High-energy ion beam three-dimensional reconstruction method and system based on secondary electron emission signal

Abstract

The invention relates to the technical field of high-current particle beam diagnosis, in particular to a high-energy ion beam three-dimensional reconstruction method and system based on secondary electron emission signals. According to the technical scheme, the method comprises the steps of data preprocessing and conversion, geometric space up-sampling, point assignment generation and model boundary division and output. According to the method, the spatial sparsity of original data is overcome through geometric upsampling, the fidelity of physical attribute reconstruction is remarkably improved by utilizing a topological structure based on a minimum spanning tree and a primary function frequency superposition model, automatic and accurate division of beam boundaries is realized according to a physical threshold value, an end-to-end automatic processing flow is formed, and the method has the advantages that the method is simple and convenient to operate and high in practicability. And outputting a visual three-dimensional model supporting interactive analysis, systematically solving the problems of data sparsity, interpolation distortion and boundary fuzziness in the prior art, and realizing high-precision and high-efficiency reconstruction and visual analysis of the three-dimensional form of the high-energy ion beam.

Description

Technical Field

[0001] This invention relates to the field of high-current particle beam diagnostic technology, and in particular to a method and system for three-dimensional reconstruction of high-energy ion beams based on secondary electron emission signals. Background Technology

[0002] In the application and research of high-current ion beams, such as inertial confinement fusion, the study of material irradiation effects, and accelerator-driven subcritical systems, accurate diagnosis of the three-dimensional spatial morphology, density distribution, and time-varying characteristics of the ion beam is crucial for optimizing beam transmission efficiency, evaluating target surface irradiation uniformity, and ensuring experimental safety. Traditional beam profile diagnosis often relies on media such as fluorescent targets, wire scanning, or imaging plates. These methods typically only provide two-dimensional projection information of the beam at a certain cross-section, making it difficult to directly obtain the three-dimensional density distribution of the beam along the propagation direction (longitudinal direction). Furthermore, they suffer from limitations such as interfering beams, limited spatial resolution, or the inability to measure in real time.

[0003] In recent years, non-interceptor diagnostic techniques based on the principle of secondary electron emission (SEE) have been developed. The basic principle is that when a high-energy ion beam grazes or bombards a specific target, secondary electrons are excited. By collecting the current or voltage signals generated by these secondary electrons, the power density information of the local beam can be deduced. By arranging one-dimensional or two-dimensional probe arrays and combining this with temporal scanning of the beam, it is theoretically possible to obtain three-dimensional spatial information of the beam with high temporal resolution. This technology has advantages such as non-interceptor capability, fast response, and long-term monitoring, providing a possible data source for three-dimensional dynamic reconstruction of the beam.

[0004] Achieving high-precision 3D reconstruction based on existing secondary electron emission signals still faces a series of key technical challenges. To avoid mutual interference between the electric fields of the probes, the secondary electron collection probes must maintain a certain interval in space (especially in the cross-sectional direction), resulting in the acquired raw signal being essentially sparse discrete sampling points in the spatial dimension. This sparsity directly limits the resolution and continuity of the direct reconstruction model, making it impossible to finely characterize the microstructure of the beam (such as beam halo, filamentary structures, etc.).

[0005] Furthermore, there is the issue of fidelity in spatial interpolation and physical property restoration. To generate a continuous 3D model from a sparse point cloud, data interpolation and densification are necessary. Existing 3D interpolation methods (such as linear interpolation based on regular grids, Kriging interpolation, or radial basis function interpolation) often focus on filling the geometric space, neglecting the specific topological structure and frequency domain characteristics of the beam's physical field (such as intensity distribution) during propagation. Simple interpolation can easily lead to poor correspondence between the physical values ​​of the generated points and the actual beam distribution, smoothing over the edges while losing key details, or introducing false artifacts, affecting the physical accuracy of the reconstructed model.

