CN113340732A - Heterogeneous material multi-parameter inversion method and device based on automatic image partitioning - Google Patents

Abstract

The invention discloses a multi-parameter inversion method and device for heterogeneous materials based on automatic image partitioning. The method includes: making a speckle test piece, fixing the speckle test piece on a testing machine, collecting an original speckle image, and passing the test The machine applies the first load to the speckle specimen, and collects M first speckle images; calculates the displacement field, excludes rigid body translation and rigid body rotation, and obtains M corrected second speckle images; on the original speckle image Select the range to be measured; calculate the strain gradient according to the displacement field, automatically divide the range to be measured into the first area to the Nth area, and calculate the elastic-plastic constitutive parameters of the first area to the Nth area respectively. The range to be measured is automatically divided into multiple regions by calculating the strain gradient, which can automatically complete the division of the range to be measured in one loading experiment, without the need for manual partitioning by methods such as hardness testing before the loading test, simplifying the elasticity of heterogeneous materials. Procedure for the measurement of plastic constitutive parameters.

Description

Multi-parameter inversion method and device for heterogeneous materials based on automatic image partitioning

technical field

The invention relates to the technical field of optical measurement in experimental mechanics, and more particularly, to a multi-parameter inversion method and device for heterogeneous materials based on automatic partitioning of images.

Background technique

In the classical parameter inversion method, the specimen is usually a homogeneous material, that is, the material parameters of the entire area of the specimen to be tested are consistent. However, in practical applications, there may be differences in the material properties of different regions of the specimen, such as welding materials. At present, the patent application publication number CN102778403A proposes a method for identifying the parameters of welding seam materials. The method first needs to divide the material according to the hardness test results, and then carry out the tensile test, and use digital image related technology equipment to obtain each area of the tensile test piece. Real-time full-field primary and secondary strain values; according to the obtained hardness values and primary and secondary strain values, the material parameters of each area of the weld are calculated by the formula of plastic mechanics, and the accuracy is verified by combining with finite element simulation. Application Publication No. CN102288499A proposes a detection method for identifying static and mechanical performance parameters of materials in different areas of welds. The method forms target responses at different areas of welds according to the simulation results of the finite element numerical model of the indentation test and the corresponding experimental results. Function optimization mathematical model; combined with optimization genetic algorithm, iterative selected objective response function to obtain static mechanical performance parameters of different regions of the weld. Application publication number CN108281193A proposes a method for processing a femoral head mechanics simulation model. The method divides the femoral head heterogeneous material simulation unit into multiple parts equally according to the density value, and uses compression linear reinforcement elastics for the femoral head simulation unit. Plastic and Tensile Linearly Strengthened Elastoplastic Constitutives. The simulation results are extracted and analyzed to obtain the strain at the compressive yield point, the strain at the tensile yield point and the strain at the tensile fracture point. However, the existing methods for the measurement of heterogeneous material parameters have problems such as complicated experimental process, manual partitioning, insufficient material parameters obtained, and low accuracy caused by only using the deformation information of the displacement field at one moment.

Therefore, it is an urgent problem to provide a multi-parameter inversion method and device for heterogeneous materials based on automatic partitioning of images, which can automatically partition heterogeneous materials quickly and accurately.

SUMMARY OF THE INVENTION

In view of this, the present invention provides a multi-parameter inversion method for heterogeneous materials based on automatic partitioning of images, including:

Making a speckle test piece, the speckle test piece includes speckle marking points;

Fix the speckle test piece on the testing machine, collect the original speckle image of the speckle test piece before deformation, apply a first load to the speckle test piece through the testing machine, and collect the speckle test piece. M first speckle images during the deformation process of the piece in the elastic-plastic segment, the elastic-plastic segment includes an elastic stage and a plastic stage, and M is a positive integer;

Calculate the displacement field of the speckle specimen during the deformation process according to the original speckle image and the M first speckle images, exclude rigid body translation and rigid body rotation, and obtain M corrected second speckle images image;

selecting a range to be measured on the original speckle image;

Calculate the strain gradient within the to-be-measured range according to the displacement field, and automatically divide the to-be-measured range into a plurality of regions according to the strain gradient, and the plurality of the regions includes the first region to the Nth region;

Select the first region to the Nth region in turn, and obtain the elastic-plastic constitutive parameters of the first region to the Nth region respectively according to the following methods:

Select the area to be measured, take the elastoplastic constitutive parameter of the speckle specimen to be measured as the quantity to be optimized, and define the initial iterative value p _i,0 of the area to be measured in the following way:

p _i,0 =[E,v,A,B,n] _i,0 ,

Wherein, p _i,0 is the initial value of the iteration, E is the elastic modulus, v is the Poisson's ratio, A is the yield strength, B is the hardening coefficient, and n is the hardening index;

According to the iterative initial value and the experimental boundary condition, respectively perform time-series affine transformation on M pieces of the second speckle images in the area to be measured, to obtain the corresponding M pieces of third speckle images before structural deformation;

Constructing an objective function, setting the termination iteration condition, substituting the M pieces of the third speckle images and the original speckle images into the objective function to iteratively optimize the elastic-plastic constitutive parameters to be measured, according to the iteratively obtained The correction modifier updates the elastic-plastic constitutive parameter correction value of the region to be measured;

If the objective function or the correction modifier satisfies the termination iteration condition, the elastic-plastic constitutive parameter correction value is output as the elastic-plastic constitutive parameter of the region to be measured.

Preferably, according to the initial iterative value and the experimental boundary conditions, the M pieces of the second speckle images are respectively subjected to time series affine transformation in the region to be measured, to obtain the corresponding M pieces of the second speckle images before structural deformation. Three speckle images including,

In the elastic stage, time-series affine transformation is performed on the second speckle image according to the following method:

x ₂ (t)=x ₁ (t)+U _Fe (E,v,F(t),x ₁ (t),y ₁ (t))+U _θ +U,

y ₂ (t)=y ₁ (t)+V _Fe (E,v,F(t),x ₁ (t),y ₁ (t))+V _θ +V,

Wherein, x ₁ (t) is the abscissa of the speckle marker point in the second speckle image, and y ₁ (t) is the vertical axis of the speckle marker point in the second speckle image Coordinate, U _Fe is the elastic deformation along the horizontal axis direction caused by the load, V _Fe is the elastic deformation along the longitudinal axis direction caused by the load, E is the elastic modulus, v is the Poisson’s ratio, F(t) is the load at time t, U _θ is the displacement component of rigid body rotation along the horizontal axis direction, V _θ is the displacement component of rigid body rotation along the vertical axis direction, U is the rigid body translation along the horizontal axis direction , V is the rigid body translation along the longitudinal axis, x ₂ (t) is the abscissa of the speckle marker in the third speckle image, y ₂ (t) is the speckle marker The ordinate in the third speckle image.

