A data processing method for gamma spectrometer spectrum stabilization

Abstract

The application discloses a kind of gamma spectrometer stable data processing method of spectrum, comprising: S1 determines reference peak and records the initial channel address of reference peak when not offset occurs;S2 when reference peak occurs offset, find the new channel address of reference peak;S3 according to the numerical relationship of single or multiple reference peak before and after offset channel address, linear relationship between initial channel address and offset channel address is established;S4 after offset, channel address segmentation processing is carried out to count spectrum, and the numerical relationship of each channel and its adjacent channel count is used to adjust the corresponding subchannel count value;S5 after adjustment, the count spectrum using the linear relationship of subchannel unit is revised;S6 according to the numerical value of each subchannel before and after revision, the count value of the subchannel is proportionally distributed to corresponding integer channel.The details of spectrum are supplemented, the error caused by spectrum distortion in processing process is reduced, and more effective data support can be provided for later data processing.

Description

Technical Field

[0001] This invention belongs to the field of gamma radiation detection technology, and more specifically, relates to a data processing method for stabilizing the spectrum of a gamma spectrometer. Background Technology

[0002] During gamma-ray detection, the scintillation crystal and photomultiplier tube, which are the main components of the detector, are susceptible to output drift due to external environmental influences. This results in a shift and distortion of the energy spectrum shape, further affecting the interpretation and analysis of the measurement results. Therefore, certain measures are needed to maintain the stability of the energy spectrum shape obtained from the measurements.

[0003] Commonly used spectral stabilization methods include feedback stabilization by adjusting the system gain and software stabilization by using algorithms to correct the spectral shape through characteristic peaks.

[0004] (1) Feedback-based spectrum stabilization: When the spectrum shifts to the right (left), it indicates that the input ADC signal voltage is too high (low). In this case, the signal voltage can be restored to normal by decreasing (increasing) the system gain, thereby restoring the spectrum to normal. A common gain adjustment method is to adjust the gain by adjusting the multiplication factor of the photomultiplier tube. However, after adjusting the gain of the photomultiplier tube, its gain is not linearly applicable to all energy values. This will cause a relative shift between some spectral peaks, which cannot be achieved by simply adjusting the gain. In fact, other gain adjustment methods also have this problem to some extent. Moreover, feedback adjustment increases the complexity of the system and increases the failure rate.

[0005] (2) Software-based spectrum stabilization: The software adjustment scheme does not require any hardware modifications. This scheme typically first calibrates the energy of the channel address, converting discrete channels into continuous energy channels. By tracking one or several reference peaks, the linear relationship before and after the energy channel shift is obtained. The reference peaks can be selected from characteristic peaks of natural elements (passive) or generated using artificial light sources (active). Then, the energy channels are recalibrated through the linear relationship to obtain a new energy channel-count spectrum, thereby achieving the correction of the spectrum. However, since the channel address-count spectrum is essentially a histogram, each count is the superposition of all energy counts in an energy range. The spectral lines of the energy-count spectrum are an approximate fit of the channel address-count spectrum. The recalibration of the energy channels is actually equivalent to the left and right translation, stretching, and compression of the energy channel counts. The accuracy is still limited by the number of bits of the ADC. At the same time, the process will also cause distortion of the spectrum, resulting in certain errors. Summary of the Invention

[0006] To address the shortcomings of existing technologies, the present invention aims to provide a data processing method for stabilizing the spectrum of a gamma spectrometer. This method addresses the problem that existing software-based stabilization methods, when processing data, directly fit the bar chart representing each channel count into a curve based on the count value-energy relationship. This only ensures the accuracy of the count point values, while the values ​​near the count point will be distorted due to discrepancies between the fitted curve and the actual situation, resulting in errors.

[0007] This invention provides a data processing method for stabilizing the spectrum of a gamma spectrometer, comprising the following steps:

[0008] S1 determines the reference peak and records the initial address of the reference peak when no offset occurs;

[0009] S2 When the reference peak shifts, find the new address of the reference peak;

[0010] S3 establishes a linear relationship between the initial address and the offset address based on the numerical relationship between the addresses before and after the offset of one or more reference peaks;

[0011] S4 performs channel address segmentation on the offset counting spectrum and adjusts the corresponding sub-channel count values ​​based on the numerical relationship between each channel and its adjacent channels.