[0006] The reconstructed 3D point cloud data contains both real beam signals and background noise. The key step in obtaining a clear and reliable 3D model is to automatically and objectively extract the effective physical boundaries of the beam from the complex point cloud and remove noise points. Existing methods often rely on manually setting global thresholds, which are insufficiently adaptable to situations with large signal-to-noise ratio variations or large beam intensity gradients, easily leading to blurred boundaries, loss of effective information, or incorrect retention of noise.

[0007] In summary, this application proposes a method and system for three-dimensional reconstruction of high-energy ion beams based on secondary electron emission signals. Summary of the Invention

[0008] The purpose of this invention is to address the problem that existing technical solutions in the background art cannot fully utilize the potential of high spatiotemporal resolution of secondary electron emission signals to achieve high-fidelity and high-precision three-dimensional morphology reconstruction of high-current ion beams. This invention proposes a method and system for three-dimensional reconstruction of high-energy ion beams based on secondary electron emission signals.

[0009] Firstly, such as Figure 4 This application provides a high-energy ion beam three-dimensional reconstruction method based on secondary electron emission signals, including the following steps:

[0010] S1. Data preprocessing and conversion: Obtain the raw data of the secondary electron emission signal collected by the detector, process it and convert it into initial sparse point cloud data containing spatial coordinates and signal strength information;

[0011] S2. Geometric space upsampling: Based on the spatial range of the initial sparse point cloud data, expand to generate points to be assigned, and merge with the initial sparse point cloud data to form a high-density hybrid point cloud;

[0012] S3. Generate point assignment, establish spatial topological relationships for the hybrid point cloud, perform block processing based on the topological relationships, and use the interpolation model constructed from the initial sparse point cloud data to calculate the signal strength values ​​for the points to be assigned, thereby generating a high-density three-dimensional beam point cloud model with continuous physical properties.

[0013] S4. Model boundary division and output: Filter and divide the high-density three-dimensional beam point cloud model according to a preset physical threshold, and output and visualize the final three-dimensional beam model.

[0014] Optionally, in step S1, data preprocessing and transformation specifically include:

[0015] Read the raw signal data and fill in and correct missing and outlier values ​​by calculating the arithmetic mean within a local sliding window;

[0016] The signal peak is located by global search, and a truncation threshold is set based on the peak intensity to extract the region of interest data whose signal intensity value is greater than or equal to the truncation threshold.

[0017] The extracted two-dimensional matrix data is mapped into discrete point cloud data containing three-dimensional spatial coordinates and intensity information to form the initial sparse point cloud data.

[0018] Optionally, in step S2, geometric space upsampling specifically includes:

[0019] Calculate the minimum bounding box of the initial sparse point cloud data in three-dimensional space;

[0020] The minimum bounding box is spatially expanded to obtain the expanded spatial range;

[0021] Within the extended space, a specified number of points are randomly generated as the points to be assigned values;

[0022] The initial sparse point cloud data and the points to be assigned are merged into the hybrid point cloud.

[0023] Optionally, in step S3, a spatial topological relationship is established for the hybrid point cloud, and block processing is performed based on this topological relationship, specifically including:

[0024] A minimum spanning tree is constructed using the spatial distance between points in the hybrid point cloud as weights.

[0025] Based on the connection relationship of the minimum spanning tree, the mixed point cloud is segmented into multiple locally connected surface patches.

[0026] Optionally, in step S3, the signal strength value is calculated for the point to be assigned using the interpolation model constructed from the initial sparse point cloud data, specifically including:

[0027] For each surface patch, project the three-dimensional points within it onto a two-dimensional plane;

[0028] Using the points in the initial sparse point cloud data as anchor points, an interpolation function model is constructed using the basis function frequency superposition method, and the superposition coefficients of the interpolation function model are calculated.

[0029] Substitute the planar projection coordinates of the point to be assigned into the interpolation function model to calculate its signal strength value, and then map it back to three-dimensional space.

[0030] Optionally, the method of constructing the interpolation function model using the basis function frequency superposition method is specifically as follows:

[0031] A linear combination of basis functions of different frequencies is used to construct an interpolation function model;

[0032] Using the true strength value of the anchor point as a constraint, the superposition coefficients of the basis functions are solved by the least squares method or the gradient descent method, so that the error between the output value of the interpolation function model at the anchor point position and the true strength value is minimized.

