Cosmic ray shower muon navigation positioning point feature extraction method

Abstract

The application discloses a cosmic ray cluster muon navigation positioning point feature extraction method, and relates to the technical field of navigation guidance and control. The steps are as follows: S1, through the upper and lower two particle detectors in the closed navigation area, accumulate and collect muon data penetrating into the navigation area, and obtain muon data for effective imaging through time sequence denoising and spatial filtering; S2, a projection horizontal plane is constructed in the closed navigation area and is divided into a grid, the number of muons penetrating into each cell of the grid and the penetrating point position are determined, the muons affecting imaging in the possible abnormal muons in each cell are identified and deleted; S3, the muon distribution density in each cell on the projection horizontal plane is re-estimated, a color heat map of the navigation area is generated through normalization and color mapping; the method realizes intuitive, clear and efficient color heat map imaging of the navigation positioning hotspot area, reduces the processing difficulty of the muon navigation positioning map, and ensures the identification efficiency and reliability of the navigation positioning feature points.

Description

Technical Field

[0001] This invention relates to the field of navigation, guidance and control technology, and in particular to a method for extracting features of navigation and positioning points from muons emitted by cosmic ray clusters. Background Technology

[0002] Muon navigation and positioning generates a large amount of muon information data, which is usually recorded in the form of positioning scatter plots. This data includes the arrival time, spatial location, and energy information of muons generated by cosmic ray showers. Currently, navigation and positioning schemes that use encoding and recognition require processing muon positioning scatter plot data within a given period. Directly using muon navigation and positioning scatter plots results in a large data volume, high processing difficulty, and low recognition accuracy.

[0003] To meet the data processing needs of novel muon navigation and positioning, it is necessary to address the problems of large amount of scattered data, high processing difficulty, and low feature extraction efficiency in the existing cosmic ray cluster-generated muon navigation and positioning process. A feature extraction method based on cosmic ray cluster-generated muon navigation and positioning points is proposed, which significantly improves the accuracy and efficiency of muon navigation and positioning, and provides a more reliable and faster positioning point scheme for muon navigation and positioning. Summary of the Invention

[0004] The purpose of this invention is to provide a method for extracting navigation and positioning points from cosmic ray cluster-emitted muons, which solves the above-mentioned technical problems.

[0005] Therefore, the technical solution of the present invention is as follows:

[0006] A method for extracting features of navigation and positioning points from muons emitted by cosmic ray showers, comprising the following steps:

[0007] S1. By deploying upper and lower particle detectors within the closed navigation area, the data of muons penetrating into the navigation area is accumulated according to a set acquisition cycle, that is, the arrival time and arrival position of each muon are collected by the upper and lower particle detectors respectively; by performing temporal denoising and spatial filtering on the muon data, muon data for effective imaging is obtained.

[0008] S2. Construct a projection plane within the closed navigation area and divide it into a grid. Determine the number of muons that penetrate each cell of the grid and the coordinates of their penetration points on the projection plane. Estimate the distribution density of muons in each cell and identify the muons that penetrate near the boundary line in each cell as possible anomalous muons. Perform local anomalous factor analysis based on the local density of each possible anomalous muon to identify and delete the muons that affect imaging among the possible anomalous muons in each cell.

[0009] S3. Based on the processing results of step S2, the muon distribution density of each cell on the projected horizontal plane is re-estimated, and normalization and color mapping are performed sequentially to generate a color muon positioning heatmap of the navigation area with high-precision feature distribution.

[0010] Further, in step S1, the temporal denoising and spatial filtering processing steps for the μ-sub-data are as follows:

[0011] (1) Temporal denoising process: Define a time window and calculate the difference in arrival time of each muon collected by the upper and lower particle detectors respectively, and retain the muon data whose absolute value of the time difference is less than the time window;

[0012] (2) Spatial filtering processing steps: Set the zenith angle threshold, and calculate the zenith angle of each muon based on the arrival position collected by the upper and lower particle detectors respectively, and retain the muon data with zenith angles less than the zenith angle threshold.

[0013] Furthermore, the time window is set to 50ns~1μs, and the zenith angle threshold is set to 10°~15°.

[0014] Furthermore, the specific implementation steps of step S2 are as follows:

[0015] S201. Construct a projection plane within the closed navigation area and divide the projection plane into grids; count the number of μs that traverse each cell based on the tracks of each μ, and the coordinates of the points on the projection plane where the tracks of each μ in each cell traverse.