Preferably, according to the initial iterative value and the experimental boundary conditions, the M pieces of the second speckle images are respectively subjected to time series affine transformation in the region to be measured, to obtain the corresponding M pieces of the second speckle images before structural deformation. Three speckle images including,

In the plastic stage, a time-series affine transformation is performed on the second speckle image according to the following method:

x ₂ (t)=x ₁ (t)+U _Fe (E,v,F(t),x ₁ (t),y ₁ (t))+U _Fp (A,B,n,F(t) ,x ₁ (t),y ₁ (t))+U _θ +U,

y ₂ (t)=y ₁ (t)+V _Fe (E,v,F(t),x ₁ (t),y ₁ (t))+V _Fp (A,B,n,F(t) ,x ₁ (t),y ₁ (t))+V _θ +V,

Wherein, x ₁ (t) is the abscissa of the speckle marker point in the second speckle image, and y ₁ (t) is the vertical axis of the speckle marker point in the second speckle image Coordinate, U _Fe is the elastic deformation along the horizontal axis direction caused by the load, V _Fe is the elastic deformation along the longitudinal axis direction caused by the load, E is the elastic modulus, v is the Poisson’s ratio, F(t) is the load at time t, U _Fp is the plastic deformation along the horizontal axis caused by the load, V _Fp is the plastic deformation along the vertical axis caused by the load, and A is the yield Strength, B is the hardening coefficient, n is the hardening index, U _θ is the displacement component of the rigid body rotating along the horizontal axis, V _θ is the displacement component of the rigid body rotating along the vertical axis, U is the displacement component along the horizontal axis , V is the rigid body translation along the longitudinal axis, x ₂ (t) is the abscissa of the speckle marker point in the third speckle image, and y ₂ (t) is the the ordinate of the speckle marker point in the third speckle image.

Preferably, the construction objective function is calculated according to the following method:

Among them, (x ₂ , y ₂ )∈Ω _i , (x ₀ ,y ₀ )∈Ω _i , (x ₀ ,y ₀ ) are the coordinates of the point in the original speckle image, (x ₂ ,y ₂ ) is the first The coordinates of the point corresponding to (x ₀ , y ₀ ) in the three-speckle image, 0<t≤S, p _i,k is the k-th correction value of the elastic-plastic constitutive parameter of the region to be measured, C(p _i,k ) is the difference value between the M third speckle images and the original speckle image at the k-th iteration, and f((x ₀ , y ₀ ), 0) is the original speckle image The gray value distribution of , g((x ₂ , y ₂ ), t) is the gray value distribution of the third speckle image at time t, Ω _i is the ith area, and S is the time period.

Preferably, the M pieces of the third speckle images and the original speckle images are substituted into the objective function to iteratively optimize the elastoplastic constitutive parameters to be measured, and update the parameters according to the iteratively obtained correction and modification amount. The elastic-plastic constitutive parameter correction value of the area to be measured is calculated according to the following method:

pi _,k+1 =pi _,k +Δpi _,k ,

Among them, p _i,k+1 is the correction value of the elastic-plastic constitutive parameter of the region to be measured at the k+1th time, and p _i,k is the correction value of the elastic-plastic constitutive parameter of the region to be measured at the kth time, Δp _i,k is the k-th correction modification amount;

Δpi _,k = -H(pi _,k ) ^-1 J(pi _,k ),

Among them, J(pi _,k ) is the first-order partial derivative of C(pi _,k ), and H(pi _,k ) is the second-order partial derivative of C(pi _,k );

H(pi _,k )=J(pi _,k ) ^T J(pi _,k ),

Among them, J(pi _,k ) ^T is the transpose matrix of J(pi _,k );

Wherein, ξ is a small amount, and C(pi _,k ) is the difference value between the M third speckle images and the k-th iteration of the original speckle images.

Preferably, the setting termination iteration condition is set according to the following method:

‖C(pi _,k+1 )-C(pi _,k )‖≤10 ^-5 ,

Wherein, C(pi _,k+1 ) is the difference value of M pieces of the third speckle image and the original speckle image at the k+1th iteration, and C(pi _,k ) is the M pieces of the difference value between the third speckle image and the original speckle image at the k-th iteration;

Alternatively, set as follows:

‖Δp _i,k ‖≤10 ^-3 ,

Among them, Δp _i,k is the k-th correction modification amount.

Preferably, calculating the strain gradient within the to-be-measured range according to the displacement field, and automatically dividing the to-be-measured range into a plurality of areas according to the strain gradient, including,

Using the first speckle image in which the range to be measured is in the plastic stage, the strain gradient in the range to be measured includes the strain, the first derivative of strain and the second derivative of strain in the range to be measured, The range to be measured is automatically divided into a plurality of the regions by taking the positive and negative changes of the second derivative of the strain as the region boundary.

The present invention provides a multi-parameter inversion device for heterogeneous materials based on automatic partitioning of images, comprising:

An image acquisition module, which is coupled with the displacement field calculation module, and is used for acquiring the original speckle image of the speckle specimen before deformation and acquiring the M image of the speckle specimen during the deformation process of the speckle specimen in the elastic-plastic section. A first speckle image is transmitted to the displacement field calculation module, where M is a positive integer;

The displacement field calculation module is respectively coupled to the image acquisition module and the to-be-measured range selection module, and is used to calculate the speckle test piece according to the original speckle image and the M first speckle images. The displacement field in the deformation process excludes rigid body translation and rigid body rotation, and obtains M corrected second speckle images and transmits them to the to-be-measured range selection module;

The to-be-measured range selection module is respectively coupled to the displacement field calculation module and the area division module, and is used to select the to-be-measured range on the original speckle image and transmit it to the area division module;

The area division module is respectively coupled with the to-be-measured range selection module and the area elastoplastic constitutive parameter calculation module, and is used for calculating the strain gradient in the to-be-measured range according to the displacement field, and according to the strain gradient automatically dividing the to-be-measured range into a plurality of regions and transmitting them to the region elastic-plastic constitutive parameter calculation module, the plurality of regions including the first region to the Nth region;

The region elastic-plastic constitutive parameter calculation module is coupled to the region division module, and is configured to select the first region to the Nth region in sequence, and obtain the first region to the Nth region respectively according to the following methods The elastic-plastic constitutive parameters of the Nth region:

Select the area to be measured, take the elastoplastic constitutive parameter of the speckle specimen to be measured as the quantity to be optimized, and define the initial iterative value p _i,0 of the area to be measured in the following way:

p _i,0 =[E,v,A,B,n] _i,0 ,

Wherein, p _i,0 is the initial value of the iteration, E is the elastic modulus, v is the Poisson's ratio, A is the yield strength, B is the hardening coefficient, and n is the hardening index;

According to the iterative initial value and the experimental boundary condition, respectively perform time-series affine transformation on M pieces of the second speckle images in the area to be measured, to obtain the corresponding M pieces of third speckle images before structural deformation;

Constructing an objective function, setting the termination iteration condition, substituting the M pieces of the third speckle images and the original speckle images into the objective function to iteratively optimize the elastic-plastic constitutive parameters to be measured, according to the iteratively obtained The correction modifier updates the elastic-plastic constitutive parameter correction value of the region to be measured;

If the objective function or the correction modifier satisfies the termination iteration condition, the elastic-plastic constitutive parameter correction value is output as the elastic-plastic constitutive parameter of the region to be measured.

Compared with the prior art, the method and device for multi-parameter inversion of heterogeneous materials based on automatic image partitioning provided by the present invention at least achieve the following beneficial effects:

1. In the multi-parameter inversion method and device for heterogeneous materials based on automatic image partitioning provided by the present invention, the strain gradient in the range to be measured is calculated according to the displacement field, and the range to be measured is automatically divided into multiple regions according to the strain gradient, which can The area division of the range to be measured is automatically completed in one loading experiment, and there is no need for manual division by methods such as hardness testing before the loading test, which simplifies the process of measuring elastic-plastic constitutive parameters of heterogeneous materials.

