Gamma photon positioning method and device for orthogonal strip-shaped tellurium-zinc-cadmium detector

Abstract

The invention relates to a gamma photon positioning method and device for an orthogonal strip-shaped tellurium-zinc-cadmium detector. The method comprises the following steps: acquiring current signals respectively generated by interaction of a plurality of gamma rays and the orthogonal strip-shaped tellurium-zinc-cadmium detector; vectorizing the plurality of current signals to obtain an input vector set; an initialized neural network is constructed based on the positions of a plurality of electrode strips in the orthogonal strip-shaped cadmium zinc telluride detector, and the number of initialized neurons is the same as that of the plurality of electrode strips; performing weight vectorization processing on the plurality of initialized neurons to construct an initialized weight vector set; and based on the input vector set and the initialized weight vector set, iteratively reducing the range of the initialized weight vector set through a self-organizing mapping algorithm to obtain a target weight vector, and determining the position of the gamma photon. The purpose of determining the position of the gamma photon based on the neural network and the self-organizing mapping algorithm is achieved, and therefore the technical effects of improving the accuracy of the position of the gamma photon and improving the position obtaining efficiency are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of tellurium zinc cadmium detector, and particularly relates to a gamma photon positioning method and device for orthogonal strip tellurium zinc cadmium detector. BACKGROUND

[0002] The tellurium zinc cadmium detector is a room temperature semiconductor nuclear radiation detector with excellent performance. The tellurium zinc cadmium material belongs to the II-VI compound, has the characteristics of large relative atomic number, large band gap, high resistivity, small leakage current, and can realize high energy resolution and spatial resolution by using a reasonable electrode structure, and is widely used in the field of nuclear radiation imaging.

[0003] In the process of nuclear radiation imaging, the gamma rays need to irradiate the tellurium zinc cadmium detector. The gamma rays interact with the tellurium zinc cadmium detector to form gamma photons, and the positions of the gamma photons are obtained, so that the final imaging result is obtained based on the positions of the gamma photons.

[0004] Since the position of the gamma photon is calculated by using a mathematical model in the related art, the position data obtained has deviation, which leads to inaccurate position of the gamma photon generated by the interaction of the gamma rays with the orthogonal strip tellurium zinc cadmium detector, and low efficiency of position acquisition.

[0005] The above problems need to be solved. SUMMARY

[0006] The present application discloses a gamma photon positioning method and device for orthogonal strip tellurium zinc cadmium detector, and aims to solve the technical problems existing in the prior art.

[0007] The present application adopts the following technical solutions:

[0008] On the one hand, the present application provides a gamma photon positioning method for orthogonal strip tellurium zinc cadmium detector, which comprises: obtaining current signals respectively generated by the interaction of multiple gamma rays with the orthogonal strip tellurium zinc cadmium detector; performing vectorization processing on the multiple current signals to obtain an input vector set; constructing an initialization neuron network based on the positions of multiple electrode strips in the orthogonal strip tellurium zinc cadmium detector, wherein the number of initialization neurons is the same as the number of the multiple electrode strips; performing weight vectorization processing on the multiple initialization neurons to construct an initialization weight vector set; based on the input vector set and the initialization weight vector set, iteratively reducing the range of the initialization weight vector set by using a self-organizing mapping algorithm to obtain a target weight vector, and determining the position of the gamma photon generated by the interaction of the gamma rays with the orthogonal strip tellurium zinc cadmium detector.

[0009] Optionally, the self-organizing mapping algorithm is used to iteratively reduce the range of the initial weight vector set based on the input vector set and the initial weight vector set to obtain a target weight vector, and the position of the gamma photon generated by the interaction of the gamma ray and the orthogonal strip CdZnTe detector is determined, comprising: constructing an initial neighborhood range of the initial weight vector set, wherein the neighborhood range is used to indicate the range of the initial weight vector set; randomly sampling a first input vector in the input vector set until all input vectors in the input vector set are sampled, iteratively updating the initial weight vector set based on the first input vector to obtain a target weight vector set, wherein the first input vector is used to indicate the currently sampled input vector; obtaining a target weight vector based on the first neighborhood range and the target weight vector set; and determining the position of the gamma photon based on the target weight vector.

[0010] Optionally, the first input vector is randomly sampled in the input vector set until all input vectors in the input vector set are sampled, and the initial weight vector set is iteratively updated based on the first input vector to obtain a target weight vector set, comprising: performing similarity matching between the first input vector and the initial weight vector set to determine a first weight vector, wherein the first weight vector is used to indicate the weight vector in the initial weight vector set that has the smallest distance interval from the first input vector; updating the initial neighborhood range based on the first input vector to obtain a first neighborhood range, wherein the first neighborhood range is used to indicate a region with the first input vector as the center; generating a plurality of first neurons in the first neighborhood range, wherein the first neuron is used to indicate the position information of the orthogonal strip CdZnTe detector divided in the first neighborhood range; and performing weight vectorization processing on a plurality of first neurons to construct a target weight vector set.

[0011] Optionally, the initial neighborhood range is updated based on the first input vector to obtain a first neighborhood range, comprising: determining a minimum distance interval based on the first input vector and the first weight vector; determining an effective width of the first neighborhood range, wherein the effective width is half of the diameter of the first neighborhood range; and determining the first neighborhood range based on the effective width and the minimum distance interval.

[0012] Optionally, the minimum distance interval is determined based on the first input vector and the first weight vector, comprising: the minimum distance interval is calculated as follows:

[0013] i(x) = arg min||X j -ω j ||

[0014] Among them, i(x) is the minimum distance interval, X j is the first input vector of the jth, ω j is the first weight vector of the jth item, and argmin is the minimum value of the function.

