Alpha radioactive source displacement measurement method

Abstract

According to the alpha radioactive source displacement measurement method, the displacement of an alpha radioactive source is judged based on a frame image which is output by a CMOS sensor and contains an alpha response signal; the measurement premise is that alpha particles generated by the alpha radioactive source vertically enter the light-sensitive surface of the CMOS sensor; the method comprises the following steps: S01, image grouping and signal superposition; s02, centroids of all signal points are calculated; s03, performing density clustering on the signal points; and S04, calculating the combination weight of each signal point. The device is used for displacement measurement of the alpha radioactive source, has the advantages of being good in stability and high in measurement precision (the 10-micrometer displacement of the alpha 1 radioactive source can be recognized), and can meet the requirements of high-precision measurement scenes (such as recognition and positioning of alpha particle polluted areas on the surface of a workpiece).

Description

Technical Field

[0001] The present invention relates to the technical field of nuclear radiation detection, in particular to a method for measuring the displacement of an alpha radiation source. Background Art

[0002] Radioactive source displacement measurement has important applications in the nuclear industry, medical imaging, and industrial inspection. In the nuclear industry, its applications are reflected in nuclear reactor safety monitoring, nuclear waste storage management, and transport safety assurance. In medical imaging, its applications are in precise radiotherapy positioning, imaging equipment calibration, and real-time dynamic tracking. In industrial inspection, its applications are in detecting radioactive contamination on workpiece surfaces.

[0003] Traditional radioactive source detection equipment mainly includes scintillator detectors, semiconductor detectors and 3 He gas detectors, but these three types of detectors often have problems such as insufficient accuracy, low sensitivity and limited resolution. They can only roughly or approximately measure the displacement of the radioactive source, which makes it difficult to meet the needs of certain high-precision measurement scenarios. Summary of the Invention

[0004] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a method for measuring the displacement of an α radiation source, which solves the problem that traditional radiation source detection equipment can only roughly or approximately measure the displacement of the radiation source and is difficult to meet certain high-precision measurement scenarios.

[0005] The technical solution of the present invention is: a method for measuring the displacement of an α radiation source, which determines the displacement of the α radiation source based on a frame image containing an α response signal output by a CMOS sensor; the measurement premise is that the α particles generated by the α radiation source are perpendicularly incident on the photosensitive surface of the CMOS sensor;

[0006] Here’s how:

[0007] Its characteristics are that the method is as follows:

[0008] S01. Image grouping and signal superposition;

[0009] S02. Calculate the centroid of all signal points;

[0010] S03. Density clustering of signal points;

[0011] S04. Calculate the combined weight of each signal point.

[0012] A further technical solution of the present invention is: the operations of each step are briefly described as follows:

[0013] S01. Image grouping and signal superposition:

[0014] First, multiple continuous frame images are converted into multiple continuous grayscale images; then, morphological operations (such as corrosion or dilation) are used to remove noise in the grayscale images; then, alpha response signals in the grayscale images are identified and extracted based on a connected component algorithm; then, all grayscale images are divided into multiple groups in order from front to back, with each group containing at least two grayscale images; finally, all alpha response signals in each group of grayscale images are extracted respectively, and are superimposed on multiple pure black images of the same size as the grayscale images, thereby obtaining superimposed images with the same number of grayscale image groups;

[0015] S02. Calculate the centroid of all signal points:

[0016] The α response signal is simplified to a single point (i.e., the pixel at the geometric centroid of the α response signal is used to represent the α response signal), which is called a signal point. The centroid of all signal points in the superimposed image is calculated. The signal points form a clustered area in the superimposed image. All signal points on the outermost side of the clustered area are sequentially connected in a clockwise or counterclockwise direction to form a closed area. The average distance from all boundary points of the closed area to the centroid is d. A circle is drawn with the centroid as the center and d as the radius. The circle is considered the clustered area of the signal points in the superimposed image.

