A textile image feature library updating system

Abstract

The invention discloses an updating system for a textile image feature library, belongs to the field of image information updating, and solves the problem of how to update the textile image feature library so that the textile image feature library can be conveniently queried and the textile image and related information can be quickly located; an information uploading module obtains textile images and text description information, adds a time stamp and sends them to an information preprocessing module; the information preprocessing module preprocesses the textile images and text description information to obtain a feature data set and a unique identification value; a cloud storage module receives and stores the textile images and text description information with a unique identification value label obtained from the information preprocessing module; and a physical node updating module generates a new physical node according to the feature data set with a unique identification value label obtained from the information preprocessing module and the file path with a unique identification value label obtained from the cloud storage module, and sends the new physical node to the feature library module for storage.

Description

[0001] This application is a divisional application of the application filed on September 19, 2023, with application number 202311207135.0 and invention name “A system for updating a textile image feature library”. Technical Field

[0002] The invention belongs to the field of picture information updating, in particular to an updating system for a textile picture feature library. Background Art

[0003] When textiles are produced and sold, it is necessary to photograph the textiles to obtain pictures of the textiles that allow consumers to feel them intuitively, thereby helping consumers make purchasing choices.

[0004] At present, textile images and corresponding features are generally stored in the same storage space, but the images need to occupy a large amount of storage space, which is not conducive to the query and analysis of textile image features. In addition, only the feature data of textile images is stored in the textile image feature library. If distributed storage is to be achieved, it is not conducive to the query and management of target images and similar images. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention proposes a textile image feature library update system, which solves the problem of how to update the textile image feature library so that the textile image feature library can be easily queried and the textile image and related information can be quickly located.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] A textile image feature library update system, comprising: an information upload module, an information preprocessing module, a cloud storage module, a feature library module and an entity node update module;

[0008] The information uploading module is used to obtain the textile image and the corresponding text description information uploaded by the user, and send the obtained textile image and the corresponding text description information to the information preprocessing module with a time stamp;

[0009] The information preprocessing module is used to preprocess the obtained textile images and corresponding text description information to obtain each feature data set and unique identification value of the corresponding textile image; the textile images and text description information are both tagged with unique identification value and sent to the cloud storage module, and each feature data set is tagged with unique identification value and sent to the entity node update module;

[0010] The cloud storage module is used to receive and store the obtained textile images and text description information marked with unique identification value tags; and mark the textile image file path and text description information file path with unique identification value tags and send them to the entity node update module;

[0011] The feature library module includes several entity nodes;

[0012] The entity node update module is used to generate a new entity node according to each feature data set, textile image file path and text description information file path that are obtained and tagged with unique identification values; and upload the created new entity node to the feature library module for storage.

[0013] Furthermore, the text description information refers to information about the textile indicated by the corresponding textile image, and the text description information includes the category, color, fabric, size, supplier information, and designer information of the textile.

[0014] Furthermore, the preprocessing process of the information preprocessing module includes:

[0015] The digital image processing technology is used to extract features of the obtained textile images, and a color feature data set C = {C1, C2, ..., Ci, ..., Cn}, a texture feature data set T = {T1, T2, ..., Tj, ..., Tm} and a shape feature data set S = {S1, S2, ..., Sh, ..., Sg} of the corresponding textile images are obtained;

[0016] Wherein, i represents the interval number of the color histogram in the color feature data set, i=1,2…n; n represents the total number of intervals of the color histogram; Ci represents the i-th interval value of the color histogram in the color feature data set; j represents the gray level number of the gray level co-occurrence matrix in the texture feature data set, j=1,2…m; m represents the total number of gray levels in the gray level co-occurrence matrix; Tj represents the element value of the j-th gray level of the gray level co-occurrence matrix in the texture feature data set; h represents the number of the shape feature in the shape feature data set, h=1,2…g; g represents the total number of shape features; Sh represents the element value of the h-th shape feature in the shape feature data set;

