A system and method for intelligently detecting fiber three-dimensional structure

Abstract

The application discloses a system and method for intelligently detecting fiber three-dimensional structure, and relates to the technical field of fiber three-dimensional structure detection; the system comprises an image acquisition module, an image preprocessing module, an image partition module and a feature extraction module; the image acquisition module is used for acquiring the image of the fiber structure; the image preprocessing module is used for preprocessing the acquired image; the image partition module is used for uniformly partitioning the preprocessed image, and the regions of each group are continuous; and the feature extraction module is used for extracting features in different ways for different groups. The application realizes intelligent extraction and analysis of fiber image features by combining image partition and multiple feature extraction modes; in particular, different feature extraction modes are adopted for different groups, a self-verification mechanism is provided, the accuracy and flexibility of feature recognition are improved, and the diversity and complexity of fiber structure are effectively coped with.

Description

Technical Field

[0001] This invention relates to the field of fiber three-dimensional structure detection technology, and in particular to a system and method for intelligently detecting fiber three-dimensional structures. Background Technology

[0002] In numerous fields such as materials science, textile engineering, and biomedicine, three-dimensional structural analysis of fiber materials is crucial for understanding their properties, optimizing production processes, and designing novel materials. Traditionally, fiber structure analysis has relied heavily on two-dimensional image processing techniques. While these methods can provide information on the surface morphology of fibers, they struggle to fully reveal key characteristics such as the three-dimensional spatial arrangement, diameter distribution, and curvature variations of the fibers.

[0003] With the rapid development of computer vision, machine learning, and 3D printing technologies, the development of a system capable of automatically and accurately detecting and reconstructing three-dimensional fiber structures has become an urgent industry need. However, existing technologies often face problems such as inaccurate feature extraction when dealing with complex fiber structures, limiting their widespread application in practice.

[0004] A search revealed Chinese patent application CN202111579611.2, which discloses a method for detecting solid fiber features based on 3D information. This method involves acquiring continuous images of solid fiber slices, preprocessing the images using digital image processing techniques to segment the background from the reconstructed target, and then using the solid fiber feature detection algorithm described in this invention to detect the reconstructed target features within a certain domain of adjacent slice images, matching targets with high feature similarity. A 3D reconstruction platform is then built, and the 3D reconstruction algorithm outputs a reconstructed model of the solid fiber. However, the feature detection method in the aforementioned patent has the following shortcomings: while it possesses detection capabilities, its self-verification ability is insufficient, and it cannot effectively detect complex fiber features. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a system and method for intelligent detection of the three-dimensional structure of fibers.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A system for intelligently detecting the three-dimensional structure of fibers, comprising:

[0008] Image acquisition module, used to acquire images of fiber structures;

[0009] The image preprocessing module preprocesses the acquired images.

[0010] The image partitioning module divides the preprocessed image into uniform partitions, with each group of regions being continuous.

[0011] The feature extraction module uses different methods to extract features for different groups.

[0012] The feature analysis module analyzes the extracted feature points sequentially by group to determine whether there are regular features, and obtains the final features based on the regular features.

[0013] The 3D reconstruction module constructs a 3D model of the fiber based on the obtained final features.

[0014] The information statistics and analysis module performs statistical analysis on the feature parameters of the generated 3D model and displays the results.

[0015] Preferably, the image partitioning module, when performing uniform partitioning, divides the image into P region arrays, denoted as... The P regions are divided into Q groups on average, with each group containing... Each region.

[0016] Preferably, the feature extraction module employs a total of T feature extraction methods, where T = Q, and extracts at least L feature points in each region, denoted as L1, L2, ..., L... j .

[0017] Preferably, the feature analysis module obtains the final features based on regular features, specifically including the following methods:

[0018] Method 1: After analysis by the feature analysis module, if regular features are found, the specific steps for obtaining the final features are as follows:

[0019] Step 1: Among the groups with regular characteristics, select one group as the standard group. The feature extraction method used in this standard group is used as the standard extraction method, and the extracted features are used as the standard features.

[0020] Step 2: Apply the standard extraction method to other groups to extract standard features from other groups; if the standard features in other groups have the same regularity as the standard features of the standard group, then determine the standard feature as the final feature.

[0021] Preferred method: In the first method, when extracting standard features from other groups in step two, if the standard features in other groups have different regularity from the standard features of the standard group, then a new group is selected from the groups with regular features as the standard group, and the above steps are repeated until the final features are obtained.

[0022] Preferably, the feature analysis module obtains the final features based on regular features, specifically including the following methods:

[0023] Method Two: If, after analysis by the feature analysis module, no regular features are found, the specific steps for obtaining the final features are as follows:

[0024] Step 1: Analyze the obtained features in groups and assign a similarity score. Statistically analyze the similarity scores of each group and use the groups with the highest similarity scores as the standard groups. The features in the standard groups together constitute the standard feature set, and the feature extraction method used is used as the standard extraction method.