2. In the multi-parameter inversion method and device for heterogeneous materials based on automatic image partitioning provided by the present invention, the gray distribution of each region of the M third speckle images and the gray distribution of the corresponding regions of the original speckle image are substituted into the objective function for treatment. The iterative optimization of the measured elastic-plastic constitutive parameters can effectively reduce the influence of random noise on the inversion of the elastic-plastic constitutive parameters of heterogeneous materials, and improve the heterogeneous Accuracy of inversion results for material elastoplastic constitutive parameters.

3. The multi-parameter inversion method and device for heterogeneous materials based on automatic image partitioning provided by the present invention can simultaneously obtain the measured area including elastic modulus, Poisson's ratio, yield strength, hardening coefficient and hardening index through one experiment. Multiple elastic-plastic constitutive parameters.

Of course, any product implementing the present invention does not necessarily need to achieve all of the above-mentioned technical effects at the same time.

Other features and advantages of the present invention will become apparent from the following detailed description of exemplary embodiments of the present invention with reference to the accompanying drawings.

Description of drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention.

Fig. 1 is a flow chart of an embodiment of a multi-parameter inversion method for heterogeneous materials based on automatic image partitioning provided by the present invention;

2 is a flow chart of another embodiment of the multi-parameter inversion method for heterogeneous materials based on automatic image partitioning provided by the present invention;

3 is a structural diagram of an embodiment of an image automatic partition-based multi-parameter inversion device for heterogeneous materials provided by the present invention;

Fig. 4 is the structural schematic diagram of the aluminum alloy friction stir welding to be tested;

Figure 5 is a comparison diagram of the calculated results of the inversion stress-strain curve and the localized digital image correlation technology DIC method;

301-image acquisition module, 302-displacement field calculation module, 303-to-be-measured range selection module, 304-region division module, 305-regional elastic-plastic constitutive parameter calculation module.

Detailed ways

Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of components and steps, the numerical expressions and numerical values set forth in these embodiments do not limit the scope of the invention unless specifically stated otherwise.

The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.

Techniques, methods, and apparatus known to those of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, such techniques, methods, and apparatus should be considered part of the specification.

In all examples shown and discussed herein, any specific values should be construed as illustrative only and not limiting. Accordingly, other instances of the exemplary embodiment may have different values.

It should be noted that like numerals and letters refer to like items in the following figures, so once an item is defined in one figure, it does not require further discussion in subsequent figures.

Example 1

A specific embodiment of the multi-parameter inversion method for heterogeneous materials based on automatic image partitioning according to the present invention is described below with reference to FIG. 1 , including:

S11: Make a speckle specimen, and the speckle specimen includes speckle marking points;

S12: Fix the speckle specimen on the testing machine, collect the original speckle image of the speckle specimen before deformation, apply the first load to the speckle specimen through the testing machine, and collect the deformation of the speckle specimen in the elastic-plastic section M first speckle images in the process, the elastic-plastic segment includes elastic stage and plastic stage, M is a positive integer;

When the external force is less than the elastic limit load, the specimen can completely recover its original shape after the external force causing the deformation is removed. This recoverable deformation is called elastic deformation, and the stage where the specimen only produces elastic deformation is called the elastic stage; When the external force exceeds the elastic limit load, and the load is removed at this time, the specimen cannot be restored to its original state, and a part of the deformation that cannot disappear is retained. This retained permanent deformation is called plastic deformation. This stage called the plastic stage.

S13: Calculate the displacement field of the speckle specimen during the deformation process according to the original speckle image and the M first speckle images, exclude rigid body translation and rigid body rotation, and obtain M corrected second speckle images;

In step S13, two-dimensional digital image correction software (Ncorr) is used to calculate the displacement field of the speckle specimen during the deformation process. The two-dimensional digital image correction software (Ncorr) is an open-source MATLAB program related to 2D digital images.

S14: Select the range to be measured on the original speckle image;

In step S14, the range to be measured may be selected randomly, or a region with research value in the original speckle image may be designated as the range to be measured.

S15: Calculate the strain gradient within the range to be measured according to the displacement field, and automatically divide the range to be measured into multiple regions according to the strain gradient, and the multiple regions include the first region to the Nth region, which can be automatically completed in one loading experiment. The area division of the range eliminates the need for manual division by methods such as hardness testing before the loading test, which simplifies the process of measuring elastic-plastic constitutive parameters of heterogeneous materials.

In step S15, the first speckle image in which the range to be measured is in the plastic stage is used, and the strain gradient in the range to be measured includes the strain, the first derivative of strain, and the second derivative of strain in the range to be measured, and the second derivative of strain is calculated as the second derivative of strain. The positive and negative changes of is the area boundary to divide the range to be measured into multiple areas.

S16: Select the first region to the Nth region in turn, and obtain the elastic-plastic constitutive parameters of the first region to the Nth region respectively according to the following methods:

S161: Select the area to be measured, take the elastoplastic constitutive parameter of the speckle specimen to be measured as the quantity to be optimized, and define the iterative initial value p _i,0 of the area to be measured in the following way:

p _i,0 =[E,v,A,B,n] _i,0 ,

Among them, p _i,0 is the initial value of the iteration, E is the elastic modulus, v is the Poisson's ratio, A is the yield strength, B is the hardening coefficient, and n is the hardening index. Multiple elastoplastic constitutive parameters for the area under test for loose ratio, yield strength, hardening coefficient and hardening exponent.

S162: According to the iterative initial value and the experimental boundary conditions, respectively perform time-series affine transformation on the M second speckle images in the area to be measured, to obtain the corresponding M third speckle images before structural deformation;

S163: Construct an objective function, set termination conditions for iteration, substitute the M third speckle images and the original speckle images into the objective function to iteratively optimize the elastic-plastic constitutive parameters to be measured, and update the area to be measured according to the iteratively obtained correction and modification amount The elasto-plastic constitutive parameter correction value is based on the elasto-plastic constitutive parameter; the gray distribution of each region of the M third speckle image and the gray distribution of the corresponding region of the original speckle image are substituted into the objective function to iteratively optimize the elasto-plastic constitutive parameters to be measured. Compared with the material parameters calculated by the strain field obtained by the related technology, it can effectively reduce the influence of random noise on the inversion of elastoplastic constitutive parameters of heterogeneous materials, and improve the accuracy of the inversion results of elasto-plastic constitutive parameters of heterogeneous materials.

In step S163, in the iterative process, the Newton-Raphson method is used to update the iterative value.

S164: If the objective function or the correction modifier satisfies the termination iteration condition, output the correction value of the elastic-plastic constitutive parameter as the elastic-plastic constitutive parameter of the region to be measured.

Example 2

Another specific embodiment of the multi-parameter inversion method for heterogeneous materials based on automatic image partitioning according to the present invention will be described below with reference to FIG. 2 , including:

S21: Make a speckle test piece, and the speckle test piece includes speckle marking points;

In step S21, the speckle test piece is made in the following manner:

The white paint is evenly sprayed on the test piece as a white base, and after the white paint is dried to form a film, the black paint is used to evenly spray black speckles on the white base.