[0017] Connecting the acquired color feature vector C, texture feature vector T and shape feature vector S together to form a comprehensive feature vector F of the textile image;

[0018] F=[C,T,S]=[C1,C2,…,Ci,…,Cn,T1,T2,…,Tj,…,Tm,S1,S2,…,Sh,…,Sg];

[0019] According to the hash function H, the obtained comprehensive feature vector F is mapped to an identification value HV; wherein HV=H(F); the hash function is selected as SHA-256 or CityHash;

[0020] After the identification value HV is stamped with the upload timestamp, a unique identification value of the textile image is formed.

[0021] Furthermore, the cloud storage module includes a picture storage unit and a text storage unit;

[0022] The cloud storage module receives a textile image and text description information marked with a unique identification value; according to the format of the textile image and text description information, the textile image marked with a unique identification value label is stored in the image storage unit, and the file path of the textile image is marked with a unique identification value label and sent to the entity node update module; the file description information marked with a unique identification value label is stored in the text storage unit, and the file path of the text description information is marked with a unique identification value label and sent to the entity node update module.

[0023] Furthermore, the entity node is used to store the unique identification value of the textile image, the textile image file path, the text description information file path, the color feature data set C, the texture feature data set T, the shape feature data set S and the unique identification value of the associated entity node.

[0024] Furthermore, the processing process of the entity node update module is as follows:

[0025] Get each feature data set and file path labeled with a unique identification value;

[0026] Compare the unique identification values ​​of each feature data set and file path; filter out feature data sets and file paths with the same unique identification value, and create a new entity node based on the unique identification value;

[0027] Add the unique identification value, each feature data set, and file path as attributes to the new entity node;

[0028] Count the number of file paths of the new entity node; if the number of file paths of the new entity node is equal to 2, determine whether the two file paths of the new entity node are the same; if they are the same, end; if they are not the same, re-acquire the corresponding textile image file path and text description information file path according to the unique identification value; if the number of file paths of the new entity node is greater than 2, re-acquire the corresponding textile image file path and text description information file path according to the unique identification value;

[0029] Calculate the comprehensive feature similarity between the textile image corresponding to the new entity node and the textile image corresponding to the entity node stored in the feature library module according to the color feature data set C, the texture feature data set T and the shape feature data set S, and then obtain the unique identification value attribute of the associated entity node of the new entity node;

[0030] The created new entity node is uploaded to the feature library module for storage.

[0031] Furthermore, the process of obtaining the unique identification value attribute of the associated entity node of the new entity node is as follows:

[0032] Step S1: number all entity nodes stored in the feature library module in order of storage time, numbered as k;

[0033] Step S2: Calculate and obtain the comprehensive feature similarity ZXk between the textile image corresponding to each entity node in the feature library module and the textile image corresponding to the new entity node, including:

[0034] Obtain a color feature data set C, a texture feature data set T, and a shape feature data set S of a new entity node; and extract a color feature data set Ck, a texture feature data set Tk, and a shape feature data set Sk of the entity node in the feature library module;

[0035] According to the calculation formula Obtain the color feature similarity X(C, Ck) between the new entity node and the entity node numbered k; where Cki represents the i-th interval value of the color histogram in the color feature data set of the entity node numbered k;

[0036] According to the calculation formula Obtain the texture feature similarity X(T,Tk) between the new entity node and the entity node numbered k; where Tkj represents the element value of the jth gray level of the gray level co-occurrence matrix in the texture feature data set of the entity node numbered k;

[0037] According to the calculation formula Obtain the shape feature similarity X(S, Sk) between the new entity node and the entity node numbered k; where Skh represents the element value of the h-th shape feature of the shape feature data set of the entity node numbered k;

[0038] According to the calculation formula ZXk=a1×X(C,Ck)+a2×X(T,Tk)+a3×X(S,Sk), the comprehensive feature similarity ZXk between the new entity node and the entity node numbered k is obtained; where a1, a2 and a3 are the weight coefficients of color feature similarity, texture feature similarity and shape similarity respectively;

