Material detection method and device and electronic equipment

Abstract

The invention relates to the technical field of material detection, in particular to a material detection method and device and electronic equipment. The method comprises the following steps: inputting a physical performance parameter vector of a target material into a first model to obtain a target signal feature vector; wherein the first model is obtained by establishing a mapping relation between a physical performance parameter vector and a signal feature vector through regression analysis. And combining the physical performance parameter vector and the target signal feature vector into a target comprehensive feature vector through feature splicing, and inputting the target comprehensive feature vector into a second model to obtain a second vector. Wherein the second model is obtained by performing multi-modal fusion training on the physical performance parameter vector, the signal feature vector and the subjective performance parameter vector based on a fusion algorithm. And determining the touch evaluation value of each dimension from the second vector based on a matrix mapping relation constructed by a multi-modal fusion algorithm. The method can realize accurate identification of material performance parameters.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of material detection, and in particular to a material detection method and device and electronic equipment. BACKGROUND

[0002] Traditional haptic feedback systems are mainly realized based on electric motors, piezoelectric elements or pneumatic devices, and simulate haptic experience through vibration, pressure change or mechanical feedback. When a user contacts the sensor surface, the device can perceive the physical parameters of the contact interface, such as the size of the acting force, the temperature characteristics of the contact object or the surface hardness. However, the existing sensor mainly captures macro mechanical parameters, and has insufficient analytical ability for fine haptic attributes such as material texture characteristics (such as fabric fiber density) and softness, and it is difficult to realize accurate identification of material performance parameters. SUMMARY

[0003] Embodiments of the present application provide a material detection method, device and electronic equipment, which are used to realize accurate identification of material performance parameters.

[0004] In a first aspect, the embodiments of the present application provide a material detection method, which comprises: inputting a physical performance parameter vector of a target material into a first model to obtain a target signal feature vector, the first model being obtained by establishing a mapping relationship between the physical performance parameter vector and the signal feature vector through regression analysis, the physical performance parameter vector representing the performance of the material under different external conditions, and the signal feature vector representing an electrical signal generated by the material due to touch; merging the physical performance parameter vector and the target signal feature vector into a target comprehensive feature vector through feature splicing, and inputting the target comprehensive feature vector into a second model to obtain a second vector, the second model being obtained by training the physical performance parameter vector, the signal feature vector and a subjective performance parameter vector in a multi-modal fusion manner based on a fusion algorithm, the subjective performance parameter vector representing common touch evaluation of different groups of people and individualized touch evaluation of a specific group of people; determining a touch evaluation value of each dimension from the second vector based on a matrix mapping relationship constructed based on the multi-modal fusion algorithm.

[0005] In the above method, real-time quantitative characterization of material performance is realized through triboelectric sensing technology. The mapping relationship among the physical performance parameter, the subjective performance parameter and the electrical signal feature is constructed by using the electrical signal generated in the haptic interaction process, and the collaborative analysis of the physical performance parameter and the subjective perception and the real-time feedback of the touch data are realized.

[0006] Optionally, the first model is trained in the following manner: a plurality of triboelectric signals generated by materials due to touch are obtained, and a signal feature vector is extracted from the triboelectric signals; determine features satisfying a preset threshold from the plurality of physical performance parameter vectors, to form a first feature set; determine covariances between different vectors in the first feature set, to construct a first covariance matrix, the first covariance matrix representing a linear correlation degree of the physical performance parameter vectors; project the physical performance parameter vectors onto a first principal component direction corresponding to a maximum eigenvalue in the first covariance matrix, to generate a first factor score matrix, each element in the first factor score matrix representing a projection value of a corresponding physical performance parameter vector in the first principal component direction; construct a regression model with the first factor score matrix as an input variable and the signal feature vector as a target variable, until the regression model satisfies a preset accuracy to obtain a first model.

[0007] In the above method, the correlation analysis is used to filter redundant parameters with low correlation with the triboelectric signal. The noise is filtered by constructing the covariance structure of the physical performance parameters. The iterative regression is used to improve the training accuracy of the first model.

[0008] Optionally, the second model is trained in the following manner: merge the physical performance parameter vector and the signal feature vector into a comprehensive feature vector through feature splicing, determine a second correlation coefficient between the comprehensive feature vector and the subjective performance parameter vector using the Pearson algorithm, to form a second correlation coefficient set; determine features in a preset threshold interval from the second correlation coefficient set, to form a second feature set; determine covariances between different variables in the second feature set, to construct a second covariance matrix based on the second feature set, the second covariance matrix representing a linear correlation degree of the signal feature vectors corresponding to the second feature set; project the comprehensive feature vector onto a second principal component direction corresponding to a maximum eigenvalue in the second covariance matrix, to generate a second factor score matrix, each element in the second factor score matrix representing a coordinate value of a corresponding signal feature in the second principal component direction; construct a regression model with the second factor score matrix as an input variable and the subjective performance parameter as a target variable, until the regression model satisfies a preset accuracy to obtain a second model.