S22: Fix the speckle specimen on the testing machine, collect the original speckle image of the speckle specimen before deformation, apply the first load to the speckle specimen through the testing machine, and collect the deformation of the speckle specimen in the elastic-plastic section M first speckle images in the process, the elastic-plastic segment includes elastic stage and plastic stage, M is a positive integer;

When the external force is less than the elastic limit load, the specimen can completely recover its original shape after the external force causing the deformation is removed. This recoverable deformation is called elastic deformation, and the stage where the specimen only produces elastic deformation is called the elastic stage; When the external force exceeds the elastic limit load, and the load is removed at this time, the specimen cannot be restored to its original state, and a part of the deformation that cannot disappear is retained. This retained permanent deformation is called plastic deformation. This stage called the plastic stage.

S23: Calculate the displacement field of the speckle specimen during the deformation process according to the original speckle image and the M first speckle images, exclude rigid body translation and rigid body rotation, and obtain M corrected second speckle images;

In step S23, a two-dimensional digital image correction software (Ncorr) is used to calculate the displacement field of the speckle specimen during the deformation process. The two-dimensional digital image correction software (Ncorr) is an open-source MATLAB program related to 2D digital images.

S24: Determine the pixel length according to the speckle specimen, and complete the unit pixel calibration on the original speckle image;

In step S24, the pixel length is determined according to the speckle specimen, that is, the scale is determined, which can reflect the characteristics of the speckle specimen in equal proportions, and the calibration result will be used in the construction of the experimental boundary conditions of the affine transformation in step S272.

S25: Select the range to be measured on the original speckle image;

In step S25, the range to be measured may be selected randomly, or a region with research value in the original speckle image may be designated as the range to be measured.

S26: Calculate the strain gradient within the range to be measured according to the displacement field, and automatically divide the range to be measured into multiple areas according to the strain gradient, the multiple areas include the first area to the Nth area, and the to-be-measured area can be automatically completed in one loading experiment The area division of the range eliminates the need for manual division by methods such as hardness testing before the loading test, which simplifies the process of measuring elastic-plastic constitutive parameters of heterogeneous materials.

In step S26, at the boundary of the two series materials, the displacement is the segmental inflection point, the strain is the midpoint of the gradient, the first derivative of the strain is the extreme point, and the second derivative of the strain is equal to 0, and the basis for dividing the area can be selected as needed, In step S26 of this embodiment, the second derivative of strain is used as the basis for dividing the regions.

Specifically, the first speckle image in which the range to be measured is in the plastic stage is used, and the strain gradient in the range to be measured includes the strain, the first derivative of strain, and the second derivative of strain in the range to be measured. The negative change is the area boundary to divide the range to be measured into multiple areas.

S27: Select the first region to the Nth region in turn, and obtain the elastic-plastic constitutive parameters of the first region to the Nth region respectively according to the following methods:

S271: Select the area to be measured, take the elastoplastic constitutive parameter of the speckle specimen to be measured as the quantity to be optimized, and define the iterative initial value p _i,0 of the area to be measured in the following way:

p _i,0 =[E,v,A,B,n] _i,0 ,

Among them, p _i,0 is the initial value of the iteration, E is the elastic modulus, v is the Poisson's ratio, A is the yield strength, B is the hardening coefficient, and n is the hardening index. Multiple elastoplastic constitutive parameters for the area under test for loose ratio, yield strength, hardening coefficient and hardening exponent.

S272: According to the iterative initial value and the experimental boundary conditions, respectively perform time series affine transformation on the M second speckle images in the area to be measured, to obtain the corresponding M third speckle images before structural deformation;

Since the elastic-plastic segment includes the elastic stage and the plastic stage, the time-series affine transformation of the M second speckle images should be discussed according to whether the second speckle image is in the elastic stage or in the plastic stage;

If in the elastic stage, perform time-series affine transformation on the second speckle image according to the following method:

x ₂ (t)=x ₁ (t)+U _Fe (E,v,F(t),x ₁ (t),y ₁ (t))+U _θ +U,

y ₂ (t)=y ₁ (t)+V _Fe (E,v,F(t),x ₁ (t),y ₁ (t))+V _θ +V,

Among them, x ₁ (t) is the abscissa of the speckle marker point in the second speckle image, y ₁ (t) is the ordinate of the speckle marker point in the second speckle image, and U _Fe is generated by the load The elastic deformation along the horizontal axis direction of , V _Fe is the elastic deformation along the longitudinal axis direction caused by the load, E is the elastic modulus, v is the Poisson's ratio, F(t) is the load at the t-th time, U _θ is the displacement component of the rigid body rotation along the horizontal axis direction, V _θ is the displacement component of the rigid body rotation along the vertical axis direction, U is the rigid body translation along the horizontal axis direction, V is the rigid body translation along the vertical axis direction , x ₂ (t) is the abscissa of the speckle marker point in the third speckle image, and y ₂ (t) is the ordinate of the speckle marker point in the third speckle image.

If in the plastic stage, perform temporal affine transformation on the second speckle image according to the following method:

x ₂ (t)=x ₁ (t)+U _Fe (E,v,F(t),x ₁ (t),y ₁ (t))+U _Fp (A,B,n,F(t) ,x ₁ (t),y ₁ (t))+U _θ +U,

y ₂ (t)=y ₁ (t)+V _Fe (E,v,F(t),x ₁ (t),y ₁ (t))+V _Fp (A,B,n,F(t) ,x ₁ (t),y ₁ (t))+V _θ +V,

Among them, x ₁ (t) is the abscissa of the speckle marker point in the second speckle image, y ₁ (t) is the ordinate of the speckle marker point in the second speckle image, and U _Fe is generated by the load The elastic deformation along the horizontal axis direction of , V _Fe is the elastic deformation along the longitudinal axis direction caused by the load, E is the elastic modulus, v is the Poisson's ratio, F(t) is the load at the t-th time, U _Fp is the plastic deformation along the horizontal axis caused by the load, V _Fp is the plastic deformation along the longitudinal axis caused by the load, A is the yield strength, B is the hardening coefficient, n is the hardening index, U _θ is the rigid body The displacement component of rotation along the horizontal axis direction, V _θ is the displacement component of the rigid body rotation along the vertical axis direction, U is the rigid body translation along the horizontal axis direction, V is the rigid body translation along the vertical axis direction, x ₂ (t) is the abscissa of the speckle marker point in the third speckle image, and y ₂ (t) is the ordinate of the speckle marker point in the third speckle image.

In step S272, the process of performing time series affine transformation on the M second speckle images to obtain the corresponding M third speckle images before structural deformation further includes: excluding the displacement of the rigid body rotation along the horizontal axis direction U _θ , the displacement V _θ along the longitudinal axis of the rigid body rotation, the rigid body translation U along the horizontal axis direction and the rigid body translation V along the longitudinal axis direction are calculated according to the following methods:

During the experiment, according to the assumption of gray invariance, the speckle marking points of the speckle specimen passively follow the deformation of the speckle specimen and deform, and the deformation field of the speckle marking points has continuity in both space and time. , so the grayscale of the speckle image before and after deformation satisfies:

为原始散斑图像在横轴方向上的灰度分布，

为图像坐标，x为图像横坐标，y为图像纵坐标，

为第t时刻的散斑试件的位移场，

为位移的坐标，u′为散斑试件的位移场的横坐标，v′为散斑试件的位移场的纵坐标，

为随机噪声导致的残差，

为构造变形前的第三散斑图像的灰度分布。in,

is the grayscale distribution of the original speckle image along the horizontal axis,

is the image coordinate, x is the abscissa of the image, y is the ordinate of the image,

is the displacement field of the speckle specimen at time t,

is the coordinate of displacement, u' is the abscissa of the displacement field of the speckle specimen, v' is the ordinate of the displacement field of the speckle specimen,

is the residual caused by random noise,

is the grayscale distribution of the third speckle image before construction deformation.

