Optical thin film quality detection method based on image processing

Abstract

The invention relates to the technical field of image processing, in particular to an optical thin film quality detection method based on image processing, which comprises the following steps: S101, acquiring a transient light field image sequence of an optical thin film in a dynamic stress field; s102, performing spectral analysis on the brightness time sequence of each pixel point, extracting a dominant frequency component as a brightness oscillation period, processing time domain fluctuation data by adopting a Fourier transform algorithm, and calculating a local optical path difference variation based on the brightness oscillation period of each pixel point in the image sequence; and S103, projecting the phase distortion field to a preset film elastic modulus distribution model, establishing a phase distortion-stress distribution-defect mapping relation, and identifying a modulus mutation region. A dynamic stress field is generated through the synergistic effect of mechanical vibration and thermal radiation, remarkable stress concentration can be generated at the micro-defect position, the deformation difference of the micro-defect is amplified, and therefore the sensitivity of defect detection is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, and particularly relates to an optical film quality detection method based on image processing. BACKGROUND

[0002] As a core component of modern optical systems, optical films are widely used in high-end fields such as laser technology, optical communication, optoelectronic display, aerospace, etc. The quality of the optical film directly affects the performance of the optical system. For example, the thickness uniformity, internal stress distribution, surface flatness and the presence of micro-defects of the film will have a significant impact on the transmission, reflection, refraction and other characteristics of light. With the development of optical technology towards high precision and high integration, higher requirements are put forward for the precision, sensitivity and real-time performance of optical film quality detection.

[0003] Traditional optical film quality detection methods mainly include static detection and destructive detection. Static detection methods such as scanning electron microscopy and atomic force microscopy have high spatial resolution, but have limitations such as slow detection speed, high cost and only obtaining static surface information, which cannot effectively evaluate the internal stress distribution and dynamic deformation of the film. Destructive detection methods such as nanoindentation can obtain material mechanical property parameters, but will cause irreversible damage to the film, which is not suitable for finished product detection and online quality control.

[0004] To solve the shortcomings of traditional methods, non-destructive detection techniques based on optical interference principles have gradually developed, such as white light interferometer and laser interferometer. These techniques measure the optical path difference to reflect the thickness and surface topography of the film, but still have difficulties in detecting micro-defects below microns and internal stress distribution of the film. The main reasons are as follows: on the one hand, the change of optical path difference caused by micro-defects is extremely weak and easy to be overwhelmed by noise; on the other hand, static detection cannot reflect the dynamic response of the film under actual working conditions, making it difficult to find potential defect hazards. SUMMARY

[0005] The present application provides an optical film quality detection method based on image processing to solve the technical problems in the prior art.

[0006] The technical solution of the present application to solve the above technical problems is as follows: an optical film quality detection method based on image processing, comprising the following steps:

[0007] S101, acquiring a transient light field image sequence of an optical film in a dynamic stress field;

[0008] S102, performing spectral analysis on the brightness time sequence of each pixel point, extracting the main frequency component as the brightness oscillation period, using Fourier transform algorithm to process the time domain fluctuation data, based on the brightness oscillation period of each pixel point in the image sequence, calculating the local optical path difference change;

[0009] S103, projecting the phase distortion field to the preset thin film elastic modulus distribution model, establishing the mapping relationship of phase distortion-stress distribution-defects, and identifying the modulus mutation area.

[0010] In a preferred embodiment, the dynamic stress field is generated by applying a frequency-adjustable first mechanical vibration source and a second thermal radiation source, the first mechanical vibration source excites resonant state elastic waves in the thin film, which can cause stress concentration effect at the micro-defects of the optical thin film, and the second thermal radiation source forms a linear expansion gradient on the surface of the thin film, which is orthogonal to the stress coupling formed by the mechanical vibration;

[0011] A high-speed camera is used to image the thin film in the dynamic stress field in real time. During the acquisition process, the timing consistency of stress field loading and image acquisition is realized through a synchronous triggering mechanism. When the first mechanical vibration source and the second thermal radiation source are started, the camera starts continuous image acquisition, ensuring that the stress state corresponding to each frame of image can be traced back. The obtained optical thin film image sequence contains the change information of light field intensity with time, wherein the brightness oscillation period of each pixel point is directly related to the local optical path difference change. The acquired optical thin film image is preprocessed, including denoising, brightness equalization and sub-pixel registration.