[0009] In the above method, the correlation analysis is used to filter redundant parameters with low correlation with the subjective performance parameter. The noise is filtered by constructing the covariance structure of the physical performance parameters and the triboelectric signal. The iterative regression is used to improve the training accuracy of the second model.

[0010] Optionally, the physical performance parameters at least include surface feature performance, bending feature performance, compression feature performance, and thermal feature performance of the target material.

[0011] Optionally, the target subjective performance parameters at least include softness, warmth, coolness, smoothness and roughness of the target material.

[0012] Optionally, the method further comprises: determining the material category of the target material based on the threshold interval in which the target signal feature vector is located.

[0013] In the method, the material category can be determined in real time or quickly based on the pre-calibrated threshold interval, without complex calculation.

[0014] In a second aspect, an embodiment of the present application provides a material detection device, which comprises: a processing module configured to input a physical performance parameter vector of the target material into a first model to obtain a target signal feature vector, the first model being obtained by regression analysis to establish a mapping relationship between the physical performance parameter vector and the signal feature vector, the physical performance parameter vector representing a performance of the material under different external conditions, and the signal feature vector representing an electrical signal generated by the material due to touch; the processing module is further configured to merge the physical performance parameter vector and the target signal feature vector into a target comprehensive feature vector through feature splicing, and input the target comprehensive feature vector into a second model to obtain a second vector, the second model being obtained by training the physical performance parameter vector, the signal feature vector and a subjective performance parameter vector based on a fusion algorithm, the subjective performance parameter vector representing common touch evaluation of different groups of people and individualized touch evaluation of a specific group of people; a determining module configured to determine a touch evaluation value of each dimension from the second vector based on a matrix mapping relationship constructed by the multi-modal fusion algorithm.

[0015] In a third aspect, an embodiment of the present application provides an electronic device, which comprises a memory, a processor and a computer program stored in the memory and executable by the processor, and when the computer program is executed by the processor, the processor implements any of the methods in the first aspect.

[0016] In a fourth aspect, an embodiment of the present application further provides a computer readable storage medium, and the computer readable storage medium stores a computer program, and when the computer program is executed by a processor, any of the methods in the first aspect is implemented.

[0017] In a fifth aspect, an embodiment of the present application further provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement any of the methods in the first aspect.

[0018] The technical effects brought by any of the implementation manners of the second aspect to the fifth aspect can be referred to the technical effects brought by the corresponding implementation manners of the first aspect, which will not be described herein. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 A flow chart of a material detection method provided in an embodiment of the present application; Figure 2 A structural schematic diagram of collecting a triboelectric signal provided in an embodiment of the present application; Figure 3 A schematic diagram of a subjective performance parameter and a physical performance parameter provided in an embodiment of the present application; Figure 4a A subgraph of a schematic diagram of a correlation between a physical performance parameter and a subjective performance parameter provided in an embodiment of the present application; Figure 4b A subgraph of a schematic diagram of a correlation between a physical performance parameter and a subjective performance parameter provided in an embodiment of the present application; Figure 4c A subgraph of a schematic diagram of a correlation between a physical performance parameter and a subjective performance parameter provided in an embodiment of the present application; Figure 4d A subgraph of a schematic diagram of a correlation between a physical performance parameter and a subjective performance parameter provided in an embodiment of the present application; Figure 4e A subgraph of a schematic diagram of a correlation between a physical performance parameter and a subjective performance parameter provided in an embodiment of the present application; Figure 4f A schematic diagram of a correlation between a physical performance parameter and a subjective performance parameter provided in an embodiment of the present application; Figure 4g A subgraph of a schematic diagram of a correlation between a physical performance parameter and a subjective performance parameter provided in an embodiment of the present application; Figure 4h A subgraph of a schematic diagram of a correlation between a physical performance parameter and a subjective performance parameter provided in an embodiment of the present application; Figure 4i A subgraph of a schematic diagram of a correlation between a physical performance parameter and a subjective performance parameter provided in an embodiment of the present application; Figure 5 A structural schematic diagram of a material detection device provided in an embodiment of the present application; Figure 6 A structural schematic diagram of a control device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0020] Hereinafter, some terms in the embodiments of the present application are explained and described, so as to facilitate the understanding of the skilled in the art.

[0021] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0022] The application scenarios described in the embodiments of the present application are used to more clearly illustrate the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. A person of ordinary skill in the art can know that, as new application scenarios appear, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems. In the description of the present application, unless otherwise specified, the meaning of “multiple” is two or more than two.

[0023] Traditional haptic feedback systems are mainly implemented based on electric motors, piezoelectric elements or pneumatic devices, and simulate haptic experience through vibration, pressure change or mechanical feedback. When a user contacts the sensor surface, the device can sense the physical parameters of the contact interface, such as the size of the acting force, the temperature characteristics of the contact object or the surface hardness. However, existing sensors mainly capture macro mechanical parameters, and have insufficient ability to analyze fine haptic attributes such as material texture characteristics (such as fabric fiber density) and softness, so it is difficult to accurately identify the material performance parameters.

