CN113763368B

CN113763368B - Multi-type damage detection characteristic analysis method for large-size test piece

Info

Publication number: CN113763368B
Application number: CN202111068223.8A
Authority: CN
Inventors: 黄雪刚; 谭旭彤; 殷春; 石安华; 雷光钰; 罗庆; 姜林
Original assignee: Ultra High Speed Aerodynamics Institute China Aerodynamics Research and Development Center
Current assignee: Ultra High Speed Aerodynamics Institute China Aerodynamics Research and Development Center
Priority date: 2021-09-13
Filing date: 2021-09-13
Publication date: 2023-06-23
Anticipated expiration: 2041-09-13
Also published as: CN113763368A

Abstract

The invention discloses a multi-type damage detection characteristic analysis method for a large-size test piece, which comprises the following steps: acquiring an infrared thermal reconstruction image of a large-size test piece from the infrared thermal image sequence; decomposing each infrared thermal reconstruction image into a basic layer infrared thermal image and a detail layer infrared thermal image; respectively acquiring a thermal amplitude fusion weight graph between corresponding base layer infrared thermal images and a thermal amplitude fusion weight graph between detail layer infrared thermal images; and fusing the detail layer thermal image information and the base layer thermal image information among typical type defect thermal reconstruction images of different areas in different detection times in the large-size test piece, and combining the weighted average base layer thermal image and the detail layer thermal image to obtain a final fused detection infrared thermal image. The method omits the step of manually identifying the defect category number and judging the category number, removes noise and then abnormal values, improves the detection performance of Shan Zhangre images, and improves the defect edge definition and contrast of the fused images.

Description

Multi-type damage detection characteristic analysis method for large-size test piece

Technical Field

The invention belongs to the technical field of equipment defect detection, and particularly relates to a multi-type damage detection characteristic analysis method for a large-size test piece.

Background

The pressure vessel is widely applied in the fields of aerospace, energy chemical industry, metallurgical machinery and the like, such as rocket fuel tanks, space station sealed cabins and the like, and is very important for safety detection because the pressure vessel is often used for containing flammable and explosive liquid or gas with certain pressure. The common defect types of the pressure vessel are fatigue crack defects, welding defects, corrosion defects and the like, and the corresponding conventional detection means are mature. However, it is very difficult to rapidly and comprehensively detect a defect in a large pressure vessel having an inner diameter of 2 m or more. The infrared thermal imaging detection technology is an effective non-contact nondestructive detection method aiming at the damage defect of a large pressure container, and the detection purpose is achieved by controlling a thermal excitation method and measuring the temperature change of the material surface to obtain the structural information of the material surface and the subsurface thereof. When acquiring structural information, a thermal infrared imager is often used to record temperature field information of the surface or subsurface of a test piece, which changes with time, and convert the temperature field information into a thermal image sequence to be displayed. By analyzing and extracting the transient thermal response of the thermal image sequence, a reconstructed image capable of characterizing and strengthening the defect characteristic is obtained, so that the defect detection and interpretation are realized. Although the reconstructed thermal image has good detectable performance when representing the characteristics of a certain type of defect damage area, when the reconstructed thermal image is applied to the defect detection of a large-size pressure container, due to the limitation of detection conditions, all defect conditions of the whole large-size pressure container cannot be obtained at the same time through single detection. Therefore, multiple infrared detection in different areas is needed for large-size pressure vessels, so that comprehensive and accurate detection results are obtained.

In the invention, after the detection precision of the defects of the current detection area extracted from a plurality of infrared thermal image sequences is improved by utilizing a grid-based self-adaptive CLIQUE clustering algorithm, how to enable the detection images to simultaneously represent the defect characteristics of different areas obtained in multiple detections needs to be further considered. In order to make up for the limitation of the single Zhang Chonggou thermal image in representing the integral defect characteristics of the large-size pressure container, it is a good way to fuse the defect thermal characteristics contained in the multiple thermal image sequences by using an infrared thermal image fusion algorithm. The infrared thermal image fusion synthesizes the thermal radiation characteristics of different areas and different types of defects in a plurality of reconstructed thermal images in different thermal image sequences, and fuses the thermal radiation characteristics into one fused thermal image, so that the capability of simultaneously representing the characteristics of the different areas and the different types of defects obtained through multiple detection of one fused thermal image is provided, and the method is an effective way for improving the capability of detecting complex types of defects of the reconstructed thermal images outside Shan Zhanggong. Therefore, how to fuse different areas and different types of thermal images with high quality is a challenging task. The common infrared thermal image fusion technology only considers the relatively obvious defect characteristic information in the thermal image when fusing the infrared thermal reconstruction images, and does not consider the situation that a plurality of small-size holes and pothole damages exist in a test piece. So that the fine crack defect in the fused thermal image is smoothed out as noise, which is fatal to the safety of the pressure vessel. In large-size pressure vessel defect feature extraction, image edge and texture information of defects are one of the very important features for quantitatively identifying defects. The smoothed fine defects directly affect the accuracy of the quantitative analysis of the defects, resulting in missing defects and reduced detection integrity. Therefore, in the infrared thermal image fusion process of large-size pressure vessel defect detection, a plurality of fusion targets and requirements should be considered simultaneously, so that not only the retention requirement of large-size defect characteristics is included, but also the detail retention and enhancement of micro defects and the background information smoothing effect of non-defect areas of the fusion image should be considered.

Therefore, the invention introduces an image fusion technology based on combination of multi-objective optimization and guide filtering to realize the fusion function of a plurality of thermal images, so that the detection image can synthesize defect information in a plurality of thermal image sequences, has the characteristic condition of characterizing different areas and different types of defects in a large-size pressure container, and realizes the high-quality imaging function of the integral defect condition of the large-size pressure container. The guided filtering is a novel edge preserving filter capable of preserving edge information of an image while smoothing the image. Therefore, the guiding filtering is very suitable for spacecraft defect detectionA need. And the multi-objective evolutionary optimization algorithm can synergistically optimize the vector optimization problem. The invention combines the multi-objective optimization and the guided filtering technology, utilizes the multi-objective simultaneous optimization of a plurality of guided filtering cost functions to obtain the targeted optimal guided filtering linear transformation coefficient a _k And b _k . Therefore, the advantages of a plurality of guide filters are combined, meanwhile, the large-size edge retaining characteristic of edge perception weighted guide filtering, the detail retaining characteristic of gradient domain guide filtering and the noise removing characteristic of LoG guide filtering are considered, so that the guide filtering after multi-objective optimization can be combined with the advantages of a plurality of different guide filtering cost functions with filtering preference, the filtered image can retain the large-size edge characteristics in an original infrared thermal image and places with severe image gradient changes to the greatest extent, and can retain some tiny crack defect textures and forms in a pressure container, and meanwhile, the background area image without defects in the infrared thermal image is smoothed and noise information is removed. The filtering performance is further improved, so that the infrared thermal image fusion performance is improved, and the detection and defect extraction performance of the algorithm on the whole defects of the large-size pressure vessel are improved.

Disclosure of Invention

It is an object of the present invention to address at least the above problems and/or disadvantages and to provide at least the advantages described below.

To achieve these objects and other advantages and in accordance with the purpose of the invention, there is provided a large-sized specimen multi-type damage detection feature analysis method including the steps of:

performing multiple infrared detection on a large-size test piece to obtain an infrared thermal image sequence of the large-size test piece, and obtaining an infrared thermal reconstruction image of the test piece from the infrared thermal image sequence by utilizing an infrared characteristic extraction and infrared thermal image reconstruction algorithm;

decomposing the infrared thermal reconstruction images of each defect area of the large-size test piece into a basic layer infrared thermal image and a detail layer infrared thermal image;

step three, acquiring a thermal amplitude fusion coarse weight graph; taking the infrared thermal reconstruction image as a guide image, taking the thermal amplitude fusion coarse weight image as an input image, and carrying out multi-target guided filtering modeling; performing multi-objective optimization problem modeling on linear transformation parameters of guide filtering; optimizing to obtain a final front approximate solution set of multi-target guided filtering linear parameters by using a multi-target optimization method based on a chebyshev decomposition method and a particle swarm, and selecting multi-target guided filtering Pareto optimal linear transformation parameters of a thermal amplitude fusion coarse weight graph from the front approximate solution set based on a weighted membership scheme; based on Pareto optimal linear transformation parameters, obtaining an expression of a final linear transformation parameter of multi-target guided filtering, thereby obtaining an expression of a multi-target guided filtering operator, and performing multi-target guided filtering on a thermal amplitude fusion rough weight graph of an infrared thermal reconstruction image of an obtained infrared detection area by utilizing the optimal guided filtering operator obtained by multi-target optimization to obtain infrared thermal amplitude fusion weight images of a corrected base layer and a corrected detail layer;

And step four, based on the obtained detailed layer heat amplitude fusion weight map and the base layer heat amplitude fusion weight map of typical type defects in each infrared detection area after finishing, fusing the detail layer heat image information and the base layer heat image information among the large-size test piece typical type defect heat reconstruction images to obtain a base layer heat image and a detail layer heat image fused with effective information of a plurality of multi-detection area reconstruction heat images, and finally combining the weighted average base layer heat image and the weighted average detail layer heat image to obtain a final fusion detection infrared heat image.

Preferably, the specific steps of acquiring the reconstructed infrared thermal image from the infrared thermal image sequence by using the infrared feature extraction and the infrared thermal image reconstruction algorithm in the first step are as follows:

step S11, a valuable transient thermal response data set X (g) is extracted from a thermal image sequence S acquired by a thermal infrared imager based on a transient thermal response data extraction algorithm of block variable step length; wherein S (I, J, T) represents pixel values of an I-th row and a J-th column of T-frame infrared thermal images of the thermal image sequence, T is a total frame number, I is a total line number, J is a total column number, t=1..; decomposing a thermal image sequence into K different data blocks by means of a threshold value _k S(i _n ,j _m T) wherein k represents the kth sub-data block, i _n 、j _m T represents the ith of the kth sub-block, respectively _n Line j _m Column, pixel value of the t frame; then defining search line step length in kth data block according to temperature variation characteristics in different data blocks _k RSS and column step size _k CSS, k=1,. -%, K; based on different search steps in different data blocks, comparing correlation coefficients between data points, and searching a series of correlation coefficients greater than a threshold THC _cr And adding the transient thermal response data set X (g);

step S12, utilizing a grid-based adaptive CLIQUE clustering algorithm to adaptively cluster transient thermal responses in a transient thermal response set X (g); dividing a T-dimension data space of the transient thermal response into rectangular grids which are not overlapped with each other by presetting a dividing interval parameter delta, and marking dense grids and sparse grids by transient thermal response data quantity in the grids; carrying out sparse grid correction on the sparse grid by utilizing a boundary correction method and a sliding grid method; according to a greedy algorithm retrieval principle, connecting the dense grids and the corrected sparse grids to form a maximum connected region; the connected dense grids and the corrected sparse grids are transient thermal response clusters, the CLIQUE clustering algorithm adaptively identifies the number of dense grid clusters in a thermal response space, so that the step of judging the defect class number is omitted, and the number of cluster clusters existing in a transient thermal response data space is adaptively identified; adaptive clustering to form cluster sets _X(g) Cluster[h]H=1, 2, |c|, where h represents a category label, |c| represents the total number of categories;

step S13, respectively extracting typical characteristic transient thermal responses from different clusters and reconstructing a thermal image based on the typical characteristic transient thermal responses; calculating a clustering center of each category in the clustering result as a typical characteristic transient thermal response of each type of defect:

wherein the method comprises the steps of

For the h (h=1, 2., |c|) clustering result _X(g) Cluster[h]H=1, …, the kth of |c| represents the transient thermal response, | _X(g) Cluster[h]The I is the total number of transient thermal responses contained in the h clustering result, and a matrix Y is formed by using typical transient thermal responses of various types of defects;

carrying out infrared thermal image reconstruction by utilizing the information of the matrices Y and S, extracting each frame of image of S into a column vector according to columns, and arranging the column vectors according to time sequence to form an I multiplied by J row and T column two-dimensional image matrix O, and obtaining a reconstruction matrix R based on the following transformation formula:

wherein,,

is a matrix of |C| x T, is a pseudo-inverse matrix of matrix Y, O ^T Is the transposed matrix of the two-dimensional image matrix O, the obtained reconstruction matrix R is I C row and I X J column, each row of the reconstruction matrix R is intercepted to form an I X J two-dimensional image, I C I X J two-dimensional images are obtained, the images are the reconstruction thermal images containing the characteristic information of different thermal response areas, and the non-defect background area reconstruction thermal images are recorded as follows _B R, recording the reconstructed thermal image corresponding to each type of defect area as _i R (i=1, …, |c|); each reconstructed thermal image contains characteristic thermal reconstruction information of one defect type of the complex type defects except the background region thermal image without defect damage.

