Article texture recognition and analysis system based on AR technology

Abstract

The invention discloses an article texture recognition and analysis system based on an AR technology, and relates to the field of texture recognition, and the system comprises a texture recognition platform which comprises a data collection module, a data classification module, a data decomposition module, a data recognition module, a virtual interaction module and a texture visualization module; the data acquisition module acquires historical article texture identification information and article acquisition information; the data classification module obtains texture feature information corresponding to the texture identification information of the historical articles and stores the texture feature information; the data decomposition module analyzes the article collection information to obtain a dimension image data subset of each plane dimension; the data identification module retrieves corresponding texture feature information according to the dimension image data subset; the virtual interaction module generates virtual texture data according to the retrieval result; the texture visualization module constructs a three-dimensional article texture visualization image according to the virtual texture data; according to the invention, the efficiency in the article texture identification process is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of texture recognition, and in particular to an article texture recognition and analysis system based on AR technology. Background Art

[0002] With the progress of technology, more and more robots and computer vision technologies are applied in various fields, and AR technology is one of them. It combines virtual information with the real world to bring users a richer and more immersive usage experience. Traditional methods for recognizing and analyzing article textures mostly rely on manual experience or simple physical measurements. However, these methods have certain limitations and are difficult to meet the requirements for comprehensive, accurate, and efficient inspection and recognition of articles. In recent years, article texture recognition and analysis systems based on AR technology have attracted wide attention from researchers.

[0003] After retrieval, the invention patent with Chinese patent number CN104568749A discloses a method, device, recognition equipment and system for recognizing the surface material of an article, including: acquiring an image of the surface of the article to be recognized collected by a camera; through texture recognition of the image, determining the material category to which the surface of the article to be recognized belongs from a preset plurality of material categories; then acquiring the spectral characteristics of the characteristic light emitted after the surface of the article to be recognized is irradiated by a specified light; and determining the material of the surface of the article to be recognized by comparing the spectral characteristics with the spectral characteristics corresponding to each material included in the material category. By adopting the solution provided by the embodiment of the present invention, compared with the prior art, the volume of the high-precision article surface material recognition equipment is reduced.

[0004] Compared with the prior art, the invention patent with Chinese patent number CN104568749A can perform texture recognition on the corresponding article surface image and obtain the material category consistent with the corresponding article surface within a plurality of material categories;

[0005] However, in the actual use process, the process of determining the article texture data by successively comparing and analyzing the existing material analogies reduces the article texture recognition efficiency to a certain extent. Summary of the Invention

[0006] The purpose of the present invention is to solve the problem of low recognition efficiency in the prior art, and to propose an article texture recognition and analysis system based on AR technology.

[0007] To achieve the above purpose, the present invention adopts the following technical solutions:

[0008] An item texture recognition and analysis system based on AR technology, including a texture recognition platform, which includes a data acquisition module, a data classification module, a data decomposition module, a data recognition module, a virtual interaction module, and a texture visualization module;

[0009] The data acquisition module is used to collect historical item texture recognition information and item acquisition information;

[0010] The data classification module is used to analyze and process the collected historical item texture recognition information, obtain corresponding texture feature information, and set up a texture feature space to classify and store the texture feature information;

[0011] The data decomposition module is used to perform multi-dimensional decomposition on the collected item acquisition information, obtain a subset of dimension image data corresponding to the corresponding planar dimension of the item according to the decomposition result, and generate a decomposition data packet and send it to the data recognition module;

[0012] The data recognition module is used to obtain the corresponding decomposition data packet, obtain the corresponding texture feature values in the subset of dimension image data, retrieve the corresponding texture feature space according to the texture feature values, and generate a retrieval data packet according to the retrieval result;

[0013] The virtual interaction module is used to obtain the retrieval data packet, analyze and process the retrieval data packet, obtain virtual texture data corresponding to each planar dimension, and send the virtual texture data to the texture visualization module;

[0014] The texture visualization module is used to obtain the virtual texture data corresponding to each planar dimension, analyze and process according to the virtual texture data, and construct a three-dimensional item texture visualization image.

[0015] The above technical solution further includes: The process of collecting historical item texture recognition information and item acquisition information includes:

[0016] Set up a historical information acquisition unit and an item information acquisition unit;

[0017] A multi-source channel set and corresponding data monitoring software are set in the historical information acquisition unit, and the multi-source channel set is monitored in real time through the data monitoring software to obtain historical item texture recognition information;

[0018] An information entry window and a device entry window are set in the item information acquisition unit. The item description information is obtained through the information entry window, and the corresponding item image information and light sensor data are collected through the device entry window. The collected data is uniformly marked to generate item acquisition information.