[0012] S5 corrects the adjusted count spectrum in sub-channels using the aforementioned linear relationship;

[0013] S6 allocates the count value of each sub-channel to the corresponding integer channels proportionally based on the corrected values ​​of the front and rear boundaries of each sub-channel.

[0014] In step S1, a characteristic peak of an artificial peak or a natural element is selected as a reference peak.

[0015] Furthermore, in step S3, the linear relationship between the initial channel address and the offset channel address is n = kn′ + b, where k = (n2 - n1) / (n2′ - n1′), b = (n2n1′ - n1n2′) / (n2′ - n1′), n′ is the channel address of any channel after offset, n is the new channel address corresponding to the corrected channel address, n1 and n2 are the initial channel addresses of the dual reference peaks, and n1′ and n2′ are the channel addresses of the dual reference peaks after offset.

[0016] Furthermore, in step S4, after the address is divided, the counts of the sub-channels in the first and second halves of each channel are redistributed according to the counts of the adjacent channels before the division, based on the principle of an arithmetic sequence.

[0017] Furthermore, the adjusted count for the xth path is as follows:

[0018]

[0019] Among them, A n For the count corresponding to channel n, M is generally an even number. After channel n is divided, the sub-channels range from (n-1)M+1 to nM channels, and the count for each channel is A. n / M, for the first half of the (n-1)M+1 to (n-1)M+M / 2 channels, the adjustment range corresponding to each sub-channel is equal to the count A of the next channel. n+1 and the previous count A n-1 Regarding the latter half of the (n-1)M+M / 2+1 to nM channels, the top adjustment of each sub-channel is related to A. n and the subsequent count A n+1 related.

[0020] Furthermore, in step S5, after the channel address is corrected, the count in each new channel is allocated to the corresponding original channel address according to the overlap between the new channel address and the original channel address. Specifically, this includes:

[0021] S51 corrects the adjusted channel address based on the linear relationship of the offset. The front and rear boundaries of each sub-channel will be corrected to the initial position, and the count of the channel will be allocated to the corresponding integer boundary channel. In this way, the spectral line can be readjusted to the correct position.

[0022] S52 The correction of the count address for each track means that the original track address is stretched, compressed and translated. The front and back boundaries of each track may change from integers to decimals. After processing, the track address will inevitably span two or more original track addresses. At this time, the count of the track cannot be directly put into a single track. The count of each track needs to be distributed to the corresponding original track address according to the overlap between the new track address and the original track address.

[0023] Furthermore, for any processed channel address, its count needs to be allocated to the corresponding channel. The specific processing method is as follows:

[0024] (1) For the nth question, its count is A. n Its left and right boundary coordinates are n-1 and n;

[0025] (2) Using linear relationship x n =kn+b Correcting the boundary of this path yields a new boundary x. n-1 x n , for x n-1 Rounding up gives y1, and with respect to x n Rounding down gives y2, which means y1 = [x n-1 ]+1、y2=[x n ];

[0026] (3) Based on the x coordinate axis of the road address n-1 ~y1 size, A nTake the corresponding proportion ΔA and put it into y1-1~y1, and do the same for the rest;

[0027] (4) The count of the corresponding track in the new track address increases by ΔA accordingly, thus forming a new count spectrum.

[0028] In step (4), if y1 = y2, then only A needs to be changed. n Divide into two groups proportionally; if y2-y1>0, then the distribution needs to be based on the number of integers between y1 and y2, as follows:

[0029]

[0030] This invention utilizes a gamma spectrometer stabilization method and proposes a method for directly correcting the count channel addresses of the energy spectrum. During the correction process, by dividing the count channel addresses and adjusting the corresponding counts, the corrected channel address data is made to closely match the actual measured values, resulting in smoother spectral lines and easier data processing.

[0031] Compared with existing technologies, the software-stabilized spectral processing method provided by this invention, through the segmentation of channel addresses in the counting spectrum and the redistribution of sub-channel counts, simulates sampling results obtained by an ADC with higher precision using an algorithm. Compared with traditional channel recalibration methods, this invention supplements the details of the spectral shape, reduces errors caused by spectral distortion during processing, and provides more effective data support for subsequent data processing. Attached Figure Description

[0032] Figure 1 This is a flowchart illustrating the specific implementation of the data processing method for stabilizing the gamma spectrometer provided in this embodiment of the invention.

