Method for determining the bias of a pixel of a pixelated detector affecting ionizing radiation

Abstract

A method for determining a bias (β) of at least one pixel of a detector (1) affecting ionizing radiation i,j ) method, the detector includes a plurality of pixels (10 i,j ), each pixel is configured to collect charge carriers (6) generated by the interaction of ionizing radiation in the detector, and a pulse signal is formed under the action of the generation and collection of charge carriers, the pixels are distributed in a matrix array, and the method comprises: a) after the interaction occurs in the detector, determining the pixel that forms a pulse exceeding the amplitude threshold during the detection time interval; b) among each pixel determined in step a), selecting a pixel of interest that produces the highest amplitude; c) selecting at least one remote pixel (10 f ), where the position of the distal pixel relative to the pixel of interest is predetermined; d) measuring the amplitude of the signal generated by each distal pixel; e) determining an offset for each distal pixel at detection time based on each measurement performed in step d).

Description

Technical Field

[0001] The present invention relates to detectors of ionizing radiation, in particular detectors of photon X-rays or gamma radiation. Background Art

[0002] Devices for detecting ionizing radiation based on gas, semiconductor or scintillation detector materials allow obtaining electrical pulses formed by the interaction of radiation in the detector material. The amplitude of each pulse depends on the energy deposited by the radiation during each interaction. These devices are often used in applications where it is necessary to know the energy of the ionizing radiation incident on the detector. The fields of application are wide and include, in particular, non-destructive testing in nuclear facilities (e.g. baggage screening), medical diagnostics or measurements. Typically, the devices used are pixelated in order to produce spatially resolved information. The detection device can be linear when the pixels are aligned to form a straight line. The detection device can also be two-dimensional when the pixels are distributed in a matrix array to form an image.

[0003] Under the influence of the interaction between the radiation and the detector material, one or more pixels generate an electrical pulse whose amplitude is related to the energy released by the radiation during the interaction. Each pixel is connected to an electronic circuit for processing the pulse. This circuit allows its amplitude to be estimated as accurately as possible. Estimating the amplitude of the pulse generated by each pixel enables an estimate of the energy of the interaction. This energy must be estimated as accurately as possible. The energy range addressed is typically 10 keV to several hundred keV or even several MeV. An energy accuracy of approximately 1%, or even lower, is desirable.

[0004] The estimation of the amplitude allows the formation of a spectrum of the radiation detected by each pixel using the well-known principle of spectrometry. Recall that a spectrum is a histogram of the amplitudes of the pulses detected during an acquisition period. If the relationship between the amplitude of a pulse and the energy corresponding to that amplitude is known, the amplitude spectrum can be converted into an energy spectrum. The estimation of the energy also allows the creation of a spectral image, i.e., an image in one or more predefined energy bands.

[0005] Energy estimation based on the amplitude of each pulse is performed by taking into account gain and bias terms to form a generally linear relationship. Gain is a multiplicative term, while bias is an additive term. Therefore, if A corresponds to the amplitude of the pulse and E corresponds to the energy released in the detector, then gain α and bias β can be determined such that E = αA - β. α and β are positive scalars. When using a pixelated detector, bias β and gain α can be determined for each pixel.

[0006] Drift in bias or gain can cause drift in the energy estimated based on the measured amplitude. Regarding bias, calibration can be performed by exposing the device to an environment that can be considered radiation-free. However, this assumes specific calibration operations outside of normal use of the device.

[0007] The invention described below allows tracking of biases affecting pixels of a pixelated detector. This tracking allows better monitoring of bias drift over time. This results in better energy estimation accuracy. One advantage of the invention is that it does not require interruption of device operation. Summary of the Invention

[0008] A subject of the invention is a method for determining a bias of at least one pixel of a detector affecting ionizing radiation, the detector comprising a plurality of pixels, each pixel being configured to collect charge carriers generated by interaction of ionizing radiation in the detector and forming pulses as a result of the generation and collection of charge carriers, the pixels being distributed in a matrix array, the method comprising:

[0009] a) determining, after the interaction occurs in the detector, at a detection time at least one pixel which forms a pulse exceeding a predefined amplitude threshold during a detection time interval established based on the detection time;

[0010] b) selecting, from each pixel determined in step a), the pixel that produces the highest amplitude during the detection time interval, the selected pixel being the pixel of interest;

[0011] c) selecting at least one distal pixel, the position of the distal pixel relative to the pixel of interest being predetermined such that the effect of the interaction detected by the pixel of interest on the distal pixel can be considered negligible;

[0012] d) measuring the amplitude of the signal generated by each distal pixel during the detection time interval;

[0013] e) determining an offset for each distal pixel at detection time based on each measurement performed in step d);

[0014] f) Update the bias affecting each distal pixel according to the bias determined at detection time.

