial geometry and / or which is formed from materials from the group of Ge, GaAs, CdTe and / or CdZnTe.

[0124] According to a further advantageous embodiment of the device according to the invention, at least two detectors are used.

[0125] According to a further advantageous embodiment of the device according to the invention, all detectors used are essentially identical in construction or at least two of the detectors used are different from one another.

[0126] According to a further advantageous embodiment of the device according to the invention, detector and system electronics are used that utilize analog and / or digital electronic components. The analog electronic components can comprise a combination of various modules, including a high-voltage power supply, a preamplifier, an amplifier, a pulse shaper, a charge integrator, a pulse height analyzer, a multi-channel analyzer (MCA), and / or a coincidence circuit.

[0127] According to a further advantageous embodiment of the device according to the invention, the digital electronic components can comprise a combination of various hardware and software components, including a high-voltage supply, an A / D converter per detector, a field programmable gate array (FPGA), a storage medium, a digital signal processor and / or evaluation software.

[0128] They show: Fig. 1 a non-scale, schematic representation of a unidirectional detector pair in a two-dimensional measurement situation, according to an embodiment of the invention; Fig. 2 a non-scale, schematic representation of a bidirectional detector pair in a two-dimensional measurement situation according to a further embodiment of the invention; Fig. 3a non-scale, schematic representation of the radiation incidence from a source onto a detector pair 1, 2 in a two-dimensional measurement situation, according to a further embodiment of the invention; Fig. 4a one Not to scale, schematic representation of a device model with a central cerium bromide detector surrounded by six annularly arranged plastic detectors; according to another embodiment of the invention; Fig. 4b a non-scale, schematic representation of a device model according to Fig. 4a in which a further cerium bromide detector is introduced, according to a further embodiment of the invention; Fig. 5 a schematic representation of a frequency distribution Y of the device model Fig. 4a in a 2-dimensional direction measurement, according to another embodiment of the invention; Fig. 6aa non-scale, schematic representation of a first symmetry class of a total of four cube-shaped detectors arranged in a square, with two horizontal and two vertical detector pairs, according to a further embodiment of the invention; Fig. 6b a non-scale, schematic representation of another symmetry class of the four cube-shaped detectors arranged in a square from Fig. 6a , according to another embodiment of the invention; Fig. 7 a schematic representation of a frequency distribution Y of the device model Fig. 4b in a 3-dimensional direction measurement, according to another embodiment of the invention; Fig. 8a A not-to-scale, schematic representation of a directional distribution for a radiation far-field whose emission points all lie in one plane. The directional distribution is defined as the azimuth angle ω dependent functionX ( ω ) modeled along the horizon circle, according to another embodiment of the invention; Fig. 8b A non-scale, schematic representation of a directional distribution for a radiation far-field whose emission points are randomly distributed in space. The directional distribution is defined as the azimuth angle ω and the elevation angle h dependent function X ( ω, h ) modeled on the celestial sphere, according to another embodiment of the invention; Fig. 9 a schematic representation of a lookup table according to another embodiment of the invention; Fig. 10 a Diagram with measured mapping functions from a lookup table of the device model from Fig. 4a for CeBr 3 - plastic detector pairs at selected asymmetry values ​​A; according to another embodiment of the invention; Fig. 11four diagrams of calculated asymmetry curves in the lookup tables of unidirectional and bidirectional detector pairs for the radionuclides Co-60 and Cs-137, according to another embodiment of the invention; Fig. 12 a non-scale, schematic representation of a measurement plan with angular positions of a reference source for creating a lookup table for the fixture model from Fig. 4a , according to another embodiment of the invention; Fig. 13 a schematic representation of a calculation scheme for an MLEM method for reconstructing the directional distribution of a radiation field, according to a further embodiment of the invention; Fig. 14 Nine diagrams of measured directional distributions for radiation fields with one, two, and three point radiation sources. The illustrations are examples of 2-dimensional directional measurements with the device model from Fig. 4a , according to another embodiment of the invention;

[0129] In the Fig. 1 is a non-scale, schematic representation of a unidirectional detector pair in a two-dimensional measurement situation. The detector with atomic number Z eff ≤ 30 is marked with the number 1, the detector with higher atomic number Z eff > 30 with the number 2. The reconstruction of the direction of incidence remains ambiguous. While the angle of the incident radiation to the line connecting the two detectors can be measured, the side on which the source lies cannot.