[0024] To solve the above problems, the embodiments of the present application provide a material detection method, device and electronic equipment, which are used to accurately identify the material performance parameters. The method comprises: inputting a physical performance parameter vector of a target material into a first model to obtain a target signal feature vector, the first model being obtained by establishing a mapping relationship between the physical performance parameter vector and the signal feature vector through regression analysis, the physical performance parameter vector representing the performance of the material under different external conditions, and the signal feature vector representing the electrical signal generated by the material due to touch. The physical performance parameter vector and the target signal feature vector are merged into a target comprehensive feature vector through feature splicing, and the target comprehensive feature vector is input into a second model to obtain a second vector, the second model being obtained by training the physical performance parameter vector, the signal feature vector and a subjective performance parameter vector through multi-modal fusion based on a fusion algorithm, the subjective performance parameter vector representing common touch evaluation of different groups of people and individualized touch evaluation of a specific group. The matrix mapping relationship constructed based on the multi-modal fusion algorithm determines the touch evaluation value of each dimension from the second vector.

[0025] In the above method, the real-time quantitative characterization of material performance is realized through triboelectric sensing technology. The mapping relationship between physical performance parameters, subjective performance parameters and electrical characteristics is constructed using the electrical signals generated during human tactile interaction, realizing the collaborative analysis of objective physical performance parameters and subjective perception and the real-time feedback of tactile data.

[0026] As shown in Figure 1 The flowchart of the material detection method provided by the embodiments of the present application. Specifically, it can include the following steps.

[0027] Step S101, input the physical performance parameter vector of the target material into the first model to obtain the target signal feature vector.

[0028] The first model is obtained by establishing the mapping relationship between the physical performance parameter vector and the signal feature vector through regression analysis. The physical performance parameter vector represents the performance of the material under different external conditions, and the signal feature vector represents the electrical signals generated by the material due to touch.

[0029] For example, the physical performance parameters at least include the surface characteristic performance, bending characteristic performance, compression characteristic performance and thermal characteristic performance of the target material. The signal feature vector includes voltage characteristic value, current characteristic value and charge characteristic value. The voltage characteristic value represents the potential difference in the target triboelectric signal. The current characteristic value represents the charge flow rate in the target triboelectric signal. The charge characteristic value represents the amount of charge transferred in the target triboelectric signal.

[0030] The training method of the first model is introduced as follows: In an optional embodiment, a plurality of triboelectric signals generated by materials due to touch are obtained, and signal feature vectors are extracted from the triboelectric signals. The physical performance parameter vectors and the signal feature vectors are subjected to data cleaning and data standardization processing. The features satisfying the preset threshold are determined from the physical performance parameter vectors of the plurality of materials to form a first feature set. The covariances between different vectors in the first feature set are determined to construct a first covariance matrix. The first covariance matrix represents the linear correlation degree of the physical performance parameter vectors. The physical performance parameter vectors are projected onto the first principal component direction corresponding to the largest eigenvalue of the first covariance matrix to generate a first factor score matrix, and each element in the first factor score matrix represents the projection value of the corresponding physical performance parameter vector in the first principal component direction. A regression model is constructed with the first factor score matrix as the input variable and the physical performance parameter vector as the target variable until the regression model meets the preset accuracy to obtain the first model.

[0031] For example, the surface characteristic performance, bending characteristic performance, compression characteristic performance, and thermal characteristic performance of different materials are obtained by a fabric feel tester. The physical performance parameter vectors can be subjected to data standardization processing such as data cleaning. For example, outliers are removed, and missing values are filled (such as median filling) to ensure data quality and eliminate dimension differences. From the physical performance parameter vectors of multiple materials, the characteristics that meet the threshold are selected according to the set threshold to form a first feature set. In this way, the characteristics that meet the threshold can be selected by comparing the physical performance parameter vectors with the threshold, preventing redundant characteristics from increasing the complexity of subsequent calculations and improving the efficiency of the virtual first model. The covariance between different vectors in the first feature set is determined to construct a first covariance matrix. The covariance matrix is subjected to eigenvalue decomposition to obtain eigenvalues and corresponding eigenvectors. The first principal component is selected according to the size of the eigenvalues, and the largest eigenvalue in the first covariance matrix is taken as the first principal component. The original physical performance parameter vectors are projected onto the selected first principal component direction to realize dimension reduction. In the data space after dimension reduction, a clustering algorithm (such as K-means, hierarchical clustering, etc.) is used to classify the physical performance parameter vectors. The reliability and stability of the classification results are verified by cross-validation, silhouette coefficient, etc. After dimension reduction and classification, a regression model (such as a linear regression, multivariate regression, etc.) is used to fit the relationship between the physical performance parameters and the triboelectric signal. The first factor score matrix is taken as the input variable, and the signal feature vector is taken as the target variable to construct a regression model until the regression model meets the preset accuracy to obtain a first model. For example, the regression coefficients are determined by minimizing the sum of squared errors to minimize the difference between the predicted value and the observed value.