[0016] S202. Using the kernel density estimation method, estimate the muon distribution density in each cell;

[0017] S203. Based on the coordinates of the penetration points of each μ in the cell on the projected horizontal plane, identify the μ that penetrates close to the cell boundary line and treat it as a possible abnormal μ.

[0018] S204. Calculate the local density of each possible anomalous muon based on the nearest neighbor number k;

[0019] S205. Calculate the LOF (Large Rank Factor) for each possible anomalous muon, and identify and delete anomalous muons by setting an LOF threshold.

[0020] Furthermore, in step S202, the method for estimating the muon distribution density within each cell is as follows:

[0021] The expression for calculating the kernel function K is:

[0022] ,

[0023] In the formula, uFor any muon in x Standardized relative deviation in the axial direction, ; v For any muon in y Standardized relative deviation in the axial direction, ; x , y These are the x and y coordinates of the cell's center point, respectively. x i , y i Each of the following is a muon within a cell. i The x and y coordinates of the points intersecting the projected horizontal plane are plotted.

[0024] Furthermore, the muon distribution density of each cell f ( m,n The calculation expression for ) is:

[0025] ,

[0026] In the formula, K i For muons i The kernel function calculation results; N N represents the total number of μ-cells embedded within the cell. μ (m,n); x , y These are the x and y coordinates of the cell's center point, respectively. x i , y i Each of the following is a muon within a cell. i The x and y coordinates of the points intersecting the projected horizontal plane are plotted. h This is the bandwidth parameter.

[0027] Furthermore, in step S203, the method for identifying potentially anomalous muons is as follows:

[0028] 1) Based on the grid division method in step S201, obtain the coordinate set of the cell boundary line;

[0029] 2) Set the recognition threshold l It sequentially determines whether the vertical distance between the point where each μ-cell passes through the projected horizontal plane and the adjacent boundary line of the cell is less than or equal to the recognition threshold. l In such cases, muons that are judged as "yes" are identified as potentially anomalous muons.

[0030] Furthermore, in step S204, the method for calculating the local density of potentially anomalous muons is as follows:

[0031] 1) Set the nearest neighbor numberk And define any possible anomalous muon P neighborhood N k ( P () is the closest k A set of possible anomalous muons;

[0032] 2) Calculate each possible anomalous muon and its neighborhood. k The reciprocal of the average distance between the points of penetration of a possible anomalous muon on the projected horizontal plane, and used as its local density (P), is expressed as:

[0033] ,

[0034] In the formula, k The nearest neighbor number, P For potentially anomalous muons, Q i For the possible anomalous muon neighborhood N r (P) Middle i A possible anomalous μ-subject, d ( P , Q i ) is a possible anomalous muon P With the possible anomalous muon neighborhood N r (P) Middle i One possible anomalous muon Qi The Euclidean distance between points on the projected horizontal plane.

[0035] Furthermore, in step S205, the method for determining the anomalous muon is as follows:

[0036] 1) Define the LOF value as the possible anomalous muon. P Compared to it k Density differences of possible anomalous muons in the neighborhood LOF k ( P The calculation formula is as follows:

[0037] ,

[0038] In the formula, N k ( p ) is a possible anomalous muon P k-neighborhood, lrd( P ) is a possible anomalous muon P Local density; Q i For the possible anomalous muon P neighborhood N r (P) Middle iA possible anomalous μ-subject, lrd( Q i ) represents the possible anomalous μ-p-neighborhood N r (P) Middle i Local density of potentially anomalous muons; |N k |For the neighborhood N r The number of potentially anomalous muons in (P);

[0039] 2) Set the LOF threshold and make a judgment: if the LOF calculation result of a possible abnormal muon is less than or equal to the LOF threshold, then the possible abnormal muon is a normal muon; if the LOF calculation result of a possible abnormal muon is greater than the LOF threshold, then the possible abnormal muon is an abnormal muon and is deleted.

[0040] Furthermore, the specific implementation steps of step S3 are as follows:

[0041] S301. Re-estimate the muon distribution density of each cell in the grid using the method in step S2;

[0042] S302. Normalize the muon distribution density of each cell in the grid to make its value normalized to the interval [0, 1].

[0043] S303. By setting the color mapping, the normalized muon distribution density is mapped to color values ​​to obtain a color muon positioning heatmap of the navigation area.