[0130] Fig. 2 shows a non-scale, schematic representation of a bidirectional detector pair in a two-dimensional measurement situation. Both detectors are made of materials with the same or similar atomic number. Z eff> 30. The labeling of the two detectors of a bidirectional pair in Detector 1 and Detector 2 is defined once and must then be applied consistently in all calculations. As with the unidirectional detector pair in Fig. 1 The reconstruction of the direction of incidence remains ambiguous. The angle of the incident radiation to the connecting line between the two detectors can be measured, but not the side on which the source lies.

[0131] Fig. 3 shows the radiation incidence of a source on a detector pair 1,2 in a two-dimensional measurement situation.

[0132] Fig. 4 shows two device models according to the invention. In the model of Abb. 4a A central cerium bromide detector is surrounded by six plastic detectors arranged in a ring. This model is suitable for 2-dimensional directional measurements. If another cerium bromide detector is placed under the central cerium bromide detector, the model of Abb. 4b , which allows 3-dimensional directional measurements.

[0133] Abb. 5 shows an example of a frequency distribution Y (the measurement data) for the device model Abb. 4a The system electronics determines which detectors were coincident and calculates the asymmetry value A = ( E 2 - E 1) / ( E 1 + E 2) from the coincident energy inputs E1 and E2. The coincidence events are classified according to their detector pair identification number and their asymmetry value A in the frequency distribution Y registered.

[0134] Abb. 6 illustrates how detector pairs are divided into symmetry classes. Four cube-shaped detectors are arranged in a square. This detector arrangement has two symmetry classes. The first symmetry class, Abb. 6a shown, comprises two horizontal and two vertical detector pairs. The two diagonal detector pairs, in Abb. 6b to see, form another symmetry class.

[0135] Abb. 7 shows an example of a frequency distribution Y (the measurement data) for the device model Abb. 4b . The measurement data Y were selected according to the three symmetry classes of the device model Abb. 4b grouped into three areas. Detector pairs I to VI belong to symmetry class 1, detector pairs VII to XII to symmetry class 2. The thirteenth detector pair forms its own symmetry class 3.

[0136] Abb. 8 shows two schematic representations of directional distributions for radiation far fields. A radiation far field is completely characterized by its directional dependence. Depending on the measurement situation, there are two types of distributions. If all radiation sources and the measuring device are in one plane, the azimuth angle ω dependent distribution X ( ω ) along the horizon circle, in Abb. 8a The general directional distribution is shown in Abb. 8b represented by the azimuth angle ω and the elevation angle h dependent distribution X ( ω, h ) on the celestial sphere. The size X represents the radiation intensity. It can either be normalized to the highest occurring value or specified with a unit of measurement.

[0137] Abb. 9 shows a schematic of a lookup table. A lookup table is a 2-dimensional frequency distribution that records coincidence counts as a function of the angular distance ϑ of a radiation source from the detector pair axis and a measured variable. The measured variable here is the energy asymmetry. A = ( E 2 - E 1) / ( E 1 + E 2) of the two energy inputs in a coincidence event.

[0138] Abb. 10 shows various measured mapping functions from a lookup table for CeBr 3 -Plastic detector pairs at selected asymmetry values AThe mapping functions represent the frequencies of radiation incident at an angle ϑ being recorded at specific measurement intervals of A. A lookup table can be thought of as a group of mapping functions. The example shows three of the functions recorded with a 10 µCi Co-60 source. Coincidence events of both Co-60 emission lines at 1173 keV and 1332 keV were stored together.