[0025] Step 2: Apply this standard extraction method to other groups to extract features from those groups;

[0026] Step 3: Re-evaluate the similarity score. If the difference between the obtained similarity score and the similarity score of the standard feature set is less than the set value, then the standard feature set is identified as the final feature set, and the final feature set is used as the unit for final feature output.

[0027] Preferred method 2: When re-scoring similarity, if the difference between the obtained similarity score and the similarity score of the standard feature group is greater than a set value, then the similarity score is deferred to the next group in descending order as the standard group for obtaining the final feature. The above steps are repeated until the final feature is obtained.

[0028] Preferably, the system further includes:

[0029] The 3D printing module performs 3D printing processing based on the established 3D model.

[0030] Preferably, the system further includes a feature combination module, which extracts at least L feature points in each region, denoted as L1, L2, ..., L... j The feature combination module combines s adjacent feature points to obtain combined feature points, denoted as Z1, Z2, ..., Zn. k ;in, j is an integer multiple of s;

[0031] When analyzing in the feature analysis module, in Method 1, the feature points directly extracted by the feature extraction module are analyzed first. If there are no regular features, the combined feature points are then included in the analysis scope. If there are still no regular features, then the analysis switches to Method 2.

[0032] Preferably, the detection method of the system includes the following steps:

[0033] S1: The image acquisition module acquires images of the fiber structure;

[0034] S2: The image preprocessing module preprocesses the acquired images;

[0035] S3: Image partitioning module, which performs uniform partitioning on the preprocessed image;

[0036] S4: The feature extraction module uses different methods to extract features for different groups;

[0037] S5: Feature analysis module, which analyzes the extracted feature points in turn by group to determine whether there are regular features, and obtains the final features based on the regular features;

[0038] S6: 3D reconstruction module, which constructs a 3D model of the fiber based on the obtained final features;

[0039] S7: Information Statistical Analysis Module. This module performs statistical analysis on the feature parameters of the generated 3D model and displays the results.

[0040] The beneficial effects of this invention are as follows:

[0041] 1. This invention achieves intelligent extraction and analysis of fiber image features by combining image partitioning with multiple feature extraction methods; in particular, it adopts different feature extraction methods for different groups, has a self-verification mechanism, improves the accuracy and flexibility of feature recognition, and effectively copes with the diversity and complexity of fiber structures.

[0042] 2. This invention, through the dual judgment mechanism of the feature analysis module, can efficiently identify and determine the regular characteristics of fiber structure. Even in the absence of clear patterns, it can find the closest set of standard features through similarity scoring, ensuring the accuracy and representativeness of the final features.

[0043] 3. This invention introduces a feature combination module to perform combined analysis on adjacent feature points, which further enhances the depth and breadth of feature analysis, helps to discover hidden fiber structure patterns, and improves the accuracy of three-dimensional reconstruction.

[0044] 4. The system of this invention integrates multiple modules such as image acquisition, preprocessing, partitioning, feature extraction, analysis, 3D reconstruction, information statistical analysis and 3D printing, realizing the full-chain automated processing from fiber image acquisition to 3D model construction and then to physical printing, which greatly improves work efficiency and accuracy. Attached Figure Description

[0045] Figure 1 This is a flowchart of a method for intelligently detecting the three-dimensional structure of fibers proposed in this invention;

[0046] Figure 2 This is a flowchart of one method by which the feature analysis module of the present invention obtains the final feature based on regular features;

[0047] Figure 3 This is a flowchart of method two for the feature analysis module of the present invention to obtain the final feature based on regular features. Detailed Implementation

[0048] The technical solution of the present invention will be further described in detail below with reference to specific embodiments.

[0049] Example 1:

[0050] A system for intelligently detecting the three-dimensional structure of fibers, comprising:

[0051] Image acquisition module, used to acquire images of fiber structures;

[0052] The image preprocessing module preprocesses the acquired images.

[0053] The image partitioning module divides the preprocessed image into P uniform regions, denoted as Pi. The P regions are divided into Q groups on average, with each group containing... Each region is preferably a continuous region within a group.

[0054] The feature extraction module employs different methods for feature extraction based on different groups, using a total of T feature extraction methods. Preferably, T = Q, extracting at least L feature points from each region, denoted as L1, L2, ..., L... j ;

[0055] The feature analysis module analyzes the extracted feature points sequentially by group to determine whether there are any regular patterns. Based on these patterns, the final features are obtained, specifically through the following two methods:

[0056] Method 1: After analysis by the feature analysis module, if regular features are found, the specific steps for obtaining the final features are as follows:

[0057] ① Among the groups with regular characteristics, select one group as the standard group, and use the feature extraction method of the standard group as the standard extraction method. The extracted features are used as standard features.