[0044] Compared with existing technologies, this method for extracting features of muon navigation and positioning points from cosmic ray clusters effectively solves the problem of massive muon scatter data and unclear direct imaging features, achieving intuitive, clear, and efficient color thermal imaging of navigation and positioning hotspot areas. In specific implementation, based on the collected raw muon data, temporal denoising and spatial filtering, as well as spatial feature enhancement and density anomaly analysis, significantly improve the positioning accuracy of the imaging map while reducing the complexity of data processing. In summary, this method effectively reduces the difficulty of processing muon navigation and positioning maps, ensuring the efficiency and reliability of navigation and positioning feature point recognition. Attached Figure Description

[0045] Figure 1 This is a flowchart of the method for extracting navigation and positioning points of muons emitted by cosmic ray clusters according to the present invention;

[0046] Figure 2(a) is a schematic diagram of the actual scene setting of the closed navigation area in an embodiment of the present invention;

[0047] Figure 2(b) is a schematic diagram of a certain muon being collected by the upper and lower particle detectors in an actual scenario in an embodiment of the present invention;

[0048] Figure 3This is a distribution feature map of muon positioning points formed from the original muon data collected in step S1 in an embodiment of the present invention;

[0049] Figure 4 This is a distribution feature map of muon positioning points formed by spatial filtering, temporal denoising and spatial denoising of the raw muon data collected in step S1 in an embodiment of the present invention.

[0050] Figure 5 In an embodiment of the present invention, a color muon positioning heatmap is obtained by processing the muon data processed in step S1 without the density feature enhancement processing in step S2, and then processing it in step S3.

[0051] Figure 6 In an embodiment of the present invention, the muon data processed in step S1 is processed sequentially in steps S2 and S3 to obtain a color muon positioning heatmap. Detailed Implementation

[0052] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but the following embodiments are by no means intended to limit the present invention.

[0053] See Figure 1 The specific implementation steps of the method for extracting navigation and positioning point features from cosmic ray cluster-generated muons for high-precision navigation and positioning are described below.

[0054] S1. By deploying upper and lower particle detectors within the closed navigation area, the muon data penetrating the navigation area is collected cumulatively according to a set acquisition cycle. The muon data is then processed by temporal denoising and spatial filtering to obtain muon data for effective imaging.

[0055] The specific implementation steps of step S1 are as follows.

[0056] S101. Two particle detectors, one above and one below, are laid in the closed navigation area to continuously collect muon data penetrating into the navigation area; in this embodiment, the particle detector is a liquid scintillator muon detector.

[0057] Figure 2(a) shows a schematic diagram of an experimental scenario set up for positioning and navigation within a subway tunnel. In the figure, two liquid scintillator muon detectors are installed at the top and bottom of the enclosed navigation area 1, respectively. Each liquid scintillator muon detector consists of three liquid scintillator muon detectors 3 arranged at equal intervals to collect muon data penetrating into the enclosed navigation area 1. In this experimental scenario, an imaging object 2 is also installed within the enclosed navigation area 1, specifically two cubic blocks connected at their vertices. Since this experimental scenario is based on a subway tunnel scenario with GNSS signal obstruction, the purpose of this embodiment is to extract the navigation positioning point features based on the muon data collected within the navigation area using the method of this invention, ultimately achieving clear imaging of the navigation area.

[0058] In this step, due to the sparsity of muon detection in actual navigation system deployment, data needs to be accumulated within a certain time range. Specifically, the muon data acquisition cycle is 30 minutes to 10 hours. Of course, if a high-precision modeling task is required, the above acquisition cycle can be extended to one day or longer, and adjusted according to the detector sensitivity and the geological structure. In this embodiment, the cumulative acquisition cycle is set to 10 hours, corresponding to a cumulative acquisition of 5 million valid muon data.

[0059] In this step, the muon data specifically refers to the arrival time of each muon, collected separately by the upper and lower particle detectors. t and arrival location X Figure 2(b) shows a schematic diagram of a muon being collected by both upper and lower particle detectors in a real-world scenario according to this embodiment. As can be seen from the figure, the arrival time of the muon collected by the upper particle detector is... t A Arrival at location X A for( x A , y A The arrival time, collected by the lower-path particle detector, is... t B Arrival at location X B for( x B , y B Based on the arrival positions of the muon collected by the upper and lower particle detectors, the trajectory 5 and its direction vector of the muon can be further obtained.

[0060] S102. A time-series denoising method is used to filter noise in the μ data collected in step S101 to reduce the impact of random noise on data imaging.

