the clothing pressures measured by the grey cloth body-attached trousers/relatively body-attached trousers and the transverse elastic cotton cloth body-attached trousers/body-attached trousers (negative loose quantity) are measured at each part for about 1 minute in the front middle, side seams and back middle, and the mode is the clothing pressure of a certain part, so that 80 real clothing pressure data can be obtained because the mode is the number with the highest occurrence frequency and can be represented at a certain part to a great extent.

Through the above measurement, a garment pressure value table of the front middle part of the gray fabric compared with the body-attached plate type is obtained, as shown in the following table 1:

table 1 white grey cloth is compared with the body-attached plate type front middle clothing compression value table unit: v (V)

The following table 2 is obtained from the data of table 1 by the following process:

(1) True clothing pressure mode: the clothing pressures measured by the front middle side seam, the rear middle side seam (the position corresponding to the buttock) and the positions of the white grey cloth body trousers/the relatively body trousers and the transverse elastic cotton cloth body trousers/the body trousers (negative loose quantity) are summarized for about 1 minute, and the clothing pressures of a certain position are determined to be taken by analyzing the parameters of the table, so that the mode is the number with the highest occurrence frequency, the clothing pressures of the certain position can be represented to a great extent, and 80 real clothing pressure data are obtained.

(2) Virtual garment press: and extracting the corresponding 80 virtual clothing pressure values by using clothing pressure data software.

(3) Unified unit: converting real clothing pressure units into units g/cm consistent with virtual clothing pressures² . The unit conversion is performed according to the following formula.

True clothing pressure measured by instrument 10=true clothing pressure (unit kpa)

Real garment pressure (unit kpa) ·10·2=real garment pressure (unit g/cm)² )

Table 2 white grey cloth is compared with the body-fitting plate type front middle clothing compression value conversion table unit: v (V)

The obtained real garment pressure and virtual garment pressure values are shown in table 3 below:

table 3 real garment pressure and virtual garment pressure value table

S4, analyzing the correlation between the real clothing pressure value and the virtual clothing pressure value and verifying that the real clothing pressure value and the virtual clothing pressure value are in a linear change relation;

in the embodiment, a pearson correlation test method is selected to be matched with a typical correlation analysis method to verify correlation so as to improve the test reliability;

s4.1, a Pelson correlation test method, wherein the calculation and the test of a correlation coefficient are carried out through a function cor.test (X, Y) function in R language, a real clothing pressure value X and a virtual clothing pressure value Y in a table 3 are input, and the correlation coefficient is output as 0.7912356; wherein the P value is also 2.2e-16< <0.05, rejecting the original assumption that the variable X is considered to be positively correlated with the variable Y, i.e. the real clothing pressure is positively correlated with the virtual clothing pressure, and from the interval point of view, the calculated interval is (0.69,0.86), and it can also be seen that X and Y are largely positively correlated.

S4.2, exemplary correlation analysis

(1) Principle of typical correlation analysis: the above analysis methods all consider real clothing pressure and virtual clothing pressure as two sets of data corresponding to each other one by one, and analyze the correlation of the two. Next, we sub-divide the two sets of data into six sets of data, two pairs (real and virtual pressure) according to heel, knee, above knee, mid thigh (crotch is not typically relevant due to missing data in lateral positions), mid thigh circumference, mid/mid front, six locations, and introduce new parameters for typical correlation analysis.

A typical correlation analysis (canonical correlation analysis) is a statistical method for analyzing the degree of correlation between two random variables, which effectively interprets the correlation between two random variables and another part of the variables. Taking this data as an example, the actual garment pressure at six positions of heel, knee, above knee, mid thigh, thigh circumference, front middle/rear middle garment pressure is (X)₁ ,X₂ ,...X₆ ) The corresponding virtual garment pressure is (Y₁ ,Y₂ ,...Y₆ )。

In general, it is assumed that there are two sets of random variables X₁ ,X₂ ,...X₆ And Y₁ ,Y₂ ,...Y_q Study of their correlation, when p=q=1, it is usually the correlation of two variables X and Y; when p > 1 and q > 1, a method similar to principal component analysis is used to find the linear combination U of the 1 st group variable and the linear combination V of the 2 nd group variable, namely

U＝a₁ X₁ +a₂ X₂ +...+a_p X_p ，

V＝b₁ Y₁ +b₂ Y₂ +...+b_q Y_q ，

The correlation problem of the two variables is then transformed into a correlation problem of the two variables, and the corresponding coefficients a, b can be adjusted appropriately so that the correlation of the variables U and V is maximized, known as such a correlation canonical correlation, and an analysis method based on this principle is known as a canonical correlation analysis. In this example, p=6 and q=6, and the correlation problem of two groups of variables was converted into a correlation problem of two variables by using a typical correlation analysis method, and the number of samples was 12.