[0039] Step S3: obtaining the comprehensive feature similarity between each entity node stored in the feature library module and the new entity node; comparing each obtained comprehensive feature similarity with a preset comprehensive feature similarity;

[0040] Screen out comprehensive feature similarities that are less than a preset comprehensive feature similarity; and count the number y of the screened comprehensive feature similarities;

[0041] If y is equal to 1, the entity node corresponding to the screened comprehensive feature similarity is used as the associated entity of the new entity node, and the unique identification value of the associated entity node is added to the new entity node as an attribute;

[0042] If y is greater than 1, each entity node corresponding to the screened comprehensive feature similarity is used as the associated entity of the new entity node, and then the unique identification value of the associated entity is arranged in ascending order according to the screened comprehensive feature similarity, and added to the new entity node as an attribute.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] 1. In the present invention, a textile image and corresponding text description information are obtained through an information uploading module, a feature data set of the textile image is extracted through an information preprocessing module, and a unique identification value of the corresponding textile image is calculated based on the proposed feature data set, a hash function and an upload timestamp, so that the obtained unique identification value can uniquely refer to the corresponding textile image, the text description letter in the cloud storage module and the corresponding entity node in the feature library module.

[0045] 2. In the present invention, the file path of the textile image and the text description information text path are tagged with a unique identification value through the cloud storage module and sent to the entity node update module, and the feature data set of the textile image is tagged with a unique identification value through the information preprocessing module and sent to the entity node update module; the entity node update module creates a new entity node according to the unique identification value, and stores the file path and feature data set belonging to the unique identification value as attribute values ​​in the new entity node, so that the subsequent user end or client end can directly retrieve the corresponding attribute through the feature library module, thereby obtaining the corresponding entity node, and then obtaining all relevant attributes related to the entity node. If the attribute is a file path, the textile image and the corresponding text description information stored in the cloud storage module are directly located; and the textile image, the corresponding text description information and the feature data set are stored separately, so that the storage space can be optimized; the query efficiency of the feature library is improved, different types of data are flexibly processed, the coupling between data is reduced, various types of data are stored independently, and as the amount of data increases, it is helpful to manage and query a large number of textile images, text description information and feature data sets.

[0046] 3. In the present invention, the associated entity nodes of the new entity node are obtained through the entity node update module, and the color feature similarity, texture feature similarity and shape feature similarity of the textile image corresponding to the new entity node and the textile image corresponding to the entity node stored in the feature library module are calculated respectively, and then the comprehensive feature similarity is calculated. By comparing each comprehensive feature similarity with the preset comprehensive feature similarity, the entity node that is relatively similar to the textile image corresponding to the new entity node is screened out as the associated entity of the new entity node, and its unique identification value is added as an attribute to the new entity node, so as to facilitate subsequent query and retrieval. When the target entity node is retrieved, other similar entity nodes can also be retrieved, so as to provide more textile choices for the production, sales and consumption ends. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 The present invention is a system block diagram for updating a textile image feature library. DETAILED DESCRIPTION

[0048] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] like Figure 1 As shown, a textile image feature library update system includes: an information uploading module, an information preprocessing module, a cloud storage module, a feature library module and a physical node update module;

[0050] The information uploading module is used to obtain the textile images and corresponding text description information uploaded by the user, and add a timestamp to the obtained textile images and corresponding text description information and send them to the information preprocessing module; wherein the timestamp includes the year, month, day and the time point accurate to microseconds;

[0051] The text description information refers to the information of the textiles indicated by the corresponding textile images, including the category, color, fabric, size, supplier information, and designer information of the textiles; the categories of textiles include jackets, T-shirts, pants, skirts, etc.;

[0052] The information preprocessing module is used to preprocess the acquired textile images and corresponding text description information, including:

[0053] The digital image processing technology is used to extract features of the obtained textile images, and a color feature data set C = {C1, C2, ..., Ci, ..., Cn}, a texture feature data set T = {T1, T2, ..., Tj, ..., Tm} and a shape feature data set S = {S1, S2, ..., Sh, ..., Sg} of the corresponding textile images are obtained;

[0054] Wherein, i represents the interval number of the color histogram in the color feature data set, i=1,2…n; n represents the total number of intervals of the color histogram; Ci represents the i-th interval value of the color histogram in the color feature data set; j represents the gray level number of the gray level co-occurrence matrix in the texture feature data set, j=1,2…m; m represents the total number of gray levels in the gray level co-occurrence matrix; Tj represents the element value of the j-th gray level of the gray level co-occurrence matrix in the texture feature data set; h represents the number of the shape feature in the shape feature data set, h=1,2…g; g represents the total number of shape features; Sh represents the element value of the h-th shape feature in the shape feature data set;

[0055] Connecting the acquired color feature vector C, texture feature vector T and shape feature vector S together to form a comprehensive feature vector F of the textile image;

[0056] F=[C,T,S]=[C1,C2,…,Ci,…,Cn,T1,T2,…,Tj,…,Tm,S1,S2,…,Sh,…,Sg];

[0057] According to the hash function H, the obtained comprehensive feature vector F is mapped to an identification value HV; wherein HV=H(F); the hash function may be SHA-256 or CityHash;

[0058] After the identification value HV is stamped with the upload timestamp, a unique identification value of the textile image is formed;

[0059] The information preprocessing module tags the textile image and text description information with unique identification value tags and sends them to the cloud storage module, and tags each feature data set of the acquired textile image with unique identification value tags and sends them to the entity node update module;

[0060] In the present invention, a textile image and corresponding text description information are obtained through an information uploading module, a feature data set of the textile image is extracted through an information preprocessing module, and a unique identification value of the corresponding textile image is calculated according to the proposed feature data set, a hash function and an upload timestamp, so that the obtained unique identification value can uniquely refer to the corresponding textile image and text description letter in the cloud storage module and the corresponding entity node in the feature library module;

[0061] The cloud storage module is used to receive and store the obtained textile images and text description information marked with unique identification value tags; the cloud storage module includes an image storage unit and a text storage unit; the process includes:

[0062] The cloud storage module receives the textile image and text description information marked with a unique identification value;

[0063] According to the format of the textile image and text description information, the textile image marked with a unique identification value tag is stored in the image storage unit, and the file path of the textile image is marked with a unique identification value tag and sent to the entity node update module; the file description information marked with a unique identification value tag is stored in the text storage unit, and the file path of the text description information is marked with a unique identification value tag and sent to the entity node update module;

[0064] The feature library module includes a plurality of entity nodes; wherein the entity nodes are used to store the unique identification value of the textile image, the textile image file path, the text description information file path, the color feature data set C, the texture feature data set T, the shape feature data set S and the unique identification value of the associated entity node;

[0065] The entity node update module is used to generate a new entity node according to the acquired color feature data set C, texture feature data set T, shape feature data set S, textile image file path and text description information file path marked with unique identification value tags; the process is as follows:

[0066] Get each feature data set and file path labeled with a unique identification value;

[0067] Compare the unique identification values ​​of each feature data set and file path; filter out feature data sets and file paths with the same unique identification value, and create a new entity node based on the unique identification value;

[0068] And add the unique identification value, each feature data set and file path as attributes to the new entity node;

[0069] Count the number of file paths of the new entity node; if the number of file paths of the new entity node is equal to 2, determine whether the two file paths of the new entity node are the same; if they are the same, end; if they are not the same, re-acquire the corresponding textile image file path and text description information file path according to the unique identification value; if the number of file paths of the new entity node is greater than 2, re-acquire the corresponding textile image file path and text description information file path according to the unique identification value;