[0061] In this step, since the penetration speed of muons is close to the speed of light, the time difference between muons passing through the dielectric layer and reaching the detector located in the navigation area is extremely small, generally on the order of nanoseconds to microseconds. That is, the detector response time and time resolution are extremely small. For example, the liquid scintillator muon detector used in this embodiment has a time resolution of tens to hundreds of nanoseconds. Therefore, according to the assumption of the continuity of physical events, if two muons come from the same cosmic ray shower event or a coherent event with a trajectory, their time difference should be very small. Thus, the muon data acquired in step S101 can be initially screened by setting time conditions.

[0062] Specifically, the processing method for step S102 is as follows:

[0063] Define time window Δt The arrival time of each muon on the upper and lower particle detectors was also recorded. t A and t B Judge and filter based on the difference:

[0064] When | t A – t B |< Δt If the data is determined to be a continuous event, it will be retained for subsequent data analysis and image processing.

[0065] When | t A – t B |≥ Δt If the data is not a coherent event, it is determined that the muon data does not belong to the coherent event and is deleted so that it will no longer participate in subsequent data analysis and image processing.

[0066] In this embodiment, the time window Δt The time window is generally defined as 50 ns to 1 μs. In this embodiment, the time window is... Δt Set to 1μs.

[0067] S103. Using spatial filtering, the angle of the retained μ data after processing in step S102 is filtered.

[0068] Specifically, the processing method for step S103 is as follows:

[0069] 1) Based on the arrival positions of each muon collected by the upper and lower particle detectors, calculate the zenith angle of each muon, which is the acute angle between the direction vector of each muon and the direction of the ground normal. θ ;

[0070] 2) Set the zenith angle threshold θ th Preserve the zenith angle θ < θ th The muon data is used to ensure that the muon data that meets the vertical incidence condition is used for subsequent imaging;

[0071] In this embodiment, the zenith angle threshold θ th The zenith angle is typically set to 10°~15°. In this embodiment, the zenith angle threshold is... θ th Set to 10°.

[0072] like Figure 3 The image shown is a distribution feature map of muon positioning points directly drawn from the raw muon data accumulated in step S1 in this embodiment. Only the raw muon position information points can be seen from the image, which is characterized by dense muon points and lacks feature recognition.

[0073] like Figure 4 The figure shows the distribution feature map of muon positioning points, which is formed by drawing the muon data obtained after temporal denoising and spatial filtering of the original muon data accumulated in step S1 in this embodiment. As can be seen from the figure, the muon data obtained after step S1 can show certain structural clustering features in spatial projection, forming a blurred image of the corresponding target area outline. Although the distribution trend of the key structure has been initially revealed, there is still a problem that the boundary is not clear enough. However, it shows that the processing of the original muon data in step S1 is effective and can be used as the data basis for subsequent feature extraction and imaging enhancement.

[0074] S2. Density outliers are identified and deleted from the μ data obtained after step S1 to achieve spatial feature enhancement processing.

[0075] The specific implementation steps of step S2 are described below.

[0076] S201. Construct a projected horizontal plane within the closed navigation area, and divide the projected horizontal plane into a grid with M×N cells; based on the tracks of each muon, count the number N muons that traverse each cell. μ (m,n), and the coordinates of the points where each μ-particle in each unit is projected onto the horizontal plane.

[0077] Specifically, in this step, the grid division accuracy on the projection horizontal plane is determined according to actual needs to meet the resolution requirements of subsequent two-dimensional imaging. In this embodiment, the projection horizontal plane is set at the bottom of the navigation area 1 so as to determine the cell layout of the muon on the projection horizontal plane based on the track information of the muon. Each cell on the projection horizontal plane is specifically a 1cm×1cm cell. The track of each muon is the line connecting the arrival positions collected by the upper and lower particle detectors, respectively, as shown in Figure 2(b).

[0078] Meanwhile, a two-dimensional plane coordinate system is constructed based on the projected horizontal plane to facilitate the representation of the boundary line position coordinate information of each cell and the position coordinate information of the penetration point of each μ particle through the projected horizontal plane, which is calculated from the track of each μ particle. In this embodiment, the two-dimensional plane coordinate system of the projected horizontal plane is constructed with the center point of the projected horizontal plane as the origin, the extension direction of the navigation path as the x-axis direction, and the width direction of the navigation path as the y-axis direction.

[0079] Furthermore, to facilitate subsequent processing, N μ-particles are inserted into each cell. μ After statistical analysis of (m,n), it is represented as a two-dimensional matrix, with the expression: φ={N μ ( m,n )∈ R M×N In the formula, m and n are the row number and column number of the cell, respectively, and M and N are the number of rows and columns of the grid divided on the projected horizontal plane, respectively.