The displacement field of the speckle specimen is controlled by the loading boundary conditions and material parameters. The boundary conditions are the loading mode of uniaxial tension, and the material parameters are the elastic-plastic constitutive parameters of different regions.

In the i-th region,

也为位移的坐标，ε_xx为横轴方向上的应变，ε_yy为纵轴方向上的应变，纵轴方向为拉伸方向，u₀为横轴方向上的刚体位移，v₀为纵轴方向上的刚体位移，通过数字图像相关方法能够求出加载过程中的刚体位移。in,

It is also the coordinate of displacement, ε _xx is the strain on the horizontal axis, ε _yy is the strain on the vertical axis, the vertical axis direction is the stretching direction, u ₀ is the rigid body displacement on the horizontal axis, and v ₀ is the vertical axis The rigid body displacement in the direction can be obtained by the digital image correlation method during the loading process.

According to Hooke's law, in the elastic phase, the strain satisfies:

纵轴方向弹性应变，

为沿横轴方向弹性应变。where E is the elastic modulus, v is the Poisson's ratio, σ is the stress,

The elastic strain along the longitudinal axis,

is the elastic strain along the transverse axis.

According to the Johnson-Cook constitutive model, in the plastic stage it should satisfy:

为沿纵轴方向上的塑性应变，

为沿横轴方向上的塑性应变。where σ is the stress, A is the yield strength, B is the hardening coefficient, n is the hardening exponent,

is the plastic strain along the longitudinal axis,

is the plastic strain along the transverse axis.

In the process of uniaxial stretching, satisfy:

S _(t) = S ₀ exp(-Sε _yy ),

Among them, σ _(t) is the stress at time t, F _(t) is the load at time t, S _(t) is the cross-sectional area of the specimen at time t, and S ₀ is the cross-sectional area of the original specimen.

According to the above formula, the elastic deformation U _Fe along the horizontal axis caused by the load, the elastic deformation V _Fe along the vertical axis direction caused by the load, and the plastic deformation U along the horizontal axis direction caused by the load can be obtained respectively. _Fp and the plastic deformation V _Fp along the longitudinal axis caused by the load, excluding the influence of rigid body translation and rigid body rotation, complete the solution of the coordinates of the speckle marker point in the third speckle image before the structural deformation, complete from Affine transformation from the second speckle image at each time instant from time t ₀ to time t to the third speckle image before structural deformation.

S273: Construct an objective function, set the termination iteration condition, substitute the M third speckle images and the original speckle images into the objective function to iteratively optimize the elastic-plastic constitutive parameters to be measured, and update the area to be measured according to the iteratively obtained correction and modification amount The elastic-plastic constitutive parameter correction value of ;

In step S273, the objective function is constructed and calculated according to the following method:

Among them, (x ₂ , y ₂ )∈Ω _i , (x ₀ ,y ₀ )∈Ω _i , (x ₀ ,y ₀ ) are the coordinates of the point in the original speckle image, (x ₂ ,y ₂ ) is the first The coordinates of the point corresponding to (x ₀ , y ₀ ) in the three-speckle image, 0<t≤S, p _i,k is the correction value of the elastic-plastic constitutive parameter of the k-th test area, C(pi _{, k} ) is the difference value between the M third speckle images and the k-th iteration of the original speckle image, f((x ₀ , y ₀ ), 0) is the gray value distribution of the original speckle image, g( (x ₂ , y ₂ ), t) is the gray value distribution of the third speckle image at time t, Ω _i is the ith area, and S is the time period.

When p _i,k are different, the displacement fields applied to the M second speckle images in the affine transformation will be different, so the gray distribution of the constructed third speckle images will be different.

In step S273, the termination iteration condition is set, which is set according to the following method:

‖C(pi _,k+1 )-C(pi _,k )‖≤10 ^-5 ,

Among them, C(pi _,k ) is the difference between the M third speckle images and the original speckle image at the k+1th iteration, and C(pi _,k ) is the M third speckle images and The disparity value at the k-th iteration of the original speckle image.

Alternatively, set as follows:

‖Δp _i,k ‖≤10 ^-3 ,

Among them, Δp _i,k is the k-th correction modification amount.

The iteration termination condition can be selected according to the requirements, which is not limited here.

In step S273, the M third speckle images and the original speckle images are substituted into the objective function to iteratively optimize the elasto-plastic constitutive parameters to be measured, and the elasto-plastic constitutive parameter correction of the area to be measured is updated according to the iteratively obtained correction modifiers value, calculated as follows:

pi _,k+1 =pi _,k +Δpi _,k ,

Among them, p _i,k+1 is the elastoplastic constitutive parameter correction value of the k+1th time to be measured, pi _,k is the elasto-plastic constitutive parameter correction value of the kth time to be measured, Δpi _,k is the k-th correction modifier;

Δpi _,k = -H(pi _,k ) ^-1 J(pi _,k ),

Among them, J(pi _,k ) is the first-order partial derivative of C(pi _,k ), and H(pi _,k ) is the second-order partial derivative of C(pi _,k );

H(pi _,k )=J(pi _,k ) ^T J(pi _,k ),

Among them, J(pi _,k ) ^T is the transpose matrix of J(pi _,k );

Among them, ξ is a small amount, and C(pi _,k ) is the difference value between the M third speckle images and the k-th iteration of the original speckle images.

Optionally, when calculating the first-order partial derivative of C(pi _,k ), you can also choose to calculate it according to the following method:

Among them, J(pi _,k ) is the Jacobian matrix, corresponding to the first-order partial derivative of C(pi, _k ), and C(pi _,k ) is the M third speckle images and the original speckle The difference value of the image at the k-th iteration, p _i,k is the correction value of the elastic-plastic constitutive parameter of the k-th test area, E is the elastic modulus, v is the Poisson's ratio, A is the yield strength, and B is the Hardening coefficient, n is the hardening exponent, T is the transpose sign.

Substituting the M third speckle images and the original speckle images into the objective function to iteratively optimize the elasto-plastic constitutive parameters to be measured can effectively reduce the influence of random noise on the inversion of the elasto-plastic constitutive parameters of heterogeneous materials, and improve the performance of heterogeneous materials. The accuracy of the inversion results of the elastic-plastic constitutive parameters of the material material.

In step S273, in the iterative process, the Newton-Raphson method is used to update the iterative value.

S274: If the objective function or the correction modifier satisfies the termination iteration condition, output the correction value of the elastic-plastic constitutive parameter as the elastic-plastic constitutive parameter of the region to be measured.

A specific embodiment of the apparatus for multi-parameter inversion of heterogeneous materials based on automatic image partitioning according to the present invention will be described below with reference to FIG. 3 .