Preferably, the step of performing multiple infrared detection on the large-size test piece to obtain multiple thermal image sequences of the large-size test piece, and obtaining multiple reconstructed infrared thermal images of the large-size test piece from the multiple thermal image sequences by using an infrared feature extraction and infrared thermal image reconstruction algorithm, wherein the specific method comprises the following steps:

step S11, using a three-dimensional matrix set { S ] for a plurality of thermal infrared image sequences acquired from the thermal infrared imager ¹ ,…,S ⁱ ,…,S ^C Represented by S, where S ⁱ Representing a thermal image sequence obtained by the thermal infrared imager in the ith infrared detection, wherein |C| represents the total thermal image sequence number; s is S ⁱ (M, N, T) represents a temperature value at an M-th row, N-th column coordinate position of a T-th frame thermal image in the i-th thermal image sequence, t=1,..t, T is a total frame number, m=1,..m, M is a total number of rows, n=1,..n, N is a total number of columns;

step S12, for the ith IR thermal image sequence S ⁱ Extracting an ith thermal image sequence S by using a transient thermal response data extraction algorithm based on block variable step length ⁱ Valuable transient thermal response data set X in (1) ⁱ (g) The method comprises the steps of carrying out a first treatment on the surface of the The ith thermal image sequence S is thresholded ⁱ Decomposition into K different data blocks _k S ⁱ (m ', n', t) wherein k represents the ith thermal image sequence S ⁱ The kth sub data block, m ', n', t respectively represent the temperature values at the coordinate positions of the (m 'th row, n' th column and t th frame of the kth sub data block), and then the ith thermal image sequence S is defined according to the temperature change characteristics in different data blocks ⁱ Search row step size in kth data block _k RSS ⁱ Sum column step size _k CSS ⁱ K=1,..k; based on different search steps in different data blocks, comparing correlation coefficients between data points, and searching a series of correlation coefficients greater than a threshold THC _cr And adding an ith thermal image sequence S ⁱ Transient thermal response data set X in (1) ⁱ (g)；

Step S13, utilizing a grid-based self-adaptive CLIQUE clustering algorithm to carry out ith thermal image sequence S ⁱ Transient thermal response set X of (2) ⁱ (g) Transient thermal response adaptive clustering in (a); dividing a T-dimension data space of transient thermal response into rectangular grids which are not overlapped with each other by presetting a dividing interval parameter delta, marking dense grids and sparse grids by transient thermal response data quantity in the grids, and carrying out sparse grid correction on the sparse grids by utilizing a boundary correction method and a sliding grid method; according to the greedy algorithm retrieval principle, connecting the dense grids and the corrected sparse grids to form a maximum connected area, wherein the connected dense grids and the corrected sparse grids are For a transient thermal response cluster, the CLIQUE clustering algorithm adaptively identifies the number of dense grid clusters in the thermal response space, so that the step of judging the defect class number is omitted, and the cluster number existing in the transient thermal response data space is adaptively identified; sequence S of thermal images ⁱ Transient thermal response set X of (2) ⁱ (g) Adaptive clustering to form cluster sets

Wherein H represents a defect type label, and H represents the total number of types of complex type defects existing in the current infrared detection area;

s14, respectively extracting representative characteristic transient thermal responses of various complex defects in the ith detection area from different clusters and reconstructing a thermal image based on the representative characteristic transient thermal responses;

calculating a clustering center of each category in the clustering result as a representative characteristic transient thermal response of each type of defect:

wherein the method comprises the steps of

For the h clustering result _X(g) Cluster[h]H=1, …, the kth transient thermal response in H, | _X(g) Cluster[h]The I is the total number of transient thermal responses contained in the h clustering result, and a matrix Y is formed by using representative transient thermal responses of various types of defects ⁱ ；

Using matrix Y ⁱ And S is ⁱ Carrying out infrared thermal image reconstruction on the information of (a) and carrying out the (i) th thermal image sequence S ⁱ Each frame of image is extracted into a column vector according to columns and arranged in time sequence to form a two-dimensional image matrix O of M multiplied by N rows and T columns ⁱ The thermal amplitude reconstruction matrix R for the ith detection is obtained based on the following transformation formula ⁱ ：

Wherein,,

is H×T matrix, which is representative transient thermal response matrix Y ⁱ Pseudo-inverse matrix of (O) ⁱ ) ^T Is a two-dimensional image matrix O ⁱ The obtained reconstruction matrix is H rows and M multiplied by N columns; intercepting reconstruction matrix R ⁱ An MxN two-dimensional image is formed, H MxN two-dimensional images are obtained, the images are the reconstructed thermal images containing the characteristic information of different thermal response areas in the thermal image sequence obtained by the ith infrared detection, and the reconstructed thermal images of the non-defective background areas are recorded as _B R, recording the reconstructed thermal image corresponding to each type of defect area as _h R, where h=1,.. recording a typical type defect reconstruction thermal image in the detected region obtained in the ith infrared detection as _Def.(i) R；

Step S15, if i < |C|, i+1, i.e. for the i+1th IR thermal image sequence S ⁱ⁺¹ Repeating the steps S12-S14 until all the typical type defect reconstruction thermal images in the current detected region are obtained from a plurality of thermal image sequences obtained by multiple detection respectively, calculating MSE values of all the type defect reconstruction images in the current region, and selecting the typical type defect reconstruction thermal images in each detected region based on the MSE minimum principle, namely obtaining a typical type defect reconstruction thermal image set { in each detected region of a large-size test piece _Def.(1) R,..., _Def.(i) R,.., _Def.(|C|) R }, wherein _Def.(i) R represents a typical type of defect reconstruction thermal image of the detected region in the i-th thermal image sequence, i=1.

Preferably, the second step decomposes the reconstructed image of each defect area of the large-size test piece into a basic infrared thermal image and a detailed layer infrared thermal imageThe body method comprises the following steps: for (C-1) sheet of infrared reconstructed image { except for the thermal image of the open background region ₁ R,…, _i R,…, _|C|-1 Each R is subjected to image decomposition, and each reconstructed image is decomposed into a base layer infrared thermal image { ₁ B,..., _i B,..., _|C|-1 B and a detail layer infrared thermal image { ₁ D,…, _i D,…, _|C|-1 D }; reconstructing thermal images with ith defective areas _i R is, for example, i=1,.., |c| -1, obtained by using the following formula _i R base layer infrared thermal image _i B and detail layer IR thermal image _i D：

_i B＝ _i R*Z

_i D＝ _i R- _i B

Wherein Z is an average filter.

Preferably, the specific method for decomposing the infrared thermal reconstruction image of each defect area of the large-size test piece into the base layer infrared thermal image and the detail layer infrared thermal image comprises the following steps: an infrared reconstructed image { of a total of |C| typical type defects in each detection region in a large-size impact test piece _Def.(1) R,..., _Def.(i) R,..., _Def.(|C|) Each of R is subjected to image decomposition, and each reconstructed image is decomposed into a base layer infrared thermal image { inf.base [ def. (1) ],...,Inf.Base[Def.(i)],...,Inf.Base[Def.(|C|)]And a detail layer infrared thermal image { Inf. Detail [ def. (1)],...,Inf.Detail[Def.(i)],...,Inf.Detail[Def.(|C|)]}；

Reconstructing thermal images with typical type defects of the ith inspection area _Def.(i) R is exemplified by the following formula _Def.(i) Typical type of defect base layer ir thermal image and detail layer ir thermal image of R inf.base [ def. (i)]And Inf. Detail [ def. (i)]：

Inf.Base[Def.(i)]＝ _Def.(i) R*Z

Inf.Detail[Def.(i)]＝ _Def.(i) R-Inf.Base[Def.(i)]

Wherein Z is an average filter.

Preferably, wherein the stepIn the third step, the corresponding infrared thermal image { of each base layer is respectively obtained by utilizing multi-objective optimized guiding filtering ₁ B, ₂ B,…, _|C|-1 B }, an infrared thermal amplitude fusion weighting graph { between ₁ W ^B , ₂ W ^B ,…, _|C|-1 W ^B And detail layer infrared thermal image { ₁ D, ₂ D,…, _|C|-1 D, fusing weight map { of infrared thermal amplitude values between the two layers ₁ W ^D , ₂ W ^D ,…, _|C|-1 W ^D The specific method is as follows:

step S31, reconstructing an image based on infrared _i R obtains thermal amplitude fusion rough weight graph _i P is as follows; an initial thermal radiation rough fusion weight map is obtained based on the following formula:

_i H＝ _i R*L

_i S＝| _i H|*GF

wherein L is a Laplacian filter, GF is a Gaussian low-pass filter, and a thermal amplitude fusion coarse weight map is obtained based on the following formula _i P：

Wherein { is as follows _i P ₁ 、,..., _i P _k ,..., _i P _I×J Is a coarse weight graph _i The thermal amplitude of each position coordinate of P fuses the weight values, _i P _k is that _i The thermal amplitude of the kth coordinate point of P fuses weight values, k=1,..i x J, _i S _k is a characteristic diagram of heat amplitude significance _i The radiation significance level value corresponding to the kth coordinate point in S, k=1,..i×j;

Step S32, modeling a relation between filtering input and filtering output of multi-target guide filtering; reconstructing an image with infrared light _i R is a guiding image, and a thermal amplitude value is used for fusing a rough weight map _i P is an input image, and multi-target guiding filtering is carried out; in the case of multi-target guided filtering, a guided filter window w is defined _k To guide the image, i.e. to reconstruct the image infrared _i At the kth coordinate point in R _i R _k For a central local rectangular window, k=1,..i×j, whose size is (2r+1) × (2r+1), the input-output relationship of the multi-objective guided filtering is:

_i O _n ＝a _k · _i R _n +b _k

wherein,, _i O _n representing images reconstructed by infrared _i R is a guiding image, and a thermal amplitude value is used for fusing a rough weight map _i P is an output image obtained by multi-target guided filtering of an input image _i The guided filtered output value corresponding to the nth coordinate point of O, n=1,..i×j; _i R _n is that _i The reconstructed image thermal amplitude corresponding to the nth coordinate point of R, n=1,/i×j; a, a _k And b _k Expressed in terms of _i R _k Centered guided filter window w _k Linear transformation parameters in, k=1,..i×j;

step S33, for obtaining the fused optimal weight value of each corresponding infrared thermal amplitude of each reconstructed infrared thermal image, the linear transformation parameter a of the guided filtering is performed _k And b _k The specific method for multi-objective optimization problem modeling is as follows:

Step S331, fusing coarse weight graphs based on thermal amplitude values _i P and infrared reconstructed images _i R, defining an infrared large-size defect edge characteristic perception weighting guide filtering cost function at each coordinate point position

Wherein,,

and->

Is composed of a large rulerAn optimal linear transformation coefficient determined by an inch defect perception filtering cost function; _i P _n is a weight graph _i An infrared heat radiation fusion weight value corresponding to the nth coordinate point of P; epsilon is a regularization factor;

