Infrared thermal image processing method and device and storage medium

Abstract

The invention relates to the technical field of infrared thermal imaging defect detection, and discloses an infrared thermal image processing method and device and a storage medium, and the processing method comprises the steps: carrying out the line-by-line detection of local minimum points of an original infrared thermal image to recognize all temperature troughs, and forming a coarse screening point set Scoarse; performing multi-iteration outlier screening on the coarse screening point set Scoarse to obtain a refined point set Sfine; fitting the refined point set Sfine to generate a non-uniform background of the infrared thermal image, and subtracting the non-uniform background from the original infrared thermal image; the original infrared thermal image without the non-uniform background is converted into a one-dimensional signal, the one-dimensional signal is decomposed into a plurality of empirical mode functions and a residual component through empirical wavelet transform, the signal is reconstructed after high and low frequency components and a residual are removed, and an enhanced infrared thermal image is obtained; according to the method, the quality of the infrared thermal image can be enhanced while the non-uniform background noise is removed.

Description

Technical Field

[0001] The present invention relates to the technical field of infrared thermal imaging defect detection, and particularly to a method, device and storage medium for processing infrared thermal images. Background Art

[0002] With the development of the modern industrial system towards high quality, the requirements for the material properties of key components in engineering are increasing day by day. Composite materials, with their excellent corrosion resistance, outstanding high-temperature stability and significant lightweight and other multi-dimensional performance advantages, have gradually replaced traditional metal alloys in many technical fields. Among them, high-performance composite materials represented by fiber reinforcement systems are the most widely used in various fields. Due to the complex preparation process and harsh application scenarios of fiber composite materials, they are prone to repeated stress and impact during equipment operation. The above problems lead to the easy occurrence of defects inside the materials. Developing non-destructive testing technologies with real-time monitoring capabilities to achieve online evaluation and early warning of the structural state of composite materials has become a key technical requirement for ensuring the safe operation of major equipment and is of great significance to actual production; in addition, many important structures in the production and equipment operation processes often have problems such as cracks and damages that are difficult to estimate. These defects will significantly reduce the strength and performance of the structure under load, and ultimately lead to structural damage and even serious safety accidents. Therefore, non-destructive testing technologies emerge as an important means to ensure the safe operation and reliability of various equipment and products.

[0003] Infrared thermal imaging technology has been applied in various industries due to its own unique advantages. This technology heats an object through external means such as optical excitation and eddy current excitation. The existence of defects hinders the heat propagation. By collecting and analyzing the thermal image sequence of the object surface, it is possible to detect whether there are defects such as cracks, delamination and corrosion on the object. Compared with traditional non-destructive testing methods, infrared thermal imaging technology is particularly suitable for defect detection of product structures. When performing infrared thermal imaging non-destructive testing, the experimental environment is non-ideal. Due to problems such as the layout of the excitation source, instrument error and precision limitation, the obtained infrared images have uneven temperature field distribution and noise. And pulsed thermography (long-pulse thermography) sometimes continues to excite, and its heating non-uniformity will continuously act on the thermal image, further increasing adverse factors. These adverse factors will bury the subtle temperature characteristics of the defect area and the non-defect area, posing challenges to defect detection. Therefore, post-processing algorithms are needed to remove these adverse factors and improve the defect detection level of thermal imaging. Summary of the Invention

[0004] The purpose of the present invention is to overcome the problem that existing infrared thermal images have uneven temperature field distribution and noise, resulting in poor defect detection capability, and to provide an infrared thermal image processing method. This processing method can effectively remove the non-uniform temperature field and noise on the infrared thermal image, improve the quality of the infrared thermal image, and thus improve the defect detection capability.

[0005] In order to achieve the above object, the present invention provides a method for processing infrared thermal images, comprising the following steps:

[0006] S1. For the original infrared thermal image collected by the infrared thermal imager, all temperature troughs are identified by detecting local minimum points line by line to form a coarse screening point set S coarse ;

[0007] S2, based on the Laida criterion to roughly screen the point set S coarse Perform multiple iterations of outlier screening to obtain a refined point set S fine ;

[0008] S3, using the polynomial model to refine the point set S fine Performing fitting to generate a non-uniform background of the infrared thermal image, and subtracting the non-uniform background from the original infrared thermal image;

[0009] S4. The original infrared thermal image with the non-uniform background removed is converted into a one-dimensional signal, decomposed into multiple empirical mode functions and a residual component through empirical wavelet transform, and the signal is reconstructed after removing the high- and low-frequency components and the residual to obtain the enhanced infrared thermal image.