[0015] Optionally, determining a first neighborhood range based on the effective width and the minimum distance interval includes calculating the first neighborhood range as follows:

[0016] h (j,i(x)) =exp(-||X j -ω j || 2 / (2σ 2 ))

[0017] Among them, h (j,i(x)) is the first neighborhood range, X j is the first input vector of the jth, ω j is the first weight vector of the jth node, ||X j -ω j || is the distance interval, and σ is the effective width.

[0018] Optionally, obtaining a target weight vector based on the first neighborhood range and the target weight vector set includes: calculating the target weight vector as follows:

[0019] w j+1 =ω j +αh (j,i(x)) (X j -ω j )

[0020] Among them, w j+1 is the target weight vector, ω j is the first weight vector of the jth node, h (j,i(x)) is the first neighborhood range, α is the learning rate parameter, X j is the first input vector of the jth.

[0021] According to another aspect of the embodiments of the present application, there is also provided a gamma photon positioning device of a orthogonal strip CdZnTe detector, comprising: an acquisition module configured to acquire current signals respectively generated by a plurality of gamma rays interacting with a orthogonal strip CdZnTe detector; an input vector module configured to vectorize the current signals to obtain an input vector set; a neural network module configured to construct an initialized neuron network based on positions of a plurality of electrode strips in the orthogonal strip CdZnTe detector, wherein the number of initialized neurons is the same as the number of the electrode strips; a weight vector module configured to vectorize weights of the initialized neurons to obtain an initialized weight vector set; and a position determination module configured to determine a position of a gamma photon generated by a gamma ray interacting with the orthogonal strip CdZnTe detector by using a self-organizing mapping algorithm to iteratively reduce a range of the initialized weight vector set based on the input vector set and the initialized weight vector set, and obtain a target weight vector.

[0022] According to another aspect of the embodiments of the present application, there is also provided a non-transitory storage medium storing a plurality of instructions adapted to be loaded and executed by a processor to perform any of the gamma photon positioning methods of the orthogonal strip CdZnTe detector.

[0023] According to another aspect of the embodiments of the present application, there is also provided a computer program product comprising a computer program which, when executed by a processor, implements the steps of any of the gamma photon positioning methods of the orthogonal strip CdZnTe detector.

[0024] The technical solutions adopted by the present application can achieve at least one of the following beneficial effects:

[0025] In the embodiment of the present application, by acquiring current signals respectively generated by interactions of multiple gamma rays and the orthogonal strip CdZnTe detector; performing vectorization processing on the multiple current signals to obtain an input vector set; constructing an initialized neuron network based on positions of multiple electrode strips in the orthogonal strip CdZnTe detector, wherein the number of initialized neurons is the same as the number of the multiple electrode strips; performing weight vectorization processing on the multiple initialized neurons to construct an initialized weight vector set; based on the input vector set and the initialized weight vector set, by using a self-organizing mapping algorithm, iteratively reducing the range of the initialized weight vector set to obtain a target weight vector, and determining the position of the gamma photon generated by the interaction of the gamma ray and the orthogonal strip CdZnTe detector. The purpose of determining the position of the gamma photon based on the neural network and the self-organizing mapping algorithm is achieved, thereby realizing the technical effects of improving the accuracy of the position of the gamma photon and improving the efficiency of position acquisition, and further solving the technical problems that in the related art, the position of the gamma photon is calculated by using a mathematical model, the obtained position data has deviation, and the efficiency of position acquisition is low. BRIEF DESCRIPTION OF DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows, which form a part of the present application. The schematic embodiments of the present application and the description and explanation thereof do not constitute an improper limitation on the present application. In the drawings:

[0027] Figure 1 is a flowchart of a gamma photon positioning method of an orthogonal strip CdZnTe detector in Embodiment 1 of the present application;

[0028] Figure 2 is a neural network diagram of a gamma photon positioning method of an orthogonal strip CdZnTe detector in Embodiment 1 of the present application;

[0029] Figure 3 is a flowchart of an optional gamma photon positioning method of an orthogonal strip CdZnTe detector in Embodiment 2 of the present application;

[0030] Figure 4 is a current signal position spectrum diagram obtained by a gamma photon positioning method of an orthogonal strip CdZnTe detector in Embodiment 2 of the present application;

[0031] Figure 5 is a structural schematic diagram of a gamma photon positioning device of an orthogonal strip CdZnTe detector in Embodiment 3 of the present application. DETAILED DESCRIPTION

[0032] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described below in connection with the specific embodiments of the present application and corresponding drawings. In the description of the present application, it should be noted that the term "or" is generally used in the sense of including "and / or" unless the context clearly indicates otherwise.

[0033] In the description of the present application, it should be noted that unless otherwise explicitly specified and limited, the terms "mounting", "connection", "connecting" should be understood in a broad sense, for example, can be fixed connection, can also be detachable connection, or integrally connected; can be mechanical connection, can also be magnetic connection; can be directly connected, can also be indirectly connected through an intermediate medium, can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances. In addition, in the description of the present application, the terms "first", "second" and the like are only used to distinguish the description and cannot be understood as indicating or implying relative importance. In the description of the present application, the meaning of "multiple" is at least two, for example, two, three or more, etc., unless otherwise explicitly specified and limited.

