Noise detection method, noise detection device, and program

Abstract

This invention provides a method for detecting noise contained in measurement results from a radiation measuring device while reducing computational complexity without causing overflow or underflow. [Solution] The noise detection method is a noise detection method performed by a computer, and includes the steps of: acquiring the count value of radiation measured per unit time over a predetermined period; calculating a value obtained by dividing the Poisson probability of one of the count values ​​measured over the predetermined period by the Poisson probability of the average value of the count values ​​over the predetermined period; and determining that the one count value is noise if the calculated value is less than or equal to a predetermined value.

Description

[Technical Field]

[0001] This disclosure relates to a noise detection method, a noise detection device, and a program. [Background technology]

[0002] A method for inspecting surface contamination of an object is known, which involves wiping the surface of the object with smear filter paper and measuring the amount of radioactive material adhering to the smear filter paper. Regarding the measurement of radioactive materials, Patent Document 1 discloses a radiation measuring device comprising a radiation detection unit that detects radiation and outputs pulses, a radiation counting unit that counts the pulses and outputs the total number of pulses counted within the measurement time, and a calculation unit that calculates the radiation dose based on the total number. When inspecting smear filter paper with this radiation measuring device, if no radioactive material is adhering to the smear filter paper, no pulses are output, and if radioactive material is adhering, pulses are output, and the value is approximately constant. In addition, noise due to electromagnetic influences from surrounding equipment may be output as pulses regardless of the contamination state of the smear filter paper. Generally, it is thought that the count value (count number) of radiation emitted from radioactive material follows a Poisson distribution. Then, for example, count values ​​where the cumulative value of the probability function of the Poisson distribution represented by the following equation (1) (referred to as the Poisson probability) exceeds a predetermined value (e.g., 95%) are judged as noise, and the degree of contamination is evaluated by correcting the count values ​​judged as noise.

[0003]

number

[0004] In equation (1), x represents the radiation count (cps), and Av represents the average value of the counts measured over a given time. The Poisson probability P(x,Av) is the probability that the count measured over a given time is x, given that the average count over a given time is Av.

[0005] However, when attempting to detect noise using this method, if the average value Av or the count value x becomes large, overflow or underflow may occur during calculation, potentially leading to calculation failure. Furthermore, when the average value Av is large, the process of sequentially calculating probabilities using equation (1) and accumulating them to obtain a count value x that reaches a predetermined value is computationally intensive and puts a strain on computing resources such as the CPU. [Prior art documents] [Patent Documents]

[0006] [Patent Document 1] Japanese Patent Publication No. 2013-174462 [Overview of the Initiative] [Problems that the invention aims to solve]

[0007] There is a need for a method to detect noise contained in the measurement results of radiation measuring devices while reducing the computational load without causing overflow or underflow.

[0008] This disclosure provides a noise detection method, a noise detection device, and a program that can solve the above-mentioned problems. [Means for solving the problem]

[0009] The noise detection method of this disclosure is a noise detection method performed by a computer, comprising the steps of: acquiring count values ​​of radiation per unit time measured over a predetermined period; calculating a value obtained by dividing the Poisson probability of one of the count values ​​measured over the predetermined period by the Poisson probability of the average value of the count values ​​over the predetermined period; and determining that the one count value is noise if the calculated value is less than or equal to a predetermined value.

[0010] The noise detection device of the present disclosure includes means for acquiring the count value of radiation per unit time measured during a predetermined period, means for calculating a value obtained by dividing the Poisson probability of one of the count values measured during the predetermined period by the Poisson probability of the average value of the count values during the predetermined period, and means for determining that the one count value is noise if the calculated value is less than or equal to a predetermined value.

[0011] Further, the program of the present disclosure causes a computer to execute steps of acquiring the count value of radiation per unit time measured during a predetermined period, calculating a value obtained by dividing the Poisson probability of one of the count values measured during the predetermined period by the Poisson probability of the average value of the count values during the predetermined period, and determining that the one count value is noise if the calculated value is less than or equal to a predetermined value.

Advantages of the Invention

[0012] According to the noise detection method, noise detection device, and program of the present disclosure, it is possible to detect noise included in the measurement result by a radiation measurement device while reducing the calculation amount without causing overflow or underflow.