[0017] S03. Density clustering of signal points:

[0018] a. Draw multiple concentric circles inside the circle to obtain multiple clustering sub-areas that are surrounded from the inside to the outside (the innermost clustering sub-area is circular, and the other clustering sub-areas are annular);

[0019] b. Count the clustered sub-regions with the largest signal point density in the superimposed image, and record the radius of the clustered sub-region with the largest density and the number of signal points within the clustered sub-region;

[0020] c. Calculate the average value R of the radius of the clustered sub-region with the largest signal point density in all superimposed images. mean ; Calculate the average number of signal points in the clustered sub-area with the largest signal point density in all superimposed images N mean ;

[0021] d. Simplify the α response signal into a point (i.e., the pixel point where the geometric centroid of the α response signal is located is used to represent the α response signal), which is called the signal point. mean As the optimal clustering radius, N mean As the minimum number of samples, the density clustering algorithm (DBSCAN) is used to perform density clustering on the signal points in the stacked images, and the point clusters and outliers in each stacked image are obtained by analysis;

[0022] S04. Calculate the combined weight of each signal point:

[0023] The entropy weight method is used to calculate the weight of each signal point in the point cluster: the weight includes linear distance weight and density weight; the density weight rule is: if the signal density around a signal point is greater than or equal to the density threshold, a relatively high weight is assigned; if the signal density around a signal point is less than the density threshold, a relatively low weight is assigned; the linear distance weight rule is: the closer the signal point is to the point cluster centroid, the higher the weight, the farther the signal point is from the point cluster centroid, the lower the weight, and the point whose distance from the point cluster centroid exceeds the distance threshold has a weight of 0;

[0024] The entropy weight method is used to calculate the combined weight of each signal point in the point cluster: first, the weight values of the two weight indicators are normalized to eliminate dimensional differences and unify the weight range; then, based on the normalized values, the information entropy is calculated for the two weight indicators respectively to measure the degree of data confusion of the two weight indicators; then, based on the normalized values, the entropy weight coefficients of the two weight indicators are calculated to determine the importance of the two weight indicators; finally, the two weight indicators are weighted and summed according to the entropy weight coefficient to realize the calculation of the combined weight;

[0025] According to the combined weights of each signal point in the point cluster, the coordinates of the weighted centroid of the point cluster in each superimposed image are calculated; finally, the moving path and moving direction of the α radiation source are obtained by combining the changes in the weighted centroid coordinates in each superimposed image.

[0026] A further technical solution of the present invention is: the connected area algorithm in step S01 is as follows:

[0027] I. Divide the grayscale value range of 0-255 in the grayscale image into multiple consecutive sub-ranges (for example: 0-5, 5-10, 10-15, etc.), calculate the increase in the number of pixels in each sub-range due to α radiation source radiation, and determine the sub-range most significantly affected by α radiation source radiation; the calculation formula is shown in Formula 1;

[0028] Formula 1:

[0029] Where A k is the increase in the number of pixels in subinterval k due to irradiation by α radiation source, is the number of pixels held by subinterval k when it is not irradiated, is the number of pixels held by subinterval k when irradiated by the α radiation source;

[0030] II. Based on formula 1, calculate the subinterval k with the most significant increase, mark all the pixels in subinterval k when irradiated by the α radiation source, and record any pixel as pixel n; calculate the number of pixels in pixel n and its 8 adjacent pixels (the 8 adjacent pixels include the 4 edge-adjacent pixels and the 4 vertex-adjacent pixels) that belong to subinterval k (including the target pixel n when counting the number), and record it as Pn ; Set 3<P n <8 and marked as α response signal, P n ≤2 and P n =9 and the corresponding connected areas are removed as noise.

[0031] A further technical solution of the present invention is: in step S02, the centroid of all signal points in the superimposed image is calculated using formula 2;

[0032] Formula 2:

[0033] Where (x, y) is the centroid coordinate, Indicates that the horizontal coordinates x of all signal points are i Add them together and divide them by the total number of signal points n to get the average value. Indicates that the ordinate y of all signal points i Add them up and divide by the total number of signal points n to get the average value.

[0034] A further technical solution of the present invention is: in step S03, the radii of the concentric circles inside the circle are 0.8d, 0.6d, 0.4d and 0.2d respectively;

[0035] A further technical solution of the present invention is: in step S03, the signal concentration of the aggregation sub-region with the largest signal density is calculated by formula 3;

[0036] Formula 3:

[0037] Where C is the signal concentration within the clustered sub-area, N is the number of signal points within the clustered sub-area, and S is the area of the clustered sub-area.

