Weighted recognition method, system and related device based on feature lightweight
By using lightweight feature processing and clustering algorithms to identify the circular outline of rivets, the problem of low efficiency in traditional detection methods is solved, and high-precision rivet quality detection is achieved.
Patent Information
- Application Number
- CN202510408582.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-02
- Publication Date
- 2026-03-03
- Estimated Expiration
- 2045-04-02
AI Technical Summary
Traditional rivet quality inspection methods are inefficient and have unstable accuracy, making it difficult to meet the high-precision requirements of large-scale production.
A weighted recognition method based on feature lightweighting is adopted. By acquiring 3D point cloud data, feature lightweighting is performed to extract contour point sets. Clustering algorithm is used to divide the data into clusters and perform circle fitting. Visualization tools are then used to identify the circular contour of the rivet.
It improves the efficiency and accuracy of rivet identification, enabling its effective application in large-scale production.
Smart Images
Figure CN120525795B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of digital inspection technology for aircraft assembly quality, and in particular to a weighted recognition method, system and related equipment based on feature lightweighting. Background Technology
[0002] As a crucial component of aircraft structure, the quality of aircraft skin rivets directly impacts structural integrity and flight safety. Therefore, accurately detecting the shape, size, and location of rivets is of paramount importance for aircraft quality inspection and maintenance. Traditional rivet quality inspection methods primarily rely on manual visual inspection or contact measurement. These methods are not only inefficient but also susceptible to human error, leading to inconsistent inspection accuracy. Furthermore, manual inspection struggles to meet the high-precision requirements of complex environments, especially in large-scale production, where the limitations of traditional methods become increasingly apparent. Summary of the Invention
[0003] To overcome the problems of low efficiency, low accuracy, and difficulty in applying traditional rivet quality inspection methods to large-scale production, this invention provides a weighted recognition method, system, and related equipment based on feature lightweighting.
[0004] In a first aspect, to address the aforementioned technical problems, this invention provides a weighted recognition method based on feature lightweighting, comprising:
[0005] Acquire the 3D point cloud data of the object to be detected;
[0006] The 3D point cloud data is subjected to feature lightweighting processing, and contour information is extracted from the feature lightweighting 3D point cloud data to determine the contour point set.
[0007] The contour point set is analyzed using a clustering algorithm to divide different contours into different clusters and determine the target cluster;
[0008] Perform circle fitting on each target cluster to determine the target circle corresponding to each target cluster;
[0009] Visualization tools are used to draw each target circle, and the objects to be detected are identified.
[0010] Secondly, the present invention provides a weighted recognition system based on feature lightweighting, comprising:
[0011] The 3D point cloud data acquisition module is used to acquire the 3D point cloud data of the object to be detected.
[0012] The contour point set determination module is used to perform feature lightweighting processing on the 3D point cloud data, extract contour information from the feature lightweighting 3D point cloud data, and determine the contour point set.
[0013] The target cluster determination module is used to analyze the contour point set using a clustering algorithm, divide different contours into different clusters, and determine the target cluster;
[0014] The target circle determination module is used to perform circle fitting on each target cluster and determine the target circle corresponding to each target cluster.
[0015] The recognition module is used to draw various target circles using visualization tools and identify the objects to be detected.
[0016] Thirdly, the present invention provides a computing device, including a memory, a processor, and a program stored in the memory and running on the processor, wherein the processor executes the program to implement the steps of the feature-lightweight weighted recognition method described above.
[0017] Fourthly, the present invention provides a computer-readable storage medium storing instructions that, when executed on a terminal device, cause the terminal device to perform the steps of the feature-lightweight weighted recognition method described above.
[0018] The beneficial effects of this invention are as follows: It acquires three-dimensional point cloud data of the object to be detected, preprocesses the three-dimensional point cloud data to obtain a contour point set, then analyzes the contour point set using a clustering algorithm to obtain target clusters, then performs circle fitting on the target clusters to obtain target circles, and finally visualizes and draws the target circles to identify the object to be detected. This application combines the circular contour recognition of rivets with point cloud processing technology, which can effectively improve recognition efficiency and accuracy, and can be applied in large-scale production. Attached Figure Description
[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be further described below in conjunction with the accompanying drawings and embodiments.