[0010] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned processing method when executing the computer program.

[0011] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above-mentioned processing method when executed by a processor.

[0012] In the technical solution provided by the present invention, the interference of most defective pixels can be simply and effectively eliminated by finding the pixel line valley method. Furthermore, non-defective pixels can be screened based on the Laida criterion, thereby improving the accuracy of non-uniform background fitting, retaining defect features while removing the non-uniformity of infrared thermal images. In addition, by setting the iterative termination condition, multiple iterations are prevented from falling into the local optimal solution, resulting in missing pixels and missing fitting items.

[0013] The present invention uses polynomial fitting as the background temperature fitting model, which conforms to the temperature distribution characteristics on the infrared thermal image; the infrared thermal image multi-modal transformation algorithm provided by the present invention can decompose the infrared thermal image into multiple modes, realize the arrangement of image components from high frequency to low frequency, and facilitate the removal of useless high and low frequency noise components.

[0014] In summary, the present invention combines the infrared thermal image non-uniform background removal algorithm with the multi-modal transformation algorithm, which can significantly enhance the quality of the infrared thermal image while removing the non-uniform background noise. Description of the Drawings

[0015] Figure 1 is a flowchart of an infrared thermal image processing method provided by the present invention;

[0016] Figure 2 is a schematic diagram of a long-pulse thermal imaging defect detection experimental system provided by an embodiment of the present invention;

[0017] Figure 3 is a dimensional drawing of a test sample plate provided by an embodiment of the present invention;

[0018] Figure 4 is a comparison diagram before and after removing the non-uniform background based on the original data provided by an embodiment of the present invention;

[0019] Figure 5 is a comparison diagram of the 3D contour effect before and after removing the non-uniform background provided by an embodiment of the present invention;

[0020] Figure 6 is a comparison of certain specific row and column pixels before and after removing the non-uniform background provided by an embodiment of the present invention;

[0021] Figure 7 is a schematic diagram of the result after 14-frequency division using the empirical wavelet transform algorithm provided by an embodiment of the present invention;

[0022] Figure 8 is a comparison of the images before and after the method of removing thermal imaging non-uniformity and enhancing defect features by modal decomposition provided by an embodiment of the present invention.

[0023] Description of the Reference Numerals

[0024] 1. Excitation source controller; 2. Halogen lamp excitation source; 3. Infrared thermal imager; 4. Residual heat shielding plate; 5. Carbon fiber composite material test sample plate; 6. Non-defect area of the test sample plate; 7. Defect area of the test sample plate; 8. Data acquisition host computer. Detailed Embodiments

[0025] In order to make the technical means, creative features, achieved objectives and effects of the present invention easy to understand, the present invention will be further clarified below in conjunction with specific embodiments and drawings. It should be understood that the specific embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0026] As Figure 1 shown, the present invention provides a method for processing infrared thermal images, including the following steps:

[0027] S1. For the original infrared thermal image collected by the infrared thermal imager, all temperature troughs are identified by detecting local minimum points row by row to form a coarse screening point set Scoarse;

[0028] S2. Based on the Grubbs' criterion, perform multiple iterations of outlier screening on the coarse screening point set Scoarse to obtain a refined point set Sfine;

[0029] S3. Use a polynomial model to fit the refined point set Sfine to generate a non-uniform background of the infrared thermal image, and subtract the non-uniform background from the original infrared thermal image;

[0030] S4. Convert the original infrared thermal image with the non-uniform background removed into a one-dimensional signal, decompose it into multiple empirical mode functions and a residual component through empirical wavelet transform, remove the high and low frequency components and the residual, and then reconstruct the signal to obtain an enhanced infrared thermal image.

[0031] It should be noted that in the processing method provided by the present invention, the image collected by the infrared thermal imager needs to meet the following conditions:

[0032] 1. The surface thermal absorptivity and shape of the test sample plate material cannot change suddenly;

[0033] 2. The entire field of view of the infrared thermal imager needs to be entirely composed of the test sample plate, without including other areas. If there are inclusions, the image needs to be cropped.

[0034] In step S2, after least squares fitting of the coarse screening point set S coarse new outliers are screened, and multiple iterations are performed to achieve fine screening of non-defective points, thereby obtaining the refined point set S fine ; among them, the number of iterations is controlled by setting the maximum number of iterations, the minimum number of remaining pixel points, and the minimum number of outliers eliminated at one time during the iteration. Through the above settings, different data can be adapted.