[0038] A further technical solution of the present invention is: in step S04, the linear distance weight and density weight of the signal point are calculated according to formula 4 and formula 5 respectively;

[0039] Formula 4:

[0040] Formula 5:

[0041] Where w l (p) is the linear distance weight of the signal point, dist(p, centroid) is the Euclidean distance from the signal point to the centroid, and R is the average distance from all boundary points of the closed area to the centroid, R = d; w d (p) is the density weight of the signal point, density t is the density threshold, max is the cutoff function; count(p) is the radius R around the signal point iThe number of signal points in the circular area is specifically set to the radius of the clustered sub-area with the highest signal point density in the current overlay image; density t is the density threshold, which is specifically set to the number of signal points in the clustered sub-region with the largest signal point density in the current superimposed image.

[0042] A further technical solution of the present invention is: in step S04, the normalization processing of the linear distance weight and the density weight of the signal point is calculated according to Formula 6 and Formula 7 respectively;

[0043] Formula 6:

[0044] Formula 7:

[0045] Where w′ l (p) is the normalized value of the linear distance weight, w′ d (p) is the density weight normalization value, w d,min is the minimum value of density weight, w d,max is the maximum value of density weight, w l,min is the minimum linear distance weight, w l,max is the maximum linear distance weight.

[0046] A further technical solution of the present invention is: in step S04, the information entropy of the linear distance weight and the density weight of the signal point are calculated according to Formula 8 and Formula 9 respectively;

[0047] Formula 8:

[0048] Formula 9:

[0049] Where, E l is the information entropy of the linear distance weight, E d is the information entropy of the density weight, n is the total number of signal points in the point cluster in the current superimposed image, and ε is the minimum value; P il is the probability ratio of the i-th signal point in the normalized linear distance weight, P id is the probability ratio of the i-th signal point in the normalized density weight, w′ l (p i ) is the normalized linear distance weight of signal point i, w′ d (p i ) is the normalized value of the density weight of signal point i.

[0050] A further technical solution of the present invention is: in step S04, the entropy weight coefficients of the linear distance weight and density weight of the signal point are calculated according to formula 10 and formula 11 respectively;

[0051] Formula 10:

[0052] Formula 11:

[0053] Where a l is the proportion of linear distance weight in the comprehensive weight, a d is the proportion of density weight in the comprehensive weight.

[0054] A further technical solution of the present invention is: in step S04, the calculation of the combined weight of the signal points is as shown in formula 12;

[0055] Formula 12: F W (p i )=a l ·w′ l (p i )+a d ·w′ d (p i )

[0056] Where, F W (p i ) is the combined weight of signal point i.

[0057] A further technical solution of the present invention is: in step S04, the coordinates of the weighted centroid of the point cluster are calculated with reference to formula 13;

[0058] Formula 13:

[0059] In the formula, (x c ,y c ) is the coordinate of the weighted centroid, (x i ,y i ) is the coordinate value of each signal point.

[0060] A further technical solution of the present invention is: the CMOS sensor belongs to a radiation source displacement measurement system; the radiation source displacement measurement system includes a CMOS sensor, a circuit board, a chip board, a housing and a PC; the photosensitive surface of the CMOS sensor is not glass-encapsulated and is attached with a layer of light-shielding material, the light-shielding material has the characteristics of shielding visible light and allowing radiation to penetrate; the CMOS sensor is fixedly mounted on the circuit board, the circuit board is provided with a power supply interface, the circuit board and the chip board are communicatively connected via a DuPont connector, the chip board is provided with a SOC chip for outputting a frame image containing a radiation response signal; the housing is made of The shell is made of radiation shielding material, with an inner cavity inside and an entrance hole on the surface of the shell. A cover plate is installed on the entrance hole, which has the characteristics of shielding visible light and allowing radiation to penetrate. The circuit board and the chip board are fixedly installed in the inner cavity of the shell, and the photosensitive surface of the CMOS sensor faces the entrance hole of the shell. There is no obstruction between the photosensitive surface and the entrance hole, and the distance between the two is less than 5mm. The size of the entrance hole is greater than or equal to the size of the photosensitive surface. The PC and the chip board are connected to each other through wired or wireless communication, which is used to adjust the parameters of the CMOS sensor and store and display frame images containing radiation response signals.