[0020] Figure 1 This is a flowchart illustrating the feature-lightweight weighted recognition method according to an embodiment of the present invention.
[0021] Figure 2 This is a diagram illustrating the cluster refinement reconstruction and circle fitting optimization process in an embodiment of the present invention.
[0022] Figure 3 This is a diagram illustrating the measured data detection process of another embodiment of the present invention;
[0023] Figure 4 This is a schematic diagram of the structure of a feature-lightweight weighted recognition system according to an embodiment of the present invention. Detailed Implementation
[0024] The following embodiments are further explanations and supplements to the present invention and do not constitute any limitation on the present invention.
[0025] The following describes, with reference to the accompanying drawings, a feature-lightweight weighted recognition method, system, and related equipment according to embodiments of the present invention.
[0026] like Figure 1 As shown, embodiments of the present invention provide a weighted recognition method based on feature lightweighting, including:
[0027] S1. Obtain the 3D point cloud data of the object to be detected.
[0028] In this embodiment, the three-dimensional point cloud data is usually acquired by three-dimensional scanning equipment (such as LiDAR, depth camera, etc.), which contains the three-dimensional coordinate information of the object surface. This method can not only be applied in the field of digital inspection of aircraft assembly quality, but also in other scenarios, such as mechanical parts inspection, automated assembly, target recognition, etc., and has a wide range of application prospects.
[0029] S2. Perform feature lightweighting on the 3D point cloud data, and extract contour information from the feature lightweighting 3D point cloud data to determine the contour point set.
[0030] S3. Analyze the contour point set using a clustering algorithm, divide different contours into different clusters, and determine the target cluster.
[0031] S4. Perform circle fitting on each target cluster to determine the target circle corresponding to each target cluster;
[0032] S5. Use visualization tools to draw each target circle and identify the object to be detected.
[0033] In this embodiment, the visualization tool can be a PCL (Point Cloud Library) tool.
[0034] In this embodiment, 3D point cloud data of the object to be detected is acquired, and the 3D point cloud data is preprocessed to obtain a contour point set. Then, a clustering algorithm is used to analyze the contour point set to obtain target clusters. Circle fitting is then performed on the target clusters to obtain target circles. Finally, the target circles are visualized and drawn to identify the object to be detected. This application combines the circular contour recognition of rivets with point cloud processing technology, which can effectively improve recognition efficiency and accuracy, and can be applied in large-scale production.
[0035] Optionally, the 3D point cloud data undergoes feature lightweighting processing, and contour information is extracted from the feature-lightweighted 3D point cloud data to determine the contour point set, including:
[0036] A voxel mesh-based downsampling method is used to downsample 3D point cloud data to determine the preprocessed data;
[0037] The PCA normal estimation algorithm is used to extract the normals of each point cloud in the preprocessed data;
[0038] Based on the normal, the contour information of the point cloud corresponding to the normal is extracted by the boundary feature extraction algorithm to determine the contour points;
[0039] Based on each contour point, determine the contour point set.
[0040] In this embodiment, a voxel-based downsampling method is used to downsample the acquired 3D point cloud data, thereby reducing the density and amount of 3D point cloud data while maintaining the key features of the 3D point cloud data.
[0041] The core idea of the PCA normal estimation algorithm in this embodiment is to use principal component analysis (PCA) to find a local plane near each point in the 3D point cloud data and estimate the normal vector of the plane. PCA is a statistical method that reduces the dimensionality of the data or extracts key features by finding the direction with the largest variance in the dataset (i.e., the principal component).
[0042] In this embodiment, the boundary feature extraction algorithm calculates the normal (normal vector) of each point in the 3D point cloud data and analyzes the angular relationship between the normal and neighboring points to determine which points are boundary points. Boundary points usually appear in locations where the normal changes significantly.
[0043] Optionally, determining the set of contour points based on each contour point also includes:
[0044] Remove contour points formed by background noise and irrelevant information to determine the contour point set.