[0032] For example, the regression model can satisfy the following formula:

[0033] wherein, is the physical performance parameter vector. is the regression coefficient. is the error term. i is greater than or equal to 0.

[0034] The regression coefficient satisfies the following formula:

[0035] wherein, X is the first factor score matrix, and Y is the target matrix composed of the physical performance parameter vector.

[0036] For example, Statistical Product and Service Solutions (SPSS) factor analysis can be used to determine the first factor score matrix: SPSS is used to screen parameters with significant correlations (e.g., |r| (correlation coefficient) ≥ 0.3 (threshold)) from the first correlation coefficient set, forming the first feature set. Then, parameters with significant correlations (e.g., setting the principal component extraction rule in SPSS to |r| (correlation coefficient) ≥ 0.6 (threshold)) are screened from the first correlation coefficient set, forming the first feature set. The physical performance parameter vector is projected onto the direction of the first principal component to generate principal component scores, constructing the first factor score matrix. Here, the first principal component is the largest eigenvalue sorted from largest to smallest.

[0037] In another possible scenario, when the target material is detected to generate a target triboelectric signal due to touch, the target signal feature vector can be directly extracted from the target triboelectric signal.

[0038] For example, a 6514 electrometer can be used to capture the triboelectric signal generated by the target material upon touch. In this way, by acquiring the triboelectric signal in real time, minute changes during the touch process can be reflected in real time, exhibiting high sensitivity and low latency.

[0039] For example, such as Figure 2 As shown in the figure, this application provides a structural schematic diagram for collecting triboelectric signals. Figure 2 In the middle, 6 is electrode layer 1, 7 is friction layer, 8 is test fabric, 9 is electrode layer 2, and 10 is 6514 electrometer.

[0040] In one alternative embodiment, the target triboelectric signal can be preprocessed. For example, wavelet thresholding or moving average filtering can be used to eliminate high-frequency noise. Z-score normalization is applied to the signal to ensure consistency in the dimensions of features across different dimensions. After preprocessing the target triboelectric signal, time-domain and frequency-domain features can be extracted, resulting in voltage, current, and charge feature vectors.

[0041] Optionally, the matrix mapping relationship constructed based on regression analysis can determine the electrical signal values ​​of each dimension from the target signal feature vector.

[0042] For example, suppose the target signal feature vector is x(0.5, 0.7, 0.5). During the training of the first model, a regression analysis algorithm is used to obtain the first factor score matrix. The elements in the first factor score matrix represent current, voltage, and charge, respectively. Therefore, it can be determined that the current when touching the target material is 0.5A, the voltage is 0.7V, and the charge is 0.5C.

[0043] In step S102, the physical performance parameter vector and the target signal feature vector are merged into a target comprehensive feature vector by feature splicing, and the target comprehensive feature vector is input into the second model to obtain a second vector.

[0044] The second model is obtained by training the physical performance parameter vector, the signal feature vector, and the subjective performance parameter vector based on a fusion algorithm.

[0045] In an optional embodiment, the subjective performance parameter at least includes a soft and stiff feeling, a warm and cool feeling, and a smooth and rough feeling of a touch target material by different crowds or different crowds of personalized touch.

[0046] The training method of the second model is described as follows: In an optional embodiment, the physical performance parameter vector and the signal feature vector are merged into a comprehensive feature vector by feature splicing, a second correlation coefficient between the comprehensive feature vector and the subjective performance parameter vector is determined by using a Pearson algorithm, and a second correlation coefficient set is constructed. The correlation coefficient ranges from -1 to 1. If the correlation coefficient is close to 1, it indicates a strong positive correlation. If the correlation coefficient is close to -1, it indicates a strong negative correlation. If the correlation coefficient is close to 0, it indicates no linear relationship. Features in a preset threshold interval are determined from the second correlation coefficient set to form a second feature set. The covariance between different variables in the second feature set is determined, a second covariance matrix is constructed based on the second feature set, and the second covariance matrix represents the linear correlation degree of the signal feature vector corresponding to the second feature set. The comprehensive feature vector is projected onto a second principal component direction corresponding to the maximum eigenvalue of the second covariance matrix to generate a second factor score matrix. Each element in the second factor score matrix represents the coordinate value of the corresponding signal feature in the second principal component direction. A regression model is constructed by taking the second factor score matrix as an input variable and taking the subjective performance parameter as a target variable, until the regression model meets a preset accuracy to obtain the second model.