[0058] ② Apply the standard extraction method to other groups to extract standard features from other groups; if the standard features in other groups have the same regularity as the standard features in the standard group, then determine the standard feature as the final feature; if the standard features in other groups have different regularity than the standard features in the standard group, then select a new group as the standard group from the groups with regular features, and repeat the above steps until the final feature is obtained.

[0059] Method Two: If, after analysis by the feature analysis module, no regular features are found, the specific steps for obtaining the final features are as follows:

[0060] ①Analyze the obtained features in groups and assign a similarity score. Statistically analyze the similarity scores of each group and use the group with the highest similarity score as the standard group. The features in the standard group together constitute the standard feature set, and the feature extraction method used is used as the standard extraction method.

[0061] ② Apply this standard extraction method to other groups, extract features from other groups, and re-score the similarity. If the difference between the obtained similarity score and the similarity score of the standard feature group is less than a set value, then the standard feature set is identified as the final feature set, and the final feature is output as the unit. If the difference between the obtained similarity score and the similarity score of the standard feature group is greater than a set value, then the next group is selected in descending order of similarity score as the standard group, and the final features are obtained. Repeat the above steps until the final features are obtained.

[0062] The 3D reconstruction module constructs a 3D model of the fiber based on the obtained final features.

[0063] The information statistics and analysis module performs statistical analysis on the characteristic parameters such as diameter and curvature of the generated 3D model to obtain information such as fiber diameter distribution and curvature distribution; and displays the statistical analysis results.

[0064] The 3D printing module performs 3D printing processing based on the established 3D model.

[0065] The system further includes a feature combination module, which extracts at least L feature points in each region, denoted as L1, L2, ..., L... j The feature combination module combines s adjacent feature points to obtain combined feature points, denoted as Z1, Z2, ..., Zn. k ;in, j is an integer multiple of s;

[0066] When analyzing in the feature analysis module, in Method 1, the feature points directly extracted by the feature extraction module are analyzed first. If there are no regular features, the combined feature points are then included in the analysis scope. If there are still no regular features, then the analysis switches to Method 2.

[0067] In 3D reconstruction, the following methods may be used, but are not limited to:

[0068] The system is based on a volume rendering 3D reconstruction algorithm. It maps the value of each voxel (volume element) to color and opacity, and then projects the volume data onto a 2D screen using ray casting to generate a 3D image. Let the volume data be V(x,y,z), the color mapping function be C(V), and the opacity mapping function be O(V), then the volume rendering formula can be expressed as:

[0069]

[0070] Where D is the depth of the volume data. It is the attenuation factor of light during its propagation.

[0071] Example 2:

[0072] A method for intelligently detecting the three-dimensional structure of fibers, such as Figure 1 It includes the following steps:

[0073] S1: The image acquisition module acquires images of the fiber structure;

[0074] S2: The image preprocessing module preprocesses the acquired images;

[0075] S3: Image partitioning module, which performs uniform partitioning on the preprocessed image;

[0076] S4: The feature extraction module uses different methods to extract features for different groups;

[0077] S5: Feature analysis module, which analyzes the extracted feature points in turn by group to determine whether there are regular features, and obtains the final features based on the regular features;

[0078] S6: 3D reconstruction module, which constructs a 3D model of the fiber based on the obtained final features;

[0079] S7: Information Statistical Analysis Module. This module performs statistical analysis on the feature parameters of the generated 3D model and displays the results.

[0080] Among them, such as Figure 2 In step S5, the specific methods for obtaining the final feature based on regularity characteristics include:

[0081] After analysis by the feature analysis module, if regular features are found, the specific steps for obtaining the final features are as follows:

[0082] S511: Among the groups with regular characteristics, select one group as the standard group, and use the feature extraction method of the standard group as the standard extraction method. The extracted features are used as standard features.

[0083] S512: Apply the standard extraction method to other groups to extract standard features from other groups; if the standard features in other groups have the same regularity as the standard features of the standard group, then determine the standard feature as the final feature; if the standard features in other groups have different regularity than the standard features of the standard group, then select a new group as the standard group from the groups with regular features, and repeat the above steps until the final feature is obtained.

[0084] Among them, such as Figure 3 In step S5, the specific method for obtaining the final feature based on the regularity features also includes:

[0085] S521: Analyze the obtained features in groups and assign a similarity score. Statistically analyze the similarity scores of each group and use the group with the highest similarity score as the standard group. The features in the standard group together constitute the standard feature set, and the feature extraction method used is used as the standard extraction method.

