PBO composite material performance analysis method and system based on CT scanning and slice three-dimensional reconstruction data analysis

Abstract

This invention discloses a method and system for performance analysis of PBO composite materials based on CT scans and 3D reconstruction data analysis of PBO slices. Specifically, it establishes a database corresponding to the relationship between CT images and 3D reconstruction models of PBO slices; it delineates bubbles in the corresponding images and models and analyzes their morphological and positional characteristics to construct a bubble relationship analysis model; and it uses this model to analyze bubbles in real-time CT images, determining the real-time 3D bubble distribution as performance analysis data. This invention can accurately obtain the 3D bubble distribution non-destructively, improving the accuracy and efficiency of PBO composite material performance analysis.

Description

Technical Field

[0001] This invention relates to the field of PBO composite material analysis technology, and in particular to a method and system for PBO composite material performance analysis based on CT scans and 3D reconstruction data analysis of slices. Background Technology

[0002] PBO fiber-reinforced composites are widely used in aerospace, armor protection, and high-performance structural components due to their excellent specific strength, specific modulus, high temperature resistance, and impact resistance. However, air bubbles (pores) inevitably occur during the molding process. The size, morphology, spatial distribution, and especially the relative position of these bubbles with PBO fibers significantly affect the interfacial bonding strength, mechanical properties, and long-term reliability of the composite material. Currently, the analysis of air bubbles inside PBO composites mainly relies on two types of methods: one is CT scanning technology, which has the advantages of being non-destructive and quickly obtaining two-dimensional slice images of the material's interior. However, due to the limitations of the projection principle and resolution of CT imaging, it is difficult to accurately reconstruct the true three-dimensional morphology, precise spatial position, and complex relative relationship with the fibers of the air bubbles, resulting in insufficient accuracy in performance evaluation. The other is a three-dimensional reconstruction method that combines physical layer-by-layer slicing with microscopic observation. Although it can obtain a high-precision three-dimensional model containing the true morphology of air bubbles and the position of fibers, it is a destructive test, cumbersome and time-consuming, and cannot meet the needs of online or real-time analysis. Summary of the Invention

[0003] The purpose of this invention is to provide a performance analysis method and system for real-time analysis and judgment of the properties of PBO composite materials using CT scanning technology.

[0004] This invention discloses a method for analyzing the properties of PBO composite materials based on CT scans and 3D reconstruction data of slices, including: Step S100: Establish a database corresponding to the relationship between CT images of PBO composite materials and three-dimensional reconstruction models of PBO slices. Step S200: In the relational database, bubble delineation is performed on the corresponding CT images and PBO 3D reconstruction models, denoted as image bubbles and model bubbles. The morphological and positional relationship features between image bubbles and model bubbles are analyzed to construct a bubble relationship analysis model. Step S300: Using the bubble relationship analysis model, the bubbles in the real-time CT image are analyzed to determine the real-time three-dimensional bubble distribution of the PBO composite material, and the real-time three-dimensional bubble distribution is used as the performance analysis data of the PBO composite material.

[0005] In some embodiments disclosed in this invention, the method for constructing a three-dimensional reconstruction model of PBO slices includes: Step S101: Perform physical slicing of the PBO composite material sample layer by layer according to preset values, observe each slice under a microscope, obtain microscopic images, and delineate model bubbles and PBO fibers in the microscopic images. Step S102: Spatial alignment and three-dimensional stacking reconstruction of several layers of microscopic images are performed according to the slice order to obtain a three-dimensional reconstruction model of PBO slices that simultaneously contains the actual bubble morphology and PBO fiber position.

[0006] In some embodiments disclosed in this invention, the method for analyzing the morphological relationship features between image bubbles and model bubbles includes: Step S201: Construct a two-dimensional positioning coordinate system for the CT image and a three-dimensional positioning coordinate system for the PBO three-dimensional reconstruction model, and establish the positional correspondence between the two-dimensional positioning coordinate system and the three-dimensional positioning coordinate system. Specifically, this includes uniformly setting several first positioning base point coordinates in the two-dimensional coordinate system and uniformly setting several second positioning base point coordinates in the three-dimensional coordinate system, and associating and recording the several second base point coordinates mapped by the first coordinate base point coordinates. Step S202: Determine the center point of the first bubble of the first bubble in the CT image, and determine the second coordinate base point sequence corresponding to the center point of the first bubble based on the coordinate mapping relationship between the positioning coordinate systems. Determine the center point of the second bubble mapped by the second coordinate base point sequence, and then determine a number of second bubbles associated with the position of the first bubble. Step S203: Analyze the edge morphology of the first bubble to determine its morphological characteristics, and analyze the edge morphology of the associated second bubble relative to the projection plane to determine its morphological characteristics, wherein the projection plane is parallel to the CT image. Step S204: Match the morphological features of the first bubble and the second bubble to determine the morphological correlation parameters between them.