[0015] Steps a) and f) may be repeated each time an interaction occurs in the detector which results in the amplitude threshold value for at least one pixel being exceeded in step a).

[0016] According to one embodiment, step a) comprises:

[0017] - determining a detection time at which a first pixel forms a first pulse exceeding an amplitude threshold;

[0018] - defining a detection time interval extending from the detection time, the duration of the detection time interval being predetermined;

[0019] - determining pixels in the detection time interval which have formed pulses exceeding an amplitude threshold.

[0020] The detection time interval may include the time before the detection time. It may then be on either side of the detection time.

[0021] According to one embodiment, step b) comprises selecting a plurality of distal pixels, each distal pixel having a predetermined position relative to the pixel of interest.

[0022] Each pixel is associated with one or more locations such that when the pixel is a pixel of interest, each associated location corresponds to a distal pixel.Two different pixels are preferably associated with at least two different locations.

[0023] In the pixel matrix array, the position of each remote pixel may be changed according to the position of the pixel of interest in the pixel matrix array.

[0024] According to one embodiment, in step f):

[0025] - Before detection time, each pixel is affected by the current bias;

[0026] - After detecting the interaction, updating the bias for each distal pixel according to the current bias and the bias determined at the time of detection.

[0027] The current bias can be initialized to a predefined value, such as zero. The bias can be updated by a weighted sum of the current bias and the bias at the detection time. According to one possibility, the bias of the far-end pixel at the detection time is influenced by a weighting factor that varies with the number of bias updates that have been performed for the far-end pixel since initialization, such that the lower the number of bias updates that have been performed since initialization, the higher the contribution of the bias measured at the measurement time. In other words, the lower the number of bias updates, the higher the weighting factor.

[0028] Each weighting factor can be initialized to the same value for all pixels, especially when the detector starts.

[0029] According to one embodiment, the method comprises, after step b):

[0030] -Select at least one pixel adjacent to the pixel of interest;

[0031] - measuring the amplitude of the pulse generated by the pixel of interest and the amplitude of the pulse generated by each neighboring pixel;

[0032] - Estimate the energy released during the interaction from the amplitudes measured for the pixel of interest and for each neighboring pixel.

[0033] The selected adjacent pixels may be pixels adjacent to the pixel of interest in two orthogonal directions.

[0034] According to one embodiment, each pixel comprises a peak detector such that an interaction is deemed to have been detected when the amplitude of a pulse resulting from the pixel collecting charge exceeds a predefined amplitude threshold.The method then comprises a periodic resetting of each peak detector.

[0035] Between two consecutive resets of the peak detector, the method may comprise estimating a drift signal that is added to the signal resulting from the charge collected by the pixel. The method may then be such that, in step e), the drift signal estimated at the detection time is subtracted from the bias.

[0036] A second subject of the invention is a method for estimating the energy corresponding to charge carriers collected by pixels of a pixelated detector for detecting ionizing radiation, each pixel being configured to collect charge carriers generated by the interaction of ionizing radiation in the detector and to form a pulse as a result of the collection of charge carriers, the pixels being distributed in a matrix array, the method comprising, when the amplitude of the pulse formed by the pixel after the collection of the charge carriers exceeds a predefined threshold:

[0037] - Determine the amplitude of the pulse;

[0038] - applying an energy quantization function to the determined amplitude;

[0039] The method is characterized in that the energy quantization function is parameterized by a bias estimated using the method according to the first subject matter of the invention.

[0040] For example, as defined with reference to the first subject matter of the present invention, the pixel may be a pixel of interest or a pixel adjacent to the pixel of interest.

[0041] A third subject of the invention is a device for detecting ionizing radiation, comprising various pixels of a pixelated detector, each pixel being configured to collect charge carriers generated by the interaction of ionizing radiation in the detector and to form pulses as a result of the collection of charge carriers, the device comprising a microprocessor configured to implement steps b) to f) of the method according to the first subject of the invention for each pulse formed by at least one pixel.