[0045] In this embodiment, CloudCompare software is used to remove contour points formed by background noise and irrelevant information.
[0046] In this embodiment, irrelevant information refers to the contour points formed when the 3D scanning device scans in debris other than the object to be detected.
[0047] In this embodiment, the impact of environmental interference factors on recognition accuracy is reduced by removing contour points formed by background noise and irrelevant information.
[0048] Optionally, the contour point set is analyzed using a clustering algorithm to divide different contours into different clusters, and the target cluster is determined, including:
[0049] Calculate the local density of each contour point in the contour point set, and take the contour points with local density greater than or equal to the density threshold as target points;
[0050] Calculate the number of contour points of each target point within a preset area, and take the target points whose number of contour points is greater than or equal to the first threshold as core points;
[0051] For each core point, find all contour points within the preset domain. If the found contour points include other core points besides itself, continue to expand the cluster until no new core points are found, and determine the target cluster corresponding to the core point.
[0052] In this embodiment, the local density ρ of each contour point i This can be represented by calculating the number of neighboring points of contour point i within a certain neighborhood radius ∈ (preset neighborhood). Assuming the 3D point cloud data has N points, P... i Let represent the coordinates of the i-th contour point. Then, the formula for calculating the local density is:
[0053]
[0054] Wherein d(P i ,P j ) represents the contour point P i and contour point P j The Euclidean distance between them, Ⅱ(·) is an indicator function, if d(P i ,P j If ≤ ∈, then the value is 1, representing point P. j At point P i If the value is within the domain of point P, then the value is 0; otherwise, it represents point P. j Not at point P i Within that field.
[0055] Based on local density ρ i Set a density threshold ρ threshold Points below the density threshold are considered noise points and removed. If ρ i <ρ threshold Then point P i It is considered a noise point and should be removed.
[0056] In this embodiment, the number of contour points within a preset area for each target point is calculated to determine whether the point is a core point. If a point P i If a point is considered a core point, the number of contour points (including itself) within its neighborhood must be greater than or equal to a specified first threshold, MinPts, i.e.:
[0057] ({P j |d(P i ,P j )≤∈})≥MinPts
[0058] For each core point P iFind all contour points within its preset neighborhood. If the preset neighborhood also contains other core points, continue to expand the cluster. Assume the core point P... i The preset domain is N(P) i The extended formula for the cluster is:
[0059]
[0060] Cluster j Represents the current cluster, CorePoint(P) k ) represents the core point P k The preset domain.
[0061] For example, assuming the core points include p1, p2, p3, p4, and p5, we first start by searching for all contour points within a preset domain, beginning with p1. We find that p2 is contained within p1's preset domain, while p3, p4, and p5 are not. We then continue searching for all contour points within p2's preset domain. Since no core points other than p1 are found within p2's preset domain, the preset domains containing p1 and p2 are grouped into the same cluster. Furthermore, since no core points other than themselves are found within p3, p4, and p5 during the search, the preset domain containing p3 is one cluster, the preset domain containing p4 is one cluster, and the preset domain containing p5 is one cluster, resulting in four target clusters.
[0062] Optionally, for each core point, all contour points within a preset neighborhood are searched. If the found contour points include other core points besides itself, the cluster is expanded until no new core points are found. Determining the target cluster corresponding to the core point also includes:
[0063] If the number of contour points contained in the generated cluster is less than the second threshold, then the cluster is removed;
[0064] Calculate the roundness of the generated clusters. If the roundness is less than the roundness threshold, remove the cluster.
[0065] In this embodiment, if the number of contour points contained within a generated cluster is less than a second threshold MinClusterSize, the cluster is considered a redundant cluster and needs to be removed.
[0066] RemoveCluster(C j = True if|C j | <MinClusterSize
[0067] Among them, |C j | is cluster C j The number of midpoints, MinClusterSize is the minimum cluster size, RemoveCluster(Cj ) indicates the removal of cluster C j .
[0068] The roundness calculation method in this embodiment is as follows:
[0069] Determine the boundaries of the cluster and find the minimum convex hull of the cluster.