[0007] In some embodiments disclosed in this invention, the method for matching and analyzing the morphological features of the first bubble and the second bubble includes: Step S2041: A number of distance detection lines are uniformly emitted at the center point of the bubble, and the edge distance from the center point of the bubble to the edge of the bubble is determined. The edge distances are summarized in order, and the order is determined according to the rotation order of the distance detection lines to form an edge distance sequence. The distance detection lines are emitted in a relative positioning coordinate system, and a distance detection line radiation template is constructed. The distance detection line radiation template includes a radiation center point and a number of distance detection lines. Step S2042: Compare the edge distance sequences corresponding to the first bubble and the second bubble respectively. The comparison method is to compare the edge distances corresponding to the detection lines of equal distance. If the distance difference between the two is less than or equal to a preset value, the corresponding edge distance is marked as the matching edge distance. Step S2043: Analyze the distribution of the matching edge distance in the edge distance sequence of the first bubble and the second bubble, determine the first continuous distribution index and the second continuous distribution index in the first bubble and the second bubble, and determine the morphological correlation parameters of the morphological features of the first bubble and the second bubble based on the first continuous distribution index and the second continuous distribution index. Among them, the continuous distribution index refers to the effective proportion of each continuous distance group in the edge distance sequence, and combines the close performance of the first continuous distribution index and the second continuous distribution index to determine the morphological correlation parameter.

[0008] In some embodiments disclosed in this invention, the method for determining morphological correlation parameters includes: Step S20431: Determine the continuous matching edge distance. If the proportion of the number of edge distances between each other is greater than or equal to a preset value, the continuous matching edge distance is considered to be interrupted. Statistically count the continuous matching edge distances to form a continuous distance group. If the proportion of the number of matching edge distances in the continuous distance group relative to the total number of edge distances is less than or equal to a preset value, the continuous distance group is removed. Step S20432: Count the number of matching edge distances in the retained continuous distance group and record them as effective matching edge distances. Calculate the effective proportion of effective matching edge distances relative to all matching edge distances and determine the continuous distribution index based on the preset proportion interval to which the effective proportion belongs. Step S20433: Calculate the index difference between the first continuous distribution index and the second continuous distribution index, and based on the index difference, correct the average value of the first continuous distribution index and the second continuous distribution index to obtain the morphological correlation parameter. The expression for calculating the morphological correlation parameter is as follows: ; Where G is the morphological correlation parameter, and K is the adjustment coefficient for the influence of exponential difference. Let X be the preset maximum exponential difference, and X be the exponential difference. It is the first continuous distribution index. It is the second continuous distribution index.

[0009] In some embodiments disclosed in this invention, the method for analyzing the positional relationship features between image bubbles and model bubbles includes: Step S205: Perform position analysis on the second bubble center point mapped by the first bubble center point. The analysis method includes determining the fiber distance between the second bubble center point and the PBO fiber, and determining the position association parameters of the corresponding model bubble based on the preset distance interval to which the fiber distance belongs.

[0010] In some embodiments disclosed in this invention, the method for constructing a bubble relationship analysis model includes: Step S206: Based on the morphological association parameters and the positional association parameters, determine whether there is an equivalent association between the image bubble and the model bubble. If so, associate the two equivalently. Step S207: Use the location nodes, shape representation, and size parameters corresponding to the image bubbles and model bubbles that are equivalently associated as retrieval data to construct a bubble relationship analysis model corresponding to the image bubbles and model bubbles.

[0011] In some embodiments disclosed in this invention, the method for constructing a bubble relationship analysis model includes: Step S208: Perform multi-dimensional feature normalization on the position nodes, shape representation, and size parameters of the image bubbles and model bubbles that are equivalently associated, to form bubble feature descriptors of a uniform scale. Specifically, this includes relative coordinate system transformation of position nodes, edge distance sequence compression encoding of shape representation, volume equivalent normalization of size parameters, and weighted fusion of the normalized descriptors according to preset weights to obtain a comprehensive feature retrieval vector. Step S209: Based on the comprehensive feature retrieval vector, construct a hierarchical retrieval index structure. Use the comprehensive feature retrieval vector of the image bubble as the query end index and the comprehensive feature retrieval vector of the model bubble as the data end index. Match and associate the two through a preset similarity threshold to form a bubble mapping relationship table that can be quickly retrieved, and construct the bubble relationship analysis model.

[0012] In some embodiments disclosed in this invention, a PBO composite material performance analysis system based on CT scans and three-dimensional reconstruction data analysis of slices is also disclosed, including: The first module is used to establish a database that corresponds to the relationship between CT images of PBO composite materials and 3D reconstruction models of PBO slices. The second module is used to delineate bubbles in the corresponding CT images and PBO 3D reconstruction models in the relational database, denoted as image bubbles and model bubbles. It analyzes the morphological and positional relationship features between image bubbles and model bubbles and constructs a bubble relationship analysis model. The third module is used to analyze bubbles in real-time CT images using a bubble relationship analysis model, determine the real-time three-dimensional bubble distribution of PBO composite materials, and use the real-time three-dimensional bubble distribution as performance analysis data of PBO composite materials.

[0013] This invention discloses a method and system for performance analysis of PBO composite materials based on CT scans and 3D reconstruction data analysis of PBO slices. Specifically, it establishes a database corresponding to the relationship between CT images and 3D reconstruction models of PBO slices; it delineates bubbles in the corresponding images and models and analyzes their morphological and positional characteristics to construct a bubble relationship analysis model; and it uses this model to analyze bubbles in real-time CT images, determining the real-time 3D bubble distribution as performance analysis data. This invention can accurately obtain the 3D bubble distribution non-destructively, improving the accuracy and efficiency of PBO composite material performance analysis.