[0042] The invention will be better understood by reading the description of the exemplary embodiment, which in the remainder of the specification is described with reference to the drawings listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1A The main components of a detector of ionizing radiation are shown schematically.

[0044] Figure 1B A matrix array of pixels is schematically shown.

[0045] Figure 2A A pulse formed by a first pixel at a detection time is shown, which causes the threshold comparator to be triggered, and takes into account a detection time interval extending from the detection time.

[0046] Figure 2B The pulse formed by the pixel of interest during the detection time interval is shown.

[0047] Figure 3 The main steps of the method according to the invention are shown.

[0048] Figure 4A and Figure 4B Two different positions of the pixel of interest are schematically shown. DETAILED DESCRIPTION

[0049] Figure 1A and Figure 1B A pixelated radiation detector 1 is shown, configured to interact with ionizing radiation 5. Ionizing radiation refers to radiation formed by particles capable of ionizing matter. This radiation may be alpha radiation, beta radiation, photon X-rays, gamma radiation, or even neutron radiation. In the example shown, the radiation is photon gamma radiation, formed by photons having energies ranging from 1 keV to 2 MeV, for example. The present invention is particularly suitable for exposing the radiation detector to isotope gamma-emitting sources.

[0050] In the example shown, the detector comprises a semiconductor detector 2 of the CdTe type, but it may also involve any semiconductor commonly used for detecting ionizing radiation (e.g. Ge, Si or CdZnTe). The semiconductor is biased by an electric field E formed between an anode 10 and a cathode 20. When particles of ionizing radiation 5 (in this case photons) interact in the detector 1, charge carriers 6 are formed and these charge carriers are directed to the electrodes divided into pixels 10. i,j The collecting electrode 10 migrates. In the example shown, the collecting electrode is the anode. i,j The amount of charge Q collected is preferably linearly dependent on the energy E released by the particles under the influence of the interaction.

[0051] Therefore, if Figure 1B As shown, the radiation detector 1 comprises a plurality of pixels 10 arranged in a plane referred to as the detection plane P. i,j . The subscripts i and j represent the coordinates of each pixel in the detection plane. The pixels are preferably arranged in a regular two-dimensional matrix array. In the example described in the present application, the detection plane P is defined by two perpendicular axes X and Y. Each pixel is, for example, a square with a side length l, where l varies, for example, from 500 μm to 1 cm. Perpendicular to the axes X and Y, the semiconductor of the detector 2 has a thickness ε parallel to the axis Z. The thickness ε includes, for example, a thickness of 500 μm to 5 mm. Along the X-axis and the Y-axis, the number of pixels can be tens or even hundreds. For example, the pixels can be grouped into subsets of 256 pixels (16×16 pixels), which are placed side by side to form the detector 1.

[0052] Other types of detectors (e.g., a scintillator coupled to a photon / charge carrier converter) are also possible, provided that they allow the collection of a certain amount of charge Q under the action of the energy E released by the ionizing radiation during the interaction process in the detector material. Detectors involving assemblies based on scintillators and photomultiplier tubes or scintillators and photodiodes may be of note.

[0053] 10 per pixel i,j Including electronic processing circuit 11 i,j , the electronic processing circuit comprises:

[0054] -Circuit 12 i,j , which is used to form a voltage pulse due to the collection of charge carriers. This involves a circuit known to those skilled in the art capable of current-to-voltage conversion. It includes a preamplifier, an amplifier, and a shaping circuit, which is often referred to as a shaper.

[0055] -Threshold comparator 13 i,j , in the circuit 12 i,j In case the amplitude of the generated pulse exceeds a predefined amplitude threshold, it addresses the trigger signal.

[0056] - Peak detector 15 i,j , which allows each pulse forming circuit 12 i,j The highest value of the generated signal is stored in a memory. This may involve the memory allowing the signal from the circuit 12 to be read for a predefined length of time (eg 15 μs). i,j The highest value of the output signal is stored in memory.

[0057] Each peak detector 15 i,jIt can be reset at regular intervals, as described below. It can also be reset at the threshold comparator 13 i,j Each peak detector is reset at the end of the detection time interval dt defined after the triggering of each peak detector. i,j Threshold comparator 13 i,j Placed in parallel.