[0033] Optionally, in step S3, when dividing the mixed point cloud into multiple surface patches, overlapping areas are retained between adjacent surface patches;

[0034] After completing the calculation and 3D mapping of all points to be assigned, the points located in the overlapping area are smoothed using a weighted average method.

[0035] Optionally, in step S4, the visualization output includes:

[0036] Pseudo-color mapping rendering is performed based on the intensity value of the point;

[0037] Generate 3D interactive views that support rotation, scaling, and slice viewing.

[0038] Secondly, this application provides a high-energy ion beam three-dimensional reconstruction system based on secondary electron emission signals, comprising:

[0039] Memory, used to store computer programs;

[0040] A processor for executing the computer program to implement the steps of the method as described in the first aspect.

[0041] Compared with the prior art, this application includes at least one of the following beneficial technical effects:

[0042] This application generates high-density point clouds in an extended space through geometric upsampling, effectively compensating for the deficiencies of insufficient spatial sampling in the original data and providing a foundation for reconstructing high-resolution 3D models.

[0043] By constructing a spatial topology using a minimum spanning tree and employing a basis function frequency superposition model for interpolation, the overall contour and local details of the beam distribution can be restored simultaneously, significantly improving the accuracy of physical values.

[0044] Automatic filtering based on physical thresholds can clearly distinguish the effective beam region from background noise, outputting a well-defined 3D solid model and solving the problem of blurred boundaries in traditional methods.

[0045] It provides an end-to-end processing flow from raw signals to 3D visualization, with a high degree of automation, significantly reducing manual intervention and improving beam diagnostic efficiency.

[0046] The output 3D model supports interactive operation and pseudo-color rendering, transforming abstract data into an intuitive form, which facilitates qualitative and quantitative analysis of beam spatiotemporal evolution.

[0047] This invention overcomes the spatial sparsity of the original data through geometric upsampling, significantly improves the fidelity of physical property reconstruction by utilizing a topological structure based on minimum spanning trees and a basis function frequency superposition model, and achieves automatic and accurate beam boundary division based on physical thresholds. The method forms an end-to-end automated processing flow and outputs a visualized 3D model that supports interactive analysis. It systematically solves the problems of data sparsity, interpolation distortion, and boundary ambiguity in existing technologies, and realizes high-precision, high-efficiency reconstruction and intuitive analysis of the 3D morphology of high-energy ion beams. Attached Figure Description

[0048] Figure 1 A flowchart of a high-energy ion beam three-dimensional reconstruction method based on secondary electron emission signals;

[0049] Figure 2 This is a block diagram illustrating the principle of the secondary electronic signal acquisition system in the embodiment.

[0050] Figure 3 This is a schematic diagram of the secondary electron probe and its installation method used in the embodiment;

[0051] Figure 4 This is a flowchart of the algorithm execution for a high-energy ion beam three-dimensional reconstruction method based on secondary electron emission signals. Detailed Implementation

[0052] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0053] Not mentioned in the following embodiments Figure 2 , Figure 3 and Figure 4 ,Replenish Figure 2 , Figure 3 and Figure 4 Write the content obtained from the diagram in the appropriate position corresponding to the embodiment.

[0054] Example: Figure 1 This invention proposes a high-energy ion beam three-dimensional reconstruction method based on secondary electron emission signals, which will be described in detail below.

[0055] Implementation Environment: Hardware: Memory >= 8GB; Software: Windows or Linux, implemented using Python, requiring Python 3.10, NumPy >= 1.20.0, SciPy >= 1.7.0, Open3D >= 0.13.0, NetworkX >= 2.6, and Matplotlib >= 3.3.0.

[0056] Input data: A schematic diagram of the secondary electronic data acquisition system is shown below. Figure 2 As shown, the installation method of the secondary electron collection probe is as follows: Figure 3 As shown, the input data is a high time-resolution signal collected by the secondary electron collection probe array, containing all information from the time period from beam exit to beam termination.