[0070] In the present invention, a unique identification value tag is added to the file path of the textile image and the text description information text path through the cloud storage module and sent to the entity node update module, and a unique identification value is added to the feature data set of the textile image through the information preprocessing module and sent to the entity node update module; the entity node update module creates a new entity node according to the unique identification value, and stores the file path and feature data set belonging to the unique identification value as attribute values ​​in the new entity node, so that the subsequent user end or client end can directly retrieve the corresponding attribute through the feature library module, thereby obtaining the corresponding entity node, and then obtaining all relevant attributes related to the entity node. If the attribute is a file path, the textile image and the corresponding text description information stored in the cloud storage module are directly located; and the textile image, the corresponding text description information and the feature data set are stored separately, so that the storage space can be optimized; the query efficiency of the feature library is improved, different types of data are flexibly processed, the coupling between data is reduced, various types of data are stored independently, and as the amount of data increases, it is helpful to manage and query a large number of textile images, text description information and feature data sets;

[0071] The comprehensive feature similarity between the textile image corresponding to the new entity node and the textile image corresponding to the entity node stored in the feature library module is calculated based on the color feature data set C, the texture feature data set T and the shape feature data set S, and then the unique identification value attribute of the associated entity node of the new entity node is obtained; the process is as follows:

[0072] Step S1: number all entity nodes stored in the feature library module in order of storage time, numbered as k;

[0073] Step S2: Calculate and obtain the comprehensive feature similarity ZXk between the textile image corresponding to each entity node in the feature library module and the textile image corresponding to the new entity node, including:

[0074] Obtain a color feature data set C, a texture feature data set T, and a shape feature data set S of a new entity node; and extract a color feature data set Ck, a texture feature data set Tk, and a shape feature data set Sk of the entity node in the feature library module;

[0075] According to the calculation formula Obtain the color feature similarity X(C, Ck) between the new entity node and the entity node numbered k; where Cki represents the i-th interval value of the color histogram in the color feature data set of the entity node numbered k;

[0076] The smaller the value of the color feature similarity X(C, Ck), the more similar the color feature of the new entity node is to the entity node numbered k;

[0077] According to the calculation formula Obtain the texture feature similarity X(T,Tk) between the new entity node and the entity node numbered k; where Tkj represents the element value of the jth gray level of the gray level co-occurrence matrix in the texture feature data set of the entity node numbered k;

[0078] Among them, the smaller the texture feature similarity X(T,Tk), the more similar the texture features of the new entity node are to the entity node numbered k;

[0079] According to the calculation formula Obtain the shape feature similarity X(S, Sk) between the new entity node and the entity node numbered k; where Skh represents the element value of the h-th shape feature of the shape feature data set of the entity node numbered k;

[0080] The smaller the shape feature similarity X(S, Sk), the more similar the shape feature of the new entity node is to the entity node numbered k.

[0081] According to the calculation formula ZXk=a1×X(C,Ck)+a2×X(T,Tk)+a3×X(S,Sk), the comprehensive feature similarity ZXk between the new entity node and the entity node numbered k is obtained; where a1, a2 and a3 are the weight coefficients of color feature similarity, texture feature similarity and shape similarity respectively;

[0082] Step S3: obtaining the comprehensive feature similarity between each entity node stored in the feature library module and the new entity node; comparing each obtained comprehensive feature similarity with a preset comprehensive feature similarity; wherein the preset comprehensive feature similarity is set by relevant technical personnel;

[0083] Screen out comprehensive feature similarities that are less than a preset comprehensive feature similarity; and count the number y of the screened comprehensive feature similarities;

[0084] If y is equal to 1, the entity node corresponding to the screened comprehensive feature similarity is used as the associated entity of the new entity node, and the unique identification value of the associated entity node is added to the new entity node as an attribute;

[0085] If y is greater than 1, each entity node corresponding to the selected comprehensive feature similarity is used as the associated entity of the new entity node, and then the unique identification values ​​of the associated entities are arranged in ascending order according to the selected comprehensive feature similarity, and added to the new entity node as an attribute;