[0058] In this example, each pixel 10 i,j The associated processing circuits 11 i,j is formed in an integrated circuit 14. The integrated circuit may take the form of an application specific integrated circuit (ASIC).

[0059] In this example, the subscripts i, j indicate that there is one processing circuit per pixel. Alternatively, a given processing circuit is common to a plurality of pixels, and the processing of the signal received from each pixel is multiplexed.

[0060] Electronic processing circuit 11 i,j An analog-to-digital converter may be included, and the ASIC then processes the digital signal. The analog-to-digital converter may be placed at the output of the integrated circuit 14.

[0061] The integrated circuit 14 is connected to a downstream processing unit 19 which is configured to perform the following operations:

[0062] - Threshold comparator 13 triggered at detection time t i,j Under the action of , consider the detection time interval dt;

[0063] - during the detection time interval dt, identifying pixels which form a pulse whose amplitude exceeds an amplitude threshold;

[0064] -Select the pixel of interest 10 from the identified pixels int , the pixel of interest is a pixel that forms a pulse whose amplitude is the highest among the pulses that have exceeded the amplitude threshold.

[0065] In this example, the processing unit 19 comprises digital components in the form of a Field Programmable Gate Array (FPGA), ie an integrated circuit consisting of a network of programmable cells.

[0066] - A module 16 for determining the bias corresponding to each pixel. The processing unit 19 may address tens or even hundreds of pixels, for example 256 pixels.

[0067] - a memory 17 for storing the value of the offset determined for each pixel respectively;

[0068] - a module 18 for generating a signal according to the processing circuit 11 i,jThe resulting amplitude is used to quantify the energy using a calibration function g parameterized in particular by a bias stored in memory for each pixel.

[0069] The processing unit 19 may also process digital data logically. In this case, the modules 16 and 18 listed above may take the form of functions that are coded into a memory and implemented by a microprocessor.

[0070] Figure 2A Schematically shows the first pixel (ie, the pixel whose amplitude exceeds the predefined threshold value A) th The time when the amplitude exceeds the threshold value is the detection time t. The detection time interval dt is defined relative to the detection time t. The detection time interval dt has a predefined duration, for example, 5 μs. The detection time interval extends from the detection time t. It can also be located on either side of the detection time t, for example, from t-2 μs to t+3 μs. As described above, during the detection time interval dt, the ASIC 14 recognizes that a pulse has been formed whose amplitude exceeds the amplitude threshold value A. th Among the pixels, the processing unit 19 selects the pixel of interest 10 int , that is, the maximum amplitude A is reached during the detection time interval max pixels.

[0071] Figure 2B The pulse formed by the pixel of interest is schematically shown. In the detection time interval dt, the amplitude A of the pulse is max Higher than the amplitude of the pulses formed by other pixels.

[0072] As described with reference to the prior art, the energy released by the interaction of the photons in the detector can be estimated from the amplitude A of the pulse formed by each pixel. To this end, a quantization function g is applied that allows the energy to be quantized according to the amplitude. The quantization function can involve a linear relationship, for example in the form of an affine quantization function g such that:

[0073] E=g α，β (A) = αA-β (1)

[0074] Symbol g α，β This means that the function g is parameterized by the gain α and the bias β.

[0075] The energy E corresponds to the energy released in the detector by the interaction of the photons when all charge carriers are collected by a single pixel.

[0076] As mentioned in the prior art, the coefficients α and β are positive real numbers: α is the gain, and β is the bias.

[0077] The coefficients α and β are variable from one pixel to the next. In the rest of this specification, α i,j and β i,j Each pixel is 10 i,j The associated gain and offset. Coefficient α i,j and β i,j It is the parameter of the quantization function defined for each pixel. i,j and β i,j So that each pixel 10 i,j Electronic Circuit 11 i,j The amplitude A of the pulse S formed is used to improve the estimate of the energy transported by the charge carriers collected by the pixel. Therefore, the coefficient α is accurately determined. i,j and β i,j is important. The coefficient α i,j and β i,j This can be determined by energy calibration, during which, as described above, each pixel is exposed to radiation (whose energy spectrum contains singularities). This calibration can be performed before the detector is started up. However, during the use of the detector, the bias β i,j Thus, in accordance with the electronic circuit 11 associated with the pixel i,j Errors arise in the energy quantization of the amplitude of the resulting pulse S.