Is an edge-aware weighting factor defined as follows:

wherein,,

representing an infrared reconstructed image _i In R, to _i R _k The variance of the heat radiation value corresponding to each coordinate point in the 3×3 window with the coordinate point as the center, ζ is a very small constant having a size of (0.001×dr #) _i P)) ² DR (·) is the dynamic range of the image, and by minimizing the cost function, the following expression of the optimal linear transform coefficient is obtained:

wherein,,

representing a representation of an infrared reconstructed image _i R and thermal amplitude fused coarse weight graph _i Hadamard product of P is found in rectangular window w _k Infrared heat amplitude mean value corresponding to each coordinate point in the graph, < + >>

Is the Hadamard product of the matrix, +.>

And->

Respectively representing infrared reconstructed images _i R and fused coarse weight map _i P is in rectangular window w _k Mean value of interior->

Representing an infrared reconstructed image _i R is in rectangular window w _k Infrared heat amplitude variance corresponding to each coordinate point in the infrared heat amplitude variance;

step S332, fusing the rough weight map based on the thermal amplitude value _i P and infrared reconstructed images _i R, defining a gradient domain infrared fine size defect detail texture guiding filtering cost function at each coordinate point position

Wherein,,

and->

The optimal linear transformation coefficient is determined by a gradient domain fine defect detail texture guide filtering cost function; epsilon is a regularization factor; v (v) _k To adjust a _k Factors of (2);

For gradient domain multi-window edge perceptual weights, it is defined as follows:

representing an infrared reconstructed image _i In R, to _i R _k Guide filter window w with coordinate point as center _k Thermal amplitude standard deviation v corresponding to each coordinate point in the graph _k Is defined as follows:

wherein eta is

Representing an infrared reconstructed image _i In R, to _i R _k The standard deviation of the infrared heat amplitude corresponding to each coordinate point in a 3X 3 window with the coordinate point as the center, n epsilon I X J,

representing an infrared reconstructed image _i In R, to _i R _k Guide filtering rectangular window w with coordinate point as center _n The standard deviation of the infrared heat amplitude corresponding to each coordinate point in the range is n epsilon I multiplied by J;

guided filtering cost function by minimizing gradient domain

Obtain->

And->

The calculation formula of (2) is as follows:

wherein,,

representing a representation of an infrared reconstructed image _i R and thermal amplitude fused coarse weight graph _i Hadamard product of P is found in rectangular window w _k The average value of the thermal amplitude value v corresponding to each coordinate point in the graph _k To adjust a _k Factors of (2);

step S333, fusing the coarse weight map based on the thermal amplitude _i P and infrared reconstructed images _i R, defining a local LoG operator space noise elimination guide filtering cost function

Wherein,,

and->

The optimal linear transformation coefficient is determined by the local Log operator space noise-oriented filtering cost function; epsilon is a regularization factor;

The local LoG edge weighting factor is defined as follows:

wherein, loG (·) is a Gaussian Laplace edge detection operator, I×J is the total coordinate point number of the infrared reconstruction image, and |·| is the absolute value operation, delta _LoG 0.1 times the maximum value of the LoG image;

guided filtering cost function by minimizing gradient domain

Obtain->

And->

The calculation formula of (2) is as follows:

wherein the method comprises the steps of

And->

Respectively representing infrared reconstructed images _i R and coarse weight map _i P is in rectangular window w _k The average value of the thermal amplitude corresponding to each coordinate point in the graph;

step S334, simultaneously optimizing 3 cost functions, and establishing the following multi-objective optimization problem:

Minimize F(a _k ')＝[ _Inf.Sig E ₁ (a _k '), _Inf.Min E ₂ (a _k '), _Inf.Noi E ₃ (a _k ')] ^T

wherein a is _k ' is the kth guided filter window w _k Is used to determine the linear transformation coefficients of the block, _Inf.Sig E ₁ (a _k ') preserving a fusion cost function for the large-size defect edges of the infrared thermal image with obvious gradient change, _Inf.Min E ₂ (a _k ') is infrared thermal image micro-scale with insignificant size and gradient changeSmall defect detail texture preserving fusion cost function, E ₃ (a _k ' is an infrared thermal image noise information sensing and eliminating cost function;

step S34, optimizing the multi-objective optimization problem by utilizing a multi-objective optimization method based on a Chebyshev decomposition method and a particle swarm, wherein the specific method comprises the following steps:

step S341, initializing multi-objective optimization related parameters; initializing the iteration times g' =0 and uniformly distributed weight vectors

Wherein l=3 is the total number of the fusion cost functions of the infrared thermal images;

initializing a reference point _i r＝{ _i r ₁ ,..., _i r ₃ }，

Is a corresponding infrared thermal image fusion reference point; _i AP (0) =Φ; maximum number of iterations g' _max The method comprises the steps of carrying out a first treatment on the surface of the Initializing related parameters of the nth infrared thermal image fusion coefficient population individual particle swarm;

step S342, utilize

Decomposing the original multi-objective problem into a series of scalar sub-objective problems using chebyshev decomposition method, using the weights vectors corresponding to the respective weights +.>

Is>

The evolution direction of each population solution is guided, and each sub-problem is as follows:

step S343, for each decomposed sheetTarget subproblems based on their corresponding weight vectors

The new infrared thermal image fusion linear transformation coefficient a is calculated according to the following formula _k ' the calculation formula:

wherein the method comprises the steps of

And->

The optimal linear change coefficients are obtained by the edge perception weighted guided filtering cost function, the gradient domain guided filtering cost function and the guided filtering cost function of the Log operator respectively; based on new a _k ' Linear transformation formula for calculating infrared thermal image fusion linear transformation parameter b _k '：

Based on new infrared thermal image fusion linear transformation parameter a _k ' and b _k ' computing and updating the individual cost function values E in the multi-objective optimization problem ₁ (a _k '),E ₂ (a _k '),E ₃ (a _k ')；

Step S344, based on the updated new linear transformation parameters a _k ' and b _k ' and a multi-objective guided filtering cost function value, for n=1,.. _P : comparing and updating speeds according to a particle swarm algorithm, and reserving a non-dominant guide filtering linear transformation coefficient solution set by using a local optimal guide filtering linear transformation coefficient solution and a global optimal guide filtering linear transformation coefficient solution; at the same time n=n+1, if N is not more than N _P G '=g' +1;

step S345, evolution termination judgment: if g 'is less than or equal to g' _max Repeating the stepsStep S343-step 344, if g '> g' _max Obtaining the final front approximate solution set of the multi-target guided filtering linear parameter _i AP；

Step S35, optimizing solution set from optimal Pareto based on weighted membership scheme _i Selecting i Zhang Re amplitude fusion coarse weight graph multi-objective guide filtering Pareto optimal linear transformation parameters from AP

Step S36, multi-objective guided filtering Pareto optimal linear transformation coefficient selected based on multi-objective optimization

Calculating the optimal linear transformation coefficient ++N of the multi-objective guided filtering of the i Zhang Re-th amplitude fused coarse weight image>

The calculation formula is as follows:

wherein,,

representing an infrared reconstructed image _i R rectangular window w _k The mean value of the infrared heat amplitude corresponding to each coordinate point in the infrared heat radiation system,

representing a coarse weight map _i P is in rectangular window w _k The average value of the infrared heat amplitude corresponding to each coordinate point in the infrared heat source;

step S37, optimal linear transformation coefficient based on Pareto

And->

Obtaining an expression of a final linear transformation parameter of the multi-objective guided filtering:

wherein, |w _n And I is the number of coordinate points in a guide filtering window with the nth coordinate as the center, and the expression of the final multi-target guide filtering operator is as follows:

wherein,, _i O _n fusing and refining weight values for the thermal amplitude values corresponding to the nth coordinate point in the multi-target guiding and filtering output image; the operation of filtering by utilizing the obtained multi-objective optimal linear transformation coefficient to obtain a multi-objective guiding filtering operator is recorded as MOGF _r,ε (P, R), wherein R is the size of a guide filter window, epsilon is a regularization parameter, P is an infrared thermal amplitude fusion coarse weight image, and R is an infrared reconstruction image;

Step S38, utilizing the optimal guiding filter operator MOGF obtained by multi-objective optimization _r,ε (P, R) performing multi-objective guided filtering on the obtained thermal amplitude fusion coarse weight graph to obtain corrected infrared thermal amplitude fusion weight images of the base layer and the detail layer:

wherein the method comprises the steps of _i W ^B And _i W ^D to fuse coarse weightsThe i-th base layer infrared heat amplitude fusion finishing weight value graph and the i-th detail layer heat radiation value fusion finishing weight value graph after the graph is subjected to multi-target guiding filtering, _i p is the i Zhang Re radiation value fusion coarse weight graph, _i r is i Zhang Chonggou infrared thermal image, R ₁ ,ε ₁ ,r ₂ ,ε ₂ And respectively obtaining parameters of the corresponding guide filters, and finally normalizing the refined thermal amplitude fusion weight map.

Preferably, the step three obtains the corresponding base layer infrared thermal image { inf.base [ def. (1) ],) and the corresponding base layer infrared thermal image { inf.base [ def. (i) ], the thermal amplitude fusion weight map { wm.base [ def. (1) ], wm.base [ def. (|c|) the }, base [ def. (i) ],) and the corresponding base layer infrared thermal image { inf.base [ def. (1) ],) and the corresponding detail layer infrared thermal image { inf.base [ def. (1) ],) and the corresponding base layer infrared thermal image { inf.base [ def. (i) ], the corresponding base [ def. (|c.) ] }, respectively.

Step S31, reconstructing an image based on infrared _Def.(i) R obtains thermal amplitude fusion rough weight graph _Def.(i) P is as follows; an initial thermal radiation rough fusion weight map is obtained based on the following formula:

_Def.(i) H＝ _Def.(i) R*L

_Def.(i) S＝| _Def.(i) H|*GF

wherein L is a Laplacian filter, GF is a Gaussian low-pass filter, and a thermal amplitude fusion coarse weight map is obtained based on the following formula _Def.(i) P：

Wherein { is as follows _Def.(i) P ₁ ,…, _Def.(i) P _k ,…, _Def.(i) P _M×N Is a coarse weight graph _Def.(i) The thermal amplitude of each position coordinate of P fuses the weight values, _Def.(i) P _k is that _Def.(i) Thermal amplitude fusion weight of kth coordinate point of PThe value of the sum of the values, _Def.(i) S _k is a characteristic diagram of heat amplitude significance _Def.(i) A radiation significance level value corresponding to a kth coordinate point in S, wherein k=1,..m×n;

step S32, modeling a relation between filtering input and filtering output of multi-target guide filtering; reconstructing an image with infrared light _Def.(i) R is a guiding image, and a thermal amplitude value is used for fusing a rough weight map _Def.(i) P is an input image, and multi-target guiding filtering is carried out; in the case of multi-target guided filtering, a guided filter window w is defined _k To guide the image, i.e. to reconstruct the image infrared _Def.(i) At the kth coordinate point in R _Def.(i) R _k A local rectangular window with a size of (2r+1) × (2r+1), k=1,..m×n, the input-output relationship of the multi-objective guided filtering is:

_Def.(i) O _n ＝a _k · _Def.(i) R _n +b _k

wherein,, _Def.(i) O _n representing images reconstructed by infrared _Def.(i) R is a guiding image, and a thermal amplitude value is used for fusing a rough weight map _Def.(i) P is a typical type defect output image of an ith detection area obtained by multi-target guided filtering of an input image _Def.(i) The guided filter output value corresponding to the nth coordinate point of O, n=1,..m×n; _Def.(i) R _n is that _Def.(i) The reconstructed image thermal amplitude corresponding to the nth coordinate point of R, n=1,..m×n; a, a _k And b _k Expressed in terms of _Def.(i) R _k Centered guided filter window w _k Linear transformation parameters in, k=1,..m×n;

step S33, for obtaining the fused optimal weight values of the infrared thermal amplitudes of the corresponding positions of the infrared thermal reconstruction images of the typical defect types of each infrared detection region, the linear transformation parameters a of the guided filtering are obtained _k And b _k The method for modeling the multi-objective optimization problem comprises the following steps:

step S331, fusing coarse weight graphs based on thermal amplitude values _Def.(i) P and infrared reconstructed images _Def.(i) R，Definition of the definitionInfrared big at each coordinate point positionSize defect edge feature perception weighting guide filtering cost function