[0057] Step-by-step implementation process:

[0058] S1. Data Preprocessing and Transformation

[0059] This step aims to convert the discrete, noisy raw voltage / current signals acquired by the detector (usually stored in .csv format) into a 3D point cloud data structure (.ply format) that can be processed by computer graphics, laying the foundation for subsequent 3D reconstruction. Figure 2 The flowchart shown is for this algorithm, which includes the following sub-steps:

[0060] Step 1.1: Raw data cleaning and outlier correction

[0061] Data Reading: Reads the raw input file. This file contains secondary electron emission signal intensity data arranged in time series.

[0062] Outlier detection: Traverse the original dataset to identify outliers. Outliers are defined as values ​​that significantly deviate from physical expectations (e.g., negative values, maximum values ​​exceeding the detector's range, or values ​​that deviate from the local window mean by more than 3δ).

[0063] Missing value and outlier imputation:

[0064] Outliers that are identified are removed.

[0065] The mean imputation method is used for correction: a sliding window (a 3×3 matrix) is set, the arithmetic mean of the effective neighborhood data around the location of the outlier or missing value is calculated, and the value at the target location is replaced with this mean.

[0066] Technical effect: Eliminates spike pulses caused by detector electronic noise or transmission interference, ensuring the continuity and smoothness of data in spatial distribution.

[0067] Step 1.2: Effective data extraction based on peak features

[0068] Background: Since the detector's scanning range is usually larger than the actual beam pulse width, the raw data contains a large amount of invalid background noise data (i.e., periods when no beam is extracted).

[0069] Peak location: After scanning and cleaning the dataset, a global search algorithm is used to find the flat-top segment of signal strength. .

[0070] Region of Interest (ROI) Extraction:

[0071] by Centered on a preset threshold strategy (peak intensity), Determine the truncation threshold for the valid signal. .

[0072] Retain all strength values Data points, The data was treated as background noise and removed.

[0073] Technical effects: It effectively reduces the amount of data processing, removes background noise from interfering with the accuracy of subsequent 3D reconstruction, and precisely locks the effective range of the beam.

[0074] Step 1.3: Point Cloud Format Conversion

[0075] Coordinate mapping: Establish a three-dimensional Cartesian coordinate system. Then, extract the truncated two-dimensional matrix data. Mapped to spatial coordinates The signal strength value at the corresponding location Mapped to third-dimensional coordinates Alternatively, it can be retained as a scalar property (ScalarField), thus forming the initial three-dimensional scatter set. .

[0076] PLY file construction: Calling a custom csv_to_ply conversion utility class:

[0077] Write header information: Define the PLY file format version (e.g., ascii 1.0), element type (vertex), attribute definition (x, y, z coordinates and intensity value), and number of points.

[0078] Write vertex data (Body): This will map the data to the body. The collection data is written to a file line by line.

[0079] Output: Generates an initial sparse point cloud file, which, although sparse, accurately preserves the geometric topology of the original measurement data.

[0080] S2. Geometric space upsampling

[0081] This step is performed by the geometric upsampling module proposed in this invention. This module aims to address the problem that the raw probe data is sparsely distributed in three-dimensional space, making it impossible to finely characterize the beam's microstructure. It does not rely on traditional grid interpolation, but instead generates randomly distributed "blank points" (i.e., points with only spatial coordinates) within an extended spatial range. An intensity value has not yet been assigned. (These points) provide a high-density spatial carrier for subsequent physical property reconstruction. The specific steps are as follows:

[0082] Step 2.1: Spatial envelope calculation and range expansion

[0083] Input: The initial sparse point cloud received from the output of step S1. .

[0084] Envelope calculation: traversal For all points in the dataset, calculate their maximum and minimum values ​​in the X, Y, and Z dimensions to determine the minimum bounding box (AABB) of the original data.

[0085] Spatial expansion: In order to avoid losing detailed information of the beam edge during reconstruction and considering the divergence characteristics of the beam during transmission, the minimum bounding box is expanded.

[0086] Set the extension factor. or fixed margin ).