[0033] Figure 2 This is a schematic diagram of the channel address segmentation and sub-channel count adjustment process provided in an embodiment of the present invention;

[0034] Figure 3 This is a schematic diagram of a counting allocation example after channel address adjustment provided in an embodiment of the present invention; the diagram shows that the sensor gain is reduced, resulting in the stretching of the obtained spectrum relative to the actual situation. At this time, it is necessary to compress the counting spectrum. (a) is the counting spectrum obtained after the sensor gain is reduced, and (b) is the counting spectrum after channel address compression.

[0035] Figure 4 This is a specific processing method for the count allocation after channel address adjustment provided in an embodiment of the present invention; wherein (a) is a count bar chart representing the nth channel before adjustment, and its count is A. n Its left and right boundary coordinates are n-1 and n, (b) is the histogram after linear adjustment, y1 and y2 are the adjusted left and right boundary xn-1 x n The new boundary after rounding, (c) is a schematic diagram that provides a detailed explanation of the histogram count allocation. Taking the left boundary as an example, the adjusted boundary x n-1 x n Divide into integers and assign A n The data is allocated proportionally to the corresponding channels. (d) is a schematic diagram of the count allocation after the final processing is completed. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0037] This invention provides a novel software-based spectrum stabilization method. This new method abandons energy calibration of channel addresses and directly performs linear correction of the channel addresses after obtaining the linear relationship before and after the shift through characteristic peaks. During the channel address correction process, measures such as channel segmentation and count redistribution are adopted to give the processed spectrum a higher number of channels, exhibiting the effect achievable only by higher-precision ADCs. This further reduces the errors generated during channel address correction, making the corrected spectrum as close as possible to the actual situation and facilitating subsequent data processing.

[0038] Figure 1 The specific implementation flow of the data processing method for stabilizing the gamma spectrometer provided in the embodiments of the present invention is illustrated. For ease of explanation, only the parts related to the embodiments of the present invention are shown, and are described in detail below:

[0039] The data processing method for stabilizing the gamma spectrometer provided in this embodiment of the invention includes the following steps:

[0040] Step S1: Select a characteristic peak of an artificial peak or a natural element as a reference peak, and confirm and record the initial address of the reference peak without any shift.

[0041] Step S2: After the offset occurs, find the new address of the reference peak.

[0042] Step S3: Establish a linear relationship between the offset address and the initial address based on the numerical relationship between the addresses before and after the offset of one or more reference peaks.

[0043] Step S4: The offset counting spectrum is processed by channel address segmentation, and the corresponding sub-channel count values ​​are adjusted according to the numerical relationship between the count values ​​of each channel and its adjacent channels.

[0044] Step S5: Correct the adjusted counting spectrum in terms of sub-channels using the linear relationship obtained in step 3.

[0045] Step S6: Based on the corrected values ​​of the front and rear boundaries of each sub-channel, proportionally allocate the count value of the sub-channel to the corresponding integer channels.

[0046] At this point, the spectral data correction is complete. Compared to the unprocessed spectral shape, the shifted spectral lines have been corrected to their initial positions based on the degree of shift of the reference peak. Furthermore, the spectral lines are smoother than the number of channels, providing more effective data support for subsequent data processing.

[0047] In this invention, when the gamma spectrometer experiences spectral distortion due to changes in the external environment, a software algorithm corrects the spectral shape. Simultaneously, during processing, the counts of each channel are segmented and redistributed, resulting in a processed data spectrum containing more count channels and forming smoother spectral lines, facilitating further data processing and analysis. Compared to traditional channel recalibration methods, this method supplements the details of the spectral shape, reduces errors caused by spectral distortion during processing, and provides more effective data support for subsequent data processing.

[0048] In this embodiment of the invention, the reference peak can be a characteristic peak of a natural element or an artificial peak generated by a flashing LED. Taking a 256-channel γ spectrum as an example, the count of the reference peak needs to be maximized in the range of 20 to 30 channels to eliminate interference from other peaks. If there are multiple reference peaks, the interval between the reference peaks must be at least 40 channels to ensure that the linear relationship is obtained as accurately as possible.