[0012] In a preferred embodiment, the preprocessed optical thin film image is used to extract image feature points based on feature point matching image registration algorithm, and then RANSAC transformation estimation algorithm is used to eliminate the global displacement caused by mechanical vibration between adjacent frames, ensuring that the same physical position corresponds to the same pixel coordinates in different frames. The brightness value I(x,y,t) of each pixel point (x,y) at different time t is extracted to form a brightness time sequence:

[0013] St={I(x,y,1),I(x,y,2),...,I(x,y,N)}

[0014] Wherein, the total frame number N needs to satisfy N≥1000 to ensure the reliability of spectral analysis. Fast Fourier transform is performed on each brightness time sequence St to convert it to frequency domain F(f), and power spectral density is calculated. The specific calculation formula is as follows:

[0015] PSD(f)=|F(f)| 2

[0016] F(f)=FFT{St}

[0017] wherein PSD(f) represents the power spectral density at frequency f, F(f) represents the Fourier transform result of the luminance time series St, FFT represents the fast Fourier transform, the main peak frequency corresponding to the main frequency component of the luminance oscillation is searched in the power spectral density, the main frequency component is determined in the power spectral density by the 3σ criterion and the luminance oscillation period T is calculated, the main peak boundary is determined by the 3σ criterion, the noise interference is excluded, and the accuracy of the main frequency component extraction is ensured, and the specific calculation formula is as follows:

[0018] f main ∈[μ f -3σ f ,μ f +3σ f ]

[0019] wherein f main represents the main frequency component, μ f represents the mean of the power spectrum peak frequency, and σ f represents the standard deviation of the power spectrum peak frequency, the luminance oscillation period is calculated according to the main frequency component, and the specific calculation formula of the luminance oscillation period is as follows:

[0020]

[0021] wherein T represents the luminance oscillation period, the local neighborhood weighted average method is used for interpolation correction on the abnormal period value (T>100 frames or T<1 frame) to ensure the rationality of the period data, and based on the luminance time series St and the oscillation period T, the instantaneous phase φ(t) is calculated, and the specific calculation formula is as follows:

[0022]

[0023] wherein t mod T represents the relative time in the period obtained by taking the modulus of the time t with respect to the period T, and 2π represents the conversion of the relative position into the relative radian, according to the interference principle of light, the phase change is converted into the change amount ΔOPD of the optical path difference, and the specific calculation formula is as follows:

[0024]

[0025] wherein λ represents the wavelength of the incident light, and Δφ(t) represents the change amount with respect to the reference phase, the spatial gradient is calculated by the Sobel operator to locate the phase mutation region, and the specific calculation formula is as follows:

[0026]

[0027] wherein, represents the partial derivative operator in x, y direction, represents the instantaneous rate of change of the OPD variation in x and y direction, compares the OPD variation trend of adjacent pixel points, marks abnormal values that do not conform to physical continuity, performs Laplacian sharpening on the calculated OPD variation ΔOPD map, enhances the phase mutation characteristics of the defect edge, applies an adaptive threshold segmentation algorithm to enhance the high gradient region, preliminarily locates the potential defects, generates an OPD variation distribution map containing the maximum OPD variation and its position, and labels the maximum OPD variation and its position as the phase distortion field.

[0028] In a preferred embodiment, the obtained phase distortion field data is converted into stress distribution data by using the photoelastic effect, and the spatial distribution of the internal stress of the optical film is obtained. The conversion formula is as follows:

[0029]

[0030] where Δφ represents the phase distortion field, C represents the photoelastic coefficient, d represents the thickness of the optical film, and σ represents the stress per unit area in the optical film. In combination with a preset elastic modulus model E(x, y) = E0·[1+δ(x, y)], where E0 represents the elastic modulus reference value and δ(x, y) represents a perturbation function of local modulus anomaly, the value range of which is preset based on historical defect data or material process specifications, such as |δ(x, y)|≤0.2, the strain of each region of the film is calculated by using Hooke's law, and the specific calculation formula is as follows:

[0031]

[0032] The corresponding relationship between stress and strain is established, the calculated strain field is compared with the measured deformation field, the least squares method is used to iteratively optimize the elastic modulus distribution, the strain predicted by the model is ensured to be consistent with the actual deformation, and the final elastic modulus mapping relationship is formed. The specific calculation formula of the mapping relationship is as follows:

[0033] S(x, y) = f[Δφ(x, y)]

[0034] where S(x, y) represents the elastic modulus distribution at the spatial coordinates (x, y), f[.] represents the mapping function of phase distortion-stress-elastic modulus, and Δφ(x, y) represents the phase distortion at the spatial coordinates (x, y);