[0080] S202. Use the kernel density estimation (KDE) method to estimate the muon distribution density in each cell.

[0081] In this step, the kernel density estimation method obtains the result by smoothing the scattered points using a kernel function K, where,

[0082] The expression for calculating the kernel function K is:

[0083] ,

[0084] In the formula, u For any muon in x Standardized relative deviation in the axial direction, ; v For any muon in y Standardized relative deviation in the axial direction, ; x , y These are the x and y coordinates of the cell's center point, respectively. x i , y iEach of the following is a muon within a cell. i The x and y coordinates of the points intersecting the projected horizontal plane are plotted.

[0085] Furthermore, the muon distribution density of each cell f ( m,n The calculation expression for ) is:

[0086] ,

[0087] In the formula, K i For muons i The kernel function calculation results; N N represents the total number of μ-cells embedded within the cell. μ (m,n); x , y These are the x and y coordinates of the cell's center point, respectively. x i , y i Each of the following is a muon within a cell. i The x and y coordinates of the points intersecting the projected horizontal plane are plotted. h This is the bandwidth parameter.

[0088] Among them, bandwidth parameter h Generally, the bandwidth parameter is set to fit the size of the cell. In this embodiment, the bandwidth parameter is... h Set it to be the same as the side length of the cell, that is h =1cm.

[0089] After step S202, the muon distribution density of each cell in the projected horizontal plane grid can be obtained.

[0090] However, due to the smoothing properties of kernel density estimation, muon penetration points near cell boundaries may simultaneously affect the density estimation results of multiple adjacent cells, leading to local ambiguity or anomalous peaks in the density estimation. These manifest as duplicate counting, misleading peaks, and / or imaging artifacts. Duplicate counting refers to the kernel function of boundary muons generating non-zero values ​​in multiple cells, causing double counting and affecting the true density distribution. Misleading peaks refer to the formation of non-physically high-density regions at the boundaries, potentially obscuring or interfering with the identification of true hotspot regions. Imaging artifacts refer to the introduction of misleading structures in subsequent feature extraction and angle accumulation imaging, reducing image clarity and accuracy. Therefore, it is necessary to identify and filter muons falling on cell boundaries to remove locally anomalous muons that affect the feature imaging effect.

[0091] S203. Identify potentially abnormal muons based on the coordinates of the points on the projected horizontal plane for each muon within the unit.

[0092] Specifically, in step S203, the method for identifying potentially abnormal muons within each cell is as follows:

[0093] 1) Based on the grid division method in step S201, obtain the coordinate set of the cell boundary line;

[0094] 2) Set the recognition threshold l It sequentially determines whether the vertical distance between the point where each μ-cell is projected onto the horizontal plane and the adjacent boundary line of the cell is less than or equal to the recognition threshold. l In this case, muons with a judgment result of "yes" are identified as potentially anomalous muons;

[0095] In step 2), l Set to 1 / 100h; in this embodiment, l Set to 0.1mm.

[0096] S204, Based on nearest neighbor number k Calculate the local density of each possible anomalous muon.

[0097] Specifically, in step S204, the method for calculating the local density of possible anomalous muons is as follows:

[0098] 1) Set the nearest neighbor number k And define any possible anomalous muon P neighborhood N k ( P ) for the possible anomalous muon P The closest k A set of possible anomalous muons;

[0099] 2) Calculate each possible anomalous muon and its neighborhood. k The reciprocal of the average distance between the points on the projected horizontal plane where a possible anomalous muon passes through is used as its local density (P), and its expression is:

[0100] ,

[0101] In the formula, k The nearest neighbor number is typically set to 5-30. In this embodiment, k Set to 8; P For potentially anomalous muons, Q i For possible anomalous muons P Neighborhood N r (P) Middle i A possible anomalous μ-subject, d ( P ,Q i ) is a possible anomalous muon P With the possible anomalous muon neighborhood N r (P) Middle i One possible anomalous muon Qi The distance between points on the projected horizontal plane.

[0102] In this embodiment, the distance is specifically the Euclidean distance, and the average distance is specifically the average value calculated based on the Euclidean distance.

[0103] S205. Calculate the LOF (Location of Abnormality) for each possible anomalous muon, and determine the anomalous muon by using the set LOF threshold.