[0139] Abb. 11 shows asymmetry curves for lookup tables calculated for the radionuclides Co-60 and Cs-137, according to one embodiment of the invention. The asymmetry curves show the regions in a lookup table with the highest abundances. If there were no measurement errors, all non-zero abundances would lie on the curves shown. A distinction is made here between lookup tables of unidirectional and bidirectional detector pairs. Lookup tables of unidirectional detector pairs ( Abb. 11 a und c ) are characterized by the presence of only one prominent asymmetry curve. In contrast, lookup tables of bidirectional detector pairs ( Abb. 11 b und d ) two striking asymmetry curves. For Co-60, the asymmetry curves are shown for both emission lines at 1173 keV and 1332 keV; Cs-137 has only one emission line at 662 keV.

[0140] Abb. 12 shows the angular positions as an example ω a measurement plan to create the lookup table for the Abb. 4a The device model shown in accordance with an embodiment of the invention is shown. According to this plan, measurements are performed in seven angular positions. The source is always at the same distance from the device center. A lookup table for the device model can be generated from the seven measurements. In this example, the lookup table has a step size of 5°. For each angle ϑ in the value range from 0° to 180°, there is at least one detector pair that has the correct angular distance from the radiation source and whose data set can be copied to the respective position in the lookup table. The measurement plan from Abb. 12 defines the assignments between the angular positions ω the reference source, the detector pairs and the angle values ​​ϑ of the lookup table.

[0141] Abb. 13 shows a calculation scheme for an MLEM method for reconstructing the directional distribution of a radiation field according to an embodiment of the invention. With each iteration step n, all Abb. 13 With increasing number n of iterations, the MLEM method leads to the directional distribution X , which has the highest probability of the observed frequency distribution Y Equations (13) and (16) include all the Abb. 13 The calculation steps shown in a compact form, including forward and back projection.

[0142] Abb. 14 shows various examples of 2-dimensional direction measurements with the device model of Abb. 4a All measurements were performed using 10 µCi Co-60 sources at a distance of 105 cm. The radial measurement scale is the same in all representations and ranges from 0 to 300 nSv rad -1< h -1< . Abb. 14a, 14b und 14c show measurement results for a Co-60 source at an angle of incidence of 125°. Subsequently, a second Co-60 source was added at 50° and finally a third Co-60 source at 90°. Abb. 14d, 14e und 14f show the measured directional distributions for the 2-source radiation field, Abb. 14g, 14h und 14i for the 3-source radiation field.

[0143] The method and device according to the invention use unidirectional and bidirectional detection processes for multidimensional gamma radiation direction measurement. The direction measurement, which partially or completely uses bidirectional detection processes, has been described in detail in the previously mentioned embodiments. The bidirectional direction measurement is based on the Abb. 11 The functional principle shown is the basis for this, which will be briefly explained here.

[0144] In contrast to unidirectional detection processes, which use a functional relationship between a measured value (here the asymmetry A) and the scattering angle ϑ for the direction measurement, bidirectional detection processes do not have such a calibration curve A (ϑ). The lookup tables of bidirectional detector pairs show two prominent structures along the Abb. 11b und 11d For a given scattering angle ϑ there are two corresponding measured values A1 and A2 , which are observed with certain frequencies. Nevertheless, the pair of values ​​( A1,A2 ) can be uniquely assigned to a specific scattering angle ϑ. Over the entire angular range from 0° to 180°, each pair of values ​​( A1,A2 ) exactly once. It cannot be used with any other pair of values ​​( A1 , A2) at a different angle ϑ. Therefore, it is also possible to measure the scattering angle ϑ with a bidirectional detector pair. Statistical reconstruction methods already possess the functionality required for directional measurement with bidirectional detector pairs. Methods such as MLEM are therefore suitable for processing the measurement data from the device models according to the invention.

[0145] Measurements are to be carried out with the device model from Abb. 4a This system consists of six 3"×3" plastic scintillation detectors and one 3"×3" cerium bromide scintillation detector. The six plastic detectors form a ring around the central cerium bromide detector. The diameter of the detector ring is 25 cm.

[0146] Each of the six plastic detectors can be combined with the central cerium bromide detector to form a unidirectional detector pair. In such a detector pair, the radiation is primarily scattered in the plastic detector and absorbed in the cerium bromide detector.