Wherein,,

and->

The optimal linear transformation coefficient is determined by a large-size defect perception filtering cost function; _Def.(i) P _n is a weight graph _Def.(i) The heat radiation fusion weight value corresponding to the nth coordinate point of P; epsilon is a regularization factor;

is an edge-aware weighting factor defined as follows:

wherein,,

representing an infrared reconstructed image _Def.(i) In R, to _Def.(i) R _k The variance of the heat radiation value corresponding to each coordinate point in the 3×3 window with the coordinate point as the center, ζ is a very small constant having a size of (0.001×dr #) _Def.(i) P)) ² DR (·) is the dynamic range of the image, and by minimizing the cost function, the following expression of the optimal linear transform coefficient is obtained:

wherein,,

representing a representation of an infrared reconstructed image _Def.(i) R and infrared thermal amplitude fusion coarse weight graph _Def.(i) Hadamard product of P is found in rectangular window w _k The mean value of the thermal amplitude corresponding to each coordinate point in the graph, < + >>

Is the Hadamard product of the matrix, +.>

And->

Respectively representing infrared reconstructed images _Def.(i) R and fused coarse weight map _Def.(i) P is in rectangular window w _k Mean value of interior->

Representing an infrared reconstructed image _Def.(i) R is in rectangular window w _k The thermal amplitude variance corresponding to each coordinate point in the graph;

step S332, fusing the rough weight map based on the thermal amplitude value _Def.(i) P and infrared reconstructed images _Def.(i) R, defining a gradient domain infrared fine size defect detail texture guiding filtering cost function at each coordinate point position

Wherein,,

and->

Representing an infrared reconstructed image _Def.(i) In R, to _Def.(i) R _k Guide filter window w with coordinate point as center _k Thermal amplitude standard deviation v corresponding to each coordinate point in the graph _k Is defined as follows:

wherein eta is

Representing an infrared reconstructed image _Def.(i) In R, to _Def.(i) R _k Standard deviation of thermal amplitude corresponding to each coordinate point in 3 x 3 window with coordinate point as center,/->

Representing an infrared reconstructed image _Def.(i) In R, to _Def.(i) R _k Guide filtering rectangular window w with coordinate point as center _n The standard deviation of the thermal amplitude corresponding to each coordinate point in the range, wherein N is E M multiplied by N;

guided filtering cost function by minimizing gradient domain

Obtain->

And->

The calculation formula of (2) is as follows:

wherein,,

representing a representation of an infrared reconstructed image _Def.(i) R and thermal amplitude fused coarse weight graph _Def.(i) Hadamard product of P is found in rectangular window w _k The average value of the thermal amplitude value v corresponding to each coordinate point in the graph _k To adjust a _k Factors of (2);

step S333, fusing the coarse weight map based on the thermal amplitude _Def.(i) P and infrared reconstructed images _Def.(i) R, defining a local LoG operator space noise elimination guide filtering cost function

Wherein,,

and->

The local LoG edge weighting factor is defined as follows:

wherein LoG (·) is a Gaussian Laplace edge detection operator, mxN is the total coordinate point number of the infrared reconstruction image, and |·| is the absolute value operation, delta _LoG 0.1 times the maximum value of the LoG image;

guided filtering cost function by minimizing gradient domain

Obtain->

And->

The calculation formula of (2) is as follows:

wherein the method comprises the steps of

And->

Respectively representing infrared reconstructed images _Def.(i) R and coarse weight map _Def.(i) P is in rectangular window w _k The average value of the thermal amplitude corresponding to each coordinate point in the graph;

wherein a is _k ' is the kth guided filter window w _k Is used to determine the linear transformation coefficients of the block, _Inf.Sig E ₁ (a _k ') preserving a fusion cost function for the large-size defect edges of the infrared thermal image with obvious gradient change, _Inf.Min E ₂ (a _k ' reserving a fusion cost function for detail textures of tiny defects of infrared thermal images with insignificant size and gradient changes, E ₃ (a _k ' is an infrared thermal image noise information sensing and eliminating cost function;

step S34, optimizing the multi-objective optimization problem by using a multi-objective optimization method based on a Chebyshev decomposition method and a particle swarm, wherein the specific method comprises the following steps:

step S341, initializing multi-objective optimization related parameters, initializing the iteration number g' =0, and uniformly distributing weight vectors

Wherein l=3 is the total number of the fusion cost functions of the infrared thermal images; initializing a reference point _i r＝{ _i r ₁ ,..., _i r ₃ }，

Is a corresponding infrared thermal image fusion reference point; _i AP (0) =Φ; maximum number of iterations g' _max The method comprises the steps of carrying out a first treatment on the surface of the Initializing related parameters of individual particle swarms of the nth thermal image fusion coefficient population;

step S342, utilize

The original multi-objective problem is decomposed into a series of scalar sub-objective problems using the chebyshev decomposition method:

step S343, for each single target sub-problem after decomposition, based on their corresponding weight vectors

wherein the method comprises the steps of

And->

step S345, evolution termination judgment: if g 'is less than or equal to g' _max Repeating steps S343 to S344, and if g '> g' _max Obtaining the final front approximate solution set of the multi-target guided filtering linear parameter _i AP；

Calculating another optimal linear transformation coefficient of multi-objective guided filtering of i Zhang Gongwai thermal amplitude fused coarse weight image

The calculation formula is as follows:

wherein,,

representing an infrared reconstructed image _Def.(i) R rectangular window w _k Infrared heat amplitude mean value corresponding to each coordinate point in the graph, < + >>

Representing a coarse weight map _Def.(i) P is in rectangular window w _k The average value of the infrared heat amplitude corresponding to each coordinate point in the infrared heat source;

step S37, optimal linear transformation coefficient based on Pareto

And->

wherein,, _Def.(i) R _n Fusing and refining weight values for thermal amplitude values corresponding to an nth coordinate point in the output image of the multi-target guide filtering, and recording the operation of filtering the weight graph of the i-th infrared detection region infrared thermal reconstruction image by utilizing the obtained multi-target optimal linear transformation coefficient as

Wherein R is the size of a guide filter window, epsilon is a regularization parameter, P is a thermal amplitude fusion coarse weight image, and R is an infrared reconstruction image;

step S38, utilizing the optimal guided filter operator obtained by multi-objective optimization

Performing multi-target guide filtering on the thermal amplitude fusion coarse weight graph of the obtained infrared thermal reconstruction image of the ith infrared detection area to obtain corrected thermal amplitude fusion weight images of the base layer and the detail layer:

wherein WM.Base [ def. (i)]Detail [ def. (i)]A base layer thermal amplitude fusion refinement weight map of the i-th infrared detection region typical type defect infrared thermal reconstruction image and a detail layer thermal radiation value fusion refinement weight map of the i-th infrared detection region infrared thermal reconstruction image after multi-target guiding filtering are fused, _Def.(i) p is a heat radiation value fusion rough weight graph of an infrared heat reconstruction image of an ith infrared detection area, _Def.(i) R is the infrared thermal reconstruction image of the ith infrared detection area, R ₁ ,ε ₁ ,r ₂ ,ε ₂ And respectively obtaining parameters of the corresponding guide filters, and finally normalizing the refined thermal amplitude fusion weight map.

Preferably, the specific method of the fourth step is as follows: fusing weight map { in the obtained fine detail layer thermal amplitude value ₁ W ^D , ₂ W ^D ,…, _|C|-1 W ^D Weight map { fused with base layer infrared thermal amplitude values } ₁ W ^B , ₂ W ^B ,…, _|C|-1 W ^B Fusing the detail layer infrared thermal image information and the base layer infrared thermal image information among the thermal reconstruction images of different defect areas except the background area to obtain a base layer infrared thermal image and a detail layer infrared thermal image fused with the effective information of a plurality of reconstruction thermal images:

finally, combining the weighted average basic layer infrared thermal image and the detail layer infrared thermal image to obtain a final fusion detection infrared thermal image:

in this way, a multi-target guide filtering fusion image which fuses the effective information of the defects of a plurality of reconstructed thermal images and simultaneously considers the reservation requirement of large-size defects and the detail texture reservation requirement of micro defects in each thermal image and the overall noise elimination reservation requirement is obtained; and inputting the high-quality infrared reconstruction fusion image F which is fused with the defect characteristics of various complex types into the steps of infrared thermal image segmentation and defect quantitative analysis so as to further extract quantitative characteristic information of various defects.

Preferably, the specific method of the fourth step is as follows: based on the obtained detail layer thermal amplitude fusion weight map { wm.detail [ def (1) ], & gt, wm.detail [ def (i) ], & gt, wm.detail [ def (|c|.) ]) and the base layer thermal amplitude fusion weight map { wm.base [ def (1) ], & gt, wm.base [ def (i) ], & gt, wm base [ def (|c|.) ], the detail layer thermal image information and the base layer thermal image information between the typical type defect thermal reconstruction images of different areas in different detection times in the large-size test piece are fused, and the base layer thermal image and the detail layer thermal image fused with the effective information of the multiple multi-detection area reconstruction thermal image are obtained.

Finally, combining the weighted average basic layer thermal image and the detail layer thermal image to obtain a final fusion detection infrared thermal image:

thus, the infrared detection fusion thermal image fused with the effective information of the reconstructed thermal image defects of the typical type defects of the plurality of infrared detection areas of the large-size test piece is obtained; the infrared fusion thermal image integrates the excellent characteristics of various guide filters by utilizing a multi-objective optimization algorithm, and the typical type defects of different areas are fused together by multiple infrared detection, so that the high-quality simultaneous imaging of the defects of the large-size pressure container is realized; and inputting the high-quality infrared reconstruction fusion image F which is fused with the typical characteristics of the defects of the detection areas into the steps of infrared thermal image segmentation and defect quantitative analysis so as to further extract quantitative characteristic information of various defects.

The invention at least comprises the following beneficial effects:

1. according to the multi-type damage detection feature analysis method for the large-size test piece, a grid-based CLIQUE clustering algorithm is combined, and a transient thermal response set is clustered rapidly and adaptively, so that various transient thermal responses corresponding to various defects in different infrared detection areas of the large-size pressure container are obtained from different thermal image sequences, thermal image reconstruction is carried out based on various typical feature thermal responses, and visual imaging of typical type defects in the current infrared detection area is achieved. After the reconstructed thermal images of typical defects in each detection area are obtained, effective information in the reconstructed thermal images of different types of defects is combined by utilizing an image fusion algorithm combining a multi-target evolutionary optimization algorithm and a guided filtering algorithm, so that the detection capability and defect characteristic characterization performance of the external thermal image of Shan Zhanggong are improved. After the original infrared thermal reconstruction image is subjected to image decomposition to obtain a basic layer image and a detail layer image of the thermal imageAnd fusing the infrared thermal images of the defects of different types on two scales of the basic layer and the detail layer. By utilizing the excellent edge retention characteristic of the guide filtering, the edge contour and detail information of various defects are reserved while the images are fused, and the detail expression capability of various defects in the images after various types of defects are fused is improved. Simultaneously combining the specific excellent performances of a plurality of guide filters together by combining a multi-objective optimization algorithm, simultaneously optimizing 3 guide filter cost functions to obtain a series of Pareto optimal non-dominant solution sets, extracting an optimal solution by utilizing a weighted membership scheme, and based on the obtained multi-objective optimal linear transformation coefficient

And->

Constructing a multi-target optimal guided filter operator MOGF _r,ε (P, R) based on a multi-objective optimal guided filter operator MOGF _r,ε (P, R) obtaining different refinement fusion weight graphs of the base layer and the detail layer in two scales. And respectively guiding to carry out weighted fusion among the base layer images and weighted fusion among the detail layer images based on the corrected weight graphs. And finally, combining the weighted average detail layer image and the base layer image to obtain a final fusion image. And performing defect segmentation positioning and quantization operation based on the final fusion thermal image.