[0035] In step S3, the expression of the polynomial model is:

[0036] T(p) = a0 + a1p + a2p 2 +…+ a np n

[0037] Among them, T(p) is the data temperature fitting background after refined screening, p is the coordinate value, and a0 to a n are fitting coefficients.

[0038] In step S4, the specific steps of the empirical wavelet transform include:

[0039] S41. Perform a Fourier transform on the one-dimensional signal and extract the unilateral spectrum. After normalization, it is divided into multiple intervals. Specifically, the unilateral spectrum is normalized to (0, π) and divided into N intervals, where the value range of N is 6-14, that is, the number of intervals can be 6, 7, 8, 9, 10, 11, 12, 13, or 14; each interval is represented as Λ k , which is expressed as follows:

[0040]

[0041] Among them, ω k-1 and ω k are the left and right boundaries of the spectrum respectively.

[0042] S42. Define the empirical scaling function φ k (ω) and the empirical wavelet function ψ k (ω) in each interval;

[0043] Among them, the construction formula of the empirical scaling function φ k (ω) is:

[0044]

[0045] The construction formula of the empirical wavelet function ψ k (ω) is:

[0046]

[0047] Among them, ω k is the left boundary of the constructed spectrum, ω k+1 is the right boundary of the constructed spectrum, ω is the current spectrum; β(x) = x 4 (35 - 84x + 70x 2 - 20x 3 ); γ satisfies the following conditions:

[0048]

[0049] Among them, γ controls the frequency support of φ k (ω) and ψ k (ω), and forms a compact support in the L 2 space.

[0050] S43. Calculate the detail coefficients through inner product and the approximation coefficients Decompose to obtain multiple empirical mode functions and a residual component.

[0051] It should be noted that its decomposition process is similar to the traditional wavelet transform:

[0052]

[0053] where k = 1, 2, …, N; is the detail coefficient of the k-th segment of the spectrum; is the approximation coefficient of the residual spectrum; ψ k is the empirical wavelet function of the k-th segment of the spectrum; φ1 is the empirical scale function of the first segment of the spectrum; x(t) is the original signal.

[0054] In the present invention, the empirical wavelet transform algorithm can decompose a one-dimensional signal into a series of empirical mode functions (EMF) m k (t) (k = 1, 2, …, N) and a residual component m0(t). Arrange them from high frequency to low frequency, and use them to reconstruct the signal into x(t):

[0055]

[0056] Furthermore, remove the unnecessary EMF and residual components according to the component thermal images, and then synthesize and reconstruct them. Here, the specific expressions of the EMFs and the residual component are:

[0057]

[0058] The following further illustrates the method for processing infrared thermal images provided by the present invention through specific embodiments.

[0059] Such as Figure 2The following shows a long-pulse thermal imaging defect detection experimental system provided by an embodiment of the present invention. The experimental system includes an excitation source controller 1, two halogen lamp excitation sources 2, an infrared thermal imager 3, a residual heat shielding plate 4, a carbon fiber composite test plate 5, and a data acquisition host computer 8. Among them, the model of the infrared thermal imager 3 is FLIR A40M, the spectral wavelength range it can detect is 7.5 - 13 μm, it can collect infrared thermal images with a resolution of 240×320 at a frame rate of 50 Hz, and its thermal sensitivity at a room temperature of 30°C is 0.08°C. The models of the two halogen lamps (halogen lamp excitation sources 2) are both Philips QVF137, their input voltage is 220V, and the single power is 1000W. The heating duration is controlled by the excitation source controller 1. The residual heat shielding plate 4 is composed of two wooden thin plates. One side surface of the carbon fiber composite test plate 5 has multiple subsurface holes to represent actual defects (where 6 indicates the non-defect area of the test plate, and 7 indicates the defect area of the test plate). The other side surface is sprayed with acrylic matte black paint to increase the surface emissivity and absorptivity.

[0060] For the specific dimensions of the carbon fiber composite test plate 5 and the corresponding defect positions, please refer to Figure 3 ; The positions and dimensions of the defects are marked in Figure 3 , with a total of 12 defects (#1 - #12).

[0061] In the present invention, the best display effect can be achieved by appropriately adjusting the focal length and viewing angle of the infrared thermal imager. After the adjustment is completed, control the halogen lamp to continuously heat the painted side surface of the test plate for 8 s. Immediately after the end, place the waste heat shielding baffle in front of the halogen lamp, and use the infrared thermal imager to record the entire process data of heating for 8 s and cooling for 8 s. Combining with Figure 1 shown below, the collected data is processed according to the following steps:

[0062] S1. Perform a line-by-line scan on the original infrared thermal images collected by the infrared thermal imager, and identify all temperature valleys by detecting local minimum points to form a rough screening point set S coarse ; Specifically, process the 16 s of original infrared thermal image data collected above. The temperature valley means a lower temperature, and all temperature valley points are used as the rough screening point set S coarse .