[0047] For example, the triboelectric signal generated by touching different materials is obtained, and a signal feature vector is extracted from the triboelectric signal. Or, the physical performance parameter vector of different materials is input into the first model to obtain the corresponding signal feature vector. The physical performance parameter vector, the signal feature vector, and the subjective performance parameter vector are subjected to data cleaning and data standardization processing. For example, outliers are removed, and missing values are filled (such as median filling) to ensure data quality and eliminate dimension differences. The physical performance parameters and the signal feature vectors are spliced by column to form a comprehensive feature vector. The Pearson correlation coefficient, Spearman correlation coefficient, and other methods are used to calculate the second correlation coefficient of the comprehensive feature vector and the subjective performance parameter vector to form a second correlation coefficient set. From the second correlation coefficient set, the features in the preset threshold interval are determined to form a second feature set (such as the correlation coefficient greater than or equal to the threshold value is the feature that meets the preset threshold interval). The covariances between different variables in the second feature set are determined, and a second covariance matrix is constructed based on the second feature set. The second covariance matrix is subjected to eigenvalue decomposition to obtain eigenvalues and corresponding eigenvectors. According to the size of the eigenvalues, the second principal component is selected, and the original physical electric signal parameters are projected onto the selected second principal component to realize dimension reduction. In the data space after dimension reduction, a clustering algorithm (such as K-means, hierarchical clustering, etc.) is used to classify the physical electric signal parameters. The reliability and stability of the classification results are verified by cross-validation, silhouette coefficient, and other methods. After dimension reduction and classification, an appropriate regression model (such as linear regression, multiple regression, etc.) is selected to fit the relationship between the physical electric signal parameters and the subjective performance parameters. The second factor score matrix is used as the input variable, and the subjective performance parameters are used as the target variable to construct a regression model until the regression model meets the preset accuracy to obtain a second model.

[0048] For example, the regression model can satisfy the following formula:

[0049] wherein, is the subjective performance parameter vector. is the regression coefficient. is the error term. i is greater than or equal to 0.

[0050] The regression coefficient satisfies the following formula:

[0051] wherein, X is the second factor score matrix, and Y is a target matrix composed of the comprehensive feature vector.

[0052] For example, SPSS can be used to determine the second factor score matrix: SPSS is used to determine the Pearson / Spearman correlation coefficients between the comprehensive feature vector and the subjective performance parameter vector, forming a second set of correlation coefficients. Parameters with significant correlations (e.g., setting the principal component extraction rule in SPSS to |r| (correlation coefficient) ≥ 0.4 (threshold)) from the second set of correlation coefficients are selected to form a second feature set. The signal feature vector is projected onto the direction of the second principal component to generate principal component scores, constructing the second factor score matrix. Here, the second principal component is the largest eigenvalue sorted from largest to smallest.

[0053] Optionally, the trained second model can be evaluated using the training data, and performance metrics (such as mean squared error, coefficient of determination R², etc.) can be calculated. If the performance metrics meet preset rules (such as being less than a preset threshold), the second model is considered to have been successfully trained.

[0054] For example, such as Figure 3 As shown in the diagram, this application provides a schematic diagram of subjective performance parameters and physical performance parameters. The subjective performance parameters include softness (soft / hard, thick / thin), warmth (firm / flexible, cool / warm, full / not full), and smoothness (rough / smooth, sticky / non-sticky, prickly / non-prickly). The physical performance parameters include triboelectric characteristic parameters (short-circuit current, open-circuit voltage, transferred charge), thermal characteristic parameters (compressive thermal conductivity, recovering thermal conductivity, maximum heat flux), compressive characteristic parameters (compressive average stiffness, recovering average stiffness, compressive recovery rate, compressive work), surface characteristic parameters (surface roughness wavelength, surface roughness, surface friction coefficient), and bending characteristic parameters (bending average stiffness, bending work).

[0055] As shown in Figure 4, this application provides a schematic diagram of the correlation between physical performance parameters and subjective performance parameters. Figure 4 includes the Pearson correlation coefficients between various subjective performance parameters and physical performance parameters. It is understood that due to the large amount of data in Figure 4 and the insufficient clarity of a single graph, Figure 4 is presented in a simplified manner. Figure 4a , Figure 4b , Figure 4c , Figure 4d , Figure 4e , Figure 4f , Figure 4g , Figure 4h and Figure 4i Together, the nine sub-images are connected in sequence to present the entire content of Figure 4.

[0056] Optionally, the trained model can be evaluated using training data, and performance metrics (such as mean squared error, coefficient of determination R², etc.) can be calculated. If the performance metrics meet preset rules (such as being less than a preset threshold), the model is considered to have completed training.

[0057] For example, the training data set is divided into a training subset and a validation subset (e.g., 80% of the training data is the training subset and 20% of the training data is the validation subset), ensuring consistent data distribution. The mean square error of the model is calculated. If the mean square error is less than a preset index threshold 1, the model is considered to be accurately fitted and the training is completed.

[0058] Optionally, after obtaining the trained first model and the second model, the correlation coefficient between the triboelectric signal and the physical performance parameter and the subjective performance parameter can be stored in the material database. In this way, the correlation of the triboelectric signal, the subjective performance parameter, and the physical performance parameter can be established for different material types, facilitating subsequent data analysis.