[0019] Further, the process of analyzing and processing the historical item texture recognition information includes:

[0020] Obtain the texture recognition information of historical items, perform semantic analysis on the item keyword information in the texture recognition information of historical items, and obtain the corresponding item attributes according to the semantic analysis results;

[0021] Set the corresponding classification system structure according to the corresponding item attributes. The classification system structure includes the classification criteria corresponding to the corresponding item attributes. Obtain the key data sets corresponding to the item keyword information according to the classification criteria, analyze and process the key data sets corresponding to each classification criterion, obtain the texture feature information corresponding to each classification criterion, and sequentially set the corresponding feature elements according to the texture feature information.

[0022] Further, the process of setting up a texture feature space for classifying and storing texture feature information includes:

[0023] Obtain each key data set, set each key data set and the included feature elements as keyword vectors, sort and combine the obtained keyword vectors, and obtain the similarity data between the corresponding keyword vectors in each combination result;

[0024] Sort the classification criteria corresponding to each keyword vector according to the obtained similarity data, and sort the feature elements in the corresponding keyword vectors; perform feature analysis on each feature element according to the corresponding classification criteria according to the sorting results, obtain the feature values corresponding to each feature element, and obtain the feature value intervals corresponding to the corresponding classification criteria according to the obtained feature values;

[0025] Set up a texture feature space, obtain the feature value intervals corresponding to the corresponding sorting results of the classification system structure, arrange them vertically from large to small in sequence, set the corresponding horizontal texture space according to the vertical arrangement results, in the horizontal texture space, obtain the sorting results of the feature values corresponding to the corresponding classification criteria, perform segmented processing on the horizontal texture space according to the sorting results of the corresponding feature values, obtain the corresponding texture feature domains, and store the corresponding historical item recognition information in the corresponding texture feature domains.

[0026] Further, the process of performing multi-dimensional decomposition on item acquisition information to generate decomposition data packets includes:

[0027] Obtain the corresponding item image information and light sensor data in the item acquisition information, perform environmental perception processing on the pixel values in the corresponding pixels in the item image information according to the light sensor data, obtain the pixel values corresponding to the item image information after environmental perception processing, analyze and process the pixel values of the corresponding pixels in the image information, obtain the amplitude information corresponding to the corresponding pixel values, and obtain the corresponding image signals;

[0028] Perform spatial discretization processing on the item image information in sequence, obtain the corresponding planar dimensions within the item image information, and mark each planar dimension in sequence;

[0029] Obtain the image signals corresponding to each planar dimension, perform two-dimensional discretization processing on the image signals, obtain the low-frequency information and high-frequency information extracted by the low-pass filter and high-pass filter respectively in the horizontal direction, perform filtering processing on the obtained low-frequency information and high-frequency information respectively in the vertical direction, obtain the corresponding dimensional image data subsets, generate decomposition data packets from the obtained dimensional image data subsets, and send the decomposition data packets to the data recognition module.

[0030] Further, the process of the data recognition module generating the retrieval data packet includes:

[0031] Obtain the dimensional image data subsets corresponding to the corresponding planar dimensions in the decomposition data packet, perform feature calculation according to the dimensional image data subsets, obtain the corresponding feature data, construct a texture feature vector for the corresponding feature data, and obtain the corresponding texture feature value according to the texture feature vector;

[0032] Obtain the range of eigenvalue intervals corresponding to each horizontal texture space in the texture feature space, obtain the corresponding median value within the eigenvalue interval range, divide the eigenvalue interval range into a left subset and a right subset according to the obtained median value, and respectively obtain the range of subset eigenvalue intervals corresponding to the left subset and the right subset;

[0033] Compare and analyze the texture feature value with the corresponding subset eigenvalue interval range, obtain the subset to which the texture feature value belongs, and obtain the median value of the corresponding subset. Repeat the division of the belonging subset until the eigenvalue corresponding to the corresponding texture feature domain is obtained. Generate retrieval data packets from the texture feature domains corresponding to each planar dimension, and send the retrieval data packets to the virtual interaction module.