[0078] Therefore, it is important to estimate the bias β associated with each pixel with as good accuracy as possible and as regularly as possible. i,j As mentioned above, for a given pixel, the bias β i,j It can change randomly over time. Therefore, the bias β is updated periodically. i,j is important.

[0079] The pixel 10 used for the estimation detector will now be described. i,j The bias β i,j The method, the steps of which are Figure 3 Generally, before using the detector 1, the gain α has been determined. i,j and bias β i,j Therefore, each pixel 10 i,j The bias β i,j is initialized to the initial value.

[0080] Step 100 : Detecting interactions

[0081] When the interaction occurs in the detector material 2, sufficient energy is released that the first pixel 10 i,jA significant charge Q is collected. After this charge is collected, the electronic processing circuit 11 associated with the pixel i,j A pulse S is formed, the amplitude of which causes the threshold comparator to be triggered. In this case, the interaction is considered to be detected at the detection time t. As described above, a detection time interval dt is defined, which extends from the interaction time t. The detection time interval dt can be located on either side of the interaction time t. The duration of the detection time interval dt is predetermined.

[0082] Step 110 : Select the pixel to focus on.

[0083] In the detection time interval dt, it is recognized that a signal has been generated that exceeds the amplitude threshold A. th The processing unit 19 selects the pixels of the pulse whose amplitude A has been formed. max The pixel with the highest pulse (this pixel is called the focus pixel 10 int ). This involves that pixels parallel to the detection plane P are considered to be closest to the interaction location.

[0084] This step is achieved by using a downstream processing unit 19 which temporarily stores the signal derived from each pixel.

[0085] At the end of step 120, the pixel of interest 10 int The coordinates (i int ,j int ) is known.

[0086] Step 120 : Select one or more distant pixels.

[0087] In this step performed by the offset determination module 16, the pixel 10 is selected as being away from the pixel of interest. int At least one pixel 10 i,j-f . Far pixel 10 i,j-f It refers to the pixels that are considered far enough away that no charge will be collected or the amount of charge collected is negligible after the interaction occurs at the detection time t. f Relative to the focus pixel 10 int Therefore, when the focus pixel 10 has been located int When the remote pixel is automatically assigned 10 based on the predefined relative position f location.

[0088] Figure 4A and Figure 4B An example is shown in which, after identifying the pixel of interest 10 int Then four far pixels 10 are selected at the same timei,j-f .exist Figure 4A and Figure 4B Focus on Pixel 10 int is different. It is represented by a black square. The far pixels are shaded. Figure 4A and Figure 4B In the example shown, if (i int ,j int ) corresponds to the focus pixel 10 int The coordinates of the far pixel 10 i,j-f Located at coordinates (i int -2, j int +1)、(i int -1,j int -2)、(i int +2,j int -1)、(i int +1,j int +2). Therefore, according to this example, the pixel of interest 10 int The coordinates (i int ,j int ) and each distal pixel 10 i,j-f The norm of the vector formed by the coordinates of is equal to 3. i,j-f In , the index i,jf means that it refers to the far pixel of coordinate (i,j).

[0089] An important aspect of the present invention is that each distal pixel 10 i,j-f Relative to the focus pixel 10 int The relative position of is predefined and depends inter alia on the size l of the pixel and the thickness of the detector material 2.

[0090] To save processing time after the detection of each pulse, the far pixel 10 i,j-f The number cannot be too high. However, it must be high enough to satisfactorily characterize the bias β i,j Therefore, the number of distal pixels considered at each detection is preferably 1 to 20, and more preferably 2 to 10. Thus, at each detection of an interaction, a small fraction of the pixels in the matrix array are considered distal pixels, and their biases are updated accordingly. After a certain number of interactions, the biases of all pixels of the matrix are considered to have been progressively updated.

[0091] Each far pixel 10 i,j-f Relative to the focus pixel 10 int The fact that the relative positions of the pixels are predetermined enables the individual positions of the remote pixels to be addressed successively. intcan vary, which results in the distal pixel 10 i,j-f At each interaction, the probability that a pixel of the matrix array is a pixel of interest can be assumed to be equal relative to all pixels of the matrix array. Under this assumption, the method enables the biases of all pixels of the matrix array to be regularly updated.