[0035] The spatial gradient of the elastic modulus field is calculated using a Gaussian difference operator, and the change intensity of the elastic modulus of each region is quantified. The specific calculation formula of the spatial gradient of the elastic modulus field is as follows:

[0036]

[0037] where This represents the spatial gradient of the elastic modulus field. represents the partial derivative of the elastic modulus in the x-direction, and represents the instantaneous rate of change of the elastic modulus in the x-direction. Let represent the partial derivative of the elastic modulus in the y-direction, and represent the instantaneous rate of change of the elastic modulus in the y-direction. The calculated gradient of the elastic modulus is then compared with a preset threshold T. E In contrast, when When this occurs, the region is determined to be a region of modulus abrupt change;

[0038] Extract the boundaries of the modulus abrupt change regions, perform polynomial fitting on the boundaries of the modulus abrupt change regions, calculate the boundary curvature K, and then select regions with curvatures exceeding a preset curvature threshold T. K The area is identified as the core defect area, thus completing the precise location of the defect.

[0039] The beneficial effects of this invention are as follows: This invention generates a dynamic stress field through the synergistic effect of mechanical vibration and thermal radiation, which can produce significant stress concentration at micro-defects, amplifying the deformation differences of micro-defects and thus improving the sensitivity of defect detection. Spectral analysis of the brightness time series of each pixel is performed, and the dominant frequency component is extracted as the oscillation period, which can accurately reflect the frequency of optical path difference changes. Combined with the Fourier algorithm, a continuous three-dimensional phase distortion field is reconstructed, achieving high-precision quantification of the phase spatial distribution of the light field. Through brightness oscillation analysis and phase reconstruction, a quantitative relationship between light intensity fluctuation and optical path difference change is established. Then, the optical signal is converted into a phase distortion field reflecting physical quantities such as film thickness and deformation. The phase distortion field is projected onto a preset film elastic modulus distribution model, establishing a mapping relationship of "phase distortion-stress distribution-elastic modulus," which can accurately invert the elastic modulus distribution of the film and provide a reliable mechanical basis for micro-defect localization. Attached Figure Description

[0040] Figure 1 This is a flowchart of the present invention;

[0041] Figure 2 This is a logic step diagram of the present invention. Detailed Implementation

[0042] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0043] In the description of the present application, the terms "first", "second" are only for descriptive purpose, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0044] In the description of the present application, the term "for example" is used to indicate "as an example, illustration or description". Any embodiment described as "for example" in the present application is not necessarily interpreted as more preferred or more advantageous than other embodiments. The following description is given in order to enable any person skilled in the art to implement and use the present application. In the following description, details are listed for the purpose of explanation. It should be understood that those skilled in the art can recognize that the present application can be implemented without using these specific details. In other examples, well-known structures and processes will not be described in detail in order to avoid unnecessary details making the description of the present application obscure. Therefore, the present application is not intended to be limited to the embodiments shown, but is consistent with the broadest scope consistent with the principles and characteristics disclosed in the present application.

[0045] As Figures 1-2 The embodiment provides an optical film quality detection method based on image processing, comprising the following steps:

[0046] S101, acquiring a transient light field image sequence of an optical film in a dynamic stress field;

[0047] Further, the dynamic stress field is generated by a first mechanical vibration source and a second thermal radiation source with adjustable frequency. The first mechanical vibration source excites resonance state elastic waves in the film, which can cause stress concentration effect at the micro-defects of the optical film. The second thermal radiation source forms a linear expansion gradient on the surface of the film, which is coupled with the mechanical vibration to form orthogonal stress.

[0048] A high-speed camera is used to image the film in the dynamic stress field in real time. During the acquisition process, the timing consistency of stress field loading and image acquisition is realized through a synchronous triggering mechanism. When the first mechanical vibration source and the second thermal radiation source are started, the camera starts continuous acquisition of the optical film image, ensuring that the stress state corresponding to each image can be traced back. The acquired optical film image sequence contains the change information of light field intensity with time, wherein the brightness oscillation period of each pixel point is directly related to the local optical path difference change. The acquired optical film image is preprocessed, including denoising, brightness equalization and sub-pixel registration.