[0061] Compared with the existing technology, the present invention has the following advantages: it is used for displacement measurement of α radiation sources, has good stability and high measurement accuracy (can identify the displacement of α radiation sources of 10μm), and can meet the needs of high-precision measurement scenarios (such as the identification and positioning of α particle contaminated areas on the surface of a workpiece).

[0062] The present invention is further described below with reference to the figures and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 is a flow chart of the method of the present invention;

[0064] Figure 2 This is a screenshot of the signal point clustering area and its internal clustering sub-area in the overlay image. DETAILED DESCRIPTION

[0065] Example 1:

[0066] The α radiation source displacement measurement method determines the displacement of the α radiation source based on the frame image containing the α response signal output by the CMOS sensor; the measurement premise is that the α particles generated by the α radiation source are vertically incident on the photosensitive surface of the CMOS sensor.

[0067] like Figure 1-2 As shown, the method is as follows:

[0068] S01. Image grouping and signal superposition:

[0069] First, multiple continuous frame images are converted into multiple continuous grayscale images; then, morphological operations (such as corrosion or expansion) are used to remove noise in the grayscale images; then, the alpha response signals in the grayscale images are identified and extracted based on the connected component algorithm; then, all the grayscale images are divided into multiple groups in order from front to back, and each group contains at least two grayscale images; finally, all the alpha response signals in each group of grayscale images are extracted separately, and are superimposed on multiple pure black images with the same size as the grayscale images, so as to obtain superimposed images with the same number of grayscale image groups.

[0070] In this step, the connected area algorithm is as follows:

[0071] I. Divide the grayscale value range of 0-255 in the grayscale image into multiple consecutive sub-ranges (for example: 0-5, 5-10, 10-15, etc.), calculate the increase in the number of pixels in each sub-range due to α radiation source radiation, and determine the sub-range most significantly affected by α radiation source radiation; the calculation formula is shown in Formula 1;

[0072] Formula 1:

[0073] Where A k is the increase in the number of pixels in subinterval k due to irradiation by α radiation source, is the number of pixels held by subinterval k when it is not irradiated, is the number of pixels held by subinterval k when irradiated by the α radiation source;

[0074] II. Based on formula 1, calculate the subinterval k with the most significant increase, mark all the pixels in subinterval k when irradiated by the α radiation source, and record any pixel as pixel n; calculate the number of pixels in pixel n and its 8 adjacent pixels (the 8 adjacent pixels include the 4 edge-adjacent pixels and the 4 vertex-adjacent pixels) that belong to subinterval k (including the target pixel n when counting the number), and record it as P n ; Set 3<P n <8 and marked as α response signal, P n ≤2 and P n =9 and the corresponding connected areas are removed as noise.

[0075] S02. Calculate the centroid of all signal points:

[0076] The α response signal is simplified to a point (i.e., the pixel point where the geometric centroid of the α response signal is located is used to represent the α response signal), which is called a signal point. The centroid of all signal points in the superimposed image is calculated. The signal points form a clustered area in the superimposed image. All signal points on the outermost side of the clustered area are connected in a clockwise or counterclockwise direction to form a closed area. The average distance from all boundary points of the closed area to the centroid is d. A circle is drawn with the centroid as the center and d as the radius. The circle is regarded as the clustered area of the signal points in the superimposed image.

[0077] In this step, the centroid of all signal points in the superimposed image is calculated using formula 2;

[0078] Formula 2:

[0079] Where (x, y) is the centroid coordinate, Indicates that the horizontal coordinates x of all signal points are i Add them together and divide them by the total number of signal points n to get the average value. Indicates that the ordinate y of all signal points i Add them up and divide by the total number of signal points n to get the average value.