2. The method combines the CLIQUE clustering algorithm to realize the efficient, rapid and self-adaptive clustering of the transient thermal response information, avoids the step of manually identifying the defect category number and judging the category number, and removes noise and abnormal values.

3. The invention adopts an image fusion strategy, and can fuse the effective information of a plurality of reconstructed thermal images. Therefore, the detection performance of the single Zhang Re image is improved, and the problem that the single detection image of the complex type test piece defect is incomplete due to ultrahigh-speed impact caused by the limitation of infrared detection performance can be solved by performing image fusion on a plurality of thermal images.

4. The invention adopts an image fusion strategy combining multi-objective optimization and guided filtering. The edge retention performance of the guide filtering is utilized to smooth the image and retain the edge, so that the definition and contrast of the defect edge of the fused image are improved. The advantages of multiple guide filters are combined together based on multi-objective optimization, so that the performance of the fused image on complex type defect contour edges and fine size defects is further improved, and image noise is smoothed.

Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention.

Drawings

FIG. 1 is a flow chart showing the implementation of the method for analyzing the multi-type damage detection characteristics of a large-size test piece in example 1;

FIG. 2 is a flow chart of an overall fusion framework for fusion of multiple (two, for example) infrared thermal images based on combining multi-objective optimization and guided filtering in accordance with example 1;

FIG. 3 is a flowchart of the modified weighted image of each image layer obtained by combining multi-objective optimization and guided filtering in detail in example 1;

FIG. 4 is a graph of the results of example 1 classifying transient thermal response sets in a thermal image sequence of a first detection region using a CLIQUE adaptive clustering algorithm;

FIG. 5 is a graph showing the results of classifying transient thermal response sets in a thermal image sequence of a second detection region using the CLIQUE adaptive clustering algorithm of example 1

FIG. 6 is a graph of a typical characteristic transient thermal response of a typical type of defect of the first detection zone extracted in example 1;

FIG. 7 is a graph of a typical characteristic transient thermal response of a typical type of defect of a second detection zone extracted in example 1;

FIG. 8 is an infrared thermal reconstruction image obtained in example 1 based on a typical characteristic transient thermal response of a typical type of defect in a first detection zone;

FIG. 9 is an infrared thermal reconstruction image obtained in example 1 based on a typical characteristic transient thermal response of a second detection zone typical type defect;

FIG. 10 is a graph showing the optimal leading edge of the fusion parameters of the infrared thermal image based on the combination of the multi-objective optimization and the plurality of guide filters and the optimal fusion parameter solution of the thermal image based on the weighted membership in the embodiment 1;

FIG. 11 is a thermal image refinement base layer image fusion weight map a corrected based on the resulting optimal multi-objective guided filter fusion operator of example 1;

FIG. 12 is a thermal image refinement base layer image fusion weight map b modified by the optimal multi-objective guided filter fusion operator of example 1;

FIG. 13 is a thermal image refinement detail layer image fusion weight map c modified by the optimal multi-objective guided filter fusion operator according to example 1;

FIG. 14 is a thermal image refinement detail layer image fusion weight map d modified by the optimal multi-objective guided filter fusion operator according to example 1;

FIG. 15 is a final infrared fused thermal image based on multi-objective optimization and guided filtering of example 1;

fig. 16 is a flowchart of a multi-type damage detection image feature extraction and recognition method of embodiment 2;

FIG. 17 is a flow chart of an overall fusion framework for fusion of multiple (two, for example) infrared thermal images based on combining multi-objective optimization and guided filtering in accordance with example 2;

FIG. 18 is a flowchart of the modified weighted image of each image layer obtained by combining multi-objective optimization and guided filtering in detail in example 2;

FIG. 19 is a graph of the results of example 2 classifying transient thermal response sets using the CLIQUE adaptive clustering algorithm;

FIG. 20 is a typical characteristic transient thermal response plot of the background region extracted in example 2;

FIG. 21 is a graph of typical characteristic transient thermal response of a first type of defect region extracted in example 2;

FIG. 22 is a typical characteristic transient thermal response plot of a second type of defect region extracted in example 2;

FIG. 23 is an infrared thermal reconstruction image of a non-defective background region from example 2 based on a typical characteristic transient thermal response of the background region;

FIG. 24 is an infrared thermal reconstruction image of a center impingement pit area obtained based on a typical characteristic transient thermal response curve for a first type of defect area of example 2;

FIG. 25 is an infrared thermal reconstruction image of an edge fine impingement sputter damage region obtained based on a typical characteristic transient thermal response curve for a second type of defect region of example 2;

FIG. 26 is a graph showing the optimal leading edge of the fusion parameters of the infrared thermal image based on the combination of the multi-objective optimization and the plurality of guide filters and the optimal fusion parameter solution of the thermal image based on the weighted membership in the embodiment 2;

FIG. 27 is a thermal image refinement base layer image fusion weight map e modified by the optimal multi-objective guided filter fusion operator of example 2;

FIG. 28 is a thermal image refinement base layer image fusion weighting map f modified by the optimal multi-objective guided filter fusion operator of example 2;

FIG. 29 is a thermal image refinement detail layer image fusion weight map g modified by the optimal multi-objective guided filter fusion operator according to example 2;

FIG. 30 is a thermal image refinement detail layer image fusion weight map h modified by the optimal multi-objective guided filter fusion operator according to embodiment 2;

FIG. 31 is a final infrared fused thermal image based on multi-objective optimization and guided filtering of example 2.

Detailed Description

The present invention is described in further detail below with reference to the drawings to enable those skilled in the art to practice the invention by referring to the description.

Example 1

As shown in fig. 1-3: the invention discloses a multi-type damage detection characteristic analysis method for a large-size test piece, which comprises the following steps of:

the method comprises the steps of firstly, carrying out infrared detection on a large-size test piece for multiple times, obtaining a plurality of thermal image sequences of the large-size test piece, and obtaining a plurality of reconstructed infrared thermal images of the large-size test piece from the plurality of thermal image sequences by utilizing an infrared characteristic extraction and infrared thermal image reconstruction algorithm, wherein the specific method comprises the following steps:

step S11, using a three-dimensional matrix set { S ] for a plurality of thermal infrared image sequences acquired from the thermal infrared imager ¹ ,…,S ⁱ ,…,S ^|C| Represented by S, where S ⁱ Representing a thermal image sequence obtained by the thermal infrared imager in the ith infrared detection, wherein |C| represents the total thermal image sequence number; s is S ⁱ (M, N, T) represents a temperature value at an M-th row, N-th column coordinate position of a T-th frame thermal image in the i-th thermal image sequence, t=1,..t, T is a total frame number, m=1,..m, M is a total number of rows, n=1,..n, N is a total number of columns;

Step S13, utilizing a grid-based self-adaptive CLIQUE clustering algorithm to carry out ith thermal image sequence S ⁱ Transient thermal response set X of (2) ⁱ (g) Transient thermal response adaptive clustering in (a); the preset partition interval parameter delta divides the T-dimension data space of the transient thermal response into The method comprises the steps of marking dense grids and sparse grids through transient thermal response data in grids by rectangular grids which are not overlapped with each other, and carrying out sparse grid correction on the sparse grids by utilizing a boundary correction method and a sliding grid method; according to a greedy algorithm retrieval principle, connecting dense grids and corrected sparse grids to form a maximum connected region, wherein the connected dense grids and corrected sparse grids are transient thermal response clusters, and the CLIQUE clustering algorithm adaptively identifies the number of dense grid clusters in a thermal response space through the steps of adaptively identifying the number of cluster clusters in the transient thermal response data space, so that the step of judging the defect category number is omitted; sequence S of thermal images ⁱ Transient thermal response set X of (2) ⁱ (g) Adaptive clustering to form cluster sets

Wherein the method comprises the steps of

Using matrix Y ⁱ And S is ⁱ Carrying out infrared thermal image reconstruction on the information of (a) and carrying out the (i) th thermal image sequence S ⁱ Each frame of (3)Two-dimensional image matrix O formed by extracting images into a column vector by columns and arranging the column vectors in time sequence ⁱ The thermal amplitude reconstruction matrix R for the ith detection is obtained based on the following transformation formula ⁱ ：

Wherein,,

Step S15, if i < |C|, i+1, i.e. for the i+1th IR thermal image sequence S ⁱ⁺¹ Repeating the steps S12-S14 until all the typical type defect reconstruction thermal images in the current detected region are obtained from a plurality of thermal image sequences obtained by multiple detection respectively, calculating MSE values of all the type defect reconstruction images in the current region, and selecting the typical type defect reconstruction thermal images in each detected region based on the MSE minimum principle, namely obtaining a typical type defect reconstruction thermal image set { in each detected region of a large-size test piece _Def.(1) R,..., _Def.(i) R,.., _Def.(|C|) R }, wherein _Def.(i) R represents a typical type defect reconstruction thermal image of the detected region in the i-th thermal image sequence, i=1,..+ -. C|;

step two, carrying out infrared reconstruction image { on a total |C| sheet of typical type defects in each detection area in the large-size impact test piece _Def.(1) R,..., _Def.(i) R,..., _Def.(|C|) Each of R is subjected to image decomposition, and each reconstructed image is decomposed into a base layer infrared thermal image { inf.base [ def. (1)],...,Inf.Base[Def.(i)],...,Inf.Base[Def.(|C|)]And a detail layer infrared thermal image { Inf. Detail [ def. (1)],...,Inf.Detail[Def.(i)],...,Inf.Detail[Def.(|C|)]}；

Reconstructing a thermal image with an i (i=1, …, |c|) th detection region typical type defect _Def.(i) R is exemplified by the following formula _Def.(i) Typical type of defect base layer ir thermal image and detail layer ir thermal image of R inf.base [ def. (i) ]And Inf. Detail [ def. (i)]：

Inf.Base[Def.(i)]＝ _Def.(i) R*Z

Inf.Detail[Def.(i)]＝ _Def.(i) R-Inf.Base[Def.(i)]

Wherein Z is an average filter.