[0104] Specifically, in step S204, the method for calculating the anomalous factor of possible anomalous muons is as follows:

[0105] 1) Define the LOF value as the possible anomalous muon. P Compared to it k Density differences of possible anomalous muons in the neighborhood LOF k ( P The calculation formula is as follows:

[0106] ,

[0107] In the formula, N k ( p ) is a possible anomalous muon P k-neighborhood, lrd( P ) is a possible anomalous muon P The local density, that is, the density(P) obtained in step S204; Q i For the possible anomalous muon P neighborhood N r (P) Middle i A possible anomalous μ-subject, lrd( Q i ) represents the possible anomalous μ-p-neighborhood N r (P) Middle i Local density of potentially anomalous muons; |N k |For the neighborhood N r The number of possible anomalous muons in (P), i.e., the value of k;

[0108] 2) Set the LOF threshold and perform the following judgment:

[0109] If the LOF calculation result of a potentially anomalous muon is less than or equal to the LOF threshold, then the potentially anomalous muon is determined to be a normal muon that does not affect imaging.

[0110] If the LOF calculation result of a potentially anomalous muon is greater than the LOF threshold, then the potentially anomalous muon is determined to be an anomalous muon that affects imaging and is deleted.

[0111] In this step, the larger the LOF value of a possible anomalous muon P, the higher its anomalousness; therefore, anomalous muons can be identified by setting a threshold; in this embodiment, the LOF threshold is set to 1.5.

[0112] S3. The μ data retained after processing in step S2 is used as μ positioning feature points for visualization imaging of the navigation area.

[0113] The specific implementation steps of step S3 are as follows:

[0114] S301. Using the method described in step S201 above, recount the number of muons that penetrate each cell and the coordinates of the penetration points of each muon track on the projected horizontal plane; then, using the method described in step S202 above, re-estimate the muon distribution density in each cell.

[0115] S302. Based on the estimation results of step S301, normalize the muon distribution density of each cell in the grid so that its value is normalized to the interval [0, 1].

[0116] In this step, the specific formula for normalization is:

[0117] ,

[0118] In the formula, f ( m,n ) represents the μ-sub-distribution density of the specified cell. f min The minimum distribution density of muons in all cells. f max The maximum value of the muon distribution density in all cells. f norm ( m,n ) represents the normalized result of the μ-sub distribution density of the specified cell.

[0119] S303. By setting the color mapping, the normalized muon distribution density is mapped to color values ​​to generate a color muon positioning heatmap (referred to as a color heatmap) of the navigation area.

[0120] Specifically, the implementation steps of step S303 are as follows:

[0121] S3031. The range of values ​​corresponding to the minimum to maximum values ​​of the muon distribution density normalization result is divided into three intervals, namely the first muon distribution density interval, the second muon distribution density interval, and the third muon distribution density interval; a hue and a corresponding color mapping value range are set for each muon distribution density interval so that the differences in muon distribution density can be clearly identified after imaging.

[0122] In this embodiment, the first muon distribution density range is a low-density range, corresponding to a cool color tone, specifically blue, with a GRB range of (0, 0, 200) ~ (50, 50, 255); the second muon distribution density range is a medium-density range, corresponding to a neutral color tone, specifically green, with a GRB range of (0, 200, 0) ~ (50, 255, 50); and the third muon distribution density range is a high-density range, corresponding to a warm color tone, specifically red, with a GRB range of (200, 0, 0) ~ (255, 50, 50).

[0123] S3032. Map the normalized muon distribution density results of each muon distribution density interval to the corresponding color mapping value range;

[0124] In this embodiment, the mapping method specifically employs a linear interpolation algorithm, and its calculation formula is as follows:

[0125] C k = C s +(( f k – f min ) / ( f max – f min ))×( C e - C s ),

[0126] In the formula, C k For the first k The color value of each cell, C s This is the starting value for the color range. C e The color terminating value of the color range. f min This represents the minimum value of the normalized result of the muon distribution density in the interval. f max This represents the maximum value of the normalized result of the muon distribution density in the interval. fk For the first k The normalized result of the μ-sub distribution density of each cell.

[0127] S3033. Based on the projected horizontal plane constructed in step S201, the normalized result of the μ-substance distribution density of each cell is mapped to the color distribution result to obtain the color heat map of the navigation area.

[0128] In the color heatmap of this navigation area, each cell represents a pixel of the image; the color heatmap can intuitively display the correspondence between density values ​​and colors, making the heatmap readable.

[0129] As a preferred technical solution of this embodiment, feature annotation can be performed on the basis of the color heat map of the navigation area generated in step S302. Specifically, feature annotation includes the outline annotation of the imaging object (i.e., the high-density hotspot area) and the center position annotation of the imaging object.