[0059] Optionally, the material category of the target material can also be determined based on the threshold interval in which the target triboelectric signal is located. The threshold of the threshold interval is determined based on samples of multiple different materials. For example, if the voltage of the triboelectric signal is between 1V and 1.5V, the target material can be determined to be a cotton and linen material. If the voltage of the triboelectric signal is between 1.5V and 2V, the target material can be determined to be a polyester material.

[0060] In step S103, the multi-modal fusion algorithm-based matrix mapping relationship is used to determine the evaluation values of each dimension of the touch sensation from the second vector.

[0061] For example, assume that the second vector is y (0.4, 0.5, 0.3). In the training process of the second model, a second factor score matrix is obtained using the multi-modal fusion algorithm. The elements in the second factor score matrix represent softness and stiffness, warmth and coolness, and smoothness, respectively. Therefore, the evaluation value of the softness and stiffness of the target material can be determined to be 0.4. The evaluation value of the warmth and coolness is 0.5, and the evaluation value of the smoothness is 0.4.

[0062] Optionally, the material category, the physical performance parameter, and the evaluation value of each dimension of the subjective performance parameter can also be displayed through a visual interface.

[0063] In the above method, the material category, the physical performance parameter, and the subjective performance parameter can be displayed through a visual interface, which can facilitate quick comparison of the physical performance and subjective performance of different materials and improve user experience.

[0064] For example, the material category, the target physical performance parameter, and the target subjective performance parameter can be displayed in the form of a table.

[0065] For example, as shown in Table 1, it is a schematic diagram of material parameter display.

[0066]

[0067] The composition of the target material is 100% polyester. The density of the target material is 280 g / m3. The current of the target material is 63.38 nA. The charge of the target material is 57.3 nC. The voltage of the target material is 167.55 V. The bending average stiffness of the target material is 173.88 gf.mm / ard. The bending work of the target material is 962.95 gf.mm.rad. The thickness of the target material is 1.26 mm. The compression work of the target material is 1102.64 gf.mm. The compression recovery rate of the target material is 0.43. The compression average stiffness of the target material is 252.88 gf.cm The recovery average stiffness of the target material is 473.64 gf.cm The compression thermal conductivity of the target material is 56.96 W / (m.K). The recovery thermal conductivity of the target material is 57.34 W / (m.K). The maximum heat flow of the target material is 515.27 W / m2.K. The surface friction coefficient of the target material is 0.26. The surface roughness amplitude of the target material is 31.58 pm. The surface roughness wavelength of the target material is 4.34 mm. The smoothness of the target material is 0.22 (the higher the score, the smoother (the score interval is 0-1). The softness of the target material is 0.48 (the higher the score, the softer (the score interval is 0-1). The warmth of the target material is 0.72 (the higher the score, the warmer (the score interval is 0-1).

[0068] Figure 5 A structure schematic diagram of a material detection device provided by the embodiment is shown in the figure. Figure 5 The device includes a processing module 501 and a determination module 502.

[0069] The processing module 501 is configured to input a physical performance parameter vector of a target material into a first model to obtain a target signal feature vector, the first model is obtained by establishing a mapping relationship between the physical performance parameter vector and the signal feature vector through regression analysis, the physical performance parameter vector represents the performance of the material under different external conditions, and the signal feature vector represents an electrical signal generated by the material due to touch. The processing module 501 is further configured to merge the physical performance parameter vector and the target signal feature vector into a target comprehensive feature vector through feature splicing, and input the target comprehensive feature vector into a second model to obtain a second vector, the second model is obtained by training the physical performance parameter vector, the signal feature vector and a subjective performance parameter vector through a multi-modal fusion algorithm, and the subjective performance parameter vector represents common touch evaluation of different crowds and individualized touch evaluation of a specific group. The determining module 502 is configured to determine the evaluation value of each dimension of the tactile sensation from the second vector based on the matrix mapping relationship constructed by the multi-modal fusion algorithm.

[0070] Optionally, the training processing module 501 of the first model is further configured to: obtain a plurality of triboelectric signals generated by the materials when touched, and extract a signal feature vector from the triboelectric signals; determine features in the plurality of physical performance parameter vectors that satisfy a preset threshold value to form a first feature set; determine the covariance between different vectors in the first feature set, and construct a first covariance matrix, the first covariance matrix representing the degree of linear correlation of the physical performance parameter vectors; project the physical performance parameter vectors to a first principal component direction corresponding to a maximum eigenvalue of the first covariance matrix to generate a first factor score matrix, each element in the first factor score matrix representing a projection value of the corresponding physical performance parameter vector in the first principal component direction; construct a regression model with the first factor score matrix as an input variable and the signal feature vector as a target variable until the regression model satisfies a preset accuracy to obtain the first model.