[0086] S522: Apply this standard extraction method to other groups to extract features from those groups;

[0087] S523: Re-evaluate the similarity score. If the difference between the obtained similarity score and the similarity score of the standard feature group is less than the set value, then the standard feature set is identified as the final feature set, and the final feature is output as the unit. If the difference between the obtained similarity score and the similarity score of the standard feature group is greater than the set value, then the next group is selected in descending order of similarity score as the standard group, and the final feature is obtained. Repeat the above steps until the final feature is obtained.

[0088] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A system for intelligently detecting the three-dimensional structure of fibers, characterized in that, include: Image acquisition module, used to acquire images of fiber structures; The image preprocessing module preprocesses the acquired images. The image partitioning module divides the preprocessed image into uniform partitions, with each group of regions being continuous. The feature extraction module uses different methods to extract features for different groups. The feature analysis module analyzes the extracted feature points sequentially by group to determine whether there are regular features, and obtains the final features based on the regular features. The 3D reconstruction module constructs a 3D model of the fiber based on the obtained final features. The information statistics and analysis module performs statistical analysis on the feature parameters of the generated 3D model and displays the results. The feature analysis module obtains the final features based on regular features, specifically including the following methods: Method 1: After analysis by the feature analysis module, if regular features are found, the specific steps for obtaining the final features are as follows: Step 1: Among the groups with regular characteristics, select one group as the standard group. The feature extraction method used in this standard group is used as the standard extraction method, and the extracted features are used as the standard features. Step 2: Apply this standard extraction method to other groups to extract standard features from other groups; if the standard features in other groups have the same regularity as the standard features of the standard group, then determine that standard feature as the final feature; The feature analysis module obtains the final features based on regular features, specifically including the following methods: Method Two: If, after analysis by the feature analysis module, no regular features are found, the specific steps for obtaining the final features are as follows: Step 1: Analyze the obtained features in groups and assign a similarity score. Statistically analyze the similarity scores of each group and use the groups with the highest similarity scores as the standard groups. The features in the standard groups together constitute the standard feature set, and the feature extraction method used is used as the standard extraction method. Step 2: Apply this standard extraction method to other groups to extract features from those groups; Step 3: Re-evaluate the similarity score. If the difference between the obtained similarity score and the similarity score of the standard feature set is less than the set value, then the standard feature set is identified as the final feature set, and the final feature set is used as the unit for final feature output. The system also includes a feature combination module, which extracts at least L feature points in each region for the feature extraction module, denoted as L. , ... The feature combination module combines s adjacent feature points to obtain combined feature points, which are denoted as: , ... ;in, j is an integer multiple of s; When analyzing in the feature analysis module, in Method 1, the feature points directly extracted by the feature extraction module are analyzed first. If there are no regular features, the combined feature points are then included in the analysis scope. If there are still no regular features, then the analysis switches to Method 2.

2. The intelligent system for detecting the three-dimensional structure of fibers according to claim 1, characterized in that, The image partitioning module, when performing uniform partitioning, divides the image into P region arrays, denoted as... P regions are divided into Q groups on average, and each group contains Each region.

3. The intelligent system for detecting the three-dimensional structure of fibers according to claim 2, characterized in that, The feature extraction module employs a total of T feature extraction methods, T=Q, and extracts at least L feature points in each region, denoted as L. , ... .

4. The intelligent system for detecting the three-dimensional structure of fibers according to claim 3, characterized in that, In Method 1, when extracting standard features from other groups in step 2, if the standard features in other groups have different regularity from the standard features of the standard group, then a new group is selected from the groups with regular features as the standard group, and the above steps are repeated until the final features are obtained.

5. The intelligent system for detecting the three-dimensional structure of fibers according to claim 4, characterized in that, In the second method, when re-scoring similarity, if the difference between the obtained similarity score and the similarity score of the standard feature group is greater than a set value, then the similarity score is deferred to the next group in descending order as the standard group for obtaining the final feature. The above steps are repeated until the final feature is obtained.

6. The intelligent system for detecting the three-dimensional structure of fibers according to claim 5, characterized in that, The system also includes: The 3D printing module performs 3D printing processing based on the established 3D model.

7. A method for intelligently detecting the three-dimensional structure of fibers. A system for intelligently detecting the three-dimensional structure of fibers according to claim 1 or 5, characterized in that, The detection method of the system includes the following steps: S1: The image acquisition module acquires images of the fiber structure; S2: The image preprocessing module preprocesses the acquired images; S3: Image partitioning module, which performs uniform partitioning on the preprocessed image; S4: The feature extraction module uses different methods to extract features for different groups; S5: Feature analysis module, which analyzes the extracted feature points in turn by group to determine whether there are regular features, and obtains the final features based on the regular features; S6: 3D reconstruction module, which constructs a 3D model of the fiber based on the obtained final features; S7: Information Statistical Analysis Module. This module performs statistical analysis on the feature parameters of the generated 3D model and displays the results.

Images