[0014] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the method steps of the performance analysis method for PBO composite materials based on CT scan and slice three-dimensional reconstruction data analysis disclosed in this embodiment of the invention. Figure 2 A schematic diagram of the principle behind edge distance sequence formation; Figure 3 This is a flowchart for real-time CT image processing and performance data acquisition. Detailed Implementation

[0016] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0017] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings and specific embodiments. It should be understood that the preferred embodiments described herein are only for illustration and explanation of the present invention and should not be construed as limiting the scope of protection of the present invention. Those skilled in the art can make some non-essential improvements and adjustments based on the following content of the present invention. In the present invention, unless otherwise expressly specified and limited, the technical terms used in the present invention should have the ordinary meaning understood by those skilled in the art.

[0018] Example: The purpose of this invention is to provide a performance analysis method and system for real-time analysis and judgment of the properties of PBO composite materials using CT scanning technology.

[0019] This invention discloses a method for performance analysis of PBO composite materials based on CT scans and 3D reconstruction data analysis of slices. (See attached document.) Figure 1 ,include: Step S100: Establish a database corresponding to the relationship between CT images of PBO composite materials and three-dimensional reconstruction models of PBO slices.

[0020] The core principle of this step lies in constructing cross-modal data correspondences, combining the rapid acquisition capabilities of non-destructive CT imaging with high-fidelity 3D information from physical slice 3D reconstruction to form a reliable prior knowledge base. Specifically, for the same or representative PBO composite material sample, a 2D projection image is first obtained using CT scanning. Then, a 3D model containing the real bubble morphology, location, and spatial distribution of PBO fibers is reconstructed through layer-by-layer physical slicing and microscopic observation. Subsequently, spatial registration and feature matching are used to associate and store the two, thereby establishing a database that reflects the inherent mapping law between CT image features and the real 3D structure. This database essentially acts as a "training sample set," providing the necessary calibration basis for inferring the complete 3D bubble distribution from limited 2D CT information, thus solving the fundamental problem of insufficient 3D reconstruction accuracy from single CT data.

[0021] Step S200: In the relational database, bubble delineation is performed on the corresponding CT images and PBO 3D reconstruction models, denoted as image bubbles and model bubbles. The morphological and positional relationship features between image bubbles and model bubbles are analyzed to construct a bubble relationship analysis model.

[0022] The core principle of this step is to perform bubble-level fine feature extraction and correlation analysis on aligned CT images and 3D reconstructed models, quantifying the morphological and positional correspondence between 2D projected bubbles and real 3D bubbles, and then abstracting a reusable relationship model. First, bubble regions are delineated in the corresponding images and models, obtaining image bubbles and model bubbles respectively. Then, a mapping relationship between 2D and 3D coordinate systems is established. Through methods such as center point localization, edge distance sequence extraction, and continuous distribution index calculation, the degree of conformity between the two in terms of edge contour shape is quantified (morphological relationship features). Simultaneously, the spatial relative position of the bubble center relative to the PBO fiber is determined (positional relationship features). Finally, the morphological correlation parameters and positional correlation parameters are fused to determine whether the bubbles are equivalently correlated, and based on this, a bubble relationship analysis model is constructed that includes position nodes, shape representation, size parameters, and comprehensive feature retrieval vectors. This model is essentially a mapping function based on multi-dimensional feature matching, capable of capturing the statistical correspondence between CT projection distortion and real 3D structures, providing a reliable decision-making basis for real-time inference.

[0023] Step S300, see Figure 3By using a bubble relationship analysis model, the bubbles in real-time CT images were analyzed to determine the real-time three-dimensional bubble distribution of PBO composite materials, and the real-time three-dimensional bubble distribution was used as the performance analysis data of PBO composite materials.

[0024] The core principle of this step is to use the established bubble relationship analysis model as prior knowledge and apply it to new real-time CT images to achieve a rapid and non-destructive inversion from two-dimensional projected bubbles to a three-dimensional true distribution. Specifically, bubble detection and feature extraction are performed on the real-time acquired CT images. Using pre-built morphology and position association rules and comprehensive feature retrieval vectors in the model, the corresponding model bubble information is quickly located through similarity matching, thereby recovering the true three-dimensional morphology, spatial position, and relative relationship with the fibers of each bubble. Finally, complete real-time three-dimensional bubble distribution data within the material is obtained. This three-dimensional distribution data directly reflects the potential impact of bubbles on the mechanical properties of PBO composite materials, and therefore can be directly used as a key input for performance analysis. It achieves near-physical slicing accuracy in obtaining three-dimensional defect information without damaging the sample, significantly improving the timeliness and accuracy of performance evaluation.