[0063] S2. Based on the Laida criterion, perform multiple iterations of outlier screening on the rough screening point set S coarse to obtain a refined point set S fine .

[0064] S3. Use a polynomial model to fit the refined point set S fine to generate the non-uniform background of the infrared thermal image, and subtract the non-uniform background from the original infrared thermal image.

[0065] Refer to Figure 4 and Figure 5 After going through the above steps, based on the two-dimensional and three-dimensional thermal image results before and after non-uniform background removal using the 3σ criterion (here, the pixel values of the image have been mapped to the range of 0 to 1); among them, Figure 4 on the left side in Figure 5 is the original image, and on the right side is the image after being processed through steps S1 - S3;

[0066] The background non-uniformity of the original infrared thermal image is relatively complex, and the three-dimensional surface contour is approximately an irregular paraboloid, resulting in the temperature of some non-defect regions being higher than that of some defect regions. After using the processing method provided by the present invention for background non-uniformity noise correction, the temperature distribution of the non-defect regions in the thermal image becomes uniform, the image becomes pure, and the background region of the three-dimensional contour map also becomes significantly flat.

[0067] As Figure 6 shown is the data comparison of several randomly selected rows and columns of pixels before and after processing, where (a) is the 33rd row of pixels, (b) is the 101st row of pixels, (c) is the 261st column of pixels, and (d) is the 231st column of pixels. Among them, (a) and (c) are non-defect regions, and it can be seen that the pixel values of green (after background removal) are basically restored to a unified baseline compared to red (before background removal); (b) and (d) are defect regions, and the protruding parts are defect features, and it can be seen that the pixel values of green (after background removal) not only return to the unified baseline but also retain the defect features compared to red (before background removal). These detailed comparisons prove the effectiveness of the processing method provided by the present invention in removing the non-uniform background. Whether in the regions containing defects or pure background regions, the processing method provided by the present invention can remove the background while retaining the distribution characteristics of the defect regions.

[0068] S4. After step S3, convert the original infrared thermal image with non-uniform background removed into a one-dimensional signal, decompose it into multiple empirical mode functions and a residual component through empirical wavelet transform, remove the high and low frequency components and the residual, and then reconstruct the signal to obtain an enhanced infrared thermal image.

[0069] It can be understood that the empirical wavelet transform algorithm can decompose a one-dimensional signal into a series of empirical mode function (EMF) components m k (t) (k = 1, 2, …, N) and a residual component m0(t), arrange them from high frequency to low frequency, remove the unnecessary EMF and residual components according to the component thermal images, and then reconstruct the signal to obtain an enhanced infrared thermal image.

[0070] As Figure 7The figure shows a schematic diagram of the result after 14 - frequency division using the empirical wavelet transform algorithm provided by the embodiment of the present invention. Figure 7 The result shown is in step S4. During the empirical wavelet transform process, the one - dimensional signal is subjected to Fourier transform and the unilateral spectrum is extracted. After normalization, it is divided into 14 intervals; from Figure 7 the decomposition result, it can be clearly seen that: the first component is mainly composed of high - frequency noise, and the subsequent components are increasingly dominated by the low - frequency background. Here, the first component and the last two components (including the residual) are removed and the data is reconstructed.

[0071] The reconstruction result is as shown in Figure 8 the right - hand side (b) in Figure 8 the left - hand side (a) in is the original image), from Figure 8 it can be seen the effect of the infrared thermal image reconstructed based on the processing method provided by the present invention. The picture is pure, and the defects are relatively obvious (especially the defects that are difficult to detect). All 12 defects of the entire test sample plate can be detected, forming a good contrast with the original thermal image.

[0072] The embodiment of the present invention also provides a computer device, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps of the above - mentioned infrared thermal image processing method are implemented.

[0073] In the present invention, the processor contains a kernel, and the kernel retrieves the corresponding program unit from the memory. One or more kernels can be set.

[0074] The memory may include non - permanent memory in a computer - readable storage medium, random access memory (RAM) and / or non - volatile memory in the form of, for example, read - only memory (ROM) or flash memory (flash RAM). The memory includes at least one storage chip.

[0075] The embodiment of the present invention also provides a computer - readable storage medium, on which a computer program is stored. When the computer program is executed by the processor, the steps of the above - mentioned infrared thermal image processing method are implemented.