[0034] Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0035] First, in order to facilitate the understanding of the embodiments of the present application, the following will explain and describe some terms or nouns involved in the present application:

[0036] The orthogonal strip-shaped cadmium zinc telluride detector adopts the design of vertically intersecting anode strips and cathode strips to form a grid-shaped electrode structure. When X-rays or gamma rays are incident, the detector collects charge signals on the orthogonal electrodes, combines with waveform analysis technology, and accurately calculates the three-dimensional coordinates (including depth information) of the ray action position.

[0037] To solve the problems in the prior art, the embodiments of the present application provide a kind of orthogonal strip-shaped cadmium zinc telluride detector gamma photon positioning method and device.

[0038] Embodiment 1

[0039] The embodiment provides a kind of orthogonal strip-shaped cadmium zinc telluride detector gamma photon positioning method, as shown in Figure 1 Figure 1 It is the flow chart of a kind of orthogonal strip-shaped cadmium zinc telluride detector gamma photon positioning method in the embodiment 1 of the present application, and the method comprises:

[0040] ​Step S102, obtaining current signals respectively generated by interactions of multiple gamma rays with the orthogonal strip-shaped CdZnTe detector;

[0041] Optionally, the interaction of the gamma ray with the orthogonal strip-shaped CdZnTe detector will generate electron-hole pairs (gamma photons) and directional drift under the action of the electric field, so that the anode and cathode (electrode strip) of the orthogonal strip-shaped CdZnTe detector generate current signals. By collecting these current signals, the data set x required for detecting the position of the gamma photon can be obtained.

[0042] Optionally, when the gamma ray interacts with the orthogonal strip-shaped CdZnTe detector, Compton scattering, photoelectric effect and other physical processes will occur, and one or more interaction positions may be generated in the orthogonal strip-shaped CdZnTe detector. For multiple interaction positions (i.e. multiple scattering events), different current signals will be obtained on the multiple anode and cathode strips of the orthogonal strip-shaped CdZnTe detector; and in a single interaction event, current signals will be generated only on the anode and cathode strips corresponding to the interaction position in the orthogonal strip-shaped CdZnTe detector, wherein the current signal size is related to the energy deposited by the ray. In a single interaction event, the collected current signals are used as data in the data set x, and when multiple interaction events are encountered, the collected multiple current signals are divided into data of multiple single interaction events, and the data set x is collected in the manner of single interaction event.

[0043] Optionally, the obtained data set x is preprocessed, wherein the preprocessing includes energy screening and data rejection exceeding the threshold, and a full-energy peak event data set x is obtained n , wherein n is the number of screened data in the data set. The full-energy peak event data contains energy, time, position and other information, and the core indicators include: peak position: corresponding to the energy of the incident particle (such as 137 Cs 0.662 MeV peak), peak area: reflecting the number of events, proportional to the activity of the radiation source, energy resolution: peak width / peak position (such as CZT detector, which can reach 0.39%), peak shape parameters: such as full width at half maximum, symmetry, used to evaluate the performance of the detector.

[0044] Step S104, vectorizing the multiple current signals to obtain an input vector set;

[0045] Optionally, the data in the preprocessed data set x of the full-energy peak event data set x n is still a current signal, but the current signal in the full-energy peak event data set x n is rejected after exceeding the threshold, and the current signal in the full-energy peak event data set x n is vectorized to obtain an input vector set X nIt should be noted that a vector includes direction and magnitude, so the input vector set X n The input vectors in include the magnitude of the current signal and the direction of current movement.

[0046] Step S106, constructing an initialized neural network based on the positions of the plurality of electrode strips in the orthogonal strip CdZnTe detector, wherein the number of the initialized neurons is the same as the number of the plurality of electrode strips;

[0047] Optionally, in order to obtain the location of the gamma photon, it is necessary to match the location of the current signal with the location of the electrode strip. The location of the current signal is unknown. Since the current signal is generated on the electrode strip, when the electrode strip corresponding to the current signal is obtained, the location of the current signal can be determined based on the location of the electrode strip. When the gamma ray acts on different depth positions of the orthogonal strip CdZnTe detector, the signals obtained by the orthogonal strip CdZnTe detector at the anode and cathode (electrode strip) are still different. Therefore, it is necessary to utilize the characteristics of the detection signal obtained at different action positions and realize the calculation of the ray interaction position by constructing a self-organizing map algorithm (SOM algorithm). Specifically, by setting the number of initialized neurons consistent with the number of electrode strips in the orthogonal strip CdZnTe detector space, the weight value of the initialized neuron is adjusted and topologically processed using the single interaction event data, so that the topological structure of the initialized neuron can directly output the probability distribution diagram of the interaction of gamma rays in the orthogonal strip CdZnTe detector space, that is, intuitively display the interaction position point of the gamma ray.

[0048] Optional, such as Figure 2 As shown, Figure 2 is a neural network diagram of a gamma photon positioning method for an orthogonal strip CdZnTe detector in Example 1 of the present invention, Figure 2 The neural network of the SOM algorithm position calculation on a plane is shown. In the figure, 16 initialization neurons are set. Each hollow circle represents an initialization neuron, and the input vector set X is used. n The SOM algorithm calculates the topology and weights of these neurons, resulting in a 4×4 array topology. By inputting the constructed SOM algorithm framework into the current signal data collected by different orthogonal strip CdZnTe detectors, a location distribution map of the interactions can be visually displayed. Furthermore, for multiple interaction events, multiple high-probability points can appear in the location distribution map.