[0025] Using a bubble relationship analysis model, bubbles in real-time CT images were analyzed to determine the real-time three-dimensional bubble distribution of the PBO composite material. This real-time three-dimensional bubble distribution was then used as performance analysis data for the PBO composite material, specifically including: Real-time CT images of PBO composite materials were acquired, and bubble delineation was performed on the real-time CT images to obtain several real-time image bubbles. The center point, edge morphology, and size parameters of each real-time image bubble were then determined. The relative coordinate system transformation of the location nodes of the real-time image bubble is performed, the edge distance sequence compression encoding is performed on the shape representation, and the volume equivalent normalization is performed on the size parameter to form a real-time bubble feature descriptor with a uniform scale. Then, the bubble is weighted and fused according to the preset weight to obtain a real-time comprehensive feature retrieval vector. Based on the hierarchical retrieval index structure in the bubble relationship analysis model, the real-time comprehensive feature retrieval vector is used as the query end index and similarity matching is performed with the data end index in the model. Model bubbles that are equivalent to real-time image bubbles are selected by using a preset similarity threshold. Based on the real three-dimensional location nodes, shape representation and size parameters corresponding to the model bubbles obtained by matching, the three-dimensional information of the real-time image bubbles is restored, the shape, position and relative relationship of each real-time image bubble in three-dimensional space with PBO fibers are determined, and then the real-time three-dimensional bubble distribution of PBO composite material is obtained. The determined real-time three-dimensional bubble distribution was directly used as the performance analysis data for the PBO composite material.

[0026] In some embodiments disclosed in this invention, the method for constructing a three-dimensional reconstruction model of PBO slices includes: Step S101: Perform physical slicing of the PBO composite material sample layer by layer according to preset values, observe each slice under a microscope, obtain microscopic images, and delineate model bubbles and PBO fibers in the microscopic images.

[0027] The core principle of this step lies in combining destructive physical layer-by-layer slicing with high-resolution microscopic imaging to obtain high-precision two-dimensional true cross-sectional information of the internal structure of PBO composite materials, thereby compensating for the shortcomings of CT imaging in terms of resolution and contrast. Specifically, the sample is precisely removed layer by layer according to a preset slice thickness, so that each newly exposed surface can be clearly observed and imaged by the microscopic equipment. Subsequently, the bubble region and PBO fiber outline are accurately delineated in the obtained microscopic image. This process directly captures the true microscopic morphology of the material, including the actual edge shape and size of the bubbles, as well as the accurate position and orientation of the fibers, providing a reliable original two-dimensional data foundation for subsequent three-dimensional reconstruction, ensuring that the final model can truly reflect the spatial relationship between internal defects and reinforcing phases in the material.

[0028] Step S102: Spatial alignment and three-dimensional stacking reconstruction of several layers of microscopic images are performed according to the slice order to obtain a three-dimensional reconstruction model of PBO slices that simultaneously contains the actual bubble morphology and PBO fiber position.

[0029] The core principle of this step is to utilize spatial registration and volume data reconstruction techniques in digital image processing to transform a discrete two-dimensional microscopic cross-sectional sequence into a continuous three-dimensional volume model, thereby restoring the complete morphology and relative positional relationships of bubbles and fibers in real three-dimensional space. By rigidly or non-rigidly aligning adjacent microscopic images to eliminate displacement and deformation errors during slicing, and then precisely stacking and interpolating along the thickness direction according to the original slicing order, a reconstructed model is finally generated that simultaneously contains the real three-dimensional morphology and position of bubbles and the three-dimensional distribution of PBO fibers. This model, as high-fidelity ground-based data, provides an accurate three-dimensional reference benchmark for establishing a correspondence with CT images, enabling the effective mapping of two-dimensional CT projection features onto real three-dimensional structural information.

[0030] In some embodiments disclosed in this invention, the method for analyzing the morphological relationship features between image bubbles and model bubbles includes: Step S201: Construct a two-dimensional positioning coordinate system for the CT image and a three-dimensional positioning coordinate system for the PBO three-dimensional reconstruction model, and establish the positional correspondence between the two-dimensional positioning coordinate system and the three-dimensional positioning coordinate system. Specifically, this includes uniformly setting several first positioning base point coordinates in the two-dimensional coordinate system and uniformly setting several second positioning base point coordinates in the three-dimensional coordinate system, and associating and recording the several second base point coordinates mapped by the first coordinate base point coordinates.

[0031] Step S202: Determine the center point of the first bubble of the first bubble in the CT image, and determine the second coordinate base point sequence corresponding to the center point of the first bubble based on the coordinate mapping relationship between the positioning coordinate systems. Determine the center point of the second bubble mapped by the second coordinate base point sequence, and then determine a number of second bubbles associated with the position of the first bubble.

[0032] Step S203: Analyze the edge morphology of the first bubble to determine its morphological characteristics, and analyze the edge morphology of the associated second bubble relative to the projection plane to determine its morphological characteristics, wherein the projection plane is parallel to the CT image.

[0033] Step S204: Match the morphological features of the first bubble and the second bubble to determine the morphological correlation parameters between them.

[0034] In some embodiments disclosed in this invention, see [link / reference]. Figure 2 The method for matching and analyzing the morphological features of the first bubble and the second bubble includes: Step S2041: A number of distance detection lines are uniformly emitted at the center point of the bubble, and the edge distance from the center point of the bubble to the edge of the bubble is determined. The edge distances are summarized in order, and the order is determined according to the rotation order of the distance detection lines to form an edge distance sequence. The distance detection lines are emitted in a relative positioning coordinate system, and a distance detection line radiation template is constructed. The distance detection line radiation template includes a radiation center point and a number of distance detection lines.