[0034] Further, the process of obtaining the virtual texture data corresponding to each planar dimension includes:

[0035] Obtain the texture feature values of each dimensional image data subset corresponding to each planar dimension, perform analysis processing on each texture feature value, obtain the corresponding initial texture data, and perform smoothing processing on each initial texture data within the planar dimension to obtain the corresponding local texture data;

[0036] Obtain the retrieval data packet, generate a corresponding interpolation fitting model according to the corresponding historical item recognition information in the retrieval data packet, and the interpolation fitting model is used for texture reconstruction of the local texture data;

[0037] Input the local texture data corresponding to each planar dimension into the interpolation fitting model, obtain the virtual texture data of each planar dimension after texture reconstruction analysis, and send the obtained virtual texture data to the texture visualization module.

[0038] Further, the process of constructing the three-dimensional object texture visualization image includes:

[0039] Obtain the virtual texture data of each planar dimension after verification processing, calculate the area of the planar dimension, sort the calculation results, and obtain the planar dimension corresponding to the maximum area;

[0040] Set up a three-dimensional space coordinate system based on the planar dimension corresponding to the maximum area;

[0041] Map the planar dimensions corresponding to the object image information to the three-dimensional space coordinate system respectively to obtain the corresponding object model, and map the corresponding virtual texture data to the corresponding positions of the object model according to the belonging planar dimension to obtain the three-dimensional object texture model;

[0042] Render the three-dimensional object texture model based on virtual display technology to obtain the three-dimensional object texture visualization image.

[0043] The present invention has the following beneficial effects:

[0044] 1. In the present invention, by constructing a classification system structure, classifying the historical object recognition information according to the corresponding classification criteria in the classification system structure, setting corresponding key data sets, setting corresponding texture feature information for each key data set, analyzing the correlation of the texture feature information corresponding to different key data sets, and sorting the feature elements corresponding to different key data sets according to the correlation analysis results, so that there is a corresponding storage rule in the storage process of the corresponding historical object recognition information, which is more convenient to quickly retrieve the corresponding historical object recognition information in the subsequent object recognition process, thereby improving the recognition efficiency in the object recognition and analysis process.

[0045] 2. In the present invention, during the object texture recognition process, corresponding planar dimensions are set, multiple filtering processing is performed from multiple dimensions according to the planar dimensions to obtain corresponding dimensional image data subsets, and each dimensional image data subset is analyzed and processed respectively to obtain corresponding texture feature values, thereby improving the accuracy in the object texture recognition process. In addition, through texture reconstruction of the interpolation fitting model constructed for the historical object recognition information corresponding to different planar dimensions, the connection of different textures in each planar dimension is more natural, which also improves the naturalness of texture recognition;

[0046] 3. In the present invention, by calculating the areas of each planar dimension and setting the corresponding three-dimensional space coordinate system according to the area calculation results, the process of constructing the three-dimensional object texture visualization image is made more convenient, and the efficiency in the process of object texture recognition is also improved to a certain extent. Brief Description of the Drawings

[0047] Figure 1 It is a schematic structural diagram of an object texture recognition and analysis system based on AR technology proposed by the present invention;

[0048] Figure 2 It is a schematic flow diagram in the present invention. Detailed Embodiments

[0049] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0050] Embodiment 1

[0051] As Figure 1 shown, an object texture recognition and analysis system based on AR technology proposed by the present invention includes a texture recognition platform, and the texture recognition platform includes a data acquisition module, a data classification module, a data decomposition module, a data recognition module, a virtual interaction module, and a texture visualization module.

[0052] In this embodiment, the texture recognition platform is used to obtain corresponding object information, analyze and process the object information, and obtain the texture features of the corresponding object information. The specific implementation process includes:

[0053] The data acquisition module is used to obtain historical object texture recognition information and object acquisition information, send the obtained historical object texture recognition information to the data classification module, and send the obtained object acquisition information to the data decomposition module. The specific implementation process includes:

[0054] Set a historical information acquisition unit and an object information acquisition unit;

[0055] The historical information acquisition unit is used to collect corresponding historical object texture recognition information, in which a multi-source channel set is set. The multi-source channel set includes corresponding scientific research institutions, university laboratories, commercial monitoring and analysis companies, data resource platforms, and professional academic resources, etc. The multi-source channels that have passed the query permission are unified and integrated to construct a multi-source channel set;