[0087] The new spatial extent is defined as:

[0088]

[0089] Step 2.2: Setting upsampling parameters

[0090] Upsampling factor setting: The user or system sets the upsampling factor (Upsampling Multiple) according to the required reconstruction accuracy. ).

[0091] Target point calculation: Calculates the total number of points generated for the target. ,in The initial number of points in the sparse point cloud.

[0092] Step 2.3: Generating random blank points

[0093] Random sampling: Within the extended space determined in step 2.1, a random number generator (such as a uniformly distributed or Gaussian distributed random function) is used to generate... Three-dimensional coordinate points.

[0094] Blank Point Definition: These newly generated points are marked as "points to be assigned" or "blank points". They have explicit spatial location information, but their beam intensity properties are initialized to 0, awaiting calculation and assignment in step S3.

[0095] Step 2.4: Point Cloud Fusion

[0096] Set merging: merging the original sparse point cloud As anchor points and newly generated blank point clouds Merge to form a hybrid point cloud set .

[0097] Output: Output the hybrid point cloud set to the S3 module. At this point, the point cloud has high-density features in geometric space, providing a foundation for subsequent construction of a fine topology (MST).

[0098] S3. Generating point assignments (based on the improved see-fsmmr algorithm module)

[0099] This step aims to assign accurate beam intensity values ​​to the "blank points" (points with only spatial coordinates but no physical properties) generated by geometric upsampling in step S2. This invention employs the see-fsmmr module from the improved FSMMR (Frequency-Selective Mesh-to-Mesh Resampling) algorithm, constructing an interpolation model through the superposition of basis function frequencies. The specific implementation steps are as follows:

[0100] Step 3.1: Construct local topological relationships (minimum spanning tree)

[0101] Input: Receive dense 3D point cloud data (including original measurement points and sampling points to be assigned values) after processing in step S2.

[0102] Distance calculation: Calculate the Euclidean distance between any two points in a point cloud.

[0103] Constructing the MST: Using spatial distance as weights, construct the minimum spanning tree (MST) of the point cloud using Prim's algorithm or Kruskal's algorithm.

[0104] Technical effect: Minimum spanning tree can effectively capture the topological structure of discrete point clouds in three-dimensional space, ensuring that subsequent data block processing can follow the natural geometric distribution of the data and avoid erroneous associations across non-adjacent areas.

[0105] Step 3.2: Patch Generation

[0106] Traversal and Segmentation: Based on the constructed minimum spanning tree structure, a specific segmentation threshold (such as node depth or path length) is set to segment the global point cloud of the entire beam into several locally connected surface patches.

[0107] Overlap processing: To ensure a smooth transition of the interpolation function at the boundary, a certain proportion of overlap area is set between adjacent patches.

[0108] Technical effect: By decomposing the complex global fitting problem into multiple simple local fitting problems, the computational complexity is significantly reduced and the accuracy of local feature restoration is improved.

[0109] Step 3.3: Dimensional Reduction Projection

[0110] Loop processing: Iterate through each divided patch.

[0111] Projection transformation: Define the projection plane (usually along the beam propagation direction). The plane whose normal vector is the axis, i.e. (Plane). The 3D points in the patch. along Projecting along the axis transforms it into a two-dimensional planar point set. .

[0112] Objective: To transform the three-dimensional surface fitting problem into a two-dimensional plane function interpolation problem, which facilitates the application of the basis function superposition algorithm.

[0113] Step 3.4: Iterative calculation of basis function superposition coefficients

[0114] Reference point selection: In the current patch, the "original measurement points" from step S1 are identified as anchor points. These points have real and reliable intensity values.

[0115] Model Construction: Constructing a basis function model based on frequency superposition. ,in These are basis functions of different frequencies (such as radial basis functions RBF or trigonometric basis functions). This is the corresponding superposition coefficient.

[0116] Iterative solution:

[0117] A system of linear equations is established using the actual values ​​of the anchor points as constraints.

[0118] The superposition coefficient is calculated iteratively using the least squares method or gradient descent method. This makes the model Minimize the error between the output value and the actual measurement value at the anchor point.