[0080] S03. Density clustering of signal points:

[0081] a. Draw multiple concentric circles inside the circle to obtain multiple clustering sub-areas that are surrounded from the inside to the outside (the innermost clustering sub-area is circular, and the other clustering sub-areas are annular);

[0082] b. Count the clustered sub-regions with the largest signal point density in the superimposed image, and record the radius of the clustered sub-region with the largest density and the number of signal points within the clustered sub-region;

[0083] c. Calculate the average value R of the radius of the clustered sub-region with the largest signal point density in all superimposed images. mean ; Calculate the average number of signal points in the clustered sub-area with the largest signal point density in all superimposed images N mean ;

[0084] d. Simplify the α response signal into a point (i.e., the pixel point where the geometric centroid of the α response signal is located is used to represent the α response signal), which is called the signal point. mean As the optimal clustering radius, N mean As the minimum number of samples, the density clustering algorithm (DBSCAN) is used to perform density clustering on the signal points in the stacked images, and the point clusters and outliers in each stacked image are obtained by analysis.

[0085] In this step, the radii of the concentric circles inside the circle are 0.8d, 0.6d, 0.4d and 0.2d respectively.

[0086] In this step, the signal concentration of the aggregation sub-region with the largest signal density is calculated using Formula 3;

[0087] Formula 3:

[0088] Where C is the signal concentration within the clustered sub-region, N is the number of signal points within the clustered sub-region (marked and counted by the connected component algorithm), and S is the area of the clustered sub-region.

[0089] S04. Calculate the combined weight of each signal point:

[0090] a. Use the entropy weight method to calculate the weight of each signal point in the point cluster: the weight includes linear distance weight and density weight; the density weight rule is: if the signal density around a signal point is greater than or equal to the density threshold, it is given a relatively high weight; if the signal density around a signal point is less than the density threshold, it is given a relatively low weight; the linear distance weight rule is: the closer the signal point is to the point cluster centroid, the higher the weight, the farther the signal point is from the point cluster centroid, the lower the weight, and the point whose distance from the point cluster centroid exceeds the distance threshold is weighted as 0;

[0091] b. Use the entropy weight method to calculate the combined weight of each signal point in the point cluster: first, normalize the weight values of the two weight indicators to eliminate dimensional differences and unify the weight range; then calculate the information entropy of the two weight indicators based on the normalized values to measure the degree of data confusion of the two weight indicators; then calculate the entropy weight coefficient of the two weight indicators based on the normalized values to determine the importance of the two weight indicators; finally, add the two weight indicators according to the entropy weight coefficient and calculate the combined weight;

[0092] c. Calculate the coordinates of the weighted centroid of the point clusters in each stacked image based on the combined weights of the signal points in the point clusters; finally, obtain the movement path and direction of the α radiation source based on the changes in the weighted centroid coordinates in each stacked image.

[0093] In this step, the linear distance weight and density weight of the signal point are calculated according to Formula 4 and Formula 5 respectively;

[0094] Formula 4:

[0095] Formula 5:

[0096] Where w l (p) is the linear distance weight of the signal point, dist(p, centroid) is the Euclidean distance from the signal point to the centroid, and R is the average distance from all boundary points of the closed area to the centroid, R = d; w d(p) is the density weight of the signal point, density t is the density threshold, max is the cutoff function; count(p) is the number of signal points in a circular area with a certain radius around the signal point, which is specifically set to the radius of the clustered sub-area with the largest signal point density in the current superimposed image; density t is the density threshold, which is specifically set to the number of signal points in the clustered sub-region with the largest signal point density in the current superimposed image.

[0097] In this step, the normalization processing of the linear distance weight and density weight of the signal point is calculated according to Formula 6 and Formula 7 respectively;

[0098] Formula 6:

[0099] Formula 7:

[0100] Where w′ l (p) is the normalized value of the linear distance weight, w′ d (p) is the density weight normalization value, w d,min is the minimum value of density weight (0.5), w d,max is the maximum value of density weight (5), w l,min is the minimum linear distance weight, w l,max is the maximum linear distance weight.