[0056] Build a data monitoring software, monitor and process the multi-source channel set through the data monitoring software, obtain the historical item texture recognition information with authorized permissions, where the historical item texture recognition information includes various texture recognition information such as corresponding item keyword information, material texture feature information, decorative texture feature information, trace texture feature information, etc., and mark the corresponding historical item texture recognition information;

[0057] The item information acquisition unit is used to obtain item acquisition information, and an information entry window and a device entry window are set therein;

[0058] The information entry window is used for corresponding staff to enter corresponding item description information, and the item description information includes description keywords, description statements, etc. for describing the corresponding item;

[0059] The device entry window is used for corresponding acquisition devices to obtain corresponding item image information. The acquisition devices include corresponding high-definition cameras, depth sensors, etc. The image information and light sensor data of the corresponding item are collected through the high-definition camera and the depth sensor;

[0060] Unify the obtained item description information and item image information for marking to generate item acquisition information, and send the obtained item acquisition information to the data decomposition module.

[0061] The data classification module is used to obtain the corresponding historical item texture recognition information, split and store the obtained historical item texture recognition information, set a texture feature space, and store the texture feature information corresponding to the split result in the texture feature space. Its specific implementation process includes:

[0062] Set a data classification unit and a data storage unit;

[0063] The data classification unit is used to obtain historical item texture recognition information, split and process the obtained historical item texture recognition information, and classify and mark the split processing result. Its specific implementation process includes:

[0064] Obtain historical item texture recognition information, perform semantic analysis on the item keyword information in the historical item texture recognition information, and obtain the corresponding item attributes according to the semantic analysis result. The item attributes include item type, use, material and other attributes;

[0065] Set a corresponding classification system structure according to the corresponding item attributes. The classification system structure includes classification criteria corresponding to the corresponding item attributes, perform data standardization processing on the classification criteria corresponding to the same item attributes, and obtain classification sub-criteria corresponding to the corresponding classification criteria. Its specific implementation process includes:

[0066] Obtain the item keyword information corresponding to the corresponding classification standard, set the obtained item keyword information as the corresponding key data set, analyze and process the key data set, and set the corresponding keyword vector in, a i For keyword vector The i-th characteristic element in the keyword vector is set according to the corresponding characteristic element in the keyword vector, and the corresponding data classification processing results are arranged and combined. The corresponding similarity data is obtained according to the arrangement and combination, and the similarity data is marked as α, and the keyword vectors corresponding to other classification standards in the combination are marked as in, b i For keyword vector The i-th characteristic element in

[0067] Analyze and process each similarity data, and sort the classification standards corresponding to the corresponding item attributes in the classification system according to the similarity data of each permutation and combination;

[0068] The similarity data of the feature elements corresponding to the keyword vector corresponding to the classification standard is also analyzed, and the corresponding feature elements are sorted in sequence according to the analysis results;

[0069] According to the sorting result, a feature mapping function corresponding to the corresponding classification standard is set, a feature value corresponding to the corresponding feature element is obtained, and the obtained feature value is matched with the corresponding classification standard and the feature element;

[0070] Acquire the feature value in the object identification process corresponding to each historical object identification information, and send the corresponding feature value to the data storage unit;

[0071] The data storage unit is used to obtain the classification and labeling results corresponding to the texture recognition information of each historical item, set the texture feature space, and store the corresponding classification and labeling results in the texture feature space. The specific implementation process includes:

[0072] Obtaining a classification system architecture sorting result corresponding to the historical item recognition information, and setting a texture feature space according to the classification system architecture sorting result;

[0073] Obtaining the characteristic value intervals corresponding to the classification criteria corresponding to the sorting results of the classification architecture and arranging them vertically from large to small, and setting the corresponding horizontal texture space according to the vertical arrangement results;

[0074] In the horizontal texture space, obtain the sorting result of the eigenvalues corresponding to the corresponding classification criteria, segment the horizontal texture space according to the sorting result of the corresponding eigenvalues, and obtain the corresponding texture feature domains;

[0075] Store the corresponding historical item identification information into the corresponding texture feature domain according to the corresponding feature element analysis result;

[0076] Integrate and mark the eigenvalue intervals corresponding to each texture feature domain, integrate the storage results of the corresponding historical item identification information according to the marking result, generate a texture feature space, and store the corresponding texture feature space.