[0092] After step 120, each distal pixel 10 i,j-f The coordinates are known.

[0093] Step 130 Determine the bias.

[0094] For each distal pixel 10 identified in step 120 i,j-f The offset determination module 16 determines the value of the signal S output from the pixel during the detection time interval dt. i，j-f The highest amplitude A i，j-f (t) to update the bias value β i,j (t). According to the first method, it can be estimated that:

[0095] β i，j (t) = A i，j-f (t) (2)

[0096] This is equivalent to considering that the pixel 10 that does not collect charge i,j-f Detection circuit 11 i,j The generated signal S i，j-f The magnitude corresponds to the pixel offset.

[0097] Before starting the detector, each pixel has 10 i,j The bias value is initialized to the initial value β init-i,j It may relate to an arbitrary value defined, for example, during energy calibration, or to a zero value, or to a value resulting from a previous use of the detector.

[0098] Pixel 10 at coordinate (i,j) i,j After being selected as a far pixel, according to the detected signal S i，j-f Amplitude A i，j-f (t) to update the bias value associated with the pixel. As mentioned above, the first option is to use the value A detected at the detection time i，j-f (t) Replace the bias value β corresponding to the pixel i,j However, the inventors believe that it is better to consider the bias β i,j (t - ) current value, that is, the value before the detection time, to gradually update the offset value. Symbol t -Corresponding to the time before the detection time t, for example, the time immediately before the detection time t. i,j (t - ) and A f-i，j The update is performed by taking the weighted sum of (t).

[0099] The following expressions can be used to perform updates:

[0100]

[0101] in:

[0102] -β i，j (t) = A i，j-f (t) (2)

[0103] -k i,j is a smoothing coefficient, which involves a strictly real number between 0 and 1.

[0104] value The value corresponding to the offset after the update is stored in the memory 17.

[0105] The smoothing factor makes it possible to modulate the inclusion of the bias value established at the detection time t relative to the current value of the bias. i，j =2 -8 .

[0106] According to one embodiment, the smoothing coefficient k i,j The value of can vary with time and pixel position depending on the number of bias updates that have been performed. i,j (t) specifies the pixel 10 before the detection time t i,j The number of updates of the bias is applied to pixel 10 i,j The smoothing coefficient k i,j for:

[0107]

[0108] Here, log2 is the logarithm with base 2.

[0109] n i,j (t) may correspond, for example, to pixel 10 having been executed since bias initialization. i,j The number of bias updates.

[0110] According to this embodiment, when at least 256 updates have been performed, the smoothing factor is equal to 2 -8 This corresponds to a number of occurrences equal to 256, where the pixel in question is 10 i,jare considered as far pixels. This embodiment amounts to overweighting the first updates performed after initialization until their number reaches 256.

[0111] After step 130, the corresponding bias of each distal pixel is updated using expression (2), and optionally taking into account expression (3). These biases are stored in memories respectively connected to each pixel in question.

[0112] Step 140 : Store the offset in memory.

[0113] In step 140 , the respective offsets updated for each pixel in step 130 are stored in memory 17 .

[0114] Step 150 : Quantify the energy collected by the pixel of interest.

[0115] In step 150, the application is applied to the pixel of interest 10 int The established quantization function g i，j , in order to estimate the energy E corresponding to the pulse collected by this pixel int .

[0116] therefore,

[0117] Among them, α int and β int Corresponding to the focus pixel 10 int The corresponding gain and offset.

[0118] Step 150 is performed by the energy quantization module 18 .

[0119] Step 160 : Consider neighboring pixels.

[0120] After the interaction, some of the charge carriers 6' may reach the pixel of interest 10 int Adjacent pixels 10 adj In this case, the method considers the 10 neighboring pixels of each adj The amplitude threshold A is exceeded during the detection interval dt th signal

[0121] therefore,

[0122] Among them, α adj and β adj Corresponding to adjacent pixels 10 adj gain and offset.

[0123] exist Figure 4Aand Figure 4B In the example shown, the nearest pixel of interest 10 is considered. int 4 adjacent pixels 10 adj .exist Figure 4A and Figure 4B In the adjacent pixel 10 adj is painted with a texture formed by small dots. Step 160 is performed for each neighboring pixel in question.