[0070] The smallest convex hull is generally a polygon, and the area A can be calculated using the coordinates of each vertex of the polygon and the polygon area formula:
[0071]
[0072] n represents the number of vertices of the polygon, x i y i This represents the coordinates of the i-th vertex.
[0073] Calculate the perimeter P of the polygon:
[0074]
[0075] Calculate roundness C:
[0076]
[0077] Set a roundness threshold C≥0.8. If the roundness of a cluster is greater than or equal to the roundness threshold, the cluster is considered a round cluster. If the roundness of a cluster is less than the roundness threshold, the cluster is filtered out.
[0078] like Figure 2 As shown, by removing noise points (contour points formed by background noise and irrelevant information) and redundant clusters (clusters containing fewer contour points than the second threshold), more accurate circles can be obtained, thus improving the accuracy of clustering.
[0079] Optionally, circle fitting is performed on each target cluster to determine the target circle corresponding to each target cluster, including:
[0080] Each target cluster is fitted using the RANSAC algorithm to determine the center and radius of the initial circle corresponding to each target cluster.
[0081] For each initial circle, the weight of each contour point outside the initial circle is determined based on the distances from all contour points outside the initial circle to the initial circle; where the weight represents the distance from each contour point outside the initial circle to the initial circle.
[0082] For each initial circle, a weighted random sampling is performed on the weights to determine the random probability of all contour points outside the initial circle;
[0083] For each initial circle, a preset number of contour points are randomly selected from all contour points outside the initial circle based on random probability, and the circle is fitted using the RANSAC algorithm to determine the target circle.
[0084] In this embodiment, the process of fitting a circle using the conventional RANSAC algorithm is as follows:
[0085] The RANSAC algorithm fits the center and radius of the initial circle by minimizing the objective function:
[0086]
[0087] Among them, P i Let E(C,R) represent the coordinates of the i-th contour point in the initial circle, C and R represent the center and radius respectively, N represent the number of contour points in the initial circle, and E(C,R) represent the objective function. i -C|| is the Euclidean distance from the i-th contour point to the center C of the circle.
[0088] In this embodiment, the distance d from all contour points outside the initial circle to the initial circle is... i The formula is as follows:
[0089] d i =|||P i -C||-R|
[0090] In this embodiment, the weight ω of each contour point outside the initial circle is determined. i The formula is as follows:
[0091]
[0092] Where σ is the standard deviation of the weight calculation, used to control the rate of weight decay, and is usually set according to the noise level of the data.
[0093] In this embodiment, based on the weight ω i Weighted random sampling is performed, and the probability of each contour point being selected is proportional to its weight, i.e.:
[0094]
[0095] Where N represents the number of contour points, P′ i Represents the contour point P i The probability of being selected.
[0096] Based on the above, the probability distribution of each contour point outside the initial circle can be obtained. A preset number of contour points are randomly selected from the probability distribution for refitting the circle, and the RANSAC algorithm is used to fit the circle again, as shown in the following formula:
[0097]
[0098] By minimizing E ω (C,R) is used to obtain the final circle fitting parameters C and R, thereby determining the target circle.
[0099] Optionally, using visualization tools to draw each target circle to identify the object to be detected also includes:
[0100] Use visualization tools to draw each target circle onto the original 3D point cloud data, and display the original 3D point cloud data and the target circles simultaneously.
[0101] In this embodiment, the PCL visualization tool is used to draw the fitted circle onto the original point cloud data, and the original point cloud and the fitted circle are displayed simultaneously to intuitively evaluate the fitting effect.
[0102] Alternatively, another embodiment of the feature-lightweight weighted recognition method is described below:
[0103] The object to be detected is scanned using a 3D scanning device to obtain... Figure 3 The three-dimensional point cloud data corresponding to 'a' in the diagram.
[0104] The 3D point cloud data undergoes feature lightweighting processing. Then, the PCA normal estimation algorithm is used to extract the normals of each point cloud. Finally, a boundary extraction algorithm is used to extract the contour information of the point cloud corresponding to the normals. Figure 3 The set of contour points shown in b.