[0071] Optionally, the training processing module 501 of the second model is further configured to: merge the physical performance parameter vector and the signal feature vector into a comprehensive feature vector by feature splicing, determine a second correlation coefficient between the comprehensive feature vector and the subjective performance parameter vector using a Pearson algorithm to form a second correlation coefficient set; determine features in the second correlation coefficient set that satisfy a preset threshold interval to form a second feature set; determine the covariance between different variables in the second feature set, and construct a second covariance matrix based on the second feature set, the second covariance matrix representing the degree of linear correlation of the signal feature vectors corresponding to the second feature set; project the comprehensive feature vector to a second principal component direction corresponding to a maximum eigenvalue of the second covariance matrix to generate a second factor score matrix, each element in the second factor score matrix representing a coordinate value of the corresponding signal feature in the second principal component direction; construct a regression model with the second factor score matrix as an input variable and the subjective performance parameter as a target variable until the regression model satisfies a preset accuracy to obtain the second model.

[0072] Optionally, the physical performance parameters at least include surface feature performance, bending feature performance, compression feature performance, and thermal feature performance of the target material.

[0073] Optionally, the subjective performance parameters at least include softness, warmth, coolness, smoothness, and roughness of the target material when touched.

[0074] Optionally, the determining module 502 is further configured to: determine the material category of the target material based on the threshold interval in which the target triboelectric signal is located.

[0075] Based on the same technical concept, the present application also provides an electronic device, which can implement the functions of the material detection apparatus.

[0076] Figure 6 The electronic device provided in the present application is shown in a structural schematic diagram.

[0077] The electronic device provided in the present application is shown in a structural schematic diagram. Figure 6 The processor 601 and the memory 602 are connected through a bus 600 in the present application. Figure 6 The bus 600 is shown in a thick line, and the connection modes between other components are only schematically illustrated, and are not limited. The bus 600 can be divided into an address bus, a data bus, a control bus, etc., for the convenience of representation, Figure 6 In the present application, only one thick line is used to represent the bus, but it does not mean that there is only one bus or one type of bus. Alternatively, the processor 601 can also be referred to as a controller, and the name is not limited.

[0078] In the present application, the memory 602 stores instructions executable by the at least one processor 601, and the at least one processor 601 can execute the material detection method discussed above by executing the instructions stored in the memory 602. The processor 601 can implement the functions of various modules in the apparatus shown in the drawings. Figure 6

[0079] The processor 601 is the control center of the apparatus, and can connect various parts of the control device through various interfaces and lines, and process data and various functions of the apparatus by running or executing instructions stored in the memory 602 and calling data stored in the memory 602, thereby monitoring the apparatus as a whole.

[0080] In a possible design, the processor 601 can include one or more processing units, and the processor 601 can integrate an application processor and a modem processor, wherein the application processor mainly processes operating systems, driver interfaces and application programs, etc., and the modem processor mainly processes wireless communication. It can be understood that the above-mentioned modem processor can also not be integrated into the processor 601. In some embodiments, the processor 601 and the memory 602 can be implemented on the same chip, and in some embodiments, they can also be implemented on independent chips respectively. ​

[0081] The processor 601 can be a general processor, such as a central processing unit (CPU), a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic component, a discrete hardware component, and can implement or execute the methods, steps and logic block diagrams disclosed in the embodiments of the present application. The general processor can be a microprocessor or any conventional processor. The steps of the material detection method disclosed in the embodiments of the present application can be directly embodied by the hardware processor for execution, or be executed by a combination of hardware and software modules in the processor.

[0082] The memory 602 is a non-volatile computer readable storage medium, which can be used to store non-volatile software programs, non-volatile computer executable programs and modules. The memory 602 can include at least one type of storage medium, such as flash memory, hard disk, multimedia card, card type memory, random access memory (RAM), static random access memory (SRAM), programmable read-only memory (PROM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), magnetic memory, magnetic disk, optical disk, etc. The memory 602 is any other medium capable of carrying or storing desired program code in the form of instructions or data structures and capable of being accessed by a computer, but is not limited to this. The memory 602 in the embodiments of the present application can also be a circuit or any other device capable of realizing the storage function, used for storing program instructions and / or data.

[0083] By designing and programming the processor 601, the code corresponding to the material detection method introduced in the foregoing embodiments can be fixed in the chip, so that the chip can execute the material detection method of the embodiments shown in the running time. Figure 1 How to design and program the processor 601 is a technology known to those skilled in the art, which will not be described here.

[0084] It should be noted that the above general electronic device provided by the embodiments of the present application can realize all method steps realized by the above method embodiments, and can achieve the same technical effects, and here the same parts and beneficial effects of the method embodiments in the embodiments will not be described in detail.

[0085] The embodiment of the present application further provides a computer readable storage medium, which stores computer executable instructions, and the computer executable instructions are used for causing a computer to execute the material detection method in the above embodiment.

[0086] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system or a computer program product. Therefore, the present application can adopt a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware aspects. Moreover, the present application can adopt a computer program product in the form of being implemented on one or more computer readable storage media containing computer usable program codes, including but not limited to disk storage, CD-ROM, optical storage and the like.

[0087] The present application is described with reference to flowcharts and / or block diagrams according to the methods, devices (systems) and computer program products of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing devices to produce a machine, so that the instructions executed by the computer or other programmable data processing devices produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks. Figure 1 The functions specified in one or more blocks.