[0049] Step S108, performing weight vectorization processing on the multiple initialized neurons to construct an initialized weight vector set;

[0050] Optionally, in the constructed SOM algorithm framework, it is necessary to perform weight vectorization processing on each initialized neuron to obtain an initialized weight vector set. Since the probability of each electrode strip generating a current signal is equal in the initial stage, the weight value of the initialized neuron corresponding to each electrode strip is equal. In order to calculate the interaction position between the orthogonal strip CdZnTe detector and the gamma ray, in the neuron topology structure constructed in the SOM algorithm, the initialized neuron is set to correspond to each electrode strip. In addition, the initialized weight vector set ω on each electrode strip is L Random selection is performed, L is the number of electrode strips, and the input vector set X n The number of times the current signal is generated on each electrode strip and the position of the electrode strip corresponding to the current signal will initialize the weight vector set ω corresponding to the initialization neuron L Iterative updates are performed, and the weight vector on each electrode strip changes. Ultimately, the neuron with the largest weight value is found, and the position of the gamma photon is determined based on the determined neuron.

[0051] Step S110, based on the input vector set and the initialization weight vector set, the range of the initialization weight vector set is iteratively narrowed by the self-organizing map algorithm to obtain the target weight vector, and determine the position of the gamma photon generated by the interaction between the gamma ray and the orthogonal strip CdZnTe detector.

[0052] Optionally, the above obtains the input vector set X corresponding to the current signal generated by the interaction between the gamma ray and the orthogonal strip CdZnTe detector n , and the initialization weight vector set ω corresponding to the electrode strip position on the orthogonal strip CdZnTe detector L , the input vector set X n and the initial weight vector set ω L Match and determine the input vector set X n The current signal and the initialization weight vector set ω L Closest to the weight vector, thus locking the initial position range of the current signal (that is, the gamma photon), based on the input vector set X n By performing multiple matches on multiple input vectors in the , the initialization weight vector set ω can be gradually reduced L The range of the position is gradually narrowed, and the input vector set X n After all input vectors in are iteratively matched, the weight vector set ω is initialized. L The range has been reduced to a small enough value. n and the initial weight vector set ω L When matching, a target weight vector of the last iteration is found, and the neuron is determined by the target weight vector, thereby determining the exact position of the gamma photon.

[0053] In some preferred embodiments, based on the input vector set and the initialized weight vector set, the range of the initialized weight vector set is iteratively reduced by a self-organizing mapping algorithm to obtain a target weight vector, and the position of a gamma photon generated by interaction of a gamma ray with the orthogonal strip CdZnTe detector is determined, comprising: constructing an initialization neighborhood range of the initialized weight vector set, wherein the neighborhood range is used to indicate the range of the region formed by the initialized weight vector set; randomly sampling a first input vector in the input vector set until all input vectors in the input vector set are sampled, iteratively updating the initialized weight vector set based on the first input vector to obtain a target weight vector set, wherein the first input vector is used to indicate the currently sampled input vector; obtaining the target weight vector based on the first neighborhood range and the target weight vector set; and determining the position of the gamma photon based on the target weight vector.

[0054] Optionally, the initialized weight vector set is a vector with direction and size, and the initialized weight vector set represents a position. When the initialized weight vector set is placed in a coordinate system, all the initialized weight vector sets will form an initialization neighborhood range. Before iteration, the initialization neighborhood range is the entire area of the orthogonal strip CdZnTe detector.

[0055] Optionally, the randomly sampled first input vector in the input vector set is matched with the initialized weight vector set, that is, the current signal is matched to the electrode strip position of the orthogonal strip CdZnTe detector. With the position as the center and the half circle set according to the experimental accuracy requirement as the radius, a new first neighborhood range is drawn, and a first neuron is updated in the first neighborhood range to obtain a new initialized weight vector. By inputting a second first input vector, the iteration is performed again on the iteratively updated first neighborhood range and the new initialized weight vector to update the new position. The iteration is performed until the last first input vector is iterated, so that the iteration is completed, and the final input position is the position of the gamma photon.

[0056] Optionally, by inputting the input vector set X n , the field function h (j,i(x)) will gradually decrease, the topology of the neuron will gradually stabilize, the weight vector of the neuron will gradually decrease, and finally the position of the orthogonal strip CdZnTe detector corresponding to the weight vector obtained by the last iteration is determined as the position of the gamma photon.

[0057] In some preferred embodiments, any first input vector in the input vector set is sampled until all input vectors in the input vector set are sampled, and the initialized weight vector set is iteratively updated based on the first input vector to obtain a target weight vector set, including: performing similarity matching between the first input vector and the initialized weight vector set to determine a first weight vector, wherein the first weight vector is used to indicate a weight vector in the initialized weight vector set having the minimum distance interval from the first input vector; updating the initialized neighborhood range based on the first input vector to obtain a first neighborhood range, wherein the first neighborhood range is used to indicate a region with the first input vector as the center; generating a plurality of first neurons in the first neighborhood range, wherein the first neurons are used to indicate position information of the orthogonal strip-shaped CdZnTe detector divided in the first neighborhood range; and performing weight vectorization processing on the plurality of first neurons to construct the target weight vector set.

[0058] Optionally, the similarity matching between the first input vector and the initialized weight vector set is performed, that is, the first input vector is compared with the plurality of weight vectors in the initialized weight vector set to determine a weight vector having the closest distance from the first input vector, that is, to determine the electrode strip position having the closest distance from the current signal, so that the weight vector having the closest distance from the first input vector is defined as the first weight vector.