[0147] Abb. 14 shows measured directional distributions for various arrangements of multiple point radiation sources. All measurements were performed with 10 µCi Co-60 sources, each spaced 105 cm apart. A first Co-60 source was placed at an angle of 125°. The dose rate generated by a 10 µCi Co-60 source at a distance of 105 cm is approximately 100 nSv h -1 . Abb. 14 a), b) und c) show the results of the directional measurement after 10 s, 20 s, and 60 s of measurement time. The number of MLEM iterations was dynamically coupled to the measurement time. With increasing measurement time, the directional distribution of the point source becomes steeper and higher. The intensity averaged over all directions is constant at 16 nSv rad -1< h -1< . Integrated over the full angle of 2π rad, this results in a dose rate of 2π rad · 16 nSv rad -1< h -1< = 100 nSv h -1< . Subsequently, a second Co-60 source was added at 50°. The results are shown in Abb. 14 d), e) und f) The presence of two sources is clearly visible after just 10 s of measurement time. The intensity averaged over all directions is now 32 nSv rad -1< h -1< . Finally, a third Co-60 source was added to the arrangement at 90°. The results in Abb. 14 g), h) und i) show that from a measurement time of 20 s, three point sources can also be resolved separately.

[0148] A methodology was developed that allows the quality of the directional measurement to be expressed as a time reference from which a reliable directional measurement is available. The goal of the procedure was to quantitatively evaluate the quality of the directional measurement of a point source as a function of the nuclide type, dose rate, and angular position. The sources used were one Co-60, one Cs-137, and one Co-57 source, each with an activity of 10 µCi.

[0149] Two quality classes were defined: ±5° and ±10° angular accuracy. At any time t During a measurement, the MLEM method provides an estimate of the direction of the radiation source. The measurements were each run over a specific measurement period and repeated 20 times. At each time point t, it is possible to determine how many of the measurements fall within the ±5° or ±10° accuracy class. A confidence level of 90% was chosen. t90 is defined as the time after which at least 18 of the total 20 measurements are accurate, i.e., within a ±5° or ±10° interval of the true value. In 90% of all measurements, the specified accuracy is met or exceeded after the minimum measurement time.

[0150] The device model of Abb. 4a has a characteristic symmetry. First, the influence of the device-specific detector arrangement on the minimum measurement times measured at different angles of incidence was investigated. The sensitivity of the directional measurement depends on the position of the source relative to the device. If the source is located in the direction of one of the six ring detectors, longer measurement times are required than if the source is located midway between two ring detectors. In this study, all sources were at a constant distance of 1 m from the device center. The minimum measurement time t90 for ±10° accuracy of the 10 µCi Co-60 source was 2.7 s in the direction of the ring detectors and shortened to 0.5 s in the middle between two ring detectors. A similar behavior was observed for Cs-137. Here, the minimum measurement time was t 90 for ±10° accuracy 5.8 s in the direction of the ring detectors and shortened to 0.7 s in the middle between two ring detectors.

[0151] Next, we investigated how the t 90 Minimum measurement times depend on the dose rate. These measurements were performed in the direction of the ring detectors. They represent the lower limit of the device's performance. The distance between the sources was varied from 1 to 5 meters. The dose rate was defined as the local dose rate. Ḣ *(10) of the source at the measuring point. It was found that above 0.05 µSv / h the product of the local dose rate Ḣ *(10) and t 90 minimum measuring time is constant.

[0152] Under this condition, the t 90 minimum measurement times with the simple analytical expression t 90 = T Nuklid σ ⋅ H ˙ * 10 μ Sv / h − 1 be parameterized. The parameters T (nuclide, σ) are device-specific parameters that are defined depending on the nuclide type and the quality class σ. For the device model with six 3"x3" plastic detectors and one 3"x3" cerium bromide detector, the parameters were determined as follows: T ( 60< Co, ±10°) = 0.32 s, T ( 60< Co, ±5°) = 0.48 s, T ( 137< Cs, ±10°) = 0.16 s and T ( 137< Cs, ±5°) = 0.24 s.