Brief Description of the Drawings

[0013] [Figure 1] It is a diagram showing an example of an inspection system according to an embodiment. [Figure 2] It is a diagram showing an example of time-series count values according to an embodiment. [Figure 3] It is a diagram showing an example of the Poisson distribution of count values. [Figure 4] It is a diagram showing an example of noise detection results of the method according to the embodiment and the conventional method. [Figure 5] It is a flowchart showing an example of noise removal processing according to an embodiment. [Figure 6] It is a diagram showing an example of the hardware configuration of a computing device according to an embodiment.

Modes for Carrying Out the Invention

[0014] <Embodiment> The radiation noise detection and correction methods described herein will be explained below with reference to Figures 1 to 6. (composition) Figure 1 shows an example of an inspection system according to an embodiment. The inspection system 100 comprises a shielding body 1, a fixed head 2, a smear filter paper 3, a radiation detector (plastic scintillator) 4 for detecting radiation emitted from the smear filter paper 3, a power supply 5 for the radiation detector 4, a logic pulse converter 6, and a computing device 10. The fixed head 2 is fixed to the lid 1a of the shielding body 1 and housed in a space 1c provided within the main body 1b of the shielding body 1. The smear filter paper 3 is attached to the end of the fixed head 2. The detection unit 4a of the radiation detector 4 is positioned toward the smear filter paper 3 in space 1c. The logic pulse converter 6 processes the measured value measured by the radiation detector 4 and outputs measurement data corresponding to the measured value. For example, the logic pulse converter 6 outputs a pulse when the measured value measured by the radiation detector 4 exceeds a threshold. The computing device 10 calculates the radiation count value from the pulses included in the measurement data. For example, the measurement data is divided into 1-second intervals, the number of pulses contained within each interval is counted, and the counted value is converted into a radiation count value (cps).

[0015] The computing device 10 includes an acquisition unit 11, a noise detection unit 12, a correction unit 13, an output unit 14, and a storage unit 15. The acquisition unit 11 acquires the measurement data output by the logic pulse converter 6 and calculates the radiation count value from the acquired measurement data (logic pulse) using a known method. The count value is also called the count number. The noise detection unit 12 detects noise from the count values ​​calculated by the acquisition unit 11. The correction unit 13 corrects the count values ​​that are considered noise when noise is detected by the noise detection unit 12. The output unit 14 outputs the count values ​​calculated by the acquisition unit 11 and the count values ​​corrected by the correction unit 13 to a display device, electronic file, etc. The memory unit 15 stores the pulse signal acquired by the acquisition unit 11, the count value, the count value after correction by the correction unit 13, and so on.

[0016] Figure 2 shows an example of time-series count values. In Figures 2(a) and 2(b), the vertical axis shows the count value per second, and the horizontal axis shows the elapsed time from the start of measurement. Figures 2(a) and 2(b) are examples of measurement results when no radiation is attached to the smear filter paper 3. Figure 2(a) is an example of measurement results when no noise is measured, and Figure 2(b) is an example of measurement results when noise is measured. The count values ​​at times t0 and t1 in Figures 2(a) and 2(b) are not due to radiation attached to the smear filter paper 3, but rather to, for example, cosmic rays passing through the shielding 1. If radiation is attached to the smear filter paper 3, a nearly constant count value (a value with a certain magnitude, for example, the dashed line L in Figure 2(a)) will be measured continuously. The count value at time t2 in Figure 2(b) is noise due to the influence of other peripheral devices (for example, motor brake operation).

[0017] Conventionally, to detect noise, it was assumed that the count value x followed a Poisson distribution. Starting from x=0, the value of x was increased sequentially, and the Poisson probability P(x) of each count value x was accumulated (P(0,Av)+P(1,Av)+P(2,Av)+...). The value of x at which the accumulated value exceeded a predetermined value was found, and count values ​​above this value were considered noise. Figure 3 shows an example of the Poisson distribution of the count value x. In the conventional method, the height of the distribution curve is accumulated sequentially starting from count value=0, and this calculation is repeated until the accumulated value reaches a predetermined value. However, this method has a high computational load and is susceptible to overflow and underflow.

[0018] Therefore, this embodiment provides a method for detecting noise with a small amount of computation. Specifically, the probability of occurrence of a count value x is calculated using equation (2) shown below, and a count value x is found for which the calculated occurrence probability P'(x, Av) is less than or equal to a predetermined value (for example, 5% or 1%), and count values ​​greater than or equal to the found x are judged to be noise. In this specification, P'(x, Av) in equation (2) is also referred to as the normalized Poisson probability. Av is the average value excluding the single point with the maximum value, and Av' is the average value Av rounded up to the nearest whole number.