[0059] Optionally, the first weight vector corresponds to a position point in the initialized neighborhood range, which is also a preliminary position of the gamma photon. However, since the first input vector is matched only once, the position of the gamma photon may be deviated, and therefore, the initialized neighborhood range needs to be reduced in range to determine a more accurate position of the gamma photon. Specifically, the initialized neighborhood range is circular with the first weight vector as the center, and a new first neighborhood range is determined with a half circle set as the radius according to the experimental accuracy requirement. It should be noted that the radius is artificially set according to the experimental accuracy requirement, and the first neurons are updated in the first neighborhood range to obtain a new weight vector set, that is, the target weight vector set.

[0060] Optionally, the completed SOM framework is used to utilize the full-energy peak event data set x n The weights of the initialized neurons are adjusted and optimized. In order to find the best matching between the input vector set X n and the initialized weight vector set ω L , the distance (generally the Euclidean distance) between each data point in the input vector set and the weight vector of each neuron in the electrode strip grid distribution of the orthogonal strip-shaped CdZnTe detector is calculated, wherein the neuron having the minimum distance represents the best matching unit (the first weight vector). The minimum interval distance between the first input vector and the initialized weight vector set is calculated according to the following formula:

[0061] i(x) = arg min ||X j - ω j ||

[0062] where i(x) is the minimum distance separation, X j is the jth first input vector, ω j is the jth first weight vector, and arg min is the minimum value of a function.

[0063] In some preferred embodiments, the initialization neighborhood range is updated based on the first input vector to obtain a first neighborhood range, comprising: determining the minimum distance separation based on the first input vector and the first weight vector; determining an effective width of the first neighborhood range, wherein the effective width is one half of a diameter of the first neighborhood range; and determining the first neighborhood range based on the effective width and the minimum distance separation.

[0064] In some preferred embodiments, the minimum distance separation is determined based on the first input vector and the first weight vector, comprising: the minimum distance separation is calculated as follows:

[0065] i(x) = arg min ||X j - ω j ||

[0066] where i(x) is the minimum distance separation, X j is the jth first input vector, ω j is the jth first weight vector, and arg min is the minimum value of a function.

[0067] In some preferred embodiments, the first neighborhood range is determined based on the effective width and the minimum distance separation, comprising: the first neighborhood range is calculated as follows:

[0068] h (j,i(x)) = exp(-||X j - ω j || 2 / (2σ 2 ))

[0069] where h (j,i(x)) is the first neighborhood range, X j is the jth first input vector, ω j is the jth first weight vector, and ||X j - ω j || is the distance separation, and σ is the effective width. It should be noted that the distance separation in the calculation process can be selected as the minimum distance separation.

[0070] Optionally, the neurons around the winning first neuron are taken as excited neurons, new neurons are generated around the excited neurons, and the range of the excited neurons is reconstituted as the first neighborhood range. When the distance between the first input vector and the weight vector set is calculated for multiple times, the reduced range of the neurons also has a smaller distance value, and in order to properly adjust the topology of the neurons, a new first neighborhood range h can be defined (j,i(x)) , the first neighborhood range h (j,i(x)) is symmetrical about the winning first neuron, and the amplitude value thereof monotonically decreases with the increase of the lateral distance.

[0071] In some preferred embodiments, based on the first neighborhood range, the first weight vector, and the target weight vector set, the target weight vector is obtained, including: the target weight vector is calculated as follows:

[0072] w j+1 =ω j +αh (j,i(x)) (X j -ω j )

[0073] wherein w j+1 is the target weight vector, ω j is the jth first weight vector, h (j,i(x)) is the first neighborhood range, α is a learning rate parameter, and X j is the jth first input vector.

[0074] Optionally, for the weight vector ω of each neuron, the weight adjustment optimization can be performed by the following formula:

[0075] w j+1 =ω j +αh (j,i(x)) (X j -ω j )

[0076] w j+1 is the target weight vector after weight vector adjustment, ω j is the first weight vector before weight vector adjustment, α is a learning rate parameter, and h (j,i(x)) is the first neighborhood range around the winning first neuron.

[0077] Through the above steps S102 to S110, the purpose of determining the position of the gamma photon based on the neural network and the self-organizing mapping algorithm is achieved, thereby realizing the technical effects of improving the accuracy of the position of the gamma photon and improving the efficiency of obtaining the position, and further solving the technical problems that the position data obtained by the mathematical model in the related art has deviation, and the efficiency of obtaining the position is low.

[0078] Embodiment 2

[0079] Based on the above embodiments and optional embodiments, the present application further proposes an optional implementation, Figure 3 is a flow chart of an optional orthogonal strip CdZnTe detector gamma photon positioning method in Embodiment 2 of the present application, as shown in the figure, the method comprises: Figure 3

[0080] In order to calculate the position of the gamma photon in the CdZnTe crystal (orthogonal strip CdZnTe detector), an initialization weight vector set (circled in the figure) is set on each electrode strip of the CdZnTe crystal, the probability of the interaction of the gamma ray on each electrode strip is calculated by multiplying the input vector and the initialization weight vector (i.e. the iteratively updated weight value), and the probability distribution (distribution of the weight value) corresponding to the electrode strip is obtained, that is, the position calculation in the CdZnTe crystal is realized. Figure 2

[0081] Step S1, initialization: randomly selecting an initialization weight vector set ω L from the input vector set, where L is the number of electrode strips, and the initialization weight vectors in ω L are different from each other.

[0082] Step S2, sampling: selecting a training sample from the input vector set with a probability of 70% to construct and train the self-organizing mapping algorithm, and selecting a verification sample from the input vector set with a probability of 30% to obtain the final self-organizing mapping algorithm; randomly selecting an input vector from the input vector set as a first input vector.