[0088] These computer program instructions can also be stored in a computer readable memory capable of guiding the computer or other programmable data processing devices to work in a specific way, so that the instructions stored in the computer readable memory produce a manufactured product including instruction devices, which implement the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks. Figure 1 The functions specified in one or more blocks.

[0089] These computer program instructions can also be loaded into a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer implemented process, so that the instructions executed on the computer or other programmable device provide a process for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one or more flows and / or blocks. Figure 1 The functions specified in one or more blocks.

[0090] Obviously, many modifications and variations of the present application are possible in light of the above teachings. It is, therefore, to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. A material testing method, characterized in that, The method includes: The target material's physical performance parameter vector is input into the first model to obtain the target signal feature vector. The first model is obtained by establishing a mapping relationship between the physical performance parameter vector and the signal feature vector through regression analysis. The physical performance parameter vector represents the material's performance under different external conditions, and the signal feature vector represents the electrical signal generated by the material due to touch. The physical performance parameter vector and the target signal feature vector are merged into a target comprehensive feature vector by feature splicing. The target comprehensive feature vector is then input into a second model to obtain a second vector. The second model is obtained by multimodal fusion training of the physical performance parameter vector, the signal feature vector, and the subjective performance parameter vector based on a fusion algorithm. The subjective performance parameter vector represents the common tactile evaluation of different groups and the personalized tactile evaluation of a specific group. The matrix mapping relationship constructed based on the multimodal fusion algorithm determines the tactile evaluation value of each dimension from the second vector.

2. The method according to claim 1, characterized in that, The first model was trained in the following way: Acquire triboelectric signals generated by multiple materials due to touch, and extract signal feature vectors from the triboelectric signals; Features that satisfy a preset threshold are determined from the physical property parameter vectors of multiple materials, forming a first feature set; Determine the covariance between different vectors in the first feature set, and construct a first covariance matrix, which represents the degree of linear correlation between the physical performance parameter vectors; The physical performance parameter vector is projected onto the first principal component direction corresponding to the largest eigenvalue in the first covariance matrix to generate a first factor score matrix. Each element in the first factor score matrix represents the projection value of the corresponding physical performance parameter vector onto the first principal component direction. Using the first factor score matrix as the input variable and the signal feature vector as the target variable, a regression model is constructed until the regression model meets the preset accuracy to obtain the first model.

3. The method according to claim 1 or 2, characterized in that, The second model was trained in the following way: The physical performance parameter vector and the signal feature vector are merged into a comprehensive feature vector by feature splicing. The Pearson algorithm is used to determine the second correlation coefficient between the comprehensive feature vector and the subjective performance parameter vector, thus forming a second correlation coefficient set. Features within a preset threshold range are determined from the second set of correlation coefficients to form a second feature set; Determine the covariance among different variables in the second feature set, and construct a second covariance matrix based on the second feature set. The second covariance matrix represents the degree of linear correlation of the signal feature vectors corresponding to the second feature set. The comprehensive feature vector is projected onto the second principal component direction corresponding to the largest eigenvalue in the second covariance matrix to generate a second factor score matrix. Each element in the second factor score matrix represents the coordinate value of the corresponding signal feature in the second principal component direction. Using the second factor score matrix as the input variable and the subjective performance parameter as the target variable, a regression model is constructed until the regression model meets the preset accuracy to obtain the second model.

4. The method according to claim 1, characterized in that, The physical performance parameters include at least the surface characteristics, bending characteristics, compression characteristics, and thermal characteristics of the target material.

5. The method according to claim 1, characterized in that, Subjective performance parameters include at least the softness and firmness, warmth and coolness, and smoothness and roughness of the target material when touched.

6. The method according to claim 1 or 2, characterized in that, The method further includes: The material category of the target material is determined based on the threshold range in which the target triboelectric signal is located.

7. A material testing device, characterized in that, The device includes: The processing module is used to input the physical performance parameter vector of the target material into the first model to obtain the target signal feature vector. The first model is obtained by establishing a mapping relationship between the physical performance parameter vector and the signal feature vector through regression analysis. The physical performance parameter vector represents the material's performance under different external conditions, and the signal feature vector represents the electrical signal generated by the material due to touch. The processing module is further configured to merge the physical performance parameter vector and the target signal feature vector into a target comprehensive feature vector by feature splicing, and input the target comprehensive feature vector into a second model to obtain a second vector. The second model is obtained by multimodal fusion training of the physical performance parameter vector, the signal feature vector and the subjective performance parameter vector based on a fusion algorithm. The subjective performance parameter vector represents the common tactile evaluation of different groups and the personalized tactile evaluation of a specific group. The determination module is used to determine the tactile evaluation values ​​of each dimension from the second vector based on the matrix mapping relationship constructed by the multimodal fusion algorithm.

8. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program for causing the computer to perform the method of any one of claims 1-6.

10. A computer program product, characterized in that, When the computer program product is invoked by a computer, it causes the computer to perform the method as described in any one of claims 1-6.

Images