[0086] The entity node update module uploads the created new entity node to the feature library module for storage;

[0087] In the present invention, the associated entity nodes of the new entity node are obtained through the entity node update module, and the color feature similarity, texture feature similarity and shape feature similarity of the textile image corresponding to the new entity node and the textile image corresponding to the entity node stored in the feature library module are calculated respectively, and then the comprehensive feature similarity is calculated. By comparing each comprehensive feature similarity with the preset comprehensive feature similarity, the entity node that is relatively similar to the textile image corresponding to the new entity node is screened out as the associated entity of the new entity node, and its unique identification value is added as an attribute to the new entity node, which is convenient for subsequent query and retrieval. When the target entity node is retrieved, other similar entity nodes can also be retrieved, providing more textile choices for the production, sales and consumption ends.

[0088] The above formulas are all calculated by removing dimensions and taking numerical values. The formula is a formula that is closest to the actual situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained by simulating a large amount of data.

[0089] In the embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation; the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the method of this embodiment.

[0090] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A system for updating a textile image feature library, characterized in that: include: Information upload module, information preprocessing module, cloud storage module, feature library module and entity node update module; The information uploading module is used to obtain the textile image and the corresponding text description information uploaded by the user, and send the obtained textile image and the corresponding text description information to the information preprocessing module with a time stamp; The information preprocessing module is used to preprocess the obtained textile images and corresponding text description information to obtain each feature data set and unique identification value of the corresponding textile image; the textile images and text description information are both tagged with unique identification value and sent to the cloud storage module, and each feature data set is tagged with unique identification value and sent to the entity node update module; including: The digital image processing technology is used to extract features of the obtained textile images, and a color feature data set C = {C1, C2, ..., Ci, ..., Cn}, a texture feature data set T = {T1, T2, ..., Tj, ..., Tm} and a shape feature data set S = {S1, S2, ..., Sh, ..., Sg} of the corresponding textile images are obtained; Wherein, i represents the interval number of the color histogram in the color feature data set, i=1,2…n; n represents the total number of intervals of the color histogram; Ci represents the i-th interval value of the color histogram in the color feature data set; j represents the gray level number of the gray level co-occurrence matrix in the texture feature data set, j=1,2…m; m represents the total number of gray levels in the gray level co-occurrence matrix; Tj represents the element value of the j-th gray level of the gray level co-occurrence matrix in the texture feature data set; h represents the number of the shape feature in the shape feature data set, h=1,2…g; g represents the total number of shape features; Sh represents the element value of the h-th shape feature in the shape feature data set; Connecting the acquired color feature vector C, texture feature vector T and shape feature vector S together to form a comprehensive feature vector F of the textile image; F=[C,T,S]=[C1,C2,…,Ci,…,Cn,T1,T2,…,Tj,…,Tm,S1,S2,…,Sh,…,Sg]; According to the hash function H, the obtained comprehensive feature vector F is mapped to an identification value HV; wherein HV=H(F); the hash function is selected as SHA-256 or CityHash; After the identification value HV is stamped with the upload timestamp, a unique identification value of the textile image is formed; The cloud storage module is used to receive and store the obtained textile images and text description information marked with unique identification value tags; and mark the textile image file path and text description information file path with unique identification value tags and send them to the entity node update module; The feature library module includes several entity nodes; The entity node update module is used to generate a new entity node according to each feature data set labeled with a unique identification value, the textile image file path and the text description information file path; upload the created new entity node to the feature library module for storage; Text description information refers to the information of the textile referred to by the corresponding textile image, including the category, color, fabric, size, supplier information and designer information of the textile; The cloud storage module includes a picture storage unit and a text storage unit; The cloud storage module receives a textile image and text description information marked with a unique identification value; according to the format of the textile image and text description information, the textile image marked with a unique identification value label is stored in the image storage unit, and the file path of the textile image is marked with a unique identification value label and sent to the entity node update module; the file description information marked with a unique identification value label is stored in the text storage unit, and the file path of the text description information is marked with a unique identification value label and sent to the entity node update module.