[0105] Contour points with local density less than a density threshold are deleted from the contour point set. Then, contour points with a number of contour points greater than or equal to a first threshold within a preset neighborhood are selected as core points. Clusters are then found using these core points. Simultaneously, clusters containing contour points with a number less than a second threshold, and clusters with roundness less than a roundness threshold, are removed. This yields... Figure 3 The effective cluster shown in c is shown in the middle.
[0106] The effective clusters undergo cluster refinement reconstruction and circle fitting optimization. First, an initial circle is fitted using the RANSAC algorithm. Then, the distances from the outer contour points of the initial fitted circle to the circle are calculated to obtain weights. Weighted random sampling is then used to calculate the random probabilities of the outer contour points of the initial fitted circle. Based on these random probabilities, three contour points are reselected to fit the circle again, resulting in the final circle. Figure 3 The target circle is shown as d in the diagram.
[0107] By using visualization tools to draw each target circle, the following can be identified: Figure 3 The rivet is positioned as shown in the middle (e).
[0108] like Figure 4 As shown, this invention provides a weighted recognition system based on feature lightweighting, comprising:
[0109] The 3D point cloud data acquisition module is used to acquire the 3D point cloud data of the object to be detected.
[0110] The contour point set determination module is used to perform feature lightweighting processing on the 3D point cloud data, extract contour information from the feature lightweighting 3D point cloud data, and determine the contour point set.
[0111] The target cluster determination module is used to analyze the contour point set using a clustering algorithm, divide different contours into different clusters, and determine the target cluster;
[0112] The target circle determination module is used to perform circle fitting on each target cluster and determine the target circle corresponding to each target cluster.
[0113] The recognition module is used to draw various target circles using visualization tools and identify the objects to be detected.
[0114] Optionally, the contour point set determination module is specifically used for:
[0115] A voxel mesh-based downsampling method is used to downsample 3D point cloud data to determine the preprocessed data;
[0116] The PCA normal estimation algorithm is used to extract the normals of each point cloud in the preprocessed data;
[0117] Based on the normal, the contour information of the point cloud corresponding to the normal is extracted by the boundary feature extraction algorithm to determine the contour points;
[0118] Based on each contour point, determine the contour point set.
[0119] Optionally, the contour point set determination module is also used for:
[0120] Remove contour points formed by background noise and irrelevant information to determine the contour point set.
[0121] Optionally, the target cluster determination module is specifically used for:
[0122] Calculate the local density of each contour point in the contour point set, and take the contour points with local density greater than or equal to the density threshold as target points;
[0123] Calculate the number of contour points of each target point within a preset area, and take the target points whose number of contour points is greater than or equal to the first threshold as core points;
[0124] For each core point, find all contour points within the preset domain. If the found contour points include other core points besides itself, continue to expand the cluster until no new core points are found, and determine the target cluster corresponding to the core point.
[0125] Optionally, the target cluster determination module is also used for:
[0126] If the number of contour points contained in the generated cluster is less than the second threshold, the cluster is removed.
[0127] Optionally, the target circle determination module is specifically used for:
[0128] Each target cluster is fitted using the RANSAC algorithm to determine the center and radius of the initial circle corresponding to each target cluster.
[0129] For each initial circle, the weight of each contour point outside the initial circle is determined based on the distances from all contour points outside the initial circle to the initial circle; where the weight represents the distance from each contour point outside the initial circle to the initial circle.
[0130] For each initial circle, a weighted random sampling is performed on the weights to determine the random probability of all contour points outside the initial circle;
[0131] For each initial circle, a preset number of contour points are randomly selected from all contour points outside the initial circle based on random probability, and the circle is fitted using the RANSAC algorithm to determine the target circle.
[0132] Optionally, the identification module is specifically used for:
[0133] Use visualization tools to draw each target circle onto the original 3D point cloud data, and display the original 3D point cloud data and the target circles simultaneously.