[0076] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer - usable storage media (including but not limited to disk storage, CD - ROM, optical storage, etc.) containing computer - usable program code.

[0077] The present invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each flow and / or block of the flowchart illustrations and / or block diagrams, and combinations of flows and / or blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to the processors of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions executed by the processors of the computer or other programmable data processing device create means for implementing the functions specified in the flowchart Figure 1 one flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0078] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable memory produce a manufacture including instruction means that implement the functions specified in the flowchart Figure 1 one flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0079] These computer program instructions may also be loaded onto a computer or other programmable data processing device, such that a series of operational steps are performed on the computer or other programmable device to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in the flowchart Figure 1 one flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in multiple blocks.

[0080] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and memory.

[0081] The memory may include non-permanent memory in the form of a computer-readable storage medium, random access memory (RAM), and / or non-volatile memory such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of a computer-readable storage medium.

[0082] A computer-readable storage medium includes permanent and non-permanent, removable and non-removable media that can implement information storage by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transitory medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer-readable storage medium does not include transitory computer-readable media, such as modulated data signals and carrier waves.

[0083] The foregoing has shown and described the basic principles, main features, and characteristics of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and what is described in the above embodiments and the specification is only to illustrate the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of the present invention claimed is defined by the appended claims and their equivalents.

Claims

1. A method for processing infrared thermal images, characterized in that, Including the following steps: S1. For the original infrared thermal image collected by the infrared thermal imager, all temperature troughs are identified by detecting local minimum points line by line to form a rough screening point set S coarse ; S2. Perform outlier screening on the roughly screened point set S based on the Pauta criterion coarse for multiple iterations to obtain the refined point set S fine ; S3. Use a polynomial model to fit the refined point set S fine to generate a non-uniform background of the infrared thermal image, and subtract the non-uniform background from the original infrared thermal image; S4. Convert the original infrared thermal image with non-uniform background removed into a one-dimensional signal, decompose it into multiple empirical mode functions and a residual component through empirical wavelet transform, remove the high and low frequency components and the residual, and then reconstruct the signal to obtain the enhanced infrared thermal image.

2. The processing method according to claim 1, characterized in that In step S2, the rough screening point set S coarse is subjected to least-squares fitting to screen new outliers, and after multiple iterations, fine screening of non-defective points is achieved, thereby obtaining the refined point set S fine ; Among them, the number of iterations is controlled by setting the maximum number of iterations, the minimum number of remaining pixel points, and the minimum number of outliers eliminated at one time during the iteration.

3. The processing method according to claim 1, characterized in that In step S3, the expression of the polynomial model is: T(p) = a0 + a1p + a2p 2 +…+ a n p n Among them, T(p) is the data temperature fitting background after refined screening, p is the coordinate value, and a0 to a n are fitting coefficients.

4. The processing method according to claim 1, characterized in that, In step S4, the specific steps of the empirical wavelet transform include: S41. Perform Fourier transform on the one-dimensional signal and extract the single-sided spectrum. After normalization, divide it into multiple intervals; S42. Define the empirical scaling function φ k (ω) and the empirical wavelet function ψ k (ω); S43. Calculate the detail coefficients through inner product and the approximation coefficients Decompose to obtain multiple empirical mode functions and a residual component.

5. The processing method according to claim 4, characterized in that In step S42, the construction formula of the empirical scaling function φ k (ω) is: Empirical wavelet function ψ k (ω) has the following construction formula: Among them, ω k is the left boundary of the constructed spectrum, ω k+1 is the right boundary of the constructed spectrum, ω is the current spectrum; β(x) = x 4 (35 - 84x + 70x 2 - 20x 3 ); γ satisfies the following conditions: Among them, γ controls φ k (ω) and ψ k (ω)'s frequency support, forms a compact support in the L 2 space.

6. The processing method according to claim 4, wherein, In step S43, the decomposition process of the detail coefficient and the approximation coefficient is as follows: where k = 1, 2, …, N; is the detail coefficient of the k-th segment of the spectrum; is the approximation coefficient of the residual spectrum; ψ k is the empirical wavelet function of the k-th segment of the spectrum; φ1 is the empirical scaling function of the first segment of the spectrum; x(t) is the original signal.

7. A computer device, comprising a memory and a processor, wherein a computer program is stored in the memory, characterized in that, When the processor executes the computer program, it implements the steps of the processing method described in any one of claims 1-6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the processing method described in any one of claims 1-6.

Images