2. The updating system of a textile image feature library according to claim 1, characterized in that: The entity node is used to store the unique identification value of the textile image, the textile image file path, the text description information file path, the color feature data set C, the texture feature data set T, the shape feature data set S and the unique identification value of the associated entity node.

3. The updating system of a textile image feature library according to claim 1, characterized in that: The processing process of the entity node update module is as follows: Get each feature data set and file path labeled with a unique identification value; Compare the unique identification values ​​of each feature data set and file path; filter out feature data sets and file paths with the same unique identification value, and create a new entity node based on the unique identification value; Add the unique identification value, each feature data set, and file path as attributes to the new entity node; Count the number of file paths of the new entity node; if the number of file paths of the new entity node is equal to 2, determine whether the two file paths of the new entity node are the same; if they are the same, end; if they are not the same, re-acquire the corresponding textile image file path and text description information file path according to the unique identification value; if the number of file paths of the new entity node is greater than 2, re-acquire the corresponding textile image file path and text description information file path according to the unique identification value; Calculate the comprehensive feature similarity between the textile image corresponding to the new entity node and the textile image corresponding to the entity node stored in the feature library module according to the color feature data set C, the texture feature data set T and the shape feature data set S, and then obtain the unique identification value attribute of the associated entity node of the new entity node; The created new entity node is uploaded to the feature library module for storage.

4. The updating system of a textile image feature library according to claim 1, characterized in that: The process of obtaining the unique identification value attribute of the associated entity node of the new entity node is as follows: Step S1: number all entity nodes stored in the feature library module in order of storage time, numbered as k; Step S2: Calculate and obtain the comprehensive feature similarity ZXk between the textile image corresponding to each entity node in the feature library module and the textile image corresponding to the new entity node, including: Obtain a color feature data set C, a texture feature data set T, and a shape feature data set S of a new entity node; and extract a color feature data set Ck, a texture feature data set Tk, and a shape feature data set Sk of the entity node in the feature library module; According to the calculation formula Obtain the color feature similarity X(C, Ck) between the new entity node and the entity node numbered k; where Cki represents the i-th interval value of the color histogram in the color feature data set of the entity node numbered k; According to the calculation formula Obtain the texture feature similarity X(T,Tk) between the new entity node and the entity node numbered k; where Tkj represents the element value of the jth gray level of the gray level co-occurrence matrix in the texture feature data set of the entity node numbered k; According to the calculation formula Obtain the shape feature similarity X(S, Sk) between the new entity node and the entity node numbered k; where Skh represents the element value of the h-th shape feature of the shape feature data set of the entity node numbered k; According to the calculation formula ZXk=a1×X(C,Ck)+a2×X(T,Tk)+a3×X(S,Sk), the comprehensive feature similarity ZXk between the new entity node and the entity node numbered k is obtained; where a1, a2 and a3 are the weight coefficients of color feature similarity, texture feature similarity and shape similarity respectively; Step S3: obtaining the comprehensive feature similarity between each entity node stored in the feature library module and the new entity node; comparing each obtained comprehensive feature similarity with a preset comprehensive feature similarity; Screen out comprehensive feature similarities that are less than a preset comprehensive feature similarity; and count the number y of the screened comprehensive feature similarities; If y is equal to 1, the entity node corresponding to the screened comprehensive feature similarity is used as the associated entity of the new entity node, and the unique identification value of the associated entity node is added to the new entity node as an attribute; If y is greater than 1, each entity node corresponding to the screened comprehensive feature similarity is used as the associated entity of the new entity node, and then the unique identification value of the associated entity is arranged in ascending order according to the screened comprehensive feature similarity, and added to the new entity node as an attribute.

Images