[0134] The present invention also provides a computing device, including a memory, a manager, and a program stored on the memory and running on the manager, wherein the manager executes the program to implement some or all of the steps of the above-described feature-lightweight weighted recognition method.
[0135] The computing device can be a computer, and the corresponding program is computer software. The parameters and steps of the computing device of the present invention can be referred to the parameters and steps in the embodiment of the feature-lightweight weighted recognition method above, and will not be repeated here.
[0136] Those skilled in the art will recognize that this invention can be implemented as a system, method, or computer program product. Therefore, this disclosure can be embodied in the following forms: it can be entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software, generally referred to herein as a "circuit," "module," or "system." Furthermore, in some embodiments, the invention can also be implemented as a computer program product contained in one or more computer-readable media, which contains computer-readable program code. Computer-readable storage media can be, for example, but not limited to—electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof.
[0137] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.
[0138] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.
Claims
1. A feature-based lightweight weighted recognition method, characterized by, The method comprises the following steps: acquiring three-dimensional point cloud data of an object to be detected; performing feature lightweight processing on the three-dimensional point cloud data, extracting contour information of the three-dimensional point cloud data after the feature lightweight processing, and determining a contour point set; analyzing the contour point set using a clustering algorithm, dividing different contours into different clusters, and determining target clusters; performing circle fitting on each target cluster to determine a target circle corresponding to each target cluster; drawing each target circle using a visualization tool to identify the object to be detected; the circle fitting on each target cluster to determine a target circle corresponding to each target cluster comprises the following steps: fitting each target cluster through a RANSAC algorithm to determine the center and radius of an initial circle corresponding to each target cluster; for each initial circle, determining the weight of each contour point outside the initial circle based on the distance of all contour points outside the initial circle to the initial circle; wherein the weight represents the distance of each contour point outside the initial circle to the initial circle; for each initial circle, performing weighted random sampling on the weight to determine the random probability of all contour points outside the initial circle; for each initial circle, randomly selecting a preset number of contour points from all contour points outside the initial circle based on the random probability, and fitting a circle through a RANSAC algorithm to determine a target circle.
2. The method of claim 1, wherein, the feature lightweight processing on the three-dimensional point cloud data, and the extraction of contour information of the three-dimensional point cloud data after the feature lightweight processing to determine a contour point set comprises the following steps: performing downsampling processing on the three-dimensional point cloud data based on a voxel grid downsampling method to determine pretreatment data; extracting the normal of each point cloud in the pretreatment data using a PCA normal estimation algorithm; extracting contour information of the point cloud corresponding to the normal based on the normal through a boundary feature extraction algorithm to determine contour points; determining a contour point set based on each contour point.
3. The method of claim 2, wherein, determining a contour point set based on each contour point further comprises the following steps: removing contour points formed by background noise and irrelevant information to determine a contour point set.
4. The method of claim 1, wherein, the analysis of the contour point set using a clustering algorithm, the division of different contours into different clusters, and the determination of target clusters comprise the following steps: calculating the local density of each contour point in the contour point set, and taking contour points with a local density greater than or equal to a density threshold value as target points; calculating the number of contour points in a preset field of each target point, and taking target points with a contour point number greater than or equal to a first threshold value as core points; for each core point, finding all contour points in a preset field of the core point, if the found contour points include other core points except itself, continuing to expand the cluster until no new core point is found, and determining a target cluster corresponding to the core point.
5. The method of claim 4, wherein, the finding of all contour points in a preset field of each core point, the continuation of the expansion of the cluster if the found contour points include other core points except itself, and the determination of a target cluster corresponding to the core point further comprise the following steps: if the number of the contour points contained in the generated cluster is less than a second threshold, the cluster is removed; a roundness of the generated cluster is calculated, and if the roundness is less than a roundness threshold, the cluster is removed.
6. The method of claim 1, wherein, The method further comprises: The method further comprises:
7. A feature-based lightweighted weighted recognition system characterized in that, The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises:
8. A computing device comprising a memory, a processor, and a program stored on the memory and running on the processor, wherein, The method further comprises:
9. A computer-readable storage medium, characterized in that, The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The method further comprises: The