[0083] Step S3, similarity matching: using the minimum distance criterion to find the most matched (winning) electrode strip, and the first neuron and the first weight vector corresponding to the electrode strip, the minimum distance criterion is:

[0084] i(x)=argmin||X j -ω j ||

[0085] Where i(x) is the minimum distance interval, X j is the jth first input vector, and ω j is the jth first weight vector.

[0086] Step S4, pixel point weight vector updating: adjusting the first weight vector by the updating formula to obtain the final target weight vector,

[0087] w j+1 =ω j +αh (j,i(x)) (X j -ω j )​​

[0088] h (j,i(x)) =exp(-||X j -ω j || 2 / (2σ 2 ))

[0089] Among them, w j+1 is the target weight vector, ω j is the first weight vector of the jth node, h (j,i(x)) is the first neighborhood range, α is the learning rate parameter, X j is the first input vector of the jth order. (j,i(x)) is the first neighborhood range, X j is the first input vector of the jth, ω j is the first weight vector of the jth node, ||X j -ω j || is the minimum distance interval, and σ is the effective width.

[0090] For best results, h (j,i(x)) The two parameters α and α change dynamically during the calculation process.

[0091] Step S5, iteration: The loop continues with steps S2 to S4 for multiple iterations until the feature map observes that the first neighborhood range has been reduced to a diameter less than 0.01 mm, or all input vectors in the input vector set have completed the iteration, and the final target weight vector is obtained. Based on the target weight vector, the neuron is obtained by reverse deduction, and the position information is determined based on the neuron to obtain the precise position of the gamma photon.

[0092] According to the self-organizing map algorithm, the waveform acquisition experimental data of the orthogonal strip CdZnTe detector Na-22 radioactive source is trained to obtain the weight value of each position on the xy plane, and the cathode-anode signal two-dimensional position spectrum is drawn as follows Figure 4 shown. Figure 4 is a current signal position spectrum obtained by a gamma photon positioning method of an orthogonal strip CdZnTe detector in Example 2 of the present invention, Figure 4 As shown in the figure, the weight potential at the center of each electrode strip is much higher than that at nearby positions, and the overall weight distribution is relatively uniform, which is consistent with the signal position distribution calculated by the signal weight method formula.

[0093] In summary, the distribution diagrams of the cathode-anode signal two-dimensional position spectrum, the cathode-depth two-dimensional position spectrum, the anode-depth two-dimensional position spectrum, etc. (x, y, z) given by the Na-22 radioactive source test waveform collection of the 1.0 mm orthogonal bar-shaped CdZnTe detector are in good agreement with the simulation results of the detector electric field and the weight potential. The calculated value of the position signal weight method of the orthogonal bar-shaped CdZnTe detector is in good agreement with the position value predicted by the self-organizing mapping algorithm based on the experimental data of the CdZnTe detector, which verifies the correctness of the proposed CdZnTe detector position calculation method.

[0094] Through the above steps S1 to S5, the software technology can be used to promote the acquisition of the position information of the gamma rays in the position sensitive orthogonal bar-shaped CdZnTe detector, reduce the cost of improving the position resolution, improve the efficiency of the position acquisition, and apply the neural network algorithm to further refine the position information of the gamma rays in the orthogonal bar-shaped position sensitive CdZnTe.

[0095] Embodiment 3

[0096] According to the embodiment of the present application, a device for implementing the above-mentioned orthogonal bar-shaped CdZnTe detector gamma photon positioning is also provided, Figure 5 is a structural schematic diagram of an orthogonal bar-shaped CdZnTe detector gamma photon positioning device in Embodiment 3 of the present application, as Figure 5 shown, the positioning device comprises an acquisition module 301, an input vector module 302, a neural network module 303, a weight vector module 304, and a position determination module 305, wherein:

[0097] The acquisition module 301 is configured to acquire current signals respectively generated by the interaction of multiple gamma rays with the orthogonal bar-shaped CdZnTe detector.

[0098] The input vector module 302 is connected to the acquisition module 301 and is configured to perform vectorization processing on the multiple current signals to obtain an input vector set.

[0099] The neural network module 303 is connected to the input vector module 302 and is configured to construct an initialized neuron network based on the positions of the multiple electrode strips in the orthogonal bar-shaped CdZnTe detector, wherein the number of initialized neurons is the same as the number of the multiple electrode strips.

[0100] The weight vector module 304 is connected to the neural network module 303 and is configured to perform weight vectorization processing on the multiple initialized neurons to construct an initialized weight vector set.

[0101] The position determination module 305, connected to the weight vector module 304, is configured to determine the position of the gamma photon generated by the interaction of the gamma ray and the orthogonal bar-shaped CdZnTe detector based on the input vector set and the initialized weight vector set, and obtain a target weight vector by iteratively narrowing the range of the initialized weight vector set through a self-organizing mapping algorithm.

[0102] It should be noted that the above modules can be implemented by software or hardware. For the latter, the modules can be located in the same processor or in different processors in any combination.

[0103] It should be noted that the above acquisition module 301, input vector module 302, neural network module 303, weight vector module 304, and position determination module 305 correspond to steps S102-S110 in the embodiments, and the modules and the corresponding steps have the same examples and application scenarios, but are not limited to the above embodiments. It should be noted that the modules as part of the device can run in a computer terminal.

[0104] It should be noted that the optional or preferred embodiments of the present embodiment can refer to the related description in the embodiments, which will not be repeated here.

[0105] The above orthogonal bar-shaped CdZnTe detector gamma photon positioning device can further include a processor and a memory, and the above acquisition module 301, input vector module 302, neural network module 303, weight vector module 304, and position determination module 305 are stored in the memory as program modules, and the processor executes the above program modules stored in the memory to realize the corresponding functions.