[0101] In this step, the information entropy of the linear distance weight and density weight of the signal point is calculated according to Formula 8 and Formula 9 respectively;

[0102] Formula 8:

[0103] Formula 9:

[0104] Where, E l is the information entropy of the linear distance weight, E d is the information entropy of the density weight, n is the total number of signal points in the point cluster in the current superimposed image, and ε is the minimum value (e.g. 10 -10 , to prevent ln(0) calculation errors); P il is the probability ratio of the i-th signal point in the normalized linear distance weight, P id is the probability ratio of the i-th signal point in the normalized density weight, w′ l (p i ) is the normalized linear distance weight of signal point i, w′ d (p i ) is the normalized value of the density weight of signal point i.

[0105] In this step, the entropy weight coefficients of the linear distance weight and density weight of the signal point are calculated according to Formula 10 and Formula 11 respectively;

[0106] Formula 10:

[0107] Formula 11:

[0108] Where a l is the proportion of linear distance weight in the comprehensive weight, a d is the proportion of density weight in the comprehensive weight.

[0109] In this step, the calculation of the combined weight of the signal points is as shown in Formula 12;

[0110] Formula 12: F W (p i )=a l ·w′ l (p i )+a d ·w′ d (p i )

[0111] Where, F W (p i ) is the combined weight of signal point i.

[0112] In this step, the coordinates of the weighted centroid of the point cluster are calculated with reference to Formula 13;

[0113] Formula 13:

[0114] In the formula, (x c ,y c ) is the coordinate of the weighted centroid, (x i ,y i ) is the coordinate value of each signal point.

[0115] The CMOS sensor belongs to a radiation source displacement measurement system. The radiation source displacement measurement system includes a CMOS sensor, a circuit board, a chip board, a housing, and a PC. The photosensitive surface of the CMOS sensor is not glass-encapsulated and is attached with a layer of light-shielding material. The light-shielding material has the characteristics of shielding visible light and allowing radiation to penetrate. The CMOS sensor is fixedly mounted on the circuit board, which has a power supply interface. The circuit board and the chip board are communicatively connected via a DuPont connector. The chip board is provided with a SOC chip for outputting a frame image containing a radiation response signal. The housing is made of radiation-shielding material, has an inner cavity within the housing, and an entrance hole is provided on the housing surface. A cover plate is mounted on the entrance hole, which has the characteristics of shielding visible light and allowing radiation to penetrate. The circuit board and the chip board are both fixedly mounted in the inner cavity of the housing. The photosensitive surface of the CMOS sensor faces the entrance hole of the housing. There is no obstacle between the photosensitive surface and the entrance hole, and the distance between them is less than 5 mm. The size of the entrance hole is greater than or equal to the size of the photosensitive surface. The PC is connected to the chip board via wired or wireless communication and is used to adjust the parameters of the CMOS sensor and store and display the frame image containing the radiation response signal.

Claims

1. Alpha radiation source displacement measurement method: The displacement of the alpha radiation source is determined based on the frame image containing the alpha response signal output by the CMOS sensor. The measurement premise is that the alpha particles generated by the alpha radiation source are perpendicular to the photosensitive surface of the CMOS sensor. Its characteristic is that the method as follows: S01. Image grouping and signal superposition; S02. Calculate the centroid of all signal points; S03. Density clustering of signal points; S04. Calculate the combined weight of each signal point.