[0019]

number

[0020] Referring to the Poisson distribution curve in Figure 3, conventionally, the count value x was calculated such that the area under the distribution curve was 95% or 99%. In contrast, with normalized Poisson probability, the ratio of the height of the distribution curve at the value x being evaluated to the height of the distribution curve at position Av' is calculated, and x is calculated such that the ratio is 5% or 1% (generally 5% is sufficient, but 1% is used when noise is extracted more strictly). This reduces the computational load. Furthermore, the normalized Poisson probability P'(x, Av) can be summarized as follows.

[0021]

number

[0022] In other words, equation (2) can be simplified to a simple multiplication as shown in equation (3), and we calculate x such that equation (3) is less than or equal to a predetermined value (e.g., 5% or 1%). Equation (4) for finding x that is less than or equal to 1% is shown below. (The process of finding the count value x such that the normalized Poisson probability is less than or equal to 1% corresponds to the process of finding the count value x such that the cumulative value of P(x, Av) is 99% or more in the conventional method.)

[0023]

number

[0024] The process of calculating x that satisfies equation (4) is relatively computationally intensive and there is no concern about overflow or underflow. To find x that is 5% or less, simply set the right-hand side of equation (4) to 0.05. (The process of finding the count value x such that the normalized Poisson probability is 5% or less corresponds to the process of finding the count value x such that the cumulative value of P(x,Av) is 95% or more in the conventional method.)

[0025] Furthermore, by dividing the numerator and denominator of the left side of equation (4) by Av and taking the reciprocal, we obtain equation (5). To find x for which the percentage is 5% or less, we can set the right side of equation (5) to 20.

[0026]

number

[0027] The process of calculating x that satisfies equation (5) is also relatively computationally intensive, and there is no need to worry about overflow or underflow.

[0028] Figure 4 shows an example of noise detection results using the conventional method and the method of this embodiment. With the conventional method, the cumulative value of P(x,Av) is 95% or higher at x=3, and 99% or higher at x=4. In contrast, with this embodiment, the normalized Poisson probability is 5% or lower at x=4, and 1% or lower at x=5. The method of this embodiment can detect noise with the same accuracy as the conventional method.

[0029] (operation) Next, the noise detection method and correction method according to this embodiment will be described using Figure 5. Figure 5 is a flowchart showing an example of noise reduction processing according to the embodiment. First, the acquisition unit 11 acquires 20 seconds of measurement data (step S1). Note that 20 seconds is just an example, and the sampling period for the measurement data may be other lengths such as 30 seconds or 1 minute. Next, the acquisition unit 11 converts the 20 seconds of measurement data into count values ​​in 1-second units (step S2). The acquisition unit 11 divides the 20 seconds of measurement data into 1-second intervals, counts how many pulses are contained within each interval, and converts the counted value into a count value (cps). The acquisition unit 11 records the converted 1-second count values ​​in the storage unit 15. Next, the noise detection unit 12 searches for the maximum value among the count values ​​recorded in the storage unit 15 (step S3). Next, the noise detection unit 12 calculates the average value Av from the remaining values ​​after excluding the maximum value from the count values ​​recorded in the storage unit 15 (step S4). The noise detection unit 12 determines whether the average value Av is 100 or more (step S5).

[0030] If the average value Av is 100 or greater (Step S5; Yes), the count value is considered sufficiently large and the system is sufficiently contaminated regardless of noise. Therefore, to reduce the computational load, noise detection and correction are not performed, and the flow shown in Figure 5 is terminated. The threshold value of 100 used for the determination in Step S5 is just an example and may be any other value.

[0031] If the average value Av is less than 100 (Step S5; No), the noise detection unit 12 initializes the counter variable i to 1 (Step S6). The noise detection unit 12 sequentially selects the count values ​​recorded in the storage unit 15 and sets the count value x i Determine whether the value is less than or equal to the average value Av (Step S7). If it is less than or equal to the average value Av (Step S7; Yes), the count value x i If it is not considered noise, the process proceeds to step S10. If it is greater than the mean value Av (step S7; No), the noise detection unit 12 calculates the normalized Poisson probability P'(x i The system determines whether (Av) is 0.01 or less (step S8). The noise detection unit 12 performs the determination in step S8 using the above equation (4) or equation (5). This allows noise detection to be performed with less computation than in conventional methods.