[0106] The processor includes a core, and the core retrieves the corresponding program modules from the memory. The above core can be one or more. The memory can include a non-persistent memory in a computer readable medium, a random access memory (RAM), and / or a non-volatile memory such as a read-only memory (ROM) or a flash memory (flash RAM), and the memory includes at least one memory chip.

[0107] According to the embodiments of the present application, an embodiment of a non-volatile storage medium is also provided. Optionally, in the present embodiment, the above non-volatile storage medium includes a stored program, wherein the above program controls the device in which the above non-volatile storage medium is located to execute the above any one orthogonal bar-shaped CdZnTe detector gamma photon positioning method when the above program is running.

[0108] Optionally, in the embodiment, the non-volatile storage medium can be located in any one of the computer terminals in the computer terminal group in the computer network, or in any one of the mobile terminals in the mobile terminal group, and the non-volatile storage medium includes the stored program.

[0109] Optionally, during the program execution, the device in which the non-volatile storage medium is located performs the following functions: obtaining current signals respectively generated by multiple gamma rays interacting with the orthogonal strip CdZnTe detector; performing vectorization processing on the multiple current signals to obtain an input vector set; constructing an initialized neuron network based on the positions of the multiple electrode strips in the orthogonal strip CdZnTe detector, wherein the number of initialized neurons is the same as the number of the multiple electrode strips; performing weight vectorization processing on the multiple initialized neurons to construct an initialized weight vector set; based on the input vector set and the initialized weight vector set, iteratively narrowing the range of the initialized weight vector set through a self-organizing mapping algorithm to obtain a target weight vector, and determining the position of the gamma photon generated by the gamma ray interacting with the orthogonal strip CdZnTe detector.

[0110] According to the embodiments of the present application, an embodiment of a processor is also provided. Optionally, in the embodiment, the processor is used to run a program, wherein the program performs any one of the above-mentioned orthogonal strip CdZnTe detector gamma photon positioning methods when running.

[0111] According to the embodiments of the present application, an embodiment of a computer program product is also provided. Optionally, in the embodiment, the computer program product includes a computer program, and the computer program is executed by a processor to implement the program steps of any one of the above-mentioned orthogonal strip CdZnTe detector gamma photon positioning methods.

[0112] Optionally, the computer program product, when executed on a data processing device, is adapted to execute the program with the following method steps: obtaining current signals respectively generated by multiple gamma rays interacting with the orthogonal strip CdZnTe detector; performing vectorization processing on the multiple current signals to obtain an input vector set; constructing an initialized neuron network based on the positions of the multiple electrode strips in the orthogonal strip CdZnTe detector, wherein the number of initialized neurons is the same as the number of the multiple electrode strips; performing weight vectorization processing on the multiple initialized neurons to construct an initialized weight vector set; based on the input vector set and the initialized weight vector set, iteratively narrowing the range of the initialized weight vector set through a self-organizing mapping algorithm to obtain a target weight vector, and determining the position of the gamma photon generated by the gamma ray interacting with the orthogonal strip CdZnTe detector.

[0113] The electronic device provided by the embodiment of the present application comprises a processor, a memory and a program stored in the memory and executable on the processor, and the processor implements the following steps when executing the program: obtaining current signals respectively generated by interactions of multiple gamma rays and a quadratically bar-shaped CdZnTe detector; performing vectorization processing on the multiple current signals to obtain an input vector set; constructing an initialized neuron network based on positions of multiple electrode strips in the quadratically bar-shaped CdZnTe detector, wherein the number of initialized neurons is the same as the number of the multiple electrode strips; performing weight vectorization processing on the multiple initialized neurons to construct an initialized weight vector set; and based on the input vector set and the initialized weight vector set, iteratively reducing the range of the initialized weight vector set through a self-organizing mapping algorithm to obtain a target weight vector, and determining the position of a gamma photon generated by the interaction of the gamma ray and the quadratically bar-shaped CdZnTe detector.

[0114] The sequence of the above-mentioned embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments.

[0115] In the above-mentioned embodiments of the present application, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0116] In the several embodiments provided by the present application, it should be understood that the disclosed technology can be implemented in other ways. Of course, the device embodiment described above is only illustrative, and for example, the division of the above modules can be a logical function division, and actual implementation can have another division manner, for example, a plurality of modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the shown or discussed mutual ones can be through some interface, indirect coupling or communication connection between modules or modules, which can be electrical or other forms.

[0117] The modules described as separate components can or can not be physically separated, and the components shown as modules can or can not be physical modules, that is, they can be located in one place or distributed to multiple modules. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0118] In addition, each functional module in each embodiment of the present application can be integrated in one processing module, or each module can exist physically, or two or more modules can be integrated in one module. The above integrated module can be realized in the form of hardware or in the form of software functional module.

[0119] If the above-mentioned integrated modules are realized in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable nonvolatile storage medium. Based on this understanding, the technical solutions of the present application, essentially or the part that contributes to the prior art, or all or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a non-volatile storage medium, including a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the embodiments of the present application. The aforementioned non-volatile storage medium includes: a U disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0120] The above is only the preferred embodiment of the present application, and it should be pointed out that for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, and these improvements and refinements should be considered as the protection scope of the present application.