[0153] The performance of directional measurements at very low local dose rates was also investigated. High-quality directional measurements are possible even below 0.05 µSv / h, but the required measurement times are then longer and may exceed those achieved with t 90 = T Nuklid σ ⋅ H ˙ * 10 μ Sv / h − 1 calculated values. With the available device model, it was possible to reliably measure the direction of a source with an accuracy of ±5°, even at dose rates of just a few nSv / h. The measurement times were then in the range of several minutes. For example, the direction measurement for a Co-60 source with a dose rate of 3 nSv / h achieved an accuracy of ±10° after 140 s and an accuracy of ±5° after 220 s. A Cs-137 source with a dose rate of 4 nSv / h at the measurement location could be located with an accuracy of ±10° after 55 s and with ±5° after 90 s. The direction measurement of a Co-57 source with a dose rate of 5 nSv / h had an accuracy of ±10° after 22 s and an accuracy of ±5° after 35 s.

[0154] Under quasi-stationary measurement conditions, the directional measurement can be performed in a timed fashion. This makes it possible, for example, to track the movement of radiation sources. The dose rate Ḣ *(10) is again a suitable value for controlling the clock frequency. For example, if the movement of a point source is to be tracked, the clock frequency can be set to the t 90 time can be set.

[0155] For an experiment under quasi-stationary measurement conditions, the device model of Abb. 4a mounted on a turntable. The clock frequency of the measuring program was set to t 90 time. At a rotation speed of 0.2 revolutions per minute, the current position of a 10 µCi Co-60 source at a distance of 1 meter could be easily tracked.

Claims

1. A method for the multidimensional directional measurement of gamma radiation in the far field by means of a group of several energy-discriminating, synchronized detectors (1, 2) for detecting radiation, wherein the method uses uni-directional and bi-directional Compton scattering processes and lookup tables LUT SK, a defined functional value f(E1, E2), a list of defined detector pairs with an identification number i for defined detector pairs and one or more frequency distributions Y for detecting the measurement values, and comprising the following steps: a) setting up the detector system for a measurement, wherein the following sequence of steps is performed: - creating a list of defined detector pairs, wherein the defined detector pairs comprise all pairs formed combinatorially from the set of detectors, each pair contains at least one detector made from a material having an atomic number of Zeff > 30 and identifying it with an ID number i; - interconnecting all detectors in a coincidence circuit such that coincidence events are detected in all defined detector pairs i=1,..., I; - identifying both detectors of each defined detector pair i with the numbers 1 and 2, respectively, wherein the detector with the lower atomic number receives the number 1 and the detector with the higher atomic number receives the number 2, both detectors are made of the same material, the identification as 1 and 2 is made arbitrarily; - defining a function f(E1, E2) = (E2-E1) / (E1+E2) which is calculated from two energy values E1 and E2; b) acquiring measured values of coincidence events if interactions take place simultaneously in two respective detectors of all defined detector pairs i, wherein the measured values originate from a radiation distribution in the far field and the measured values are the interaction energies E1 and E2 of the radiation measured in the detectors; c) associating coincidence events with an identification number i; d) calculating the functional value f(E1, E2) from two energy values E1, E2 per coincidence event by means of the function f(E1, E2) defined in step a); e) detecting the coincidence events according to their identification number i and their functional values f(E1, E2) in one or more frequency distributions Y, wherein a separate frequency distribution Y is provided for each radionuclide, f) calculating one or more directional distributions X from the frequency distributions Y by means of a statistical image reconstruction method of emission tomography using lookup tables LUTSK, wherein a separate directional distribution X is provided for each radionuclide.

2. The method according to claim 1, characterized in that the radiation sources emit a discrete and / or continuous distribution of radiation; and / or in that the radiation sources emit, in addition to gamma radiation, particle radiation from a particle group of electrons, positrons, protons, ions and / or neutrons; and / or in that the radiation originates from the radioactive decay of one or more radionuclides; and / or in that the radiation is the prompt gamma radiation produced during the absorption of proton or ion radiation in target materials; and / or in that the radiation is of low intensity.