[0049] In this embodiment of the invention, the position of the reference peak in its initial state and its position after the shift are obtained to obtain the linear relationship between the shifted channel address and the original channel address. In a 256-channel γ spectrum, the shift of the reference peak is generally only 20 to 30 channels. Taking a dual-reference peak as an example, the instrument analyzes the obtained count spectrum and finds the maximum value in the range of 20 to 30 channels to obtain the channel addresses n1′ and n2′ of the reference peak after the shift. Then, combined with the initial channel addresses n1 and n2 of the reference peak, the linear relationship from the initial channel address to the shifted channel address can be obtained as n = kn′ + b, where: k = (n2 - n1) / (n2′ - n1′), b = (n2n1′ - n1n2′) / (n2′ - n1′), n′ is the channel address of any channel after the shift, and n is the new channel address corresponding to the corrected channel address.

[0050] In this embodiment of the invention, the ideal energy spectrum line should be a smooth continuous curve. However, due to the limitations of ADC accuracy, the actual sampled count spectrum is a discrete histogram. After the channel address is divided, the counts of the first and second halves of each channel can be redistributed according to the counts of the adjacent channels before the division, following the principle of an arithmetic sequence, so that the counts are as close as possible to the actual situation.

[0051] like Figure 2 As shown, the counter corresponding to the nth channel is A. n For ease of processing, M is generally taken as an even number. After the nth channel is divided, the range of sub-channels is (n-1)M+1 to nM channels, and each channel is counted as A. n / M, for the first half of the (n-1)M+1 to (n-1)M+M / 2 channels, the adjustment range corresponding to each sub-channel is equal to the count A of the next channel. n+1 and the previous count A n-1 Regarding the latter half of the (n-1)M+M / 2+1 to nM channels, the top adjustment of each sub-channel is related to A. n and the subsequent count A n+1 Regarding the xth track, the adjusted count is as follows:

[0052]

[0053] In this embodiment of the invention, since the process of track address correction is equivalent to stretching, compressing and translating the original track address, the track address after processing will inevitably not coincide with the original track address. There will be a situation where the new track address spans two or more original track addresses. The count in each new track needs to be allocated to the corresponding original track address according to the overlap between the new track address and the original track address.

[0054] like Figure 3 As shown, after the drift, the counts of channels 1 to 4 in the counting spectrum are A1, A2, A3, and A4, respectively. After adjusting the channel addresses, the range of channels 1 to 4 becomes 0 to 0.75, 0.75 to 1.5, 1.5 to 2.25, and 2.25 to 3. The counts of the four channels need to be redistributed to channels 1 to 3, and the counts of each channel are A1+A2 / 3, (A2+A3) / 3, and A3 / 3+A4, respectively.

[0055] For any processed track address, its count needs to be assigned to the corresponding track. The following section combines... Figure 4 The specific processing method is explained in detail below:

[0056] 1) For the nth question, its count is A. n Its left and right boundary coordinates are n-1 and n;

[0057] 2) Using a linear relationship, x n =kn+b Correcting the boundary of this path yields a new boundary x. n-1 x n , for x n-1 Rounding up gives y1, and with respect to x n Rounding down gives y2, which means y1 = [x n-1 ]+1、y2=[x n ];

[0058] 3) On the track address coordinate axis, according to xn-1 ~y1 size, A n Take the corresponding proportion ΔA and put it into y1-1~y1, and do the same for the rest;

[0059] 4) The count of the corresponding track in the new track address increases by ΔA accordingly, thus forming a new count spectrum. If y1 = y2, then only A needs to be increased. n Divide the material into two sections proportionally. If y2 - y1 > 0, then the distribution should be based on the number of integers between y1 and y2, as follows:

[0060]

[0061] The method described in this invention uses an algorithm to divide each channel of the original spectrum count into several sub-channels, and then performs channel address adjustment and count correction on the sub-channels to make the final spectrum data as close as possible to the actual situation.

[0062] In step S4 of this embodiment of the invention, the initial counting spectrum is divided into channels, and the channel count values ​​are evenly distributed into sub-channels. Then, the count values ​​of each sub-channel after division are adjusted according to the count values ​​of adjacent channels.

[0063] (1) The channel is divided and a higher-order ADC is simulated using software. Each sub-channel is divided equally among the channel counts. When the channel address is corrected later, the denser sub-channels are beneficial to maintaining the details of the spectrum.

[0064] (2) The counts of the sub-channels after the division are redistributed based on the counts of the preceding and following channels. The count spectrum is fitted to increase the detail of the spectrum and make the spectrum smoother.

[0065] In step S5 of this embodiment of the invention, after the channel address is corrected, the count in each new channel is allocated to the corresponding original channel address according to the overlap between the new channel address and the original channel address:

[0066] (1) The adjusted channel address is corrected according to the linear relationship of the offset. The front and back boundaries of each sub-channel will be corrected to the initial position, and the count of the channel will be allocated to the corresponding integer boundary channel. In this way, the spectral line can be readjusted to the correct position.