[0124] Step 160 is optional and is implemented by the energy quantization module 18 when the size of the pixel is such that at least one adjacent pixel is likely to collect a non-negligible percentage of the charge carriers.

[0125] Step 170 : Estimate the energy released by a photon.

[0126] The magnitude of the energy released by an incident gamma photon in the detector can be estimated based on the energy collected by the pixel of interest and any energy collected by each neighboring pixel. In general, the energy released at the detection time E(t) is considered to correspond to the energy collected by the pixel of interest 10 int and each adjacent pixel 10 adj The amplitude of the collected data exceeds the amplitude threshold A. th The sum of the energies of the pulses S.

[0127] Step 180 : Positioning interactions

[0128] Optionally, the device 1 can be used to estimate the position of the interaction in the detector material and in particular in a plane parallel to the detection plane. If neighboring pixels are not taken into account, the position of the interaction corresponds to the pixel of interest 10 int When the neighboring pixels are taken into account, the location of the interaction can be refined in step 160 by calculations such as a centroid calculation. The centroid calculation consists in weighting the coordinates of each pixel by the energy collected by each pixel. This involves a method known to those skilled in the art, namely estimating the location of the interaction with a resolution better than the size of each pixel. This step is performed by the processing unit 19.

[0129] Steps 100 to 180 are repeated each time a new interaction is detected.

[0130] Reset the peak detector.

[0131] According to one embodiment, the device is configured to separately connect each pixel 10 i,j Each peak detector associated with 15 i,j Specifically, as long as the pixel does not collect charge carriers generated by the interaction, the drift potential Vdrift It gradually accumulates at each peak detector 15 i,j After a certain period of time, the drift potential V drift To avoid this drift, all peak detectors 15 can be periodically commanded to i,j The reset period may be, for example, 1 ms. Such a reset enables the peak detector 15 to be reset. i,j A reset is performed to prevent the respective peak detectors 15 i,j Drift potential V drift Become too big.

[0132] variants

[0133] In the peak detector 15 i,j Between two consecutive resets, it is preferred to consider the drift potential V drift During the calibration phase, in the absence of incident radiation, the drift potential V can be determined for each pixel. drift Thus, for each pixel, a time-dependent drift function V is obtained. drift(t') , where t' is the peak detector 15 connected to the pixel i,j The time-dependent drift function of each pixel can be averaged to obtain the average function It is considered to apply to every pixel.

[0134] When considering this function, expression (2) is replaced by:

[0135]

[0136] In the estimation of the bias, it is optional to consider the drift charge. After digitization, the drift charge is considered to correspond to 30 to 350 LSBs (least significant bits), while the bias is considered to correspond to 3000 to 4000 LSBs (least significant bits).

[0137] The above method has the following advantages:

[0138] - estimation of the bias of the pixels during use of the detector. This is due to the fact that, at each event, the method includes updating the bias of pixels that are distant and therefore considered unimportant for the processing of the event;

[0139] - The distribution of pixels whose biases are updated varies according to the position of the pixel of interest in the pixel matrix array. This enables a gradual update of all pixels of the detector.

[0140] - In each event, the bias of a small number of pixels is updated, which means that the method is not resource intensive and remains compatible with fast acquisition rates.

[0141] The method can be implemented such that after detecting an event and determining the pixel of interest, a read mask is applied to the matrix array in order to identify the far end pixels. int As the center, and defines the far pixel 10 i,j-f Each position and any adjacent pixel 10 adj Therefore, the number of pixels read on each event ranges from 2 (the pixel of interest and a single far-end pixel, not considering neighboring pixels) to 10 or even 15 (one pixel of interest and 4 or 6 neighboring pixels, the others being far-end pixels). When the pixel of interest is on or near a boundary of the matrix, the number and position of the far-end pixels and any neighboring pixels are adjusted.

[0142] The invention may be implemented in measurement applications for locating radioactive sources, for example in nuclear facilities or in the environment.