Let x= (X)₁ ,X₂ ,...,X_p )^T ,Y＝(Y₁ ,Y₂ ,...,Y_q )^T As a random vector, a linear combination a of X and Y is used^T X and b^T Correlation between Y to study the correlation between X and Y and hope to find a and b, ρ (a^T X,b^T Y) is maximum.

By the definition of the correlation coefficient(s),

for any of α, β and c, d, there are

ρ(α(a^T X)+β,c(b^T Y)+d)＝ρ(a^T X,b^T Y)。

The above description maximizes a of the correlation coefficient^T X and b^T Y is not unique. Thus, in integrating variables, one can define

var(a^T X)＝1,var(b^T Y)＝1

Let x= (X)₁ ,X₂ ,...,X_p )^T ,T＝(Y₁ ,Y₂ ,...,Y_p )^T P+q-dimensional random vector

The mean value of (2) is 0 and the covariance matrix Σ is positive. If a is present₁ ＝(a₁₁ ,a₁₂ ,...,a_1p )^T And b₁ ＝(b₁₁ ,b₁₂ ,...,b_1q )^T Make->

Is a constraint problem

maxρ(a^T X,b^T Y)

s.t.var(a^T X)＝1

var(b^T Y)＝1

The maximum of the objective function, called,

the first pair (set) of typical variables (canomical variates) being X, Y, is called the correlation coefficient ρ (U)₁ ,V₁ ) Is a first typical correlation coefficient (canonical correlation).

If a is present_k ＝(a_k1 ,a_k2 ,...,a_kp )^T And b_k ＝(b_k1 ,b_k2 ,...,b_kq )^T Such that:

(a)

and k-1 above is not related to the typical variable;

(b)

(c)

and->

The correlation coefficient is the largest.

Then call for

The k-th pair (group) of typical variables of X, Y is called the correlation coefficient ρ (U)_k ,V_k ) For k (k=2, 3., min { p, q }) typical correlation coefficients.

(2) Typical correlation analysis of real and virtual garment pressure R software: the real garment pressure and the virtual garment pressure are analyzed for correlation using typical correlation coefficients, wherein the garment pressures are divided into 6 groups of heel, knee, upper knee, mid thigh, thigh circumference, front middle/rear middle for corresponding analysis. Specific programming languages and test results are shown in the appendix:

(3) Typical correlation analysis results of real and virtual garment pressure R software: where cor is a typical correlation coefficient, it is known that the typical correlation coefficient of real clothing pressure and virtual clothing pressure is 1, that is to say very typically correlated, xcoef is a coefficient corresponding to data X, also referred to as typical load (canonical loadings) for data X, that is to say the transpose of the coefficient matrix a of the sample typical variable U; ycoef is a coefficient corresponding to data Y, also referred to as a typical load for data Y, i.e. a transpose of the coefficient matrix B of sample typical variables V; the $ xcenter is the center of the data X, i.e. the sample mean of the data X

Ycenter is the center of data Y, i.e. the sample mean of data Y +.>

Since the data has been normalized, the sample mean calculated here is 0.

For real and virtual garment pressure data, the mathematical meaning corresponding to the calculation result is:

wherein the method comprises the steps of

i=1, 2,3 is normalized data, and the corresponding correlation coefficients are:

ρ(U₁ ,V₁ )＝0.796,ρ(U₂ ,V₂ )＝0.201,ρ(U₃ ,V₃ )＝0.0726

it will be appreciated that the coefficients are not unique and may be any multiple of them.

The number of samples is calculated below based on the scores under the typical variables. Since u=ax, v=by, the R procedure for calculating the score is:

U<-as.matrix(test[,1:6])％*％ca$xcoef

>V<-as.matrix(test[,7:12])％*％ca$ycoef

the above formula is to calculate the number of samples according to the scores under the typical variables, and the embodiment draws the related variables U₁ ,V₁ And U3, V₃ A data scatter plot of coordinates, as shown in fig. 3, the plot representing a first representative variable; as shown in fig. 4, a scatter plot of coordinates in the figure represents a third exemplary variable;

as can be seen from the two graphs, the first exemplary variable is that the points in the coordinate scatter diagram are all on a straight line (the corresponding cor exemplary correlation coefficient is 1.0000), while the third exemplary variable is that the points in the coordinate scatter diagram are relatively scattered (the corresponding cor exemplary correlation coefficient is 0.9608820), but are basically near a straight line, so that the exemplary correlation of the real and virtual clothing pressures can be more judged to be very remarkable.