[0119] Technical effect: By utilizing the concept of frequency domain superposition, it is possible to simultaneously capture the low-frequency contour information (overall trend) and high-frequency detail information (local fluctuations) of the beam distribution, making the generated point cloud effect closer to the real value.

[0120] Step 3.5: Generate point numerical calculations

[0121] Interpolation assignment: The planar coordinates of the "blank points" generated in step S2. Substitute the function model obtained in step 3.4 In the calculation, the predicted beam intensity value at that location is obtained. .

[0122] Assignment operation: Assign corresponding blank points to complete the transformation from "geometric points" to "physical points".

[0123] Step 3.6: 3D Structure Reconstruction

[0124] Reverse mapping: Mapping a set of two-dimensional points that has been assigned values ​​back to three-dimensional space to restore its original spatial topology.

[0125] Merging: All processed patches are merged, and points in overlapping areas are smoothed using a weighted average method to ultimately form a complete, high-density 3D beam point cloud model with continuous physical properties.

[0126] S4. Model Boundary Delineation and Output

[0127] This step aims to perform final trimming and visualization of the reconstructed 3D model, clarify the effective physical boundaries of the beam, and provide an intuitive interactive interface.

[0128] Step 4.1: 3D Visualization Rendering

[0129] Visualization engine: Calls computer graphics visualization libraries (in specific embodiments, the Visualization class of the Open3D open source library can be used).

[0130] Rendering beamlines:

[0131] Load the final point cloud file.

[0132] Color mapping: Applying pseudo-color mapping (such as Jet or Viridis spectrum) based on the intensity value of points to visually display the distribution differences between the beam core region (highlight) and the edge region (cool color).

[0133] Interactive Display: Generates interactive 3D views, supporting users to perform operations such as rotation, zoom, and slice viewing, enabling intuitive monitoring and analysis of the 3D structure of high-energy ion beams.

[0134] This embodiment also provides a high-energy ion beam three-dimensional reconstruction system based on secondary electron emission signals, including:

[0135] Memory, used to store computer programs;

[0136] A processor is used to execute the computer program to implement the steps of a high-energy ion beam three-dimensional reconstruction method based on secondary electron emission signals as described above.

[0137] like Figure 4 The flowchart below illustrates the algorithm of this invention. It is worth noting that this invention achieves 3D reconstruction based on secondary electron emission signals. The principle involves converting time-series acquired voltage signals into an initial sparse point cloud containing spatial location and intensity through coordinate mapping. By calculating and expanding the spatial envelope of the point cloud, points to be assigned are randomly generated within the expanded range, achieving density enhancement in geometric space. To accurately restore physical properties, a minimum spanning tree is constructed using the Euclidean distance between points as weight. Based on topological relationships, the point cloud is segmented into locally connected surface patches. Projection reduces the 3D fitting problem to 2D. Using points in the initial sparse point cloud as anchor points, an interpolation function is established using the frequency superposition of basis functions. After solving for the coefficients, intensity values ​​are assigned to the points to be assigned, and then mapped back to 3D space, forming a 3D model with both high spatial density and continuous physical properties. Finally, background noise points are automatically filtered out based on a preset intensity threshold, achieving clear beam boundary delineation and outputting a visualized 3D model that supports interactive operation. The implementation of this method can directly reconstruct high-resolution beam three-dimensional structures from sparse signals, improve the accuracy of physical quantity reconstruction, and achieve objective and automatic boundary determination.

[0138] The above specific embodiments are merely several optional embodiments of the present invention. Based on the technical solutions of the present invention and the relevant teachings of the above embodiments, those skilled in the art can make various alternative improvements and combinations to the above specific embodiments.