[0035] The core principle of this step lies in transforming the two-dimensional edge contour of the bubble into a rotation-invariant and easily quantifiable one-dimensional feature sequence, thereby achieving a digital description of the irregular bubble morphology. Specifically, a distance detection line radiation template is constructed at the bubble's center point, and multiple rays are uniformly emitted along the circumference. The length of each ray from the center to the edge is measured, and these length values ​​are arranged sequentially according to the rotation order of the rays to form an edge distance sequence. This sequence fully encodes the radial dimension changes of the bubble in all directions, preserving both the local details of the morphology and eliminating the influence of absolute coordinate positions, providing a unified and computable feature representation for subsequent morphological similarity comparisons between two bubbles.

[0036] Step S2042: Compare the edge distance sequences corresponding to the first bubble and the second bubble respectively. The comparison method is to compare the edge distances corresponding to the detection lines of equal distance. If the distance difference between the two is less than or equal to a preset value, the corresponding edge distance is marked as the matching edge distance.

[0037] The core principle of this step is to identify locally matching parts in terms of shape by comparing the edge distance sequences of two bubbles point by point, thereby quantifying the similarity between the two in terms of contour shape. Specifically, the difference between the edge distances corresponding to the detection lines in the same direction in the first and second bubbles is calculated. When the difference does not exceed a preset threshold, the pair of distances is marked as matching edge distances. This process is essentially a local matching after rigid alignment of the two contours in polar coordinates, which can effectively filter out differences caused by projection distortion, noise, or minor deformations, retaining only edge segments with truly similar shapes, laying the data foundation for subsequent continuous distribution analysis.

[0038] Step S2043: Analyze the distribution of the matching edge distance in the edge distance sequence of the first bubble and the second bubble, determine the first continuous distribution index and the second continuous distribution index in the first bubble and the second bubble, and determine the morphological correlation parameters of the morphological features of the first bubble and the second bubble based on the first continuous distribution index and the second continuous distribution index.

[0039] The core principle of this step lies in further examining the spatial continuity and concentration of the matching edge distance throughout the entire sequence to more accurately reflect the correlation strength of the overall morphology of the two bubbles, rather than relying solely on the number of local matching points. By dividing the matching edge distance into continuous groups and eliminating invalid segments that are too short or have too low a proportion, the proportion of effective continuous groups in all edge distances is calculated to obtain the continuous distribution index. This index is then corrected by combining the proximity of the continuous distribution indices of the two bubbles, ultimately generating morphological correlation parameters. This process can distinguish the essential difference between "scattered matching" and "large-segment continuous matching," making the morphological matching results more consistent with the physical meaning of the correspondence between the actual bubble projections, thereby improving the reliability of subsequent equivalent correlation determinations.

[0040] Among them, the continuous distribution index refers to the effective proportion of each continuous distance group in the edge distance sequence, and combines the close performance of the first continuous distribution index and the second continuous distribution index to determine the morphological correlation parameter.

[0041] In some embodiments disclosed in this invention, the method for determining morphological correlation parameters includes: Step S20431: Determine the continuous matching edge distance. If the proportion of the number of edge distances between each other is greater than or equal to a preset value, the continuous matching edge distance is considered to be interrupted. Statistically count the continuous matching edge distances to form a continuous distance group. If the proportion of the number of matching edge distances in the continuous distance group relative to the total number of edge distances is less than or equal to a preset value, the continuous distance group is removed.

[0042] The core principle of this step lies in using a set interval threshold to filter and remove noise from the sequence of matching edge distances, ensuring that subsequent morphological correlation calculations are based only on continuous matching segments with actual physical meaning. Specifically, the marked matching edge distances are iterated. When the proportion of adjacent matching points separated by non-matching distances reaches or exceeds a preset value, the continuous segment is considered interrupted, thus dividing the entire sequence into several independent continuous distance groups. Subsequently, the proportion of matching edge distances within each continuous distance group to the total edge distance is calculated. If this proportion is lower than a preset threshold, the group is considered as scattered noise and is removed. This process effectively filters out isolated matching points caused by local noise, projection artifacts, or minor deformations, retaining only continuous regions that truly reflect large similar segments of the bubble contour, providing high-quality input data for accurate calculation of morphological correlation parameters.

[0043] Step S20432: Count the number of matching edge distances in the retained continuous distance group and record them as effective matching edge distances. Calculate the effective proportion of effective matching edge distances relative to all matching edge distances and determine the continuous distribution index based on the preset proportion interval to which the effective proportion belongs.

[0044] The core principle of this step is to quantify the effective matching information after continuous screening into a continuous distribution index, thereby using a single value to characterize the concentration and completeness of bubble morphology in the matching region. Specifically, the number of matching edge distances within the retained continuous distance groups is accumulated to obtain the total number of effective matching edge distances, and then its effective proportion relative to all matching edge distances is calculated; subsequently, based on the preset interval in which this effective proportion falls, the corresponding continuous distribution index is mapped to obtain. This process transforms discrete matching results into index values ​​with hierarchical significance, which can not only reflect the overall scale of the matching region, but also distinguish morphological matching results of different qualities, so that the degree of morphological correlation between subsequent two bubbles has a comparable and calculable unified measurement standard.