Example 3

This is a specific embodiment of the multi-parameter inversion device for heterogeneous materials based on automatic image partitioning according to the present invention, including:

The image acquisition module 301, which is coupled with the displacement field calculation module 302, is used to acquire the original speckle image of the speckle specimen before deformation and to acquire M images of the speckle specimen during the deformation process in the elastic-plastic section. The first speckle image is transmitted to the displacement field calculation module 302, where M is a positive integer;

The displacement field calculation module 302 is respectively coupled to the image acquisition module 301 and the to-be-measured range selection module 303, and is used for calculating the displacement field of the speckle specimen during the deformation process according to the original speckle image and the M first speckle images, Excluding rigid body translation and rigid body rotation, M corrected second speckle images are obtained and transmitted to the to-be-measured range selection module 303;

The range-to-be-measured selection module 303 is coupled to the displacement field calculation module 302 and the area division module 304 respectively, and is used for selecting the range to be measured on the original speckle image and transmitting it to the area division module 304;

The area division module 304 is respectively coupled with the to-be-measured range selection module 303 and the area elastic-plastic constitutive parameter calculation module 305, and is used to calculate the strain gradient within the to-be-measured range according to the displacement field, and automatically divide the to-be-measured range into two parts according to the strain gradient. Multiple regions are transmitted to the regional elastic-plastic constitutive parameter calculation module 305, the multiple regions include the first region to the Nth region, and the region division of the to-be-measured range can be automatically completed in one loading experiment, without the need to pass the hardness before the loading test. Test and other methods are used for manual partitioning, which simplifies the process of measuring elastic-plastic constitutive parameters of heterogeneous materials.

The regional elastic-plastic constitutive parameter calculation module 305 is coupled to the region dividing module 304, and is used to sequentially select the first region to the Nth region, and obtain the elastic-plastic constitutive parameters of the first region to the Nth region respectively according to the following methods:

Select the area to be measured, take the elastoplastic constitutive parameters of the speckle specimen to be measured as the quantity to be optimized, and define the initial iterative value p _i,0 of the area to be measured in the following way:

p _i,0 =[E,v,A,B,n] _i,0 ,

Among them, p _i,0 is the initial value of the iteration, E is the elastic modulus, v is the Poisson's ratio, A is the yield strength, B is the hardening coefficient, and n is the hardening index. Multiple elastoplastic constitutive parameters for the area under test for loose ratio, yield strength, hardening coefficient and hardening exponent.

According to the iterative initial value and the experimental boundary conditions, the M second speckle images are respectively subjected to time series affine transformation in the area to be tested, and the corresponding M third speckle images before structural deformation are obtained;

The objective function is constructed, the termination iteration condition is set, the M third speckle images and the original speckle images are substituted into the objective function to iteratively optimize the elasto-plastic constitutive parameters to be measured, and the elastic-plastic constitutive parameters of the to-be-measured area are updated according to the iteratively obtained correction and modification amount. The plastic constitutive parameter correction value; the gray distribution of each region of the M third speckle images and the gray distribution of the corresponding regions of the original speckle image are substituted into the objective function to iteratively optimize the elastic-plastic constitutive parameters to be measured. Compared with the calculated material parameters of the obtained strain field, it can effectively reduce the influence of random noise on the inversion of elastoplastic constitutive parameters of heterogeneous materials, and improve the accuracy of the inversion results of elasto-plastic constitutive parameters of heterogeneous materials.

If the objective function or the correction modifier satisfies the termination iteration condition, the correction value of the elastic-plastic constitutive parameter is output as the elastic-plastic constitutive parameter of the region to be measured.

When the k-th iteration, the objective function or the correction modifier satisfies the termination iteration condition, output according to the following methods:

P=pi _,k =[E,v,A,B,n] _i,k ,

P is the elastic-plastic constitutive parameter of the area to be measured, p _i,k is the correction value of the elastic-plastic constitutive parameter of the k-th area to be measured, E is the elastic modulus, v is the Poisson's ratio, A is the yield strength, B is the hardening coefficient and n is the hardening exponent.

Example 4

Another specific embodiment of the multi-parameter inversion method for heterogeneous materials based on automatic image partitioning according to the present invention will be described with reference to FIG. 4 and FIG. 5 , including:

This embodiment takes friction stir welding as an example. Through the observation of microstructure, the vicinity of the friction stir welding joint can be divided into weld nugget, thermo-mechanical affect zone, heat affected zone (weld nugget) according to the microscopic results. heat-affect zone). The mechanical properties of the base metal, weld nugget zone, heat-mechanically affected zone and heat-affected zone are also different. The base metal is the material to be welded in the welding process, and the base metal is not affected by welding. Affected by the welding process, the mechanical properties of the three zones of the weld are significantly different from those of the base metal. In this embodiment, the displacement field in the tensile experiment is used as the basis for material partitioning, combined with the time series integrated image correlation algorithm, and the identification of the elastoplastic constitutive parameter distribution is completed through the best matching of affine.

The friction stir welding material used in this example is aluminum alloy 6061T6, the plate thickness is 3mm, the modulus is 60MPa, and the yield strength is 250MPa. The process parameters of the friction stir welding are shown in Table 1:

Table 1

Welding speed Rotating speed shoulder Needle length Press-in amount inclination 300mm/min 1200rpm 10mm 2.8mm 2.8mm 2.5°

The friction stir welding is made into a test piece, and white paint is evenly sprayed on the test piece as a white base. After the white paint is dried to form a film, black speckle is evenly sprayed on the white base with black paint to make a speckle test piece.

The speckle specimen was fixed on the uniaxial tensile testing machine. The experimental settings were as follows: the loading speed was 1 mm/min, the image acquisition was 1 fps, and the image resolution was 2448pixel×1942pixel. The original speckle image of the speckle specimen before deformation was collected. Apply a tensile load to collect M first speckle images during the deformation process of the speckle specimen in the elastic-plastic section, where M is a positive integer;

Calculate the displacement field of the speckle specimen during the deformation process according to the original speckle image and M first speckle images, exclude rigid body translation and rigid body rotation, and obtain M corrected second speckle images;

Determine the pixel length according to the speckle specimen, complete the unit pixel calibration on the original speckle image; select the range to be measured on the original speckle image;

Select the second speckle image at a certain moment after the speckle specimen begins to yield, within the selected range to be measured, use the digital image correlation software to calculate the strain gradient at this moment, and use the second derivative of the strain as the division of the weld area. in accordance with. Since there are obvious extreme values in different areas in the area to be measured, the second derivative of strain is binarized with 0 as the boundary to achieve rapid area division, and multiple areas are obtained, including the first area to the Nth area;

Select the first region to the Nth region in turn, and obtain the elastic-plastic constitutive parameters of the first region to the Nth region according to the following methods: Select the region to be measured, and take the elastic-plastic constitutive parameters of the speckle specimen to be measured as For the quantity to be optimized, define the iterative initial value p _i,0 of the area to be tested in the following way:

p _i,0 =[E,v,A,B,n] _i,0 ,

Among them, p _i,0 is the initial value of the iteration, E is the elastic modulus, v is the Poisson's ratio, A is the yield strength, B is the hardening coefficient, and n is the hardening index;

According to the iterative initial value and the experimental boundary conditions, the M second speckle images are respectively subjected to time series affine transformation in the area to be tested, and the corresponding M third speckle images before structural deformation are obtained;

The objective function is constructed, the termination iteration condition is set, the M third speckle images and the original speckle images are substituted into the objective function to iteratively optimize the elasto-plastic constitutive parameters to be measured, and the elastic-plastic constitutive parameters of the to-be-measured area are updated according to the iteratively obtained correction and modification amount. Plastic constitutive parameter correction value;

If the objective function or the correction modifier satisfies the termination iteration condition, the correction value of the elastic-plastic constitutive parameter is output as the elastic-plastic constitutive parameter of the region to be measured.