[0130] like Figure 5 The image shown is a color heatmap obtained in this embodiment by directly processing the muon data obtained after step S1 without the density feature enhancement processing in step S2, using the method in step S3; from Figure 5 As can be seen, the color heatmap generated without step S2 has low overall contrast, blurred boundaries, insignificant hot spots, and a lot of background noise. It is prone to producing false high-density areas, which makes it difficult to accurately identify the target structure outline and will affect the subsequent positioning point feature imaging and recognition effect.

[0131] like Figure 6 The image shown is a color heatmap obtained in this embodiment by further processing the μ data processed in step S1, then in step S2, and finally in step S3. Figure 6 As can be seen from the image, the normalized value corresponding to the brightest (red) part is close to 1, indicating a high-density region; the normalized value corresponding to the darkest (blue) part is close to 0, indicating a low-density region. Based on this, the readability features of the image include: the red pixels representing high-density regions are high-density muon regions in the image, possessing significant navigation and positioning performance; the gradient region from red to green pixels represents a density gradient, reflecting the navigation and positioning distribution characteristics of the detector; the blue pixel region is a low-density muon region, representing the region with the most muon scattering and absorption, with sparse point distribution within it, not participating in navigation and positioning feature recognition, which is why the imaging image within the navigation region can ultimately obtain a clear boundary contour; relative to Figure 5 This demonstrates the importance and effectiveness of the spatial feature enhancement processing method in step S2 in highlighting the structural contours of the imaged object.

[0132] Therefore, in this invention, the color heatmap generated by processing the raw muon data collected by the particle detector through steps S1 and S2 significantly improves the spatial resolution and contrast of the muon density distribution, making the outline of the target area clearer and the hotspot features more concentrated and prominent, effectively suppressing background noise and boundary blurring. On the other hand, a significant edge effect caused by scattering can be clearly seen at the edge of the imaged object, namely the red pixels that appear around the imaged object in the figure. This is an anomaly that will appear concentrated in the edge imaging of high atomic number materials, proving the effectiveness of the heatmap imaging and helping to improve the overall navigation and recognition accuracy.

[0133] In summary, the method of this invention transforms complex muon scatter data based on cosmic ray showers into a heat map with high-precision feature distribution. Furthermore, this two-dimensional heat map can effectively identify key feature points for navigation and positioning, successfully pinpoint hotspot areas, and provide a reliable basis for navigation, positioning, and subsequent analysis.

Claims

1. A method for extracting features of navigation and positioning points from muons emitted by cosmic ray clusters, characterized in that, The steps are as follows: S1. By deploying upper and lower particle detectors within the closed navigation area, the data of muons penetrating into the navigation area is accumulated according to a set acquisition cycle, that is, the arrival time and arrival position of each muon are collected by the upper and lower particle detectors respectively; by performing temporal denoising and spatial filtering on the muon data, muon data for effective imaging is obtained. S2. Construct a projection plane within the closed navigation area and divide it into a grid. Determine the number of muons that penetrate each cell of the grid and the coordinates of their penetration points on the projection plane. Estimate the distribution density of muons in each cell and identify the muons that penetrate near the boundary line in each cell as possible anomalous muons. Perform local anomalous factor analysis based on the local density of each possible anomalous muon to identify and delete the muons that affect imaging among the possible anomalous muons in each cell. S3. Based on the processing results of step S2, the distribution density of muons in each cell on the projected horizontal plane is re-estimated, and normalization and color mapping are performed sequentially to generate a color muon positioning heatmap of the navigation area with high-precision feature distribution. In step S1, the temporal denoising and spatial filtering processes for the μ-sub-data are as follows: (1) Temporal denoising process: Define a time window and calculate the difference in arrival time of each muon collected by the upper and lower particle detectors respectively, and retain the muon data whose absolute value of the time difference is less than the time window; (2) Spatial filtering processing steps: Set the zenith angle threshold, and calculate the zenith angle of each muon based on the arrival position collected by the upper and lower particle detectors respectively, and retain the muon data with zenith angles less than the zenith angle threshold.

2. The method for extracting navigation and positioning point features from cosmic ray shower muons according to claim 1, characterized in that, The time window was set to 50ns~1μs, and the zenith angle threshold was set to 10°~15°.