[0045] Step S20433: Calculate the index difference between the first continuous distribution index and the second continuous distribution index, and based on the index difference, correct the average value of the first continuous distribution index and the second continuous distribution index to obtain the morphological correlation parameter.

[0046] The core principle of this step lies in using an exponential difference to weight and correct the continuous distribution indices of the two bubbles, generating the final morphological correlation parameter to more accurately reflect the true similarity between their shapes. Specifically, the difference between the first and second continuous distribution indices is first calculated as the exponential difference. Then, the arithmetic mean of the two indices is dynamically corrected based on the ratio of this difference to a preset maximum difference (the smaller the difference, the smaller the correction; the larger the difference, the lower the average). This ultimately yields the morphological correlation parameter. This correction mechanism avoids the bias that might result from simple averaging. A higher correlation score is given when the continuous distribution characteristics of the two bubbles are highly consistent, and vice versa, thus making the morphological correlation parameter more realistically reflect the reliability of the matching between the first and second bubbles in their overall contour shape.

[0047] The expression for calculating the morphological correlation parameter is as follows: .

[0048] Where G is the morphological correlation parameter, and K is the adjustment coefficient for the influence of exponential difference. Let X be the preset maximum exponential difference, and X be the exponential difference. It is the first continuous distribution index. It is the second continuous distribution index.

[0049] The index difference affects the adjustment coefficient. This is a preset positive adjustment parameter used to control the weight of the "index difference" on the final form-related parameters. By adjusting the value of K, the contribution ratio of "the closeness between two continuous distribution indices" and "the average continuous distribution level" in the final score can be flexibly balanced.

[0050] The preset maximum exponential difference. It is a manually set upper limit threshold, representing the maximum allowable difference between the first and second continuous distribution exponents, and serves to suppress excessive differences.

[0051] In some embodiments disclosed in this invention, the method for analyzing the positional relationship features between image bubbles and model bubbles includes: Step S205: Perform position analysis on the second bubble center point mapped by the first bubble center point. The analysis method includes determining the fiber distance between the second bubble center point and the PBO fiber, and determining the position association parameters of the corresponding model bubble based on the preset distance interval to which the fiber distance belongs.

[0052] In some embodiments disclosed in this invention, the method for constructing a bubble relationship analysis model includes: Step S206: Based on the morphological association parameters and the positional association parameters, determine whether there is an equivalent association between the image bubble and the model bubble. If so, associate the two equivalently. Step S207: Use the location nodes, shape representation, and size parameters corresponding to the image bubbles and model bubbles that are equivalently associated as retrieval data to construct a bubble relationship analysis model corresponding to the image bubbles and model bubbles.

[0053] In some embodiments disclosed in this invention, the method for constructing a bubble relationship analysis model includes: Step S208 involves performing multi-dimensional feature normalization on the position nodes, shape representation, and size parameters of the image bubbles and model bubbles that are equivalently associated, forming bubble feature descriptors of a unified scale. Specifically, this includes performing relative coordinate system transformation on the position nodes, edge distance sequence compression encoding on the shape representation, volume equivalent normalization on the size parameters, and weighted fusion of the normalized descriptors according to preset weights to obtain a comprehensive feature retrieval vector.

[0054] The core principle of this step lies in transforming the original features of image bubbles and model bubbles across three heterogeneous dimensions—position, shape, and size—into a unified, directly comparable comprehensive feature vector. This eliminates interference from different units and numerical ranges, laying the foundation for efficient retrieval and matching. Specifically, relative coordinate transformation is performed on position nodes to remove absolute coordinate dependencies; shape representation (edge ​​distance sequence) is compressed and encoded to retain key morphological information and reduce dimensionality; and size parameters are normalized using volume equivalence to unify scale standards. Then, the three types of normalized descriptors are weighted and fused according to preset weights, ultimately generating a comprehensive feature retrieval vector that fully represents the multidimensional attributes of a single bubble. This processing provides a unified mathematical expression for heterogeneous features that were previously difficult to compare directly, significantly improving the accuracy and robustness of subsequent similarity calculations.

[0055] Step S209: Based on the comprehensive feature retrieval vector, construct a hierarchical retrieval index structure. Use the comprehensive feature retrieval vector of the image bubble as the query end index and the comprehensive feature retrieval vector of the model bubble as the data end index. Match and associate the two through a preset similarity threshold to form a bubble mapping relationship table that can be quickly retrieved, and construct the bubble relationship analysis model.