Step three, respectively acquiring a thermal amplitude fusion weight map { WM.Base [ Def (1) ], an angle, WM.Base [ Def (i) ], an angle of the corresponding base layer infrared thermal image { Inf.Base [ Def (1) ], an angle of Inf.Base [ Def (i) ], an angle of Inf.Base [ Def ] (|C|) ] }, an angle of the corresponding base layer infrared thermal image { WM.Base [ Def (1) ], an angle of the corresponding base layer infrared thermal image { Base [ Def ] (i) ], wm.base [ def. (|c|) and detail layer infrared thermal image { inf.detail [ def. (1) ],) inf.detail [ def. (i) ],) a thermal amplitude fusion weight map { wm.detail [ def. (1) ], }, wm.detail [ def. (|c|) between inf.detail [ def. ] }, a specific method comprising

_Def.(i) H＝ _Def.(i) R*L

_Def.(i) S＝| _Def.(i) H|*GF

Wherein { is as follows _Def.(i) P ₁ ,…, _Def.(i) P _k ,…, _Def.(i) P _M×N Is a coarse weight graph _Def.(i) The thermal amplitude of each position coordinate of P fuses the weight values, _Def.(i) P _k is that _Def.(i) The thermal amplitude value of the kth coordinate point of P fuses the weight values, _Def.(i) S _k is a characteristic diagram of heat amplitude significance _Def.(i) A radiation significance level value corresponding to a kth coordinate point in S, wherein k=1,..m×n;

_Def.(i) O _n ＝a _k · _Def.(i) R _n +b _k

step S331, fusing coarse weight graphs based on thermal amplitude values _Def.(i) P and infrared reconstructed images _Def.(i) R，Definition of the definitionInfrared large-size defect edge feature perception weighting guide filtering cost function at each coordinate point position

Wherein,,

and->

is an edge-aware weighting factor defined as follows:

wherein,,

wherein,,

Is the hadamard product of the matrix,

and->

Representing an infrared reconstructed image _Def.(i) R is in rectangular window w _k In the inner partThe thermal amplitude variance corresponding to each coordinate point;

Wherein,,

and->

wherein eta is

guided filtering cost function by minimizing gradient domain

Obtain->

And->

The calculation formula of (2) is as follows:

wherein,,

Wherein,,

and->

The local LoG edge weighting factor is defined as follows:

guided filtering cost function by minimizing gradient domain

Obtain->

And->

The calculation formula of (2) is：

Wherein the method comprises the steps of

And->

step S342, utilize

wherein the method comprises the steps of

And->

Cost function of edge perception weighted guided filtering, cost function of gradient domain guided filtering and guided filtering of Log operator The optimal linear change coefficient obtained by the wave cost function. Based on new a _k ' Linear transformation formula for calculating infrared thermal image fusion linear transformation parameter b _k '：

The calculation formula is as follows:

wherein,,

step S37, optimal linear transformation coefficient based on Pareto

And->

wherein the method comprises the steps of， _Def.(i) R _n Fusing and refining weight values for thermal amplitude values corresponding to an nth coordinate point in the output image of the multi-target guide filtering, and recording the operation of filtering the weight graph of the i-th infrared detection region infrared thermal reconstruction image by utilizing the obtained multi-target optimal linear transformation coefficient as

Performing multi-target guide filtering on the thermal amplitude fusion coarse weight graph of the obtained infrared thermal reconstruction image of the ith infrared detection area to obtain corrected thermal amplitude fusion weight images of the base layer and the detail layer: />

Wherein WM.Base [ def. (i)]Detail [ def. (i)]A base layer thermal amplitude fusion refinement weight map of the i-th infrared detection region typical type defect infrared thermal reconstruction image and a detail layer thermal radiation value fusion refinement weight map of the i-th infrared detection region infrared thermal reconstruction image after multi-target guiding filtering are fused, _Def.(i) p is a heat radiation value fusion rough weight graph of an infrared heat reconstruction image of an ith infrared detection area, _Def.(i) r is the infrared thermal reconstruction image of the ith infrared detection area, R ₁ ,ε ₁ ,r ₂ ,ε ₂ Respectively corresponding parameters of the guide filter, and finally normalizing the refined thermal amplitude fusion weight mapAnd (5) performing chemical treatment.

And step four, fusing the detail layer thermal image information and the base layer thermal image information between the typical defect thermal reconstruction images of different areas in different detection times in the large-size test piece based on the obtained detail layer thermal amplitude fusion weight map { WM.detail [ def (1) ], & gt, WM.detail [ def (i) ], & gt, WM.detail [ def (|C|.) ] }, and the base layer thermal amplitude fusion weight map { WM.base [ def (1) ], & gt, WM.base [ def (i) ], & gt, WM.base [ def (|C|.) ] }, so as to obtain a base layer thermal image and a detail layer thermal image fused with the effective information of the reconstructed thermal images of a plurality of detection areas.

In this embodiment, two areas of defects on the test piece need to be detected, namely an artificial surface hole digging defect 1 in the first area and an artificial filling defect 2 in the second area.

A flow chart of an overall fusion framework based on fusion of multiple (two, for example) infrared thermal images in combination with multi-objective optimization and guided filtering is shown in fig. 2.

A flowchart of the modified weighted image of each image layer obtained by specifically combining multi-objective optimization and guided filtering is shown in fig. 3.

In this example, a result chart obtained by classifying the transient thermal response set of the first detection area by using the CLIQUE adaptive clustering algorithm is shown in fig. 4, and a result chart obtained by classifying the transient thermal response set of the second detection area is shown in fig. 5.

Based on CLIQUE self-adaptive clustering algorithm, a clustering center corresponding to each transient thermal response set is obtained and used as typical characteristic transient thermal response of typical type defects of each region _Def.(1) R and _Def.(2) r is defined as the formula. Their respective typical characteristic transient thermal response curves are shown in fig. 6, 7.

After obtaining typical characteristic transient thermal response curves of typical type defects of each region of a test piece, carrying out an infrared thermal image reconstruction algorithm based on the typical characteristic transient thermal response curves to obtain artificial surface hole digging of a first region of a material _Def.(1) R-corresponding reconstructed thermal image and artificial filling defect of second region _Def.(2) R corresponds to the reconstructed thermal image, as shown in FIGS. 8 and 9, and their respective highlighted defect types are shown.

By the method for solving the optimal guided filtering linear transformation parameters by combining multi-objective optimization and guided filtering, a series of Pareto optimal non-dominant solutions are obtained, a Pareto optimal front face (PF) is obtained based on the Pareto optimal non-dominant solutions, and an optimal guided filtering thermal image fusion parameter solution is selected based on an optimal weighted membership principle, and is shown in figure 10.

Obtaining optimal guide filtering thermal image fusion parameters based on multi-objective optimization and guide filtering, obtaining a multi-objective guide filtering optimal operator, and respectively corresponding to a base layer image and a detail layer image obtained after infrared thermal reconstruction image decompositionThe weighted image is subject to a multi-objective guided filtering operation. And obtaining refined weight diagrams on each image level after multi-target guided filtering correction. In W ₁ ^B A, W representing the refined base layer weight map ₂ ^B Representing the refined base layer weight map b, W ₁ ^D Weight map c, W representing detail layer after refinement ₂ ^D The base layer weight map d after finishing is shown in fig. 11, 12, 13, and 14, respectively.

And carrying out infrared thermal image fusion operation on each layer of weight image corrected based on the multi-target optimal guiding filter operator, wherein the obtained infrared fusion thermal images of each region of the large-size pressure vessel are shown in fig. 15. The damage condition characteristics of the defect 1 and the defect 2 can be clearly and simultaneously characterized with high quality in the figure, and the subsequent image segmentation and defect identification quantitative operation can be better carried out.

In this embodiment, the extracted features fusing large-sized pressure vessel defects are shown in fig. 15.

It can be seen that the final fused infrared detection image obtained in this embodiment has better detectability for defects in various areas of a large-sized pressure vessel.

Example 2

As shown in fig. 16-19: the invention discloses a multi-type damage detection image feature extraction and identification method, which comprises the following steps:

the method comprises the following specific steps of:

step S11, a valuable transient thermal response data set X (g) is extracted from a thermal image sequence S acquired by a thermal infrared imager based on a transient thermal response data extraction algorithm of block variable step length; where S (I, J, T) represents the pixel value of the I (i=1, …, I) th row (I is the total number of rows) and the J (j=1, … J) th column (J is the total number of columns) of the T (t=1, …, T) frame infrared thermal image (T is the total number of frames) of the thermal image sequence; decomposing a thermal image sequence into K different data blocks by means of a threshold value _k S(i _n ,j _m T) wherein k represents the kth sub-data block, i _n 、j _m T represents the ith of the kth sub-block, respectively _n Line j _m Column, pixel value of the t frame; then defining the search line step length in the (k=1, …, K) data blocks according to the temperature change characteristics in the different data blocks _k RSS and column step size _k CSS; based on different search steps in different data blocks, comparing correlation coefficients between data points, and searching a series of correlation coefficients greater than a threshold THC _cr And adding the transient thermal response data set X (g);

Wherein the method comprises the steps of

For h (h=1, 2, |C|) clustering results _X(g) Cluster[h]H=1, …, the kth of |c| represents the transient thermal response, | _X(g) Cluster[h]The I is the total number of transient thermal responses contained in the h clustering result, and a matrix Y is formed by using typical transient thermal responses of various types of defects;

wherein,,

Step two, carrying out { on (C-1) pieces of infrared reconstruction images except for the thermal images of the open background areas ₁ R,…, _i R,…, _|C|-1 Each R is subjected to image decomposition, and each reconstructed image is decomposed into a base layer infrared thermal image { ₁ B,..., _i B,..., _|C|-1 B and a detail layer infrared thermal image { ₁ D,…, _i D,…, _|C|-1 D }; reconstructing a thermal image with the i (i=1, …, |c| -1) th defective region _i R is exemplified by the following formula _i R base layer infrared thermal image _i B and detail layer IR thermal image _i D

_i B＝ _i R*Z

_i D＝ _i R- _i B

Wherein Z is an average filter.

Step three, respectively obtaining corresponding infrared thermal images { of each base layer by utilizing multi-objective optimized guiding filtering ₁ B, ₂ B,…, _|C|-1 B }, an infrared thermal amplitude fusion weighting graph { between ₁ W ^B , ₂ W ^B ,…, _|C|-1 W ^B And detail layer infrared thermal image { ₁ D, ₂ D,…, _|C|-1 D, fusing weight map { of infrared thermal amplitude values between the two layers ₁ W ^D , ₂ W ^D ,…, _|C|-1 W ^D The specific method is as follows:

_i H＝ _i R*L

_i S＝| _i H|*GF

Wherein { is as follows _i P ₁ 、,..., _i P _k ,..., _i P _I×J Is a coarse weight graph _i The thermal amplitude of each position coordinate of P fuses the weight values, _i P _k (k=1, …, i×j) is _i The thermal amplitude value of the kth coordinate point of P fuses the weight values, _i S _k (k=1, …, i×j) is a thermal amplitude saliency map _i A radiation significance level value corresponding to a kth coordinate point in S;

step S32, performing multiple processesModeling a relation between a filtering input and a filtering output of target-oriented filtering; reconstructing an image with infrared light _i R is a guiding image, and a thermal amplitude value is used for fusing a rough weight map _i P is an input image, and multi-target guiding filtering is carried out; in the case of multi-target guided filtering, a guided filter window w is defined _k To guide the image, i.e. to reconstruct the image infrared _i At the kth coordinate point in R _i R _k (k=1, …, i×j) as a central local rectangular window, whose size is (2r+1) × (2r+1), the input-output relationship of the multi-objective guided filtering is:

_i O _n ＝a _k · _i R _n +b _k

wherein,, _i O _n (n=1, …, i×j) denotes reconstructing the image in infrared _i R is a guiding image, and a thermal amplitude value is used for fusing a rough weight map _i P is an output image obtained by multi-target guided filtering of an input image _i A guide filtering output value corresponding to the nth coordinate point of O; _i R _n (n=1, …, i×j) is _i The reconstructed image thermal amplitude corresponding to the nth coordinate point of R; a, a _k And b _k Expressed in terms of _i R _k (k=1, …, i×j) centered guide filter window w _k Linear transformation parameters in;

Wherein,,

and->

The optimal linear transformation coefficient is determined by a large-size defect perception filtering cost function; _i P _n is a weight graph _i An infrared heat radiation fusion weight value corresponding to the nth coordinate point of P; epsilon is a regularization factor;

Is an edge-aware weighting factor defined as follows:

wherein,,

wherein,,

representing a representation of an infrared reconstructed image _i R and thermal amplitude fused coarse weightsDrawing of the figure _i Hadamard product of P is found in rectangular window w _k Infrared heat amplitude mean value corresponding to each coordinate point in the graph, < + >>

Is the Hadamard product of the matrix, +.>

And->

Wherein,,

and->

wherein eta is

Representing an infrared reconstructed image _i In R, to _i R _k The standard deviation of the infrared heat amplitude corresponding to each coordinate point in the 3 x 3 window with the coordinate point as the center,

representing an infrared reconstructed image _i In R, to _i R _k Guide filtering rectangular window w with coordinate point as center _n Infrared heat amplitude standard deviation corresponding to each coordinate point in the range;

guided filtering cost function by minimizing gradient domain

Obtain->

And->

The calculation formula of (2) is as follows:

Wherein,,

Wherein,,

and->

Is a local LoG (Laplacian-of-Gaussian) edge weighting factor, defined as follows:

guided filtering cost function by minimizing gradient domain

Obtain->

And->

The calculation formula of (2) is as follows:

wherein the method comprises the steps of

And->

wherein,,a _k ' is the kth guided filter window w _k Is used to determine the linear transformation coefficients of the block, _Inf.Sig E ₁ (a _k ') preserving a fusion cost function for the large-size defect edges of the infrared thermal image with obvious gradient change, _Inf.Min E ₂ (a _k ' reserving a fusion cost function for detail textures of tiny defects of infrared thermal images with insignificant size and gradient changes, E ₃ (a _k ' is an infrared thermal image noise information sensing and eliminating cost function;

initializing a reference point _i r＝{ _i r ₁ ,…, _i r ₃ }，

step S342, utilize

Is>

wherein the method comprises the steps of

And->

Step S344, based on the updated new linear transformation parameters a _k ' and b _k ' and multi-objective guided filtering cost function value, for n=1, …, N _P : comparing and updating speeds according to a particle swarm algorithm, and reserving a non-dominant guide filtering linear transformation coefficient solution set by using a local optimal guide filtering linear transformation coefficient solution and a global optimal guide filtering linear transformation coefficient solution; at the same time n=n+1, if N is not more than N _P G '=g' +1;

step S345, evolution termination judgment: if g 'is less than or equal to g' _max Repeating steps S343 to 344, if g '> g' _max Obtaining the final front approximate solution set of the multi-target guided filtering linear parameter _i AP；

The calculation formula is as follows:

wherein,,

step S37, optimal linear transformation coefficient based on Pareto

And->

wherein the method comprises the steps of _i W ^B And _i W ^D the i-th basic layer infrared heat amplitude fusion finishing weight value graph and the i-th detail layer heat radiation value fusion finishing weight value graph after the multi-target guiding filtering are fused with the coarse weight graph, _i p is the i Zhang Re radiation value fusion coarse weight graph, _i r is i Zhang Chonggou infrared thermal image, R ₁ ,ε ₁ ,r ₂ ,ε ₂ And respectively obtaining parameters of the corresponding guide filters, and finally normalizing the refined thermal amplitude fusion weight map.

Step four, fusing weight map { based on the obtained refined detail layer thermal amplitude ₁ W ^D , ₂ W ^D ,…, _|C|-1 W ^D Weight map { fused with base layer infrared thermal amplitude values } ₁ W ^B , ₂ W ^B ,…, _|C|-1 W ^B Fusing the detail layer infrared thermal image information and the base layer infrared thermal image information among the thermal reconstruction images of different defect areas except the background area to obtain a base layer infrared thermal image and a detail layer infrared thermal image fused with the effective information of a plurality of reconstruction thermal images:

In this example, there are two types of defects on the test piece, namely a center impact damage defect 1 and a surrounding sputter type fine damage defect 2.

A flowchart of an overall fusion framework based on fusion of multiple (two, for example) infrared thermal images in combination with multi-objective optimization and guided filtering is shown in fig. 17.

A flowchart of the combination of multi-objective optimization and guided filtering to obtain the corrected weighted image of each image layer is shown in fig. 18.

In this example, a result graph of classifying the transient thermal response set using CLIQUE adaptive clustering algorithm is shown in fig. 19.

Based on CLIQUE self-adaptive clustering algorithm, clustering centers corresponding to various transient thermal response sets are obtained and used as typical characteristic transient thermal response of various types of damaged areas _X(g) C _Cluster [1]、 _X(g) C _Cluster [2]And _X(g) C _Cluster [3]. Their respective typical characteristic transient thermal response curves are shown in fig. 20, 21 and 22.

After typical characteristic transient thermal response curves of each damaged area of the test piece are obtained, carrying out an infrared thermal image reconstruction algorithm based on the typical characteristic transient thermal response curves to obtain a reconstructed thermal image of a non-damaged background area of the material ₁ R and defect 1 temperature point corresponding reconstructed thermal image ₂ R and reconstructed thermal image corresponding to defect 2 temperature point ₃ R, as shown in FIGS. 23, 24 and 25, each of which highlights the defect type as indicated by the label.

By the method for solving the optimal guided filtering linear transformation parameters by combining multi-objective optimization and guided filtering, a series of Pareto optimal non-dominant solutions are obtained, a Pareto optimal front face (PF) is obtained based on the Pareto optimal non-dominant solutions, and an optimal guided filtering thermal image fusion parameter solution is selected based on an optimal weighted membership principle, and is shown in figure 26.

And obtaining optimal guide filtering thermal image fusion parameters based on multi-target optimization and guide filtering, obtaining a multi-target guide filtering optimal operator, and performing multi-target guide filtering operation on weight images corresponding to the base layer image and the detail layer image obtained after the infrared thermal reconstruction image is decomposed. And obtaining refined weight diagrams on each image level after multi-target guided filtering correction. In W ₁ ^B Representing the refined base layer weight map e, W ₂ ^B Representing the refined base layer weight map f, W ₁ ^D A detailed layer weight graph g, W after finishing ₂ ^D The base layer weight map h after finishing is shown in fig. 27, 28, 29, and 30, respectively.

And carrying out infrared thermal image fusion operation on each layer of weight image corrected based on the multi-target optimal guide filtering operator, wherein the obtained final complex type defect infrared fusion thermal image is shown in fig. 31. The damage condition characteristics of the defect 1 and the defect 2 can be clearly and simultaneously characterized with high quality in the figure, and the subsequent image segmentation and defect identification quantitative operation can be better carried out.

In the present embodiment, the extracted features fusing defects of various types are shown in fig. 31.

It can be seen that the final fused infrared detection image obtained in this embodiment has better detectability for various types of lesions.

The number of equipment and the scale of processing described herein are intended to simplify the description of the present invention. Applications, modifications and variations of the present invention will be readily apparent to those skilled in the art.

Although embodiments of the present invention have been disclosed above, it is not limited to the details and embodiments shown and described, it is well suited to various fields of use for which the invention would be readily apparent to those skilled in the art, and accordingly, the invention is not limited to the specific details and illustrations shown and described herein, without departing from the general concepts defined in the claims and their equivalents.

Claims

1. The characteristic analysis method for detecting the multi-type damage of the large-size test piece is characterized by comprising the following steps of:

2. The method for analyzing the multi-type damage detection characteristics of the large-size test piece according to claim 1, wherein the specific steps of acquiring the reconstructed infrared thermal image from the infrared thermal image sequence by using the infrared characteristic extraction and the infrared thermal image reconstruction algorithm in the first step are as follows:

step S12, utilizing a grid-based adaptive CLIQUE clustering algorithm to adaptively cluster transient thermal responses in a transient thermal response set X (g); s of preset dividing interval parameter delta to transient thermal response _w Dividing the dimensional data space into rectangular grids which are not overlapped with each other, and marking dense grids and sparse grids through transient thermal response data quantity in the grids; utilizing boundary correction sums for sparse gridsPerforming sparse grid correction by a sliding grid method; according to a greedy algorithm retrieval principle, connecting the dense grids and the corrected sparse grids to form a maximum connected region; the connected dense grids and the corrected sparse grids are transient thermal response clusters, the CLIQUE clustering algorithm adaptively identifies the number of dense grid clusters in a thermal response space, so that the step of judging the defect class number is omitted, and the number of cluster clusters existing in a transient thermal response data space is adaptively identified; adaptive clustering to form cluster sets _X(g) Cluster[h]H=1, 2, …, |c '|, where h represents a category label, |c' | represents the total number of categories;

wherein the method comprises the steps of

For the h clustering result _X(g) Cluster[h]H=1, 2, …, the kth of |c' | represents the transient thermal response, | _X(g) Cluster[h]The I is the total number of transient thermal responses contained in the h clustering result, and a matrix Y is formed by using typical transient thermal responses of various types of defects;

wherein,,

is a matrix of |C' | x T, is a pseudo-inverse matrix of matrix Y, O ^T Is the transposed matrix of the two-dimensional image matrix O, the obtained reconstruction matrix R is I C 'row and I X J column, each row of the reconstruction matrix R is intercepted to form an I X J two-dimensional image, I C' I X J two-dimensional images are obtained, the images are the reconstruction thermal images containing the characteristic information of different thermal response areas, and the reconstruction thermal images of the non-defect background areas are recorded as follows _B R, recording the reconstructed thermal image corresponding to each type of defect area as _i R, i=1, |c' |; each reconstructed thermal image contains characteristic thermal reconstruction information of one defect type of the complex type defects except the background region thermal image without defect damage.

3. The method for analyzing the characteristics of multi-type damage detection of a large-size test piece according to claim 1, wherein the step of performing infrared detection on the large-size test piece for a plurality of times to obtain a plurality of thermal image sequences of the large-size test piece, and obtaining a plurality of reconstructed infrared thermal images of the large-size test piece from the plurality of thermal image sequences by using an infrared characteristic extraction and infrared thermal image reconstruction algorithm comprises the following steps:

Step S12, for the ith IR thermal image sequence S ⁱ Extracting an ith thermal image sequence S by using a transient thermal response data extraction algorithm based on block variable step length ⁱ A valuable transient thermal response dataset Xi (g); the ith thermal image sequence S is thresholded ⁱ Decomposition into K different data blocks _k Si (m ', n', t) thereink represents the ith thermal image sequence S ⁱ The kth sub data block, m ', n', t respectively represent the temperature values at the coordinate positions of the (m 'th row, n' th column and t th frame of the kth sub data block), and then the ith thermal image sequence S is defined according to the temperature change characteristics in different data blocks ⁱ Search row step size in kth data block _k RSS ⁱ Sum column step size _k CSS ⁱ K=1,..k; based on different search steps in different data blocks, comparing correlation coefficients between data points, and searching a series of correlation coefficients greater than a threshold THC _cr And adding an ith thermal image sequence S ⁱ Transient thermal response data set X in (1) ⁱ (g)；

Step S13, utilizing a grid-based self-adaptive CLIQUE clustering algorithm to carry out ith thermal image sequence S ⁱ Transient thermal response set X of (2) ⁱ (g) Transient thermal response adaptive clustering in (a); s of preset dividing interval parameter delta to transient thermal response _w Dividing the dimensional data space into rectangular grids which are not overlapped with each other, marking dense grids and sparse grids through transient thermal response data quantity in the grids, and carrying out sparse grid correction on the sparse grids by using a boundary correction method and a sliding grid method; according to a greedy algorithm retrieval principle, connecting dense grids and corrected sparse grids to form a maximum connected region, wherein the connected dense grids and corrected sparse grids are transient thermal response clusters, and the CLIQUE clustering algorithm adaptively identifies the number of dense grid clusters in a thermal response space through the steps of adaptively identifying the number of cluster clusters in the transient thermal response data space, so that the step of judging the defect category number is omitted; sequence S of thermal images ⁱ Transient thermal response set X of (2) ⁱ (g) Adaptive clustering to form cluster sets _X(g) Cluster[h]H=1, 2, …, H, where H represents a defect class label and H represents the total number of classes of complex type defects present in the current infrared detection region;

wherein the method comprises the steps of

For the h clustering result _X(g) Cluster[h]H=1, 2, …, the kth transient thermal response in H, | _X(g) Cluster[h]The I is the total number of transient thermal responses contained in the h clustering result, and a matrix Y is formed by using representative transient thermal responses of various types of defects ⁱ ；

Wherein,,

is H×T matrix, which is representative transient thermal response matrix Y ⁱ Pseudo-inverse matrix of (O) ⁱ ) ^T Is a two-dimensional image matrix O ⁱ The obtained reconstruction matrix is H rows and M multiplied by N columns; intercepting reconstruction matrix R ⁱ An MxN two-dimensional image is formed, H MxN two-dimensional images are obtained, the images are the reconstructed thermal images containing the characteristic information of different thermal response areas in the thermal image sequence obtained by the ith infrared detection, and the reconstructed thermal images of the non-defective background areas are recorded as _B R, recording the reconstructed thermal image corresponding to each type of defect area as _h R，Wherein h=1, H-1, each reconstructed thermal image contains characteristic thermal reconstruction information of one type of defect of the complex type defects in the current detection area except for the background area thermal image without defect damage, and the typical type defect reconstruction thermal image in the detection area obtained in the ith infrared detection is recorded as _Def.(i) R；

4. The method for analyzing the multi-type damage detection characteristics of the large-size test piece according to claim 2, wherein the specific method for decomposing the reconstructed image of each defect area of the large-size test piece into the basic infrared thermal image and the detail layer infrared thermal image is as follows: for (|C' | -1) sheet of infrared reconstructed image { other than the open background region thermal image ₁ R,…, _i R,…, _|C′|-1 Each R is subjected to image decomposition, and each reconstructed image is decomposed into a base layer infrared thermal image { ₁ B,…, _i B,…, _|C′|-1 B and a detail layer infrared thermal image { ₁ D,…, _i D,…, _|C′|-1 D }; reconstructing thermal images with ith defective areas _i R is exemplified by i=1, …, |c' | -1, obtained using the following formula _i R base layer infrared thermal image _i B and detail layer IR thermal image _i D：

_i B＝ _i R*Z

_i D＝ _i R- _i B

Wherein Z is an average filter.