[0077] The data decomposition module is used to obtain item acquisition information, perform multi-dimensional decomposition on the obtained item acquisition information, sequentially set corresponding dimensional image data subsets according to the multi-dimensional decomposition result, and send the obtained dimensional image data subsets to the data recognition module. Its specific implementation process includes:

[0078] Set a data processing unit and a data decomposition unit;

[0079] The data processing unit is used to obtain the corresponding item image information in the item acquisition information, perform environmental perception processing on the item image information, correct the item image information according to the environmental perception processing result, obtain the corrected item image information, and send it to the data decomposition unit;

[0080] Obtain the image information and light sense information of the item image information, perform analysis processing on the image information and light sense information, and perform white balance processing based on the light sense information obtained from the item image information;

[0081] Obtain the pixel values in the corresponding image information, perform analysis processing on the pixel values of each pixel point, judge whether the corresponding pixel point is white, and if the corresponding pixel point is white, mark the pixel point as a reference point;

[0082] Obtain the average values corresponding to the three color channels of the corresponding reference point, and record them as R ref、 G ref and B ref ; Obtain the color channel values of the corresponding pixel points in the image information, and record them as R(x,y) 、 G(x,y and B(x,y); Analyze and process the color channel values of the corresponding pixel points in the image information according to the channel average values, and obtain the corrected color channel values, which are respectively recorded as R j (x,y) 、 G j (x,y) and B j where:

[0083]

[0084] Send the corrected image information to the data decomposition unit;

[0085] The data decomposition unit obtains the pixel values corresponding to the image information after correction processing, analyzes and processes the pixel values of the corresponding pixel points in the image information, performs mapping processing on the corresponding pixel values, obtains the amplitude information corresponding to the corresponding pixel values, and performs mapping processing according to the amplitude information corresponding to each pixel point in the image information to obtain the corresponding image signal;

[0086] Perform spatial discretization processing on the item image information in sequence, obtain the corresponding planar dimensions in the item image information, and mark the item positions where the corresponding planar dimensions are located;

[0087] Obtain and analyze the image signals corresponding to the planar dimensions of the corresponding marking results. The specific implementation process includes:

[0088] Perform two-dimensional discretization processing on the image signals corresponding to the corresponding marking results, mark the corresponding image as I(x, y), and pass the corresponding horizontal direction in the image signal through the low-pass filter h L (n) and the high-pass filter h H (n) to extract the corresponding low-frequency information and high-frequency information, and mark the image processed by horizontal direction filtering as L r (x, y) and H r (x, y);

[0089] Perform filtering processing on the corresponding low-frequency information and high-frequency information in the horizontal direction in the vertical direction through the low-pass filter h L (n) and the high-pass filter h H (n);

[0090] For example, the process of performing low-pass filtering processing on L r (x, y) corresponding to the low-frequency information in the vertical direction to obtain the corresponding LL sub-band information set includes:

[0091] LL(x, y) = ∑ n h L (n)L r (x - n, y). The methods for analyzing and processing other information in the vertical direction are the same as this method to obtain the corresponding sub-band information sets. The sub-band information sets respectively include LL sub-band information set, LH sub-band information set, HH sub-band information set, and HL sub-band information set;

[0092] Mark each sub-band information set corresponding to the decomposition result of the corresponding planar dimension in the item image information corresponding to the device as the corresponding dimensional image data subset;

[0093] Mark the item image information to which the obtained subset of dimensional image data belongs, generate a decomposition data packet according to the marking result, and send the obtained decomposition data packet to the data recognition module.

[0094] The data recognition module is used to obtain the corresponding decomposition data packet, obtain the corresponding texture feature values in the subset of dimensional image data, retrieve the corresponding texture feature space according to the texture feature values, and generate a retrieval data packet according to the retrieval result. The specific implementation process includes:

[0095] Set up a texture recognition unit and a data retrieval unit;

[0096] The texture recognition unit is used to obtain the corresponding decomposition data packet, analyze and process the obtained decomposition data packet, obtain the subset of dimensional image data corresponding to each dimension, analyze and process the subset of dimensional image data, obtain the corresponding sub-band information set in the corresponding dimensional image data, perform feature calculations on the sub-band information set respectively, and obtain the corresponding standard deviation feature data, skewness feature data and kurtosis feature data;

[0097] Construct a texture feature vector from the obtained standard deviation feature data, skewness feature data and kurtosis feature data;

[0098] Analyze and process the obtained texture feature vector to obtain the corresponding texture feature value;

[0099] Mark the texture feature values corresponding to the corresponding plane dimensions in the item image information and send them to the data retrieval unit;