Claims

1. A method for determining a bias (β) of at least one pixel of a detector (1) affecting ionizing radiation i,j ) method, wherein the detector includes a plurality of pixels (10 i,j ), each pixel being configured to collect charge carriers (6) generated by interaction (5) with ionizing radiation in the detector and to form pulses (S) as a result of the generation and collection of the charge carriers, the pixels being distributed in a matrix array, the method comprising: a) determining, after the interaction in the detector, at a detection time (t) at least one pixel which forms an amplitude exceeding a predefined amplitude threshold (A) during a detection time interval (dt) established based on the detection time (t) th ) pulse; b) selecting, from each pixel determined in step a), the pixel which produces the highest amplitude (A) during said detection time interval max ) of the pixel, the selected pixel is the focus pixel (10 int ); c) Select at least one far pixel (10 i,j-f ), the distal pixel is relative to the focus pixel (10 int ) is predetermined such that the effect of the interaction detected by the pixel of interest on the distal pixel is considered negligible; d) measuring the amplitude of the signal generated by each distal pixel during the detection time interval; e) determining, based on each measurement performed in step d), for each distal pixel at said detection time, a bias (β i,j(t) ); f) Updating the bias affecting each distal pixel according to the bias determined at said detection time.

2. The method according to claim 1, wherein Step a) comprises: - determining a detection time at which the first pixel forms an amplitude exceeding the amplitude threshold (A th )'s first pulse; - defining said detection time interval (dt), said detection time interval extending from said detection time (t), said detection time interval having a predetermined duration; - determining pixels forming pulses exceeding said amplitude threshold value during said detection time interval.

3. The method according to claim 1, wherein Step b) comprises selecting a plurality of distal pixels (10 i,j-f ), the position of each distal pixel relative to the focus pixel is predetermined.

4. The method according to claim 1, wherein The position of each distal pixel in the pixel matrix array changes according to the position of the focus pixel in the pixel matrix array.

5. The method according to claim 1, wherein In step f): Before the detection time, each pixel is subject to the current bias (β i,j (t - ))’s impact; - After detecting the interaction, according to the current bias (β i,j (t - )) and the bias (β i,j (t)) Update the bias for each far pixel.

6. The method according to claim 5, wherein: By the current bias (β i,j (t - )) and the bias of the detection time (β i,j (t)) to update the bias.

7. The method according to claim 6, wherein: At the far pixel (10 i,j-f ) bias (β i,j (t)) is affected by the weighting factor (k i,j ) influence, the weighting factor increases with the bias (n i,j ) such that the lower the number of updates of the bias that have been performed since initialization, the higher the contribution of the bias measured at the measurement time.

8. The method according to claim 1, comprising, after step b): - Select at least one pixel adjacent to the pixel of interest (10 adj ); - measuring the amplitude of the pulse generated by the pixel of interest and the amplitude of the pulse generated by each neighboring pixel; - Estimating the energy released during the interaction from the amplitudes measured for the pixel of interest and for each neighboring pixel.

9. The method according to claim 1, wherein Each pixel includes a peak detector (15 i,j ), so that when the amplitude of the pulse (S) generated by the pixel collecting charge exceeds the predefined amplitude threshold (A th ), an interaction is deemed to have been detected, the method comprising periodic resetting of each peak detector.

10. The method according to claim 9, comprising: i,j ) between two consecutive resets of the bias (β ) and the method is such that, in step e), the drift signal added to the signal resulting from the charge collected by the pixel is estimated. i,j (t)) is subtracted from the drift signal estimated at the detection time.

11. A method for estimating the energy corresponding to the energy of a pixel (10) of a pixelated detector (1) for detecting ionizing radiation. i,j ), each pixel being configured to collect charge carriers generated by interaction (5) with ionizing radiation in the detector and to form a pulse (S) under the effect of the collection of the charge carriers, the pixels being distributed in a matrix array, the method comprising, when, after the collection of the charge carriers, the amplitude of the pulse formed by the pixel exceeds a predefined threshold (A th )hour: - determining the amplitude of said pulse; - applying an energy quantization function (g) to the determined amplitude; The method is characterized in that the energy quantization function is parameterized by a bias, which is estimated using the method according to claim 1.

12. A device for detecting ionizing radiation, comprising various pixels of a pixelated detector, each pixel being configured to collect charge carriers generated by the interaction of ionizing radiation in the detector and to form a pulse as a result of the collection of the charge carriers, the device comprising a processing unit configured to implement steps b) to f) of the method according to claim 1 for each pulse formed by at least one pixel.

Images