3. The method according to claim 1 or 2, wherein the method in step a) additionally comprises, in part or in whole, the following sequence of steps: - calibrating the signals from all detectors as absorbed radiation energy E; - determining a suitable coordinate system; - detecting the directions of the defined detector pairs, wherein in the 2-dimensional directional measurement the direction of detector pair i is detected with the azimuth angle ϕi and wherein in the 3-dimensional directional measurement the direction of detector pair i is detected with the azimuth angle ϕi and the height angle βi; - dividing the measurement range for the functional value f(E1, E2) into a number J of equidistant measured value channels j = 1,..., J; - creating one or more 2-dimensional arrays with I·J fields as data structures for storing the frequency distributions Y, in which coincidence events are recorded according to their ID number i and their measured value channel j; for each radionuclide, a separate 2-dimensional array Y, is created; - dividing the detector pairs i into symmetry classes SK, wherein the detector pairs i which are imaged on other identical detector pairs in case of rotation or displacement are combined to form a respective symmetry class SK(i); - creating one or more lookup tables LUTSK for each symmetry class SK and each radionuclide, wherein they cover an angular range ϑ of 0° to 180° and are divided into equidistant angular steps; - creating the lookup tables LUTSK by measurements with the detector system or by Monte Carlo simulations or by means of a theoretical model; - transferring all lookup tables LUTSK for all symmetry classes and all radionuclides to an image reconstruction algorithm that processes the measurement data and calculates the directional distributions X.

4. The method according to claim 3, characterized in that the lookup tables LUTSK are generated by measurements with the detector system by performing the following sequence of steps: - creating a specification for the generation and validation of the lookup tables LUTSK from reference measurements, which selects the reference sources, takes into account the nuclide type, and defines the measurement conditions under which the measurements are to be carried out; - performing reference measurements for all steps previously defined in the specification; - creating and validating the lookup tables LUTSK according to the previously defined specification for evaluating the measurement data from the reference measurements; - detecting the natural radiation background bSK(i) for all symmetry classes SK; - excluding the natural radiation background bSK from the lookup tables LUTSK.

5. The method according to any one of claims 1-4, characterized in that in step e) a selection condition is applied for the energy sum E1+E2 of the energies detected in both detectors of a pair; and / or in that for each radionuclide a separate frequency distribution Y is created; and / or in that for each radionuclide a separate directional distribution X is calculated.

6. The method according to any one of claims 1-5, characterized in that, in step e), in the case of radionuclides having multiple gamma energies, multiple selection conditions are applied with respect to energy sum E1+E2; and / or in that for such radionuclides with multiple gamma energies one or more frequency distributions Y are created.

7. The method according to any one of claims 1-6, characterized in that in step f) the statistical image reconstruction method is part or all of the Maximum Likelihood Expectation Maximization (MLEM) method, the Ordered Subset Expectation Maximization (OSEM) method, the List Mode - Maximum Likelihood Expectation Maximization (LM-MLEM) method and / or the List Mode - Ordered Subset Expectation Maximization (LM-OSEM) method.

8. The method according to any one of claims 1-7, characterized in that in step f) for each detected radionuclide a 2-dimensional directional distribution Xk = X(ωk) is calculated according to X k n + 1 = X k n K ∑ i , j Y ij ∑ i , j LU T j SK i ϑ ω k φ i Y ij ∑ k ′ = 1 K LU T j SK i ϑ ω k ′ φ i X k ′ n from an initial distribution ∀ k = 1 , … , K X k 0 = 1 using lookup tables LUTSK and an angular distance function in accordance with ϑ ω k φ i = arccos cos ω k − φ i wherein K is a user-defined number of pixels for the azimuth angle ωk.

9. The method according to any one of claims 1-8, characterized in that in step f) for each detected radionuclide a 3-dimensional directional distribution Xkl = X(ωk, hl) is calculated in accordance with X kl n + 1 = X kl n KL ∑ i , j Y ij ∑ i , j LU T j SK i ϑ ω k h l φ i β i Y ij ∑ k ′ = 1 K ∑ l ′ = 1 L LU T j SK i ϑ ω k ′ h l ′ φ i β i X k ′ l ′ n from an initial distribution ∀ k = 1 , … , K , l = 1 , … , L X kl 0 = 1 using lookup tables LUTSK and an angular distance function in accordance with wherein K and L are user-defined numbers of pixels for the azimuth angle ωk and the height angle hl.