[0100] The data retrieval unit is used to obtain the texture feature space and the corresponding texture feature values, obtain the corresponding texture feature information in the texture feature space according to the texture feature values, and send the obtained texture feature information and the corresponding plane dimension of the item image information to the virtual interaction module. The specific implementation process includes:

[0101] Obtain the eigenvalue intervals corresponding to each texture feature domain in the texture feature space, analyze and process each eigenvalue interval, obtain the eigenvalue interval range corresponding to all eigenvalue intervals in the texture feature space, analyze and process the eigenvalue interval range, obtain the median value corresponding to the eigenvalue interval range, divide the eigenvalue interval range into a left subset and a right subset according to the obtained median value, and respectively obtain the subset eigenvalue interval ranges corresponding to the left subset and the right subset;

[0102] Obtain the texture feature values corresponding to the corresponding planar dimensions within the item image information, compare and analyze the obtained texture feature values with the subset feature value range corresponding to the left subset and the right subset respectively, and judge the subset feature value range to which the texture feature value corresponding to the corresponding planar dimension belongs according to the comparison and analysis results;

[0103] If the corresponding texture feature value belongs to the subset feature value range corresponding to the left subset, analyze and process the subset feature value range corresponding to the left subset, obtain the median value corresponding to the subset feature value range corresponding to the left subset, and divide the subset feature value range corresponding to the left subset again according to the corresponding median value to obtain the left subset and the right subset corresponding to the left subset, and recursively retrieve within the corresponding subset until the corresponding texture feature domain corresponding to the texture feature value is obtained, and obtain the corresponding historical item recognition information within the texture feature domain;

[0104] Generate a retrieval data packet for the historical item recognition information corresponding to the texture feature values corresponding to each planar dimension of the item image information, and send the retrieval data packet corresponding to the item image information to the virtual interaction module for analysis and processing.

[0105] The texture visualization module is used to obtain the virtual texture data corresponding to each planar dimension, analyze and process according to the virtual texture data, and construct a three-dimensional item texture visualization image. The specific implementation process includes:

[0106] Obtain the texture feature values of each subset of image data corresponding to each planar dimension, analyze and process each texture feature value to obtain the corresponding initial texture data, and smooth each initial texture data within the planar dimension to obtain the corresponding initial virtual texture data within the planar dimension;

[0107] Preset a texture smoothing index, and compare and analyze the obtained initial virtual texture data with the texture smoothing index:

[0108] If it does not meet the standard corresponding to the texture smoothing index, mark the corresponding pixel points as local texture data;

[0109] If it meets the standard corresponding to the texture smoothing index, do not mark the corresponding pixel points as local texture data;

[0110] Obtain the local texture data marked accordingly;

[0111] Obtain a retrieval data packet, generate a corresponding interpolation fitting model according to the corresponding historical item recognition information in the retrieval data packet, and the interpolation fitting model is used to reconstruct the texture of local texture data. Obtain the corresponding texture data within adjacent pixel points and perform linear interpolation and polynomial fitting on the local texture data of the corresponding pixel points using the historical item recognition information in the interpolation fitting model to obtain the corresponding texture data;

[0112] Input the local texture data corresponding to each plane dimension into the interpolation fitting model, obtain the virtual texture data after texture reconstruction analysis for each plane dimension, and send the obtained virtual texture data to the texture visualization module;

[0113] It should be further noted that in the specific implementation process, the virtual texture data includes the local texture data after analysis and processing and the initial virtual texture data.

[0114] The texture visualization module is used to obtain the virtual texture data corresponding to each plane dimension, perform analysis and processing according to the virtual texture data, and construct a three-dimensional item texture visualization image. Its specific implementation process includes:

[0115] Obtain the virtual texture data that has passed the verification process for each plane dimension, calculate the area of the plane dimension, and sort the calculation results to obtain the plane dimension corresponding to the maximum area;

[0116] Set up a three-dimensional space coordinate system based on the plane dimension corresponding to the maximum area;

[0117] Map the plane dimensions corresponding to the item image information to the three-dimensional space coordinate system respectively to obtain the corresponding item model, and map the corresponding virtual texture data to the corresponding positions of the item model according to the plane dimension to obtain a three-dimensional item texture model;

[0118] Render the three-dimensional item texture model based on virtual display technology to obtain a three-dimensional item texture visualization image.