Claims

1. A method for three-dimensional reconstruction using a high-energy ion beam based on secondary electron emission signals, characterized in that, Includes the following steps: S1. Data preprocessing and conversion: Obtain the raw data of the secondary electron emission signal collected by the detector, process it and convert it into initial sparse point cloud data containing spatial coordinates and signal strength information; S2. Geometric space upsampling: Based on the spatial range of the initial sparse point cloud data, expand to generate points to be assigned, and merge with the initial sparse point cloud data to form a high-density hybrid point cloud; S3. Generate point assignment, establish spatial topological relationships for the hybrid point cloud, perform block processing based on the topological relationships, and use the interpolation model constructed from the initial sparse point cloud data to calculate the signal strength values ​​for the points to be assigned, thereby generating a high-density three-dimensional beam point cloud model with continuous physical properties. S4. Model boundary division and output: Filter and divide the high-density three-dimensional beam point cloud model according to a preset physical threshold, and output and visualize the final three-dimensional beam model.

2. The high-energy ion beam three-dimensional reconstruction method based on secondary electron emission signals according to claim 1, characterized in that, In step S1, data preprocessing and transformation specifically include: Read the raw signal data and fill in and correct missing and outlier values ​​by calculating the arithmetic mean within a local sliding window; The signal peak is located by global search, and a truncation threshold is set based on the peak intensity to extract the region of interest data whose signal intensity value is greater than or equal to the truncation threshold. The extracted two-dimensional matrix data is mapped into discrete point cloud data containing three-dimensional spatial coordinates and intensity information to form the initial sparse point cloud data.

3. The high-energy ion beam three-dimensional reconstruction method based on secondary electron emission signals according to claim 1, characterized in that, In step S2, geometric space upsampling specifically includes: Calculate the minimum bounding box of the initial sparse point cloud data in three-dimensional space; The minimum bounding box is spatially expanded to obtain the expanded spatial range; Within the extended space, a specified number of points are randomly generated as the points to be assigned values; The initial sparse point cloud data and the points to be assigned are merged into the hybrid point cloud.

4. The high-energy ion beam three-dimensional reconstruction method based on secondary electron emission signals according to claim 1, characterized in that, In step S3, a spatial topological relationship is established for the hybrid point cloud, and block processing is performed based on this topological relationship, specifically including: A minimum spanning tree is constructed using the spatial distance between points in the hybrid point cloud as weights. Based on the connection relationship of the minimum spanning tree, the mixed point cloud is segmented into multiple locally connected surface patches.

5. The high-energy ion beam three-dimensional reconstruction method based on secondary electron emission signals according to claim 4, characterized in that, In step S3, the signal strength values ​​for the points to be assigned are calculated using the interpolation model constructed from the initial sparse point cloud data. Specifically, this includes: For each surface patch, project the three-dimensional points within it onto a two-dimensional plane; Using the points in the initial sparse point cloud data as anchor points, an interpolation function model is constructed using the basis function frequency superposition method, and the superposition coefficients of the interpolation function model are calculated. Substitute the planar projection coordinates of the point to be assigned into the interpolation function model to calculate its signal strength value, and then map it back to three-dimensional space.

6. The high-energy ion beam three-dimensional reconstruction method based on secondary electron emission signals according to claim 5, characterized in that, The method of constructing the interpolation function model using the basis function frequency superposition method is as follows: A linear combination of basis functions of different frequencies is used to construct an interpolation function model; Using the true strength value of the anchor point as a constraint, the superposition coefficients of the basis functions are solved by the least squares method or the gradient descent method, so that the error between the output value of the interpolation function model at the anchor point position and the true strength value is minimized.

7. The high-energy ion beam three-dimensional reconstruction method based on secondary electron emission signals according to claim 5, characterized in that, In step S3, when dividing the mixed point cloud into multiple surface patches, overlapping areas are preserved between adjacent surface patches; After completing the calculation and 3D mapping of all points to be assigned, the points located in the overlapping area are smoothed using a weighted average method.

8. The high-energy ion beam three-dimensional reconstruction method based on secondary electron emission signals according to claim 1, characterized in that, In step S4, the visualization output includes: Pseudo-color mapping rendering is performed based on the intensity value of the point; Generate 3D interactive views that support rotation, scaling, and slice viewing.

9. A high-energy ion beam three-dimensional reconstruction system based on secondary electron emission signals, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the method steps as described in any one of claims 1 to 8.

Images