10. A device for carrying out the method of multidimensional directional measurement of gamma radiation in the far field according to any one of claims 1-9 comprising: - a group of a plurality of synchronized detectors (1, 2) for detecting radiation, wherein at least one detector material has an atomic number of Zeff > 30 and all detectors measure the energies E which occur in interactions of the radiation with the detector materials; - a system electronics which records coincidence events when interactions take place simultaneously in two detectors each from a list of defined detector pairs i, wherein the list of defined detector pairs comprises all pairs which are formed combinatorially from the set of all detectors, and the defined detector pairs comprise at least one detector made of a material having an atomic number of Zeff > 30; - a data acquisition system which determines a ranking for both detectors involved in a coincidence event, which ranking defines a first and a second detector, and orders the energies (E1, E2) measured in coincidence events according to their identification 1, 2 and stores them in a chronological list with the attributes {i, E1, E2} and the detection time t; wherein in each detector pair i the detector with the lower order number is given the number 1 and the detector with the higher order number is given the number 2; if both detectors of a pair have the same order number, the identification as 1 and 2 is arbitrarily taken; and - an analysis unit which calculates a function value f(E1, E2) = (E2-E1) / (E1+E2) for each coincidence event from the energies (E1, E2) stored by the data acquisition system and detects the coincidence events according to their detector pair number i and their functional values f(E1, E2) in one or more frequency distributions Y and using a statistical image reconstruction method of emission tomography using lookup tables LUTSK calculates from each frequency distribution Y a respective directional distribution X of the radiation field, wherein a respective frequency distribution Y is provided for each radionuclide and for each directional distribution X.

11. The device according to claim 10, characterized in that the device is partially or completely formed as a Compton camera, a Compton telescope, a Single Plane Compton camera, a neutron camera and / or a dual gamma / neutron camera; and / or wherein all the detectors are arranged in a ring and / or wherein inside the ring there are no central detectors, there is one central detector or there are multiple central detectors; and / or the ring preferably contains four or five and particularly preferably six plastic scintillation detectors and / or one or two scintillation detectors of Nal, Csl, CeBr3 and / or LaBr3 are preferably located inside the ring; and / or the scintillation detectors are preferably designed in the sizes 1"×1", 1.5"×1.5", 2"×2" and / or 3"×3".

12. The device according to claim 10 or 11, characterized in that a scintillation detector is used as detector and / or the scintillator is formed as monolithic block or as a pixelated scintillator module and / or is formed from pure or doped materials of the group of PVT, anthracene, stilbene, p-terphenyl, CaF2, BaF2, Nal, CeBr3, LaBr3, LaCl3, La(BrxCl1-x)3, CsI, SrI2, CLYC, CLBC, CLCB, CLLB, BGO, LSO, LYSO, GAGG, YAP and / or YAG; and / or in that a semiconductor detector is used as the detector and / or the semiconductor is designed as a segmented or unsegmented semiconductor and / or has a planar or coaxial geometry and / or is formed from materials of the group of Ge, GaAs, CdTe and / or CdZnTe.

13. The device according to any one of claims 10-12, characterized in that at least two detectors are used; and / or in that all the detectors used are substantially identical in construction or that at least two of the detectors used are different from one another.

14. The device according to any one of claims 10-13, characterized in that detector and system electronics used use analog and / or digital electronic components; and / or in that the analog electronic components comprise a combination of different modules including a high voltage supply, a preamplifier, an amplifier, a pulse shaper, a charge integrator, a pulse height analyzer, a multichannel analyzer (MCA) and / or a coincidence circuit; and / or in that the digital electronic components comprise a combination of different hardware and software components, which comprise a high-voltage supply, an A / D converter per detector, a field programmable gate array (FPGA), a storage medium, a digital signal processor and / or an evaluation software.