The correlation between the values of the real clothing pressure and the virtual clothing pressure is checked by the method for checking the correlation to be consistent, which shows that the two values have very obvious correlation, so that the next operation can be performed, the two sets of numbers are linearly regressed, the linear relation between the two numbers is analyzed, a linear regression relation is obtained, and the obvious effect of the relation is checked.

The present embodiment verifies that the two are in a linear change relationship by a correlation coefficient test method:

s4.3, let beta₁ Representing the rate of change of E (Y) with X in linear fashion, if beta₁ E (Y) does not actually vary linearly with X, only when beta₁ Not 0, E (Y) varies linearly with X, and only then does the unified linear regression equation make sense. Thus assume that test is

H₀ :β₁ ＝0,H₁ :β₁ ≠0

Recording device

Referring to R as a sample correlation coefficient, for a given significance level alpha, looking up a correlation coefficient threshold table to obtain R_α (n-2), the reject field of the test is

|R|＞r_α (n-2)

When rejecting H₀ When linear regression equations are considered significant.

(2) Linear regression analysis of real and virtual garment pressure R software: the R software is used for analyzing the program language of the unitary linear regression relation between the real clothing pressure and the virtual clothing pressure and the result is shown in the annex.

(3) Linear regression analysis results of real and virtual garment pressure R software: the formula of the corresponding regression model is listed in the first part (call) of the calculation result: lm (formula=y to 1+x). The second step (residues:) lists the minimum point of the residual-28.954, 1/4 quantile-7.091, the number of bits 3.825, 3/4 quantile 3.825, and the maximum point 51.108.

In the third part of the calculation (coeffients:), estime represents the regression equation parameters, i.e

(standard deviation) means the standard deviation of regression parameters, i.e.>

value is the value of t, namely:

pr (> |t|) represents the P value, i.e., the probability value P { T > |T ]_zhi I. There are significant labels, where "/indicates extremely significant,"/indicates highly significant, "/indicates less significant, and no label is given asIs not significant. The data test result shows that the data is very significant, namely the test shows that the regression relationship between the real clothing pressure and the virtual clothing pressure is significant.

In the fourth part of the calculation, residual standard error represents the standard deviation of the residual, i.e

The degree of freedom is n-2.Multiple R-Squared is the square of the correlation coefficient, i.e. +.>

The R-Squared test shows that the regression relationship between the real and virtual clothing pressures is obvious, and the correlation test is passed.

F-statics represents the F statistic, i.e. by correlation checking

The degree of freedom of the data in this paper is (1, 78), and the result obtained by F test is 130.2, namely

F≥F_0.99 (1,78)

In the check rejection domain with the significance level of 99%, the check shows that the regression relationship between the real clothing pressure and the virtual clothing pressure is obvious.

The P-value is the value P, i.e. the probability value P { F > |F_zhi I, pass the correlation check. From the calculation results, it can be seen that the regression equation passes the inspection of the regression parameters and the inspection of the regression equation.

S5, carrying out regression analysis on the real clothing pressure value and the virtual clothing pressure value, establishing a regression relation equation of the real clothing pressure value and the virtual clothing pressure value, and writing the equation into software;

the equation for establishing the regression relation between the real clothing pressure value and the virtual clothing pressure value is as follows:

Y＝μ(x)+ε

wherein x represents an independent variable virtual clothing pressure value; y represents the real clothing pressure value of the dependent variable; epsilon is a random error, which is the sum of various factors that have an influence on Y except for an independent variable; the regression equation is obtained after the real clothing pressure value and the virtual clothing pressure value are brought into Y=9.4024+4.4852X.

S6, importing the virtual clothing parameters to be detected and the virtual human model parameters to obtain clothing pressure measurement values.