[0049] It should be noted that the dynamic stress field is generated by the cooperation of the first mechanical vibration source with adjustable frequency and the second thermal radiation source, and the core is to amplify the stress concentration effect of the micro-defects of the film through the composite stress. The frequency of the first mechanical vibration source needs to be less than 50% of the resonance frequency of the film, which can avoid the resonance caused by the film deformation, and ensure the controllability and non-destructiveness of the stress loading; the temperature gradient of the second thermal radiation source is controlled to be not more than 10℃ / cm, and the thermal stress is generated in the film by the gradient thermal field, which forms a coupling with the mechanical vibration stress, and can make the micro-defect area produce a unique dynamic deformation characteristic;

[0050] The design principle of the composite stress field is that the elastic modulus of the micro-defects (such as cracks and impurities) is different from that of the surrounding matrix, and under the action of the dynamic stress, the defect boundary will produce local stress concentration, and then cause abnormal fluctuation of the light field phase, which provides observable physical signals for subsequent defect detection. The mechanical vibration source adopts a low-frequency excitation mode, and the frequency is set to be less than 50% of the resonance frequency of the film. The parameter selection is based on the principle of material dynamics: when the external excitation frequency is far away from the resonance frequency, the overall vibration amplitude of the film is in the linear response interval, which can avoid the interference of nonlinear vibration to the defect characteristics. By adjusting the frequency and amplitude of the vibration source, a periodic mechanical stress wave is formed on the surface of the film. The stress wave will be reflected and refracted when it encounters micro-defects during propagation, which will cause the vibration mode of the defect area to be different from that of the matrix. This difference is reflected in the change of the light field phase, for example, the vibration amplitude or phase of the defect edge will form a phase difference with the surrounding area, thereby providing a basis for subsequent phase distortion analysis.

[0051] The second thermal radiation source forms a stable thermal stress field on the surface of the film by controlling the temperature gradient to be not more than 10℃ / cm. The accurate control of the temperature gradient is based on the heat conduction theory: when there is a temperature difference between the two sides of the film, the inhomogeneity of the thermal expansion coefficient will cause internal thermal stress. The thermal conduction characteristics of the micro-defects (such as holes and debonding) are different from those of the matrix, which will form a thermal stress concentration at the defect boundary. By limiting the temperature gradient to be within 10℃ / cm, the loading rate of the thermal stress can be matched with the frequency of the mechanical vibration, which can avoid the thermal damage of the film caused by excessive thermal stress, and at the same time ensure that the cooperative action of the thermal stress and the mechanical stress can effectively stimulate the dynamic response of the micro-defects, so that the light field phase change of the defect area is detectable.

[0052] The principle and parameter matching of the transient light field image sequence acquisition are that a high-speed camera is used to image the film placed in the dynamic stress field in real time. The frame rate of the camera needs to be ≥1000fps to ensure that the light field transient fluctuation in the stress field change period can be captured. The spatial resolution of the image acquisition is set to be ≥5μm / pixel to ensure that the details of the micro-defects (usually in microns) can be clearly distinguished. During the acquisition process, the time sequence consistency of the stress field loading and the image acquisition is realized through a synchronous triggering mechanism.

[0053] Under the action of dynamic stress field, the deformation of the thin film will lead to the change of light propagation path, and then produce the change of optical path difference. For the uniform thin film, the change of optical path difference presents periodicity, while in the micro-defect area, the non-uniform deformation caused by stress concentration will make the change of optical path difference deviate from the periodicity, showing phase distortion. The brightness oscillation signal collected by the high-speed camera essentially reflects the change of optical path difference: when the thin film deformation leads to the change of optical path difference, the interference intensity of light will change, forming a periodic brightness oscillation. By analyzing the brightness oscillation period of each pixel point in the image sequence, the dynamic change of local optical path difference can be deduced, and then the mapping relationship between the light field phase and the stress distribution is established, providing optical basis for the positioning of micro-defects.

[0054] S102, performing frequency spectrum analysis on the brightness time sequence of each pixel point, extracting the main frequency component as the brightness oscillation period, using Fourier transform algorithm to process the time domain fluctuation data, based on the brightness oscillation period of each pixel point in the image sequence, calculating the local optical path difference change;

[0055] Further, the image feature points of the preprocessed optical thin film image are extracted by using the image registration algorithm based on feature point matching, and then the RANSAC transformation estimation algorithm is used to eliminate the global displacement caused by mechanical vibration between adjacent frames, so as to ensure that the same physical position corresponds to the same pixel coordinates in different frames. The brightness value I(x,y,t) of each pixel point (x,y) at different time t is extracted to form a brightness time sequence:

[0056] St={I(x,y,1),I(x,y,2),...,I(x,y,N)}

[0057] Wherein, the total frame number N needs to satisfy N≥1000, so as to ensure the reliability of the spectrum analysis. The fast Fourier transform is performed on each brightness time sequence St to convert it to the frequency domain F(f), the power spectral density is calculated, and the specific calculation formula is as follows:

[0058] PSD(f)=|F(f)| 2

[0059] F(f)=FFT{St}

[0060] Wherein, PSD(f) represents the power spectral density at frequency f, F(f) represents the Fourier transform result of the brightness time sequence St, FFT represents the fast Fourier transform, the main peak frequency in the power spectral density is searched, which corresponds to the main frequency component of the brightness oscillation, the main frequency component is determined in the power spectral density by 3σ criterion and the brightness oscillation period T is calculated, the main peak boundary is determined by 3σ criterion, the noise interference is excluded, and the accuracy of the main frequency component extraction is ensured. The specific calculation formula is as follows:

[0061] f main ∈[μ f -3σ f ,μ f +3σ f ]

[0062] Among them, f main Represents the dominant frequency component, μ f σ represents the mean of the peak frequency of the power spectrum. f The standard deviation of the peak frequency of the power spectrum is used to calculate the brightness oscillation period based on the dominant frequency component. The specific formula for calculating the brightness oscillation period is as follows:

[0063]

[0064] Where T represents the brightness oscillation period, and the local neighborhood weighted average method is used to interpolate and correct abnormal period values ​​(T>100 frames or T<1 frame) to ensure the rationality of the period data. Based on the brightness time series St and the oscillation period T, the instantaneous phase φ(t) is calculated. The specific calculation formula is as follows:

[0065]

[0066] Where t mod T represents taking the modulus of time t with respect to period T to obtain the relative time within the period, and 2π represents converting the relative position into relative radians. According to the principle of light interference, the phase change is converted into the change in optical path difference ΔOPD. The specific calculation formula is as follows:

[0067]

[0068] Where λ represents the incident light wavelength, Δφ(t) represents the change in phase relative to the reference phase, and the spatial gradient is calculated using the Sobel operator. The specific calculation formula for locating the phase abrupt change region is as follows:

[0069]

[0070] in, The partial derivative operator with respect to the x and y directions represents the instantaneous rate of change of optical path difference in the x and y axes. By comparing the optical path difference change trends of adjacent pixels, outliers that do not conform to physical continuity are marked. The calculated optical path difference change ΔOPD map is Laplacian sharpened to enhance the phase abrupt change features of the defect edges. An adaptive threshold segmentation algorithm is applied to enhance high gradient regions, preliminarily locate potential defects, and generate an optical path difference change distribution map containing the maximum optical path difference change and its location. The maximum optical path difference change and its location are marked as the phase distortion field.

[0071] It should be noted that the phase distortion can directly reflect the optical path difference distribution, and the optical path difference is positively correlated with the thickness of the film, so as to determine whether the thickness is uniform. Stress can cause elastic deformation of the film material, and then affect the propagation path of light. Through the mapping relationship between phase distortion and elastic modulus, the stress distribution can be deduced. The spatial distribution of phase distortion can directly reflect the undulation of the film surface or interface, and the flatness error can be quantified.

[0072] S103, projecting the phase distortion field to the preset film elastic modulus distribution model to establish a mapping relationship among phase distortion, stress distribution and defects, and identifying the modulus mutation area;

[0073] Further, by using the photoelastic effect, the obtained phase distortion field data is converted into stress distribution data to obtain the spatial distribution of the internal stress of the optical film. The conversion formula is as follows:

[0074]

[0075] Where Δφ represents the phase distortion field, C represents the photoelastic coefficient, d represents the thickness of the optical film, and σ represents the stress per unit area in the optical film. In combination with the preset elastic modulus model E(x,y)=E0·[1+δ(x,y)], wherein E0 represents the elastic modulus reference value, and δ(x,y) represents the perturbation function of local modulus anomaly, the value range thereof is preset based on historical defect data or material process specification, such as |δ(x,y)|≤0.2. The strain of each region of the film is calculated by Hooke's law, and the specific calculation formula is as follows:

[0076]

[0077] The corresponding relationship between stress and strain is established, the calculated strain field is compared with the measured deformation field, the least square method is used to iteratively optimize the elastic modulus distribution, the strain predicted by the model is ensured to be consistent with the actual deformation, and the final elastic modulus mapping relationship is formed. The specific calculation formula of the mapping relationship is as follows:

[0078] S(x,y)=f[Δφ(x,y)]

[0079] Where S(x,y) represents the elastic modulus distribution at the spatial coordinates (x,y), f[.] represents the mapping function of phase distortion-stress-elastic modulus, and Δφ(x,y) represents the phase distortion at the spatial coordinates (x,y).