[0056] The core principle of this step is to utilize comprehensive feature retrieval vectors to construct an efficient indexing and matching mechanism. This solidifies the offline established correspondence between image bubbles and model bubbles into a rapidly searchable mapping table, thus forming a complete bubble relationship analysis model. Specifically, the comprehensive feature retrieval vectors of image bubbles serve as the query-end index, and the comprehensive feature retrieval vectors of model bubbles serve as the data-end index. A hierarchical retrieval index structure (such as a tree or hash-based hierarchical structure) accelerates similarity calculation, and successfully matched correspondences are filtered and recorded based on a preset similarity threshold, ultimately generating a bubble mapping table. This table serves as the core knowledge base of the model, enabling rapid retrieval of corresponding 3D model bubble information based on new CT image bubble features during real-time application, achieving efficient inference from 2D projection to 3D true distribution.

[0057] In some embodiments disclosed in this invention, a PBO composite material performance analysis system based on CT scans and three-dimensional reconstruction data analysis of slices is also disclosed, including: The first module is used to establish a database that corresponds to the relationship between CT images of PBO composite materials and 3D reconstruction models of PBO slices.

[0058] The second module is used to delineate bubbles in the corresponding CT images and PBO 3D reconstruction models in the relational database, denoted as image bubbles and model bubbles. It analyzes the morphological and positional relationship features between image bubbles and model bubbles to construct a bubble relationship analysis model.

[0059] The third module is used to analyze bubbles in real-time CT images using a bubble relationship analysis model, determine the real-time three-dimensional bubble distribution of PBO composite materials, and use the real-time three-dimensional bubble distribution as performance analysis data of PBO composite materials.

[0060] This invention discloses a method and system for performance analysis of PBO composite materials based on CT scans and 3D reconstruction data analysis of PBO slices. Specifically, it establishes a database corresponding to the relationship between CT images and 3D reconstruction models of PBO slices; it delineates bubbles in the corresponding images and models and analyzes their morphological and positional characteristics to construct a bubble relationship analysis model; and it uses this model to analyze bubbles in real-time CT images, determining the real-time 3D bubble distribution as performance analysis data. This invention can accurately obtain the 3D bubble distribution non-destructively, improving the accuracy and efficiency of PBO composite material performance analysis.

[0061] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented in hardware or by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, portable hard drive, etc.) and includes several instructions to cause a computer device (such as a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0062] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for performance analysis of PBO composite materials based on CT scan and 3D reconstruction data analysis of slices, characterized in that, include: Step S100: Establish a database corresponding to the relationship between CT images of PBO composite materials and three-dimensional reconstruction models of PBO slices. Step S200: In the relational database, bubble delineation is performed on the corresponding CT images and PBO 3D reconstruction models, denoted as image bubbles and model bubbles. The morphological and positional relationship features between image bubbles and model bubbles are analyzed to construct a bubble relationship analysis model. Step S300: Using the bubble relationship analysis model, the bubbles in the real-time CT image are analyzed to determine the real-time three-dimensional bubble distribution of the PBO composite material, and the real-time three-dimensional bubble distribution is used as the performance analysis data of the PBO composite material.

2. The method for performance analysis of PBO composite materials based on CT scan and slice three-dimensional reconstruction data analysis according to claim 1, characterized in that, Methods for constructing a 3D reconstruction model of PBO slices include: Step S101: Perform physical slicing of the PBO composite material sample layer by layer according to preset values, observe each slice under a microscope, obtain microscopic images, and delineate model bubbles and PBO fibers in the microscopic images. Step S102: Spatial alignment and three-dimensional stacking reconstruction of several layers of microscopic images are performed according to the slice order to obtain a three-dimensional reconstruction model of PBO slices that simultaneously contains the actual bubble morphology and PBO fiber position.

3. The method for performance analysis of PBO composite materials based on CT scan and slice three-dimensional reconstruction data analysis according to claim 1, characterized in that, Methods for analyzing the morphological relationship between image bubbles and model bubbles include: Step S201: Construct a two-dimensional positioning coordinate system for the CT image and a three-dimensional positioning coordinate system for the PBO three-dimensional reconstruction model, and establish the positional correspondence between the two-dimensional positioning coordinate system and the three-dimensional positioning coordinate system. Specifically, this includes uniformly setting several first positioning base point coordinates in the two-dimensional coordinate system and uniformly setting several second positioning base point coordinates in the three-dimensional coordinate system, and associating and recording the several second base point coordinates mapped by the first coordinate base point coordinates. Step S202: Determine the center point of the first bubble of the first bubble in the CT image, and determine the second coordinate base point sequence corresponding to the center point of the first bubble based on the coordinate mapping relationship between the positioning coordinate systems. Determine the center point of the second bubble mapped by the second coordinate base point sequence, and then determine a number of second bubbles associated with the position of the first bubble. Step S203: Analyze the edge morphology of the first bubble to determine its morphological characteristics, and analyze the edge morphology of the associated second bubble relative to the projection plane to determine its morphological characteristics, wherein the projection plane is parallel to the CT image. Step S204: Match the morphological features of the first bubble and the second bubble to determine the morphological correlation parameters between them.