The embodiment also provides a clothing pressure measurement system based on the virtual human model, which is based on 3DRaways software and comprises: the system comprises a virtual human model construction module, a color map processing module, a real clothing pressure value acquisition module, a correlation analysis and verification module, a regression analysis module and a clothing pressure measured value output module;

in this embodiment, the virtual mannequin construction module is configured to construct a virtual mannequin having the same size as the physical mannequin, and derive a color chart of the virtual clothing compression condition;

in this embodiment, the color map processing module is configured to process a color map, and obtain a virtual clothing pressure value of a specific location according to a map mapping function;

in this embodiment, the real clothing compression value obtaining module is configured to obtain a real clothing compression value of the physical mannequin;

in this embodiment, the correlation analysis and verification module is configured to perform correlation analysis on a real clothing pressure value and a virtual clothing pressure value and verify that the real clothing pressure value and the virtual clothing pressure value are in a linear change relationship;

in this embodiment, the regression analysis module is configured to perform regression analysis on the real clothing pressure value and the virtual clothing pressure value, and establish a regression relation equation between the real clothing pressure value and the virtual clothing pressure value;

in this embodiment, the clothing pressure measurement value output module is configured to obtain a virtual clothing parameter to be detected and a virtual human model parameter, and output a clothing pressure measurement value.

As shown in fig. 5, the virtual garment pressure module of 3 drawax shows the approximate distribution of garment pressures. The right side of the picture is provided with a scale which displays the range of the virtual clothing pressure (101.8 g/cm 2-0.3 g/cm 2), the scale is red to blue from top to bottom, and the red represents the maximum pressure (101.8 g/cm)² ) The blue color represents the minimum pressure (-0.3 g/cm)² ) The middle color is changed from red to blue, and the corresponding clothing pressure is from 101.8g/cm² ～-0.3g/cm² But not specifically. Virtual garment with virtual wearing back trousersThe pressing and the color of the ruler mark are in one-to-one correspondence.

Through testing of 10 pictures (including clothing pressure in different ranges of trousers, knitwear, evening wear, business wear and the like), the software processes the picture information, so that the accuracy of obtaining the virtual clothing pressure value reaches more than 95%, the sensitivity is very good, the requirement of extracting the virtual clothing pressure value can be met, and the software can be used for paper research and industrial production design. As shown in fig. 6, the virtual garment pressure display interface is shown, when The mouse is moved to The position where The garment pressure is required to be displayed, the virtual garment pressure specific value of The corresponding scale of The point is displayed in The pressure of The color dialog box, and as shown in fig. 7, the garment pressure value of The point is 48.34g/cm² 。

The embodiment quantifies the fuzzy pressure distribution condition into an accurate numerical value applicable to industrial production design based on the existing 3DRaways virtual clothing pressure module, and the testing method has small calculated amount and low complexity.

The above examples are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above examples, and any other changes, modifications, substitutions, combinations, and simplifications that do not depart from the spirit and principle of the present invention should be made in the equivalent manner, and the embodiments are included in the protection scope of the present invention.

Claims

1. The clothing pressure measuring method based on the virtual human model is characterized by comprising the following steps:

obtaining a real clothing compression value of a physical mannequin;

the method comprises the steps that an air bag type clothing pressure measuring method is adopted for obtaining an actual clothing pressure value of an entity human model, an air bag with the thickness smaller than 3mm is placed at a position to be measured, and dynamic clothing pressure in a state of being mounted is measured through resistance change of a semiconductor pressure sensor connected with the air bag;

2. The garment pressure measurement method of claim 1, wherein the physical mannequin employs a real person or intelligent pressure test mannequin simulating softness of a human body.

3. The garment pressure testing method of claim 1, wherein the real garment pressure measurement portion of the physical mannequin comprises: lower abdomen, thigh root, crotch, thigh lower part, knee, ankle, hip.

4. The garment pressure testing method according to claim 1, wherein the method for performing a correlation analysis of real garment pressure values and virtual garment pressure values uses any one or more of a rank correlation test, a KENDALL correlation test, a pearson correlation test, or a typical correlation analysis.

5. The garment pressure testing method according to claim 4, wherein the representative correlation analysis method performs a correspondence analysis of each part of the real garment pressure measurement and each part of the virtual human model measurement.

6. The clothing pressure measurement method according to claim 1, wherein the analysis of the correlation between the real clothing pressure value and the virtual clothing pressure value and the verification of the correlation are linear change, and the specifically adopted analysis method is any one of a t-test method, an F-test method and a correlation coefficient test method.

7. The garment pressure testing method according to claim 1, wherein the establishing a regression equation of real garment pressure values and virtual garment pressure values is specifically as follows:

Y＝μ(x)+ε

8. A garment pressure measurement system based on a virtual human model, comprising: the system comprises a virtual human model construction module, a color map processing module, a real clothing pressure value acquisition module, a correlation analysis and verification module, a regression analysis module and a clothing pressure measured value output module;