[0080] The spatial gradient of the elastic modulus field is calculated using the Gaussian difference operator to quantify the change degree of the elastic modulus of each region. The specific calculation formula of the spatial gradient of the elastic modulus field is as follows:

[0081]

[0082] wherein, represents the spatial gradient of the elastic modulus field, represents the elastic modulus x-direction partial derivative, represents the instantaneous rate of change of the elastic modulus in the x-direction, represents the elastic modulus y-direction partial derivative, represents the instantaneous rate of change of the elastic modulus in the y-direction, the calculated elastic modulus gradient is compared with a preset threshold T E In contrast, when , the region is determined as a modulus abrupt change region;

[0083] The boundary of the modulus abrupt change region is extracted, the boundary of the modulus abrupt change region is polynomially fitted, the curvature K of the boundary is calculated, and the region with the curvature exceeding a preset curvature threshold T K is determined as a defect core region, and the accurate positioning of the defect is completed.

[0084] It should be noted that the modulus region is the defect core region because the optical film defect will cause local material structure discontinuity, and the elastic modulus E thereof is significantly different from that of the surrounding normal region. For example, the material separates at the crack, and the elastic modulus tends to 0. The modulus of the debonding region is significantly lower than that of the matrix due to the weakened interfacial bonding force.

[0085] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0086] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt 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.

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

[0088] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions Figure 1 function specified in the flow Figure 1 block or blocks.

[0089] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions Figure 1 function specified in the flow Figure 1 block or blocks.

[0090] While the preferred embodiments of the application have been described, additional variations and modifications can be made to the embodiments by those skilled in the art once they learn of the basic inventive concepts. Therefore, the appended claims are intended to encompass within their scope all possible variations and modifications of the preferred embodiments.

[0091] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. A method for detecting the quality of optical thin films based on image processing, characterized in that, Includes the following steps: S101. Obtain the transient optical field image sequence of the optical thin film in the dynamic stress field; S102. Perform spectral analysis on the brightness time series of each pixel, extract the main frequency component as the brightness oscillation period, use the Fourier transform algorithm to process the time domain fluctuation data, and calculate the local optical path difference change based on the brightness oscillation period of each pixel in the image sequence. S103. Project the phase distortion field onto the preset thin film elastic modulus distribution model to establish the mapping relationship between phase distortion, stress distribution and defects, and identify regions of modulus abrupt change.

2. The optical thin film quality inspection method based on image processing according to claim 1, characterized in that, The dynamic stress field is generated by applying a first mechanical vibration source and a second thermal radiation source with adjustable frequency. The first mechanical vibration source excites resonant elastic waves inside the thin film, which can generate stress concentration effect at the micro-defects of the optical thin film. The second thermal radiation source forms a linear expansion gradient on the surface of the thin film, which is orthogonally coupled with the mechanical vibration.

3. The optical thin film quality detection method based on image processing according to claim 2, characterized in that, A high-speed camera is used to image a thin film placed in a dynamic stress field in real time. During the acquisition process, a synchronous triggering mechanism is used to ensure the temporal consistency between stress field loading and image acquisition. When the first mechanical vibration source and the second thermal radiation source are activated, the camera is synchronously triggered to start continuously acquiring optical thin film images, ensuring that the stress state corresponding to each frame of the image is traceable. The acquired optical thin film image sequence contains information on the change of light field intensity over time. The brightness oscillation period of each pixel is directly related to the change of local optical path difference. The acquired optical thin film images are preprocessed.

4. The optical thin film quality inspection method based on image processing according to claim 1, characterized in that, Image feature points are extracted from the preprocessed optical thin film image using an image registration algorithm based on feature point matching; The RANSAC transform estimation algorithm is used to eliminate global displacement caused by mechanical vibration between adjacent frames, ensuring that the same physical location corresponds to the same pixel coordinates in different frames. Extract the brightness value I(x,y,t) of each pixel (x,y) at different times t to form a brightness time series; Perform a fast Fourier transform on each brightness time series and calculate the power spectral density; The dominant frequency component was