[0032] P´(x i , Av) ≤ 0.01 (step S8; Yes), the noise detection unit 12 determines that x i is noise and instructs the correction unit 13 to perform correction. The correction unit 13 corrects x i according to the following formula (6) (step S9). x i = (x i+1 + x i-1 ) ÷ 2 ···(6) However, when P´(x i+1 , Av) ≤ 0.01, x i ) = x i-1 , and when P´(x i-1 , Av) ≤ 0.01, x i = x i+1 . The correction unit 13 records the corrected x i in the storage unit 15.

[0033] When P´(x i , Av) > 0.01 (step S8; No), the noise detection unit 12 determines that x i is not noise and proceeds to step S10. Next, the noise detection unit 12 adds 1 to the counter variable i (i = i + 1) (step S10). Next, the noise detection unit 12 determines whether the counter variable i is greater than 20 (step S11). When i ≤ 20 (step S11; No), the process from step S7 is repeated. When i > 20 (step S11; Yes), the output unit 14 outputs the count value x i (i = 1 to 20) (step S12). If the original count value x i was noise, the corrected x i is output.

[0034] (Effect) As described above, according to this embodiment, without causing overflow or underflow, the calculation amount can be reduced, and noise detection and correction included in the measurement result by the radiation measurement device can be performed at high speed. Thereby, reexamination due to false detection can be avoided, and the time required for the contamination inspection can be significantly shortened.

[0035] Figure 6 shows an example of the hardware configuration of a computing device. Computer 900 comprises a CPU 901, main memory 902, auxiliary memory 903, input / output interface 904, and communication interface 905. The above-mentioned computing device 10 is implemented in computer 900. The above-mentioned functions are stored in auxiliary memory 903 in the form of programs. The CPU 901 reads the program from auxiliary memory 903, expands it in main memory 902, and executes the above processing according to the program. The CPU 901 also allocates memory area in main memory 902 according to the program. The CPU 901 also allocates memory area in auxiliary memory 903 to store data being processed according to the program.

[0036] Furthermore, a program to implement all or part of the functions of the computing device 10 may be recorded on a computer-readable recording medium, and the program recorded on this recording medium may be loaded into a computer system and executed to perform processing by each functional unit. Here, "computer system" includes hardware such as the OS and peripheral devices. Also, if a WWW system is used, "computer system" also includes the homepage provisioning environment (or display environment). Furthermore, "computer-readable recording medium" refers to portable media such as CDs, DVDs, USBs, and storage devices such as hard disks built into the computer system. In addition, if this program is distributed to computer 900 via a communication line, computer 900 that receives the program may load it into main memory 902 and execute the above processing. Furthermore, the above program may be for implementing part of the functions described above, and may also be able to implement the above functions in combination with programs already recorded in the computer system.

[0037] As described above, several embodiments relating to this disclosure have been explained, but all of these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be carried out in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents.

[0038] <Note> The noise detection method, noise detection device, and program described in the embodiment can be understood, for example, as follows.

[0039] (1) A noise detection method according to the first embodiment is a noise detection method performed by a computer, comprising the steps of: acquiring count values ​​of radiation measured per unit time over a predetermined period; calculating a value obtained by dividing the Poisson probability of one of the count values ​​measured over the predetermined period by the Poisson probability of the average value of the count values ​​over the predetermined period; and determining that the one count value is noise if the calculated value is less than or equal to a predetermined value. This allows for the detection of noise contained in the measurement results from radiation measuring devices while reducing the computational load without causing overflow or underflow.

[0040] (2) The evaluation method relating to the second aspect is the noise detection method of (1), wherein when one of the count values ​​is x, the average value of the count values ​​excluding the one point with the maximum is Av, and the average value of the count value rounded up to the nearest whole number is Av', in the calculation step, the following formula is calculated:

number

[0041] (3) The evaluation method relating to the third aspect is the noise detection method of (1), wherein when one of the count values ​​is x, the average value of the count values ​​excluding the one point with the maximum is Av, and Av' is the value of the average value of the count values ​​rounded up to the nearest whole number, in the calculation step, the following formula is calculated:

number

[0042] (4) The evaluation method relating to the fourth aspect is the noise detection method of (1) to (3), wherein the predetermined value is a value of 0.05 or less. This allows for noise detection with the same accuracy as before.