2. The method for measuring the displacement of an α-radiation source according to claim 1, wherein: The operations of each step are briefly described as follows: S01. Image grouping and signal superposition: First, multiple continuous frame images are converted into multiple continuous grayscale images; morphological operations are then used to remove noise in the grayscale images; then, alpha response signals in the grayscale images are identified and extracted based on a connected component algorithm; then, all grayscale images are divided into multiple groups in order from front to back, with each group containing at least two grayscale images; finally, all alpha response signals in each group of grayscale images are extracted separately, and are superimposed on multiple pure black images of the same size as the grayscale images, thereby obtaining superimposed images with the same number of grayscale image groups; S02. Calculate the centroid of all signal points: The α response signal is simplified to a single point, called a signal point. The centroid of all signal points in the superimposed image is calculated. The signal points form a clustered area in the superimposed image. All signal points on the outermost side of the clustered area are sequentially connected in a clockwise or counterclockwise direction to form a closed area. The average distance from all boundary points of the closed area to the centroid is d. A circle is drawn with the centroid as the center and d as the radius. The circle is considered the clustered area of the signal points in the superimposed image. S03. Density clustering of signal points: a. Draw multiple concentric circles inside the circle to obtain multiple clustered sub-areas that are sequentially surrounded from the inside to the outside; b. Count the clustered sub-regions with the largest signal point density in the superimposed image, and record the radius of the clustered sub-region with the largest density and the number of signal points within the clustered sub-region; c. Calculate the average value R of the radius of the clustered sub-region with the largest signal point density in all superimposed images. mean ; Calculate the average number of signal points in the clustered sub-area with the largest signal point density in all superimposed images N mean ; d. Simplify the α response signal into a point, called the signal point, and use R mean As the optimal clustering radius, N mean As the minimum number of samples, the density clustering algorithm is used to perform density clustering on the signal points in the stacked images, and the point clusters and outliers in each stacked image are obtained by analysis; S04. Calculate the combined weight of each signal point: a. Use the entropy weight method to calculate the weight of each signal point in the point cluster: the weight includes linear distance weight and density weight; the density weight rule is: if the signal density around a signal point is greater than or equal to the density threshold, it is given a relatively high weight; if the signal density around a signal point is less than the density threshold, it is given a relatively low weight; the linear distance weight rule is: the closer the signal point is to the point cluster centroid, the higher the weight, the farther the signal point is from the point cluster centroid, the lower the weight, and the point whose distance from the point cluster centroid exceeds the distance threshold is weighted as 0; b. Use the entropy weight method to calculate the combined weight of each signal point in the point cluster: first, normalize the weight values of the two weight indicators to eliminate dimensional differences and unify the weight range; then calculate the information entropy of the two weight indicators based on the normalized values to measure the degree of data confusion of the two weight indicators; then calculate the entropy weight coefficient of the two weight indicators based on the normalized values to determine the importance of the two weight indicators; finally, add the two weight indicators according to the entropy weight coefficient and calculate the combined weight; c. Calculate the coordinates of the weighted centroid of the point clusters in each stacked image based on the combined weights of the signal points in the point clusters; finally, obtain the movement path and direction of the α radiation source based on the changes in the weighted centroid coordinates in each stacked image.

3. The method for measuring the displacement of an α-radiation source according to claim 2, wherein: The connected area algorithm in step S01 is as follows: I. Divide the grayscale value range of 0-255 in the grayscale image into multiple consecutive sub-ranges, calculate the increase in the number of pixels in each sub-range caused by α-radiation source radiation, and determine the sub-range most significantly affected by α-radiation source radiation; The calculation formula is shown in Formula 1; Formula 1: Where A k is the increase in the number of pixels in subinterval k due to irradiation by α radiation source, is the number of pixels held by subinterval k when it is not irradiated, is the number of pixels held by subinterval k when irradiated by the α radiation source; II. Based on formula 1, calculate the subinterval k with the most significant increase, mark all the pixels in subinterval k when irradiated by the α radiation source, and record any pixel as pixel n; calculate the number of pixels in pixel n and its 8 adjacent pixels (the 8 adjacent pixels include the 4 edge-adjacent pixels and the 4 vertex-adjacent pixels) that belong to subinterval k, and record it as P n ; Set 3<P n <8 and marked as α response signal, P n ≤2 and P n =9 and the corresponding connected areas are removed as noise.

4. The method for measuring the displacement of an α radiation source according to claim 3, wherein: In step 02, the centroid of all signal points in the superimposed image is calculated using formula 2; Formula 2: Where (x, y) is the centroid coordinate, Indicates that the horizontal coordinates x of all signal points are i Add them together and divide them by the total number of signal points n to get the average value. Indicates that the ordinate y of all signal points i Add them up and divide by the total number of signal points n to get the average value.

5. The method for measuring the displacement of an α-radiation source according to claim 4, wherein: In step S03, the radii of the concentric circles inside the circle are 0.8d, 0.6d, 0.4d and 0.2d respectively; In step S03, the signal concentration of the clustered sub-region with the largest signal density is calculated using formula 3; Formula 3: Where C is the signal concentration within the clustered sub-area, N is the number of signal points within the clustered sub-area, and S is the area of the clustered sub-area.