[0119] Embodiment 2

[0120] As Figure 2 shown, based on Embodiment 1, the process of item texture recognition and analysis is summarized as follows:

[0121] In this embodiment, the data acquisition module is used to acquire historical item texture recognition information and item acquisition information;

[0122] The data classification module is used to analyze and process the acquired historical item texture recognition information, obtain the corresponding texture feature information, and set up a texture feature space to classify and store the texture feature information;

[0123] The data decomposition module is used to perform multi-dimensional decomposition on the collected item collection information, obtain the subset of dimension image data corresponding to the corresponding plane dimension of the item according to the decomposition result, and generate a decomposition data packet to be sent to the data recognition module;

[0124] The data recognition module is used to obtain the corresponding decomposition data packet, obtain the corresponding texture feature value in the subset of dimension image data, retrieve the corresponding texture feature space according to the texture feature value, and generate a retrieval data packet according to the retrieval result;

[0125] The virtual interaction module is used to obtain the retrieval data packet, analyze and process the retrieval data packet, obtain the virtual texture data corresponding to each plane dimension, and send the virtual texture data to the texture visualization module;

[0126] The texture visualization module is used to obtain the virtual texture data corresponding to each plane dimension, analyze and process according to the virtual texture data, and construct a three-dimensional item texture visualization image;

[0127] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An object texture recognition and analysis system based on AR technology, including a texture recognition platform, characterized in that The texture recognition platform includes a data acquisition module, a data classification module, a data decomposition module, a data recognition module, a virtual interaction module, and a texture visualization module; The data acquisition module is used to acquire historical item texture recognition information and item acquisition information; The data classification module is used to analyze and process the acquired historical item texture recognition information, obtain corresponding texture feature information, and set up a texture feature space to classify and store the texture feature information; The data decomposition module is used to perform multi-dimensional decomposition on the acquired item acquisition information, obtain a subset of dimension image data corresponding to the corresponding planar dimension of the item according to the decomposition result, and generate a decomposition data packet to send to the data recognition module; The data recognition module is used to obtain the corresponding decomposition data packet, obtain the corresponding texture feature values in each subset of dimension image data, retrieve the corresponding texture feature space according to the texture feature values, and generate a retrieval data packet according to the retrieval result; The virtual interaction module is used to obtain the retrieval data packet, analyze and process the retrieval data packet, and obtain the virtual texture data corresponding to each planar dimension; The texture visualization module is used to obtain the virtual texture data corresponding to each planar dimension, analyze and process according to the virtual texture data, and construct a three-dimensional item texture visualization image.

2. The texture recognition and analysis system of an article based on AR technology according to claim 1, characterized in that, The process of acquiring historical item texture recognition information and item acquisition information includes: Set up a historical information acquisition unit and an item information acquisition unit; A multi-source channel set and corresponding data monitoring software are set in the historical information acquisition unit, and the multi-source channel set is monitored in real time through the data monitoring software to obtain historical item texture recognition information; An information entry window and a device entry window are set in the item information acquisition unit. The item description information is obtained through the information entry window, and the corresponding item image information and light sensor data are acquired through the device entry window. The acquired data is uniformly marked to generate item acquisition information.

3. An article texture recognition and analysis system based on AR technology according to claim 2, characterized in that, The process of analyzing and processing historical item texture recognition information includes: Obtain historical item texture recognition information, perform semantic analysis on the item keyword information in the historical item texture recognition information, and obtain the corresponding item attributes according to the semantic analysis result; Set up a corresponding classification system structure according to the corresponding item attributes. The classification system structure includes classification criteria corresponding to the corresponding item attributes. Obtain the key data sets corresponding to the item keyword information according to the classification criteria, analyze and process the key data sets corresponding to each classification criteria, obtain the texture feature information corresponding to each classification criteria, and sequentially set the corresponding feature elements according to the texture feature information.

4. An article texture recognition and analysis system based on AR technology according to claim 3, characterized in that, The process of setting up a texture feature space to classify and store texture feature information includes: Obtain each key data set, set the key data sets and the included feature elements as keyword vectors, sort and combine the obtained keyword vectors, and obtain the similarity data between the corresponding keyword vectors in each combination result; Sort the classification criteria corresponding to each keyword vector according to the obtained similarity data, and sort the characteristic elements within the corresponding keyword vector; perform characteristic analysis on each characteristic element according to the corresponding classification criteria based on the sorting results, obtain the characteristic values corresponding to each characteristic element, and obtain the characteristic value range corresponding to the corresponding classification criteria according to the obtained characteristic values; Set up a texture feature space, obtain the characteristic value ranges corresponding to the corresponding sorting results of the classification system structure, arrange them vertically from large to small in sequence, set the corresponding horizontal texture space according to the vertical arrangement results, within the horizontal texture space, obtain the sorting results of the characteristic values corresponding to the corresponding classification criteria, perform segmented processing on the horizontal texture space according to the sorting results of the corresponding characteristic values, obtain the corresponding texture feature domain, and store the corresponding historical item recognition information in the corresponding texture feature domain.