Claims

1. A method for positioning gamma photons of an orthogonal strip-shaped cadmium zinc telluride detector, characterized in that: include: Acquire current signals generated by the interaction of multiple gamma rays with orthogonal strip-shaped CdZnTe detectors; performing vectorization processing on the plurality of current signals to obtain an input vector set; constructing an initialized neural network based on positions of a plurality of electrode strips in the orthogonal strip-shaped CdZnTe detector, wherein the number of initialized neurons is the same as the number of the plurality of electrode strips; Performing weight vectorization processing on the plurality of initialized neurons to construct an initialized weight vector set; Based on the input vector set and the initialization weight vector set, a self-organizing map algorithm is used to iteratively narrow the range of the initialization weight vector set to obtain a target weight vector, and determine the position of a gamma photon generated by the interaction between gamma rays and an orthogonal strip cadmium zinc telluride detector.

2. The method for positioning gamma photons of an orthogonal strip-shaped CdZnTe detector according to claim 1, characterized in that: The method of iteratively narrowing the range of the initialization weight vector set based on the input vector set and the initialization weight vector set by a self-organizing map algorithm to obtain a target weight vector and determining the position of a gamma photon generated by the interaction between the gamma ray and the orthogonal strip cadmium zinc telluride detector comprises: Constructing an initialization neighborhood range of the initialization weight vector set, wherein the neighborhood range is used to indicate a region range constituted by the initialization weight vector set; Randomly sampling a first input vector from the input vector set until all input vectors in the input vector set are sampled, and iteratively updating the initialized weight vector set based on the first input vector to obtain a target weight vector set, wherein the first input vector is used to indicate a currently sampled input vector; Obtaining a target weight vector based on the first neighborhood range and the target weight vector set; Based on the target weight vector, a position of the gamma photon is determined.

3. The method for positioning gamma photons of an orthogonal strip-shaped CdZnTe detector according to claim 2, characterized in that: The method of arbitrarily sampling a first input vector from the input vector set until all input vectors in the input vector set are sampled, and iteratively updating the initialized weight vector set based on the first input vector to obtain a target weight vector set includes: Performing similarity matching between the first input vector and the initialization weight vector set to determine a first weight vector, wherein the first weight vector is used to indicate a weight vector in the initialization weight vector set that has the smallest distance interval with the first input vector; Based on the first input vector, the initialized neighborhood range is updated to obtain a first neighborhood range, wherein the first neighborhood range is used to indicate an area with the first input vector as the center; generating, within the first neighborhood, first neurons having the same number as the plurality of electrode strips, wherein the first neurons are used to indicate position information of the orthogonal strip-shaped cadmium zinc telluride detectors divided within the first neighborhood; Perform weight vectorization processing on the plurality of first neurons to construct a target weight vector set.

4. The method for positioning gamma photons of an orthogonal strip-shaped CdZnTe detector according to claim 3, characterized in that: The updating of the initialized neighborhood range based on the first input vector to obtain a first neighborhood range includes: Determining a minimum distance interval based on the first input vector and the first weight vector; Determine an effective width of the first neighborhood range, wherein the effective width is half of a diameter of the first neighborhood range; A first neighborhood range is determined based on the effective width and the minimum distance interval.

5. The method for positioning gamma photons of an orthogonal strip-shaped CdZnTe detector according to claim 4, characterized in that: The determining of the minimum distance interval based on the first input vector and the first weight vector includes: The minimum distance interval is calculated as follows: i(x)=arg min||X j -oh j || Among them, i(x) is the minimum distance interval, X j is the first input vector of the jth, ω j is the first weight vector of the jth item, and argmin is the minimum value of the function.

6. The method for positioning gamma photons of an orthogonal strip-shaped CdZnTe detector according to claim 5, characterized in that: The determining of a first neighborhood range based on the effective width and the minimum distance interval includes: The first neighborhood range is calculated as follows: h (j,i(x)) =exp(-||X j -oh j || 2 / (2σ 2 )) Among them, h (j,i(x)) is the first neighborhood range, X j is the first input vector of the jth, ω j is the first weight vector of the jth node, ||X j -ω j || is the distance interval, and σ is the effective width.

7. The method for positioning gamma photons of an orthogonal strip-shaped CdZnTe detector according to claim 6, characterized in that: The obtaining of a target weight vector based on the first neighborhood range and the target weight vector set includes: The target weight vector is calculated as follows: w j+1 =ω j +ah (j,i(x)) (X j -oh j ) Among them, w j+1 is the target weight vector, ω j is the first weight vector of the jth node, h (j,i(x)) is the first neighborhood range, α is the learning rate parameter, X j is the first input vector of the jth.

8. An orthogonal strip-shaped cadmium zinc telluride detector gamma photon positioning device, characterized in that: include: An acquisition module, used for acquiring current signals respectively generated by the interaction of multiple gamma rays with orthogonal strip-shaped CdZnTe detectors; An input vector module, configured to perform vector processing on the plurality of current signals to obtain an input vector set; A neural network module, configured to construct an initialized neural network based on positions of a plurality of electrode strips within the orthogonal strip-shaped CdZnTe detector, wherein the number of initialized neurons is the same as the number of the plurality of electrode strips; A weight vector module, configured to perform weight vectorization processing on the plurality of initialized neurons to construct an initialized weight vector set; A position determination module is used to iteratively narrow the range of the initialization weight vector set based on the input vector set and the initialization weight vector set through a self-organizing map algorithm to obtain a target weight vector and determine the position of the gamma photon generated by the interaction of the gamma ray with the orthogonal strip cadmium zinc telluride detector.

9. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executed by a gamma photon positioning method for an orthogonal strip cadmium zinc telluride detector according to any one of claims 1 to 7.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the gamma photon positioning method of an orthogonal strip cadmium zinc telluride detector described in any one of claims 1 to 7 are implemented.

Images