6. The method for measuring the displacement of an α-radiation source according to claim 5, wherein: In step S04, the linear distance weight and density weight of the signal point are calculated according to formula 4 and formula 5 respectively; Formula 4: Formula 5: Where w l (p) is the linear distance weight of the signal point, dist(p, centroid) is the Euclidean distance from the signal point to the centroid, and R is the average distance from all boundary points of the closed area to the centroid, R = d; w d (p) is the density weight of the signal point, density t is the density threshold, max is the cutoff function; count(p) is the number of signal points in a circular area with a certain radius around the signal point, which is specifically set to the radius of the clustered sub-area with the largest signal point density in the current superimposed image; density t is the density threshold, which is specifically set to the number of signal points in the clustered sub-region with the largest signal point density in the current superimposed image.

7. The method for measuring the displacement of an α-radiation source according to claim 6, wherein: In step S04, the normalization processing of the linear distance weight and density weight of the signal point is calculated according to Formula 6 and Formula 7 respectively; Formula 6: Formula 7: Where w l ′(p) is the normalized value of the linear distance weight, w′ d (p) is the density weight normalization value, w d,min is the minimum value of density weight, w d,max is the maximum value of density weight, w l,min is the minimum linear distance weight, w l,max is the maximum linear distance weight; In step S04, the information entropy of the linear distance weight and density weight of the signal point is calculated according to Formula 8 and Formula 9 respectively; Formula 8: Formula 9: Where, E l is the information entropy of the linear distance weight, E d is the information entropy of the density weight, n is the total number of signal points in the point cluster in the current superimposed image, and ε is the minimum value; P il is the probability ratio of the i-th signal point in the normalized linear distance weight, P id is the probability ratio of the i-th signal point in the normalized density weight, w l ′(p i ) is the normalized linear distance weight of signal point i, w′ d (p i ) is the normalized value of the density weight of signal point i.

8. The method for measuring the displacement of an α-radiation source according to claim 7, wherein: In step S04, the entropy weight coefficients of the linear distance weight and density weight of the signal point are calculated according to formula 10 and formula 11 respectively; Formula 10: Formula 11: Where a l is the proportion of linear distance weight in the comprehensive weight, a d is the proportion of density weight in the comprehensive weight; In step S04, the calculation of the combined weight of the signal points is as shown in formula 12; Enclosure 12:F W (p. 100) i )r l ·w l ′(p i )+a d ·w′ d (p. 100) i ) Where, F W (p i ) is the combined weight of signal point i; In step S04, the coordinates of the weighted centroid of the point cluster are calculated with reference to formula 13; Formula 13: In the formula, (x c ,y c ) is the coordinate of the weighted centroid, (x i ,y i ) is the coordinate value of each signal point.

9. The method for measuring the displacement of an α-radiation source according to claim 8, wherein: The CMOS sensor belongs to a radiation source displacement measurement system. The radiation source displacement measurement system includes a CMOS sensor, a circuit board, a chip board, a housing, and a PC. The photosensitive surface of the CMOS sensor is not glass-encapsulated and is attached with a layer of light-shielding material. The light-shielding material has the characteristics of shielding visible light and allowing radiation to penetrate. The CMOS sensor is fixedly mounted on the circuit board, which has a power supply interface. The circuit board and the chip board are communicatively connected via a DuPont connector. The chip board is provided with a SOC chip for outputting a frame image containing a radiation response signal. The housing is made of radiation-shielding material, has an inner cavity within the housing, and an entrance hole is provided on the housing surface. A cover plate is mounted on the entrance hole, which has the characteristics of shielding visible light and allowing radiation to penetrate. The circuit board and the chip board are both fixedly mounted in the inner cavity of the housing. The photosensitive surface of the CMOS sensor faces the entrance hole of the housing. There is no obstacle between the photosensitive surface and the entrance hole, and the distance between them is less than 5 mm. The size of the entrance hole is greater than or equal to the size of the photosensitive surface. The PC is connected to the chip board via wired or wireless communication and is used to adjust the parameters of the CMOS sensor and store and display the frame image containing the radiation response signal.

Images