5. An article texture recognition and analysis system based on AR technology according to claim 4, characterized in that, The process of multidimensionally decomposing the item collection information to generate a decomposition data packet includes: Obtain the corresponding item image information and light sensor data in the item collection information, perform environmental perception processing on the pixel values within the corresponding pixel points in the item image information according to the light sensor data, obtain the pixel values corresponding to the item image information after environmental perception processing, analyze and process the pixel values of the corresponding pixel points in the image information, obtain the amplitude information corresponding to the corresponding pixel values, and obtain the corresponding image signal; Perform spatial discretization processing on the item image information in sequence, obtain the corresponding planar dimensions within the item image information, and mark each planar dimension in sequence; Obtain the image signals corresponding to each planar dimension, perform two-dimensional discretization processing on the image signals, obtain the low-frequency information and high-frequency information extracted by the low-pass filter and the high-pass filter respectively in the horizontal direction, perform filtering processing on the obtained low-frequency information and high-frequency information respectively in the vertical direction, obtain the corresponding dimensional image data subsets, generate a decomposition data packet from the obtained dimensional image data subsets, and send the decomposition data packet to the data recognition module.

6. An article texture recognition and analysis system based on AR technology according to claim 5, characterized in that, The process of the data recognition module generating a retrieval data packet includes: Obtain the dimensional image data subsets corresponding to the corresponding planar dimensions in the decomposition data packet, perform characteristic calculation according to the dimensional image data subsets, obtain the corresponding characteristic data, construct a texture feature vector for the corresponding characteristic data, and obtain the corresponding texture feature value according to the texture feature vector; Obtain the range of the characteristic value intervals corresponding to each horizontal texture space in the texture feature space, obtain the corresponding median value within the range of the characteristic value intervals, divide the range of the characteristic value intervals into a left subset and a right subset according to the obtained median value, and respectively obtain the range of the subset characteristic value intervals corresponding to the left subset and the right subset; Compare and analyze the texture feature value with the corresponding range of the subset characteristic value intervals, obtain the subset to which the texture feature value belongs, and obtain the median value of the corresponding subset. Repeat the division of the belonging subset until the characteristic value corresponding to the corresponding texture feature domain is obtained. Generate a retrieval data packet from the texture feature domains corresponding to each planar dimension, and send the retrieval data packet to the virtual interaction module.

7. An article texture recognition and analysis system based on AR technology according to claim 6, characterized in that, The process of obtaining the virtual texture data corresponding to each planar dimension includes: Obtain the texture feature values of the subsets of image data corresponding to each plane dimension, analyze and process each texture feature value to obtain the corresponding initial texture data, smooth each initial texture data within the plane dimension to obtain the corresponding local texture data; Obtain a retrieval data packet, generate a corresponding interpolation fitting model according to the corresponding historical item identification information in the retrieval data packet, and the interpolation fitting model is used for texture reconstruction of the local texture data; Input the local texture data corresponding to each plane dimension into the interpolation fitting model, obtain the virtual texture data after texture reconstruction analysis for each plane dimension, and send the obtained virtual texture data to the texture visualization module.

8. An article texture recognition and analysis system based on AR technology according to claim 7, characterized in that, The process of constructing a three-dimensional item texture visualization image includes: Obtain the virtual texture data that has passed the verification process for each plane dimension, calculate the area of the plane dimension, sort the calculation results, and obtain the plane dimension corresponding to the maximum area; Set up a three-dimensional space coordinate system based on the plane dimension corresponding to the maximum area; Map the plane dimensions corresponding to the item image information to the three-dimensional space coordinate system respectively to obtain the corresponding item model, and map the corresponding virtual texture data to the corresponding positions of the item model according to the plane dimension to obtain a three-dimensional item texture model; Render the three-dimensional item texture model based on virtual display technology to obtain a three-dimensional item texture visualization image.

Images