4. The method for performance analysis of PBO composite materials based on CT scan and slice three-dimensional reconstruction data analysis according to claim 3, characterized in that, Methods for matching the morphological features of the first bubble and the second bubble include: Step S2041: A number of distance detection lines are uniformly emitted at the center point of the bubble, and the edge distance from the center point of the bubble to the edge of the bubble is determined. The edge distances are summarized in order, and the order is determined according to the rotation order of the distance detection lines to form an edge distance sequence. The distance detection lines are emitted in a relative positioning coordinate system, and a distance detection line radiation template is constructed. The distance detection line radiation template includes a radiation center point and a number of distance detection lines. Step S2042: Compare the edge distance sequences corresponding to the first bubble and the second bubble respectively. The comparison method is to compare the edge distances corresponding to the detection lines of equal distance. If the distance difference between the two is less than or equal to a preset value, the corresponding edge distance is marked as the matching edge distance. Step S2043: Analyze the distribution of the matching edge distance in the edge distance sequence of the first bubble and the second bubble, determine the first continuous distribution index and the second continuous distribution index in the first bubble and the second bubble, and determine the morphological correlation parameters of the morphological features of the first bubble and the second bubble based on the first continuous distribution index and the second continuous distribution index. Among them, the continuous distribution index refers to the effective proportion of each continuous distance group in the edge distance sequence, and combines the close performance of the first continuous distribution index and the second continuous distribution index to determine the morphological correlation parameter.

5. The method for performance analysis of PBO composite materials based on CT scan and slice three-dimensional reconstruction data analysis according to claim 4, characterized in that, Methods for determining morphological correlation parameters include: Step S20431: Determine the continuous matching edge distance. If the proportion of the number of edge distances between each other is greater than or equal to a preset value, the continuous matching edge distance is considered to be interrupted. Statistically count the continuous matching edge distances to form a continuous distance group. If the proportion of the number of matching edge distances in the continuous distance group relative to the total number of edge distances is less than or equal to a preset value, the continuous distance group is removed. Step S20432: Count the number of matching edge distances in the retained continuous distance group and record them as effective matching edge distances. Calculate the effective proportion of effective matching edge distances relative to all matching edge distances and determine the continuous distribution index based on the preset proportion interval to which the effective proportion belongs. Step S20433: Calculate the index difference between the first continuous distribution index and the second continuous distribution index, and based on the index difference, correct the average value of the first continuous distribution index and the second continuous distribution index to obtain the morphological correlation parameter. The expression for calculating the morphological correlation parameter is as follows: ； Where G is the morphological correlation parameter, and K is the adjustment coefficient for the influence of exponential difference. Let X be the preset maximum exponential difference, and X be the exponential difference. It is the first continuous distribution index. It is the second continuous distribution index.

6. The method for performance analysis of PBO composite materials based on CT scan and slicing three-dimensional reconstruction data analysis according to claim 3, characterized in that, Methods for analyzing the positional relationship features between image bubbles and model bubbles include: Step S205: Perform position analysis on the second bubble center point mapped by the first bubble center point. The analysis method includes determining the fiber distance between the second bubble center point and the PBO fiber, and determining the position association parameters of the corresponding model bubble based on the preset distance interval to which the fiber distance belongs.

7. The method for performance analysis of PBO composite materials based on CT scan and 3D reconstruction data analysis of slices according to claim 6, characterized in that, Methods for constructing bubble relationship analysis models include: Step S206: Based on the morphological association parameters and the positional association parameters, determine whether there is an equivalent association between the image bubble and the model bubble. If so, associate the two equivalently. Step S207: Use the location nodes, shape representation, and size parameters corresponding to the image bubbles and model bubbles that are equivalently associated as retrieval data to construct a bubble relationship analysis model corresponding to the image bubbles and model bubbles.

8. The method for performance analysis of PBO composite materials based on CT scan and slice three-dimensional reconstruction data analysis according to claim 7, characterized in that, Methods for constructing bubble relationship analysis models include: Step S208: Perform multi-dimensional feature normalization on the position nodes, shape representation, and size parameters of the image bubbles and model bubbles that are equivalently associated, to form bubble feature descriptors of a uniform scale. Specifically, this includes relative coordinate system transformation of position nodes, edge distance sequence compression encoding of shape representation, volume equivalent normalization of size parameters, and weighted fusion of the normalized descriptors according to preset weights to obtain a comprehensive feature retrieval vector. Step S209: Based on the comprehensive feature retrieval vector, construct a hierarchical retrieval index structure. Use the comprehensive feature retrieval vector of the image bubble as the query end index and the comprehensive feature retrieval vector of the model bubble as the data end index. Match and associate the two through a preset similarity threshold to form a bubble mapping relationship table that can be quickly retrieved, and construct the bubble relationship analysis model.

9. A performance analysis system for PBO composite materials based on 3D reconstruction data analysis of CT scans and slices, characterized in that, The method for performing the PBO composite material performance analysis according to any one of claims 1-8 includes: The first module is used to establish a database that corresponds to the relationship between CT images of PBO composite materials and 3D reconstruction models of PBO slices. The second module is used to delineate bubbles in the corresponding CT images and PBO 3D reconstruction models in the relational database, denoted as image bubbles and model bubbles. It analyzes the morphological and positional relationship features between image bubbles and model bubbles and constructs a bubble relationship analysis model. The third module is used to analyze bubbles in real-time CT images using a bubble relationship analysis model, determine the real-time three-dimensional bubble distribution of PBO composite materials, and use the real-time three-dimensional bubble distribution as performance analysis data of PBO composite materials.

Images