[0067] (2) The correction of the count address of each channel means that the original channel address is stretched, compressed and translated. The front and back boundaries of each channel may change from integers to decimals. After the processing, the channel address will inevitably span two or more original channel addresses. At this time, the count of the channel cannot be directly put into a single channel. The count of each channel needs to be distributed to the corresponding original channel address according to the overlap between the new channel address and the original channel address.

[0068] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A data processing method for stabilizing the spectrum of a gamma spectrometer, characterized in that, Includes the following steps: S1 determines the reference peak and records the initial address of the reference peak when no offset occurs; S2 When the reference peak shifts, find the new address of the reference peak; S3 establishes a linear relationship between the initial address and the offset address based on the numerical relationship between the addresses before and after the offset of one or more reference peaks; S4 performs channel address segmentation on the offset counting spectrum and adjusts the corresponding sub-channel count values ​​based on the numerical relationship between each channel and its adjacent channels. S5 corrects the adjusted count spectrum in sub-channels using the aforementioned linear relationship; S6 allocates the count value of each sub-channel to the corresponding integer channel proportionally based on the corrected values ​​of the front and rear boundaries of each sub-channel. In step S4, after the address is divided, the sub-channels of each sub-channel are redistributed according to the counts of adjacent channels before the division, based on the principle of an arithmetic sequence. The adjusted count for the xth path is as follows: Among them, A n For the count corresponding to channel n, M is generally an even number. After channel n is divided, the sub-channels range from (n-1)M+1 to nM channels, and the count for each channel is A. n / M, for the first half of the (n-1)M+1 to (n-1)M+M / 2 channels, the adjustment range corresponding to each sub-channel is equal to the count A of the next channel. n+1 and the previous count A n-1 Regarding the latter half of the (n-1)M+M / 2+1 to nM channels, the top adjustment of each sub-channel is related to A. n and the subsequent count A n+1 related; In step S5, after the channel address is corrected, the count in each new channel is allocated to the corresponding original channel address according to the overlap between the new channel address and the original channel address. Specifically, this includes: S51 corrects the adjusted channel address based on the linear relationship of the offset. The front and rear boundaries of each sub-channel will be corrected to the initial position, and the count of the channel will be allocated to the corresponding integer boundary channel. In this way, the spectral line can be readjusted to the correct position. S52 The correction of the count address for each track means that the original track address is stretched, compressed and translated. The front and back boundaries of each track may change from integers to decimals. After processing, the track address will inevitably span two or more original track addresses. At this time, the count of the track cannot be directly put into a single track. The count of each track needs to be distributed to the corresponding original track address according to the overlap between the new track address and the original track address. For any processed track address, its count needs to be assigned to the corresponding track. The specific processing method is as follows: (1) For the nth question, its count is A. n Its left and right boundary coordinates are n-1 and n; (2) Using linear relationship x n =kn+b Correcting the boundary of this path yields a new boundary x. n-1 x n , for x n-1 Rounding up gives y1, and with respect to x n Rounding down gives y2, which means y1 = [x n-1 ]+1、y2=[x n ]; (3) Based on the x coordinate axis of the road address n-1 ~y1 size, A n Take the corresponding proportion ΔA and put it into y1-1~y1, and do the same for the rest; (4) The count of the corresponding track in the new track address increases by ΔA accordingly, thus forming a new count spectrum.

2. The data processing method as described in claim 1, characterized in that, In step S1, select a characteristic peak of an artificial peak or a natural element as a reference peak.

3. The data processing method as described in claim 1, characterized in that, In step S3, the linear relationship between the initial address and the offset address is n = kn′ + b. Where, k = (n2-n1) / (n2′-n1′), b = (n2n1′-n1n2′) / (n2′-n1′), n′ is the address of any path after offset, n is the new address corresponding to the corrected address, n1 and n2 are the initial addresses of the dual reference peaks, and n1′ and n2′ are the addresses of the dual reference peaks after offset.

4. The data processing method as described in claim 1, characterized in that, In step (4), if y1 = y2, then only A needs to be changed. n Divide it into two sections proportionally; If y2-y1>0, then the allocation needs to be based on the number of integers between y1 and y2, as follows:

Images