5. The method for analyzing the characteristics of multi-type damage detection of the large-size test piece according to claim 3, wherein the specific method for decomposing the infrared thermal reconstruction image of each defect area of the large-size test piece into the basic layer infrared thermal image and the detail layer infrared thermal image comprises the following steps: an infrared reconstructed image { of a total of |C| typical type defects in each detection region in a large-size impact test piece _Def.(1) R,..., _Def.(i) R,..., _Def.(|C|) Each of R is subjected to image decomposition, and each reconstructed image is decomposed into a base layer infrared thermal image { inf.base [ def. (1)],...,Inf.Base[Def.(i)],...,Inf.Base[Def.(|C|)]And a detail layer infrared thermal image { Inf. Detail [ def. (1)],...,Inf.Detail[Def.(i)],...,Inf.Detail[Def.(|C|)]}；

Inf.Base[Def.(i)]＝ _Def.(i) R*Z

Inf.Detail[Def.(i)]＝ _Def.(i) R-Inf.Base[Def.(i)]

Wherein Z is an average filter.

6. The method for analyzing the multi-type damage detection characteristics of the large-size test piece according to claim 4, wherein in the third step, the corresponding infrared thermal images { of the base layers are respectively obtained by utilizing multi-objective optimized guiding filtering ₁ B, ₂ B,…, _|C′|-1 B }, an infrared thermal amplitude fusion weighting graph { between ₁ W ^B , ₂ W ^B ,…, _|C′|-1 W ^B And detail layer infrared thermal image { ₁ D, ₂ D,…, _|C′|-1 D, fusing weight map { of infrared thermal amplitude values between the two layers ₁ W ^D , ₂ W ^D ,…, _|C′|-1 W ^D The specific method is as follows:

_i H＝ _i R*L

_i S＝| _i H|*GF

_i O _n ＝a _k · _i R _n +b _k

Wherein,,

and->

Is an edge-aware weighting factor defined as follows:

wherein,,

wherein,,

Is the Hadamard product of the matrix, +.>

And->

Wherein,,

and->

wherein eta is

Representing an infrared reconstructed image _i In R, to _i R _k Infrared heat amplitude standard deviation corresponding to each coordinate point in 3 x 3 window with coordinate point as center, n E I x J,/I->

guided filtering cost function by minimizing gradient domain

Obtain->

And->

The calculation formula of (2) is as follows:

wherein,,

Wherein,,

and->

The local LoG edge weighting factor is defined as follows:

wherein, loG (·) is a Gaussian Laplace edge detection operator, I×J is the total coordinate point number of the infrared reconstruction image, and |·| is the absolute value operation, delta _LoG 0.1 times the maximum value of the image;

guided filtering cost function by minimizing gradient domain

Obtain->

And->

The calculation formula of (2) is as follows:

wherein the method comprises the steps of

And->

N _p Wherein l=3 is the total number of the infrared thermal image fusion cost functions;

initializing a reference point _i r＝{ _i r ₁ ,…, _i r ₃ }，

step S342, utilize

Decomposing the original multi-objective problem into a series of scalar sub-objective problems using chebyshev decomposition method, using the weights vectors corresponding to the respective weights +. >

Is>

wherein the method comprises the steps of

And->

step S345, evolution termination judgment: if g 'is less than or equal to g' _max Repeating steps S343-344, if g '> g' _max Obtaining the final front approximate solution set of the multi-target guided filtering linear parameter _i AP；

The calculation formula is as follows:

wherein,,

representing an infrared reconstructed image _i R rectangular window w _k Infrared thermal amplitude mean value corresponding to each coordinate point in the infrared thermal power meter，

step S37, optimal linear transformation coefficient based on Pareto

And->

step S38, utilizing the optimal guiding filter operator MOG obtained by multi-objective optimizationF _r,ε (P, R) performing multi-objective guided filtering on the obtained thermal amplitude fusion coarse weight graph to obtain corrected infrared thermal amplitude fusion weight images of the base layer and the detail layer:

_i W ^B ＝MOGF _r1,ε1 ( _i P, _i R),(i＝1,…,|C′|-1)

_i W ^D ＝MOGF _r2,ε2 ( _i P, _i R),(i＝1,…,|C′|-1)

7. The large-size specimen multi-type lesion detection feature analysis method according to claim 5, characterized in that the step three obtains the corresponding base layer infrared thermal image { inf.base [ def. (1) ], inf.base [ def. (i) ], inf. (c|) the thermal amplitude fusion weight map { wm base [ def. (1) ], wm base [ def. (i) ], wm.c.) ] and detail layer infrared thermal image { inf [ def. (1) ], inf.detail [ def. (i) ], inf. (C.) }) the thermal amplitude fusion weight map { wm [ def. (1.) ] and the detail layer infrared thermal image { inf [ def. (i.) ], wm.base [ def. (C.) ] the method comprises }).

_Def.(i) H＝ _Def.(i) R*L

_Def.(i) S＝| _Def.(i) H|*GF

_Def.(i) P＝{ _Def.(i) P ₁ ,…, _Def.(i) P _k ，…, _Def.(i) P _M×N }

_Def.(i) O _n ＝a _k · _Def.(i) R _n +b _k

wherein,, _Def.(i) O _n representing images reconstructed by infrared _Def.(i) R is a guiding image, and a thermal amplitude value is used for fusing a rough weight map _Def.(i) P is typical of the ith detection area obtained by multi-objective guided filtering of the input imageType defect output image _Def.(i) The guided filter output value corresponding to the nth coordinate point of O, n=1,..m×n; _Def.(i) R _n is that _Def.(i) The reconstructed image thermal amplitude corresponding to the nth coordinate point of R, n=1,..m×n; a, a _k And b _k Expressed in terms of _Def.(i) R _k Centered guided filter window w _k Linear transformation parameters in, k=1,..m×n;

step S331, fusing coarse weight graphs based on thermal amplitude values _Def.(i) P and infrared reconstructed images _Def.(i) R, defining an infrared large-size defect edge characteristic perception weighting guide filtering cost function at each coordinate point position

Wherein,,

and->

The optimal linear transformation coefficient is determined by a large-size defect perception filtering cost function; _Def.(i) P _n is a weight graph _Def.(i) The heat radiation fusion weight value corresponding to the nth coordinate point of P; epsilon is a regularization factor; Γ -shaped structure _(Def.(i)Rk) Is an edge-aware weighting factor defined as follows:

wherein,,

wherein,,

Is the Hadamard product of the matrix, +.>

And

Wherein,,

and->

wherein eta is

Representing an infrared reconstructed image _Def.(i) In R, to _Def.(i) R _k Guide filtering rectangular window w with coordinate point as center _n The standard deviation of the thermal amplitude corresponding to each coordinate point in the range, wherein N is E M multiplied by N; />

Guided filtering cost function by minimizing gradient domain

Obtain->

And->

The calculation formula of (2) is as follows:

wherein,,

Wherein,,

and->

The local LoG edge weighting factor is defined as follows:

wherein, loG (.cndot.) is Gaussian praat The edge detection operator, M multiplied by N is the total coordinate point number of the infrared reconstruction image, and I/I is the absolute value operation, delta _LoG 0.1 times the maximum value of the LoG image;

guided filtering cost function by minimizing gradient domain

Obtain->

And->

The calculation formula of (2) is as follows:

wherein the method comprises the steps of

And->

wherein a is _k ' is the kth guided filter window w _k Is a wire in (a)The coefficient of the transformation is changed such that, _Inf.Sig E ₁ (a _k ') preserving a fusion cost function for the large-size defect edges of the infrared thermal image with obvious gradient change, _Inf.Min E ₂ (a _k ' reserving a fusion cost function for detail textures of tiny defects of infrared thermal images with insignificant size and gradient changes, E ₃ (a _k ' is an infrared thermal image noise information sensing and eliminating cost function;

Wherein l=3 is the total number of the fusion cost functions of the infrared thermal images; initializing reference point- >

step S342, utilize

step S343, for each single target sub-problem after decomposition, based on their correspondingWeight vector of (2)

wherein the method comprises the steps of

And->

step S345, evolution termination judgment: if g' is less than or equal tog' _max Repeating steps S343 to S344, and if g '> g' _max Obtaining the final front approximate solution set of the multi-target guided filtering linear parameter _i AP；

Calculating the multi-objective guide filtering another optimal linear transformation coefficient of the i Zhang Gongwai hot amplitude fused coarse weight image +.>

The calculation formula is as follows:

wherein,,

representing an infrared reconstructed image _Def.(i) R rectangular window w _k The mean value of the infrared heat amplitude corresponding to each coordinate point in the infrared heat radiation system,

representing a coarse weight map _Def.(i) P is in rectangular window w _k The average value of the infrared heat amplitude corresponding to each coordinate point in the infrared heat source; />

Step S37, optimal linear transformation coefficient based on Pareto

And->

Performing multi-objective guide filtering on the thermal amplitude fusion coarse weight map of the i-th infrared detection region infrared thermal reconstruction image to obtain a corrected baseThermal amplitude fusion weight image of layer and detail layer:

8. The method for analyzing the multi-type damage detection characteristics of the large-size test piece according to claim 6, wherein the specific method in the fourth step is as follows: fusing weight map { in the obtained fine detail layer thermal amplitude value ₁ W ^D , ₂ W ^D ,…, _|C′|-1 W ^D Weight map { fused with base layer infrared thermal amplitude values } ₁ W ^B , ₂ W ^B ,…, _|C′|-1 W ^B Fusing the detail layer infrared thermal image information and the base layer infrared thermal image information among the thermal reconstruction images of different defect areas except the background area to obtain a base layer infrared thermal image and a detail layer infrared thermal image fused with the effective information of a plurality of reconstruction thermal images:

9. The method for analyzing the multi-type damage detection characteristics of the large-size test piece according to claim 7, wherein the specific method in the fourth step is as follows:

based on the obtained detail layer thermal amplitude fusion weight map { WM.Detail [ Def (1) ], & gt, WM.Detail [ Def (i) ], & gt, WM.Detail [ Def (C|)) ] }, and the obtained base layer thermal amplitude fusion weight map { WM.Base [ Def (1) ], & gt, WM.Base [ Def (i) ], & gt, WM.Bas e [ Def ] (|C|)) ] }, fusing the detail layer thermal image information and the base layer thermal image information between the typical type defect thermal reconstruction images of different areas in the large-size test piece to obtain a base layer thermal image and a detail layer thermal image fused with effective information of a plurality of multi-detection area reconstruction thermal images.