[0043] (5) An evaluation method relating to the fifth aspect is a noise detection method of (1) to (4), wherein the one count value is x i , the x during the predetermined period i The count value measured immediately before x i-1 , the aforementioned x i The count measured immediately after the first one is x i+1 Let x i If it is determined to be noise, then x i-1 and the aforementioned x i+1 The average value of x i By setting the value to the above x i It further includes a step to correct the error. This allows for correction of count values ​​that were previously identified as noise.

[0044] (6) A noise detection device according to the sixth embodiment includes means for acquiring the count value of radiation per unit time measured over a predetermined period; means for calculating a value obtained by dividing the Poisson probability of one of the count values ​​measured over the predetermined period by the Poisson probability of the average value of the count values ​​over the predetermined period; and means for determining that the one count value is noise if the calculated value is less than or equal to a predetermined value.

[0045] (7) A program according to the seventh aspect is a program that causes a computer to perform the steps of: acquiring count values ​​of radiation measured per unit time during a predetermined period; calculating a value obtained by dividing the Poisson probability of one of the count values ​​measured during the predetermined period by the Poisson probability of the average value of the count values ​​during the predetermined period; and determining that if the calculated value is less than or equal to a predetermined value, the one count value is noise. [Explanation of symbols]

[0046] 1...shielding body 2. Fixed head 3. Smear filter paper 4. Radiation detector 5...Power supply 6. Logic pulse converter 10...calculating device 11...Acquisition part 12. Noise detection unit 13. Correction section 14.. Output section 15...Storage section 100... Inspection System 900... Computer 901···CPU 902...Main memory 903...Auxiliary storage device 904... Input / Output Interface 905...Communication Interface

Claims

1. A noise detection method performed by a computer, A step of obtaining the radiation count value per unit time measured over a predetermined period, The steps include: calculating the value obtained by dividing the Poisson probability of one of the count values ​​measured during the predetermined period by the Poisson probability of the average value of the count values ​​during the predetermined period; If the calculated value is less than or equal to a predetermined value, the step of determining that one of the count values ​​is noise, A noise detection method having the following characteristics.

2. If x is one of the aforementioned count values, Av is the average of the aforementioned count values ​​excluding the one with the maximum value, and Av' is the average of the aforementioned average value rounded up to the nearest whole number, In the calculation step described above, the following formula is calculated: [Math 1] In the determination step, if the calculation result of the formula is less than or equal to the predetermined value, it is determined that one of the count values ​​is noise. The noise detection method according to claim 1.

3. If x is one of the aforementioned count values, Av is the average of the aforementioned count values ​​excluding the one with the maximum value, and Av' is the average of the aforementioned average value rounded up to the nearest whole number, In the calculation step described above, the following formula is calculated: [Math 2] In the determination step described above, if the calculation result of the formula is greater than or equal to the reciprocal of the predetermined value, it is determined that one of the count values ​​is noise. The noise detection method according to claim 1.

4. The predetermined value is 0.05 or less. A noise detection method according to any one of claims 1 to 3.

5. Let the one count value be x ｉ , and let the count value measured one before the x ｉ during the predetermined period be x ｉ－１ , and let the count value measured one after the x ｉ be x ｉ＋１ . When the x ｉ is determined to be noise, by taking the average value of the x ｉ－１ and the x ｉ＋１ as the value of the x ｉ , the step of correcting the x ｉ is performed A noise detection method according to any one of claims 1 to 3, further comprising the above.

6. A means for acquiring radiation count values ​​per unit time measured over a predetermined period, A means for calculating the value obtained by dividing the Poisson probability of one of the count values ​​measured during the predetermined period by the Poisson probability of the average value of the count values ​​during the predetermined period, If the calculated value is less than or equal to a predetermined value, means for determining that one of the count values ​​is noise, A noise detection device equipped with the following features.

7. On the computer, A step of obtaining the radiation count value per unit time measured over a predetermined period, The steps include: calculating the value obtained by dividing the Poisson probability of one of the count values ​​measured during the predetermined period by the Poisson probability of the average value of the count values ​​during the predetermined period; If the calculated value is less than or equal to a predetermined value, the step of determining that one of the count values ​​is noise, A program that executes the command.

Images