3. The method for extracting navigation and positioning point features from cosmic ray shower muons according to claim 1, characterized in that, The specific implementation steps of step S2 are as follows: S201. Construct a projection plane within the closed navigation area and divide the projection plane into grids; count the number of μs that traverse each cell based on the tracks of each μ, and the coordinates of the points on the projection plane where the tracks of each μ in each cell traverse. S202. Using the kernel density estimation method, estimate the muon distribution density in each cell; S203. Based on the coordinates of the penetration points of each μ in the cell on the projected horizontal plane, identify the μ that penetrates close to the cell boundary line and treat it as a possible abnormal μ. S204. Calculate the local density of each possible anomalous muon based on the nearest neighbor number k; S205. Calculate the LOF (Large Rank Factor) for each possible anomalous muon, and identify and delete anomalous muons by setting an LOF threshold.

4. The method for extracting features of navigation and positioning points from cosmic ray cluster-emitted muons according to claim 3, characterized in that, In step S202, the method for estimating the muon distribution density within each cell is as follows: The expression for calculating the kernel function K is: ， In the formula, u For any muon in x Standardized relative deviation in the axial direction, ; v For any muon in y Standardized relative deviation in the axial direction, ; x , y These are the x and y coordinates of the cell's center point, respectively. x i , y i Each of the following is a muon within a cell. i The x and y coordinates of the points intersecting the projected horizontal plane are plotted. Furthermore, the muon distribution density of each cell f ( m,n The calculation expression for ) is: ， In the formula, K i For muons i The kernel function calculation results; N N represents the total number of μ-cells embedded within the cell. μ (m,n); x , y These are the x and y coordinates of the cell's center point, respectively. x i , y i Each of the following is a muon within a cell. i The x and y coordinates of the points intersecting the projected horizontal plane are plotted. h This is the bandwidth parameter.

5. The method for extracting features of navigation and positioning points from cosmic ray shower muons according to claim 3, characterized in that, In step S203, the method for identifying possible anomalous muons is as follows: 1) Based on the grid division method in step S201, obtain the coordinate set of the cell boundary line; 2) Set the recognition threshold l It sequentially determines whether the vertical distance between the point where each μ-cell passes through the projected horizontal plane and the adjacent boundary line of the cell is less than or equal to the recognition threshold. l In such cases, muons that are judged as "yes" are identified as potentially anomalous muons.

6. The method for extracting features of navigation and positioning points from cosmic ray shower muons according to claim 3, characterized in that, In step S204, the method for calculating the local density of possible anomalous muons is as follows: 1) Set the nearest neighbor number k And define any possible anomalous muon P neighborhood N k ( P () is the closest k A set of possible anomalous muons; 2) Calculate each possible anomalous muon and its neighborhood. k The reciprocal of the average distance between the points of penetration of a possible anomalous muon on the projected horizontal plane, and used as its local density (P), is expressed as: ， In the formula, k The nearest neighbor number, P For potentially anomalous muons, Q i For the possible anomalous muon neighborhood N r (P) Middle i A possible anomalous μ-subject, d ( P , Q i ) is a possible anomalous muon P With the possible anomalous muon neighborhood N r (P) Middle i One possible anomalous muon Qi The Euclidean distance between points on the projected horizontal plane.

7. The method for extracting features of navigation and positioning points from cosmic ray shower muons according to claim 3, characterized in that, In step S205, the method for determining the anomalous muon is as follows: 1) Define the LOF value as the possible anomalous muon. P Compared to it k Density differences of possible anomalous muons in the neighborhood LOF k ( P The calculation formula is as follows: ， In the formula, N k ( p ) is a possible anomalous muon P k-neighborhood, lrd( P ) is a possible anomalous muon P Local density; Q i For the possible anomalous muon P neighborhood N r (P) Middle i A possible anomalous μ-subject, lrd( Q i ) represents the possible anomalous μ-p-neighborhood N r (P) Middle i Local density of potentially anomalous muons; |N k |For the neighborhood N r The number of potentially anomalous muons in (P); 2) Set the LOF threshold and make a judgment: if the LOF calculation result of a possible abnormal muon is less than or equal to the LOF threshold, then the possible abnormal muon is a normal muon; if the LOF calculation result of a possible abnormal muon is greater than the LOF threshold, then the possible abnormal muon is an abnormal muon and is deleted.

8. The method for extracting features of navigation and positioning points from cosmic ray shower muons according to claim 1, characterized in that, The specific implementation steps of step S3 are as follows: S301. Re-estimate the muon distribution density of each cell in the grid using the method in step S2; S302. Normalize the muon distribution density of each cell in the grid to make its value normalized to the interval [0, 1]. S303. By setting the color mapping, the normalized muon distribution density is mapped to color values ​​to obtain a color muon positioning heatmap of the navigation area.

Images