When the elastoplastic constitutive parameters from the first region to the Nth region are all identified, the automatic region division of the material weld is realized in one experiment, and the parameters including elastic modulus, Poisson's ratio, yield strength, hardening coefficient and hardening exponent are identified. Elastoplastic constitutive parameters.

In order to further verify the effectiveness, according to the material elastic-plastic constitutive parameters identified by the algorithm, the stress-strain curves of a total of 14 points in different regions were drawn, and compared with the calculation results of the localized digital image correlation technology DIC method, the comparison results are shown in Figure 5. Show. It can be seen that the stress and strain curves obtained by the multi-parameter inversion method of heterogeneous materials based on automatic image partitioning are in good agreement with the calculation results of the localized digital image correlation technique DIC method, which proves the effectiveness and accuracy of this method. sex.

It can be seen from the above embodiments that the method and device for multi-parameter inversion of heterogeneous materials based on automatic image partitioning provided by the present invention at least achieve the following beneficial effects:

1. In the multi-parameter inversion method and device for heterogeneous materials based on automatic image partitioning provided by the present invention, the strain gradient in the range to be measured is calculated according to the displacement field, and the range to be measured is automatically divided into multiple regions according to the strain gradient, which can The area division of the range to be measured is automatically completed in one loading experiment, and there is no need for manual division by methods such as hardness testing before the loading test, which simplifies the process of measuring elastic-plastic constitutive parameters of heterogeneous materials.

2. In the multi-parameter inversion method and device for heterogeneous materials based on automatic image partitioning provided by the present invention, the gray distribution of each region of the M third speckle images and the gray distribution of the corresponding regions of the original speckle image are substituted into the objective function for treatment. The iterative optimization of the measured elastic-plastic constitutive parameters can effectively reduce the influence of random noise on the inversion of the elastic-plastic constitutive parameters of heterogeneous materials, and improve the heterogeneous Accuracy of inversion results for material elastoplastic constitutive parameters.

3. The multi-parameter inversion method and device for heterogeneous materials based on automatic image partitioning provided by the present invention can simultaneously obtain the measured area including elastic modulus, Poisson's ratio, yield strength, hardening coefficient and hardening index through one experiment. Multiple elastic-plastic constitutive parameters.

Although some specific embodiments of the present invention have been described in detail by way of examples, those skilled in the art should understand that the above examples are provided for illustration only and not for the purpose of limiting the scope of the present invention. Those skilled in the art will appreciate that modifications may be made to the above embodiments without departing from the scope and spirit of the present invention. The scope of the invention is defined by the appended claims.

Claims (8)

1. a heterogeneous material multi-parameter inversion method based on image automatic partition, is characterized in that, comprises: Making a speckle test piece, the speckle test piece includes speckle marking points; Fix the speckle test piece on the testing machine, collect the original speckle image of the speckle test piece before deformation, apply a first load to the speckle test piece through the testing machine, and collect the speckle test piece. M first speckle images during the deformation process of the piece in the elastic-plastic segment, the elastic-plastic segment includes an elastic stage and a plastic stage, and M is a positive integer; Calculate the displacement field of the speckle specimen during the deformation process according to the original speckle image and the M first speckle images, exclude rigid body translation and rigid body rotation, and obtain M corrected second speckle images image; selecting a range to be measured on the original speckle image; Calculate the strain gradient within the to-be-measured range according to the displacement field, and automatically divide the to-be-measured range into a plurality of regions according to the strain gradient, and the plurality of the regions includes the first region to the Nth region; Select the first region to the Nth region in turn, and obtain the elastic-plastic constitutive parameters of the first region to the Nth region respectively according to the following methods: Select the area to be measured, take the elastoplastic constitutive parameter of the speckle specimen to be measured as the quantity to be optimized, and define the initial iterative value p _i,0 of the area to be measured in the following way: p _i,0 =[E,v,A,B,n] _i,0 , Wherein, p _i,0 is the initial value of the iteration, E is the elastic modulus, v is the Poisson's ratio, A is the yield strength, B is the hardening coefficient, and n is the hardening index; According to the iterative initial value and the experimental boundary condition, respectively perform time-series affine transformation on M pieces of the second speckle images in the area to be measured, to obtain the corresponding M pieces of third speckle images before structural deformation; Constructing an objective function, setting the termination iteration condition, substituting the M pieces of the third speckle images and the original speckle images into the objective function to iteratively optimize the elastic-plastic constitutive parameters to be measured, according to the iteratively obtained The correction modifier updates the elastic-plastic constitutive parameter correction value of the region to be measured; If the objective function or the correction modifier satisfies the termination iteration condition, the elastic-plastic constitutive parameter correction value is output as the elastic-plastic constitutive parameter of the region to be measured. 2 . The multi-parameter inversion method for heterogeneous materials based on automatic image partitioning according to claim 1 , wherein, according to the iterative initial value and the experimental boundary conditions, respectively, in the area to be measured, The M second speckle images are subjected to time series affine transformation to obtain the corresponding M third speckle images before structural deformation, including, In the elastic stage, time-series affine transformation is performed on the second speckle image according to the following method: x ₂ (t)=x ₁ (t)+U _Fe (E,v,F(t),x ₁ (t),y ₁ (t))+U _θ +U, y ₂ (t)=y ₁ (t)+V _Fe (E,v,F(t),x ₁ (t),y ₁ (t))+V _θ +V, Wherein, x ₁ (t) is the abscissa of the speckle marker point in the second speckle image, and y ₁ (t) is the vertical axis of the speckle marker point in the second speckle image Coordinate, U _Fe is the elastic deformation along the horizontal axis direction caused by the load, V _Fe is the elastic deformation along the longitudinal axis direction caused by the load, E is the elastic modulus, v is the Poisson’s ratio, F(t) is the load at time t, U _θ is the displacement component of rigid body rotation along the horizontal axis direction, V _θ is the displacement component of rigid body rotation along the vertical axis direction, U is the rigid body translation along the horizontal axis direction , V is the rigid body translation along the longitudinal axis, x ₂ (t) is the abscissa of the speckle marker in the third speckle image, y ₂ (t) is the speckle marker The ordinate in the third speckle image. 3. The multi-parameter inversion method for heterogeneous materials based on automatic image partitioning according to claim 1, characterized in that, according to the iterative initial value and the experimental boundary conditions, in the region to be measured, The M second speckle images are subjected to time series affine transformation to obtain the corresponding M third speckle images before structural deformation, including, In the plastic stage, a time-series affine transformation is performed on the second speckle image according to the following method: x ₂ (t)=x ₁ (t)+U _Fe (E,v,F(t),x ₁ (t),y ₁ (t))+U _Fp (A,B,n,F(t) ,x ₁ (t),y ₁ (t))+U _θ +U, y ₂ (t)=y ₁ (t)+V _Fe (E,v,F(t),x ₁ (t),y ₁ (t))+V _Fp (A,B,n,F(t) ,x ₁ (t),y ₁ (t))+V _θ +V, Wherein, x ₁ (t) is the abscissa of the speckle marker point in the second speckle image, and y ₁ (t) is the vertical axis of the speckle marker point in the second speckle image Coordinate, U _Fe is the elastic deformation along the horizontal axis direction caused by the load, V _Fe is the elastic deformation along the longitudinal axis direction caused by the load, E is the elastic modulus, v is the Poisson’s ratio, F(t) is the load at time t, U _Fp is the plastic deformation along the horizontal axis caused by the load, V _Fp is the plastic deformation along the vertical axis caused by the load, and A is the yield Strength, B is the hardening coefficient, n is the hardening index, U _θ is the displacement component of the rigid body rotating along the horizontal axis, V _θ is the displacement component of the rigid body rotating along the vertical axis, U is the displacement component along the horizontal axis , V is the rigid body translation along the longitudinal axis, x ₂ (t) is the abscissa of the speckle marker point in the third speckle image, and y ₂ (t) is the the ordinate of the speckle marker point in the third speckle image. 4. The multi-parameter inversion method for heterogeneous materials based on automatic image partitioning according to claim 1, wherein the construction objective function is calculated according to the following method:

Among them, (x ₂ , y ₂ )∈Ω _i , (x ₀ ,y ₀ )∈Ω _i , (x ₀ ,y ₀ ) are the coordinates of the point in the original speckle image, (x ₂ ,y ₂ ) is the first The coordinates of the point corresponding to (x ₀ , y ₀ ) in the three-speckle image, 0<t≤S, p _i,k is the k-th correction value of the elastic-plastic constitutive parameter of the region to be measured, C(p _i,k ) is the difference value between the M third speckle images and the original speckle image at the k-th iteration, and f((x ₀ , y ₀ ), 0) is the original speckle image The gray value distribution of , g((x ₂ , y ₂ ), t) is the gray value distribution of the third speckle image at time t, Ω _i is the ith area, and S is the time period. 5 . The multi-parameter inversion method for heterogeneous materials based on automatic image partitioning according to claim 4 , wherein the M third speckle images and the original speckle images are substituted into the The objective function iteratively optimizes the elastic-plastic constitutive parameters to be measured, and updates the correction value of the elastic-plastic constitutive parameters of the region to be measured according to the iteratively obtained correction modification, and calculates according to the following method: pi _,k+1 =pi _,k +Δpi _,k , Among them, p _i,k+1 is the correction value of the elastic-plastic constitutive parameter of the region to be measured at the k+1th time, and p _i,k is the correction value of the elastic-plastic constitutive parameter of the region to be measured at the kth time, Δp _i,k is the k-th correction modification amount; Δpi _,k = -H(pi _,k ) ^-1 J(pi _,k ), Among them, J(pi _,k ) is the first-order partial derivative of C(pi _,k ), and H(pi _,k ) is the second-order partial derivative of C(pi _,k ); H(pi _,k )=J(pi _,k ) ^T J(pi _,k ), Among them, J(pi _,k ) ^T is the transpose matrix of J(pi _,k );

Wherein, ξ is a small amount, and C(pi _,k ) is the difference value between the M third speckle images and the k-th iteration of the original speckle images. 6. The multi-parameter inversion method for heterogeneous materials based on automatic image partitioning according to claim 5, wherein the setting termination iteration condition is set according to the following method: ‖C(pi _,k+1 )-C(pi _,k )‖≤10 ^-5 , Wherein, C(pi _,k+1 ) is the difference value of M pieces of the third speckle image and the original speckle image at the k+1th iteration, and C(pi _,k ) is the M pieces of the difference value between the third speckle image and the original speckle image at the k-th iteration; Alternatively, set as follows: ‖Δp _i,k ‖≤10 ^-3 , Among them, Δp _i,k is the k-th correction modification amount. 7 . The multi-parameter inversion method for heterogeneous materials based on automatic image partitioning according to claim 1 , wherein, calculating the strain gradient within the range to be measured according to the displacement field, and calculating the strain gradient according to the strain The gradient automatically divides the range to be measured into multiple regions, including, Using the first speckle image in which the range to be measured is in the plastic stage, the strain gradient in the range to be measured includes the strain, the first derivative of strain and the second derivative of strain in the range to be measured, The range to be measured is automatically divided into a plurality of the regions by taking the positive and negative changes of the second derivative of the strain as the region boundary. 8. A heterogeneous material multi-parameter inversion device based on image automatic partitioning is characterized in that, comprising, An image acquisition module, which is coupled with the displacement field calculation module, and is used for acquiring the original speckle image of the speckle specimen before deformation and acquiring the M image of the speckle specimen during the deformation process of the speckle specimen in the elastic-plastic section. A first speckle image is transmitted to the displacement field calculation module, where M is a positive integer; The displacement field calculation module is respectively coupled to the image acquisition module and the to-be-measured range selection module, and is used to calculate the speckle test piece according to the original speckle image and the M first speckle images. The displacement field in the deformation process excludes rigid body translation and rigid body rotation, and obtains M corrected second speckle images and transmits them to the to-be-measured range selection module; The to-be-measured range selection module is respectively coupled to the displacement field calculation module and the area division module, and is used to select the to-be-measured range on the original speckle image and transmit it to the area division module; The area division module is respectively coupled with the to-be-measured range selection module and the area elastoplastic constitutive parameter calculation module, and is used for calculating the strain gradient in the to-be-measured range according to the displacement field, and according to the strain gradient automatically dividing the to-be-measured range into a plurality of regions and transmitting them to the region elastic-plastic constitutive parameter calculation module, the plurality of regions including the first region to the Nth region; The region elastic-plastic constitutive parameter calculation module is coupled to the region division module, and is configured to select the first region to the Nth region in sequence, and obtain the first region to the Nth region respectively according to the following methods The elastic-plastic constitutive parameters of the Nth region: Select the area to be measured, take the elastoplastic constitutive parameter of the speckle specimen to be measured as the quantity to be optimized, and define the initial iterative value p _i,0 of the area to be measured in the following way: p _i,0 =[E,v,A,B,n] _i,0 , Wherein, p _i,0 is the initial value of the iteration, E is the elastic modulus, v is the Poisson's ratio, A is the yield strength, B is the hardening coefficient, and n is the hardening index; According to the iterative initial value and the experimental boundary condition, respectively perform time-series affine transformation on M pieces of the second speckle images in the area to be measured, to obtain the corresponding M pieces of third speckle images before structural deformation; Constructing an objective function, setting the termination iteration condition, substituting the M pieces of the third speckle images and the original speckle images into the objective function to iteratively optimize the elastic-plastic constitutive parameters to be measured, according to the iteratively obtained The correction modifier updates the elastic-plastic constitutive parameter correction value of the region to be measured; If the objective function or the correction modifier satisfies the termination iteration condition, the elastic-plastic constitutive parameter correction value is output as the elastic-plastic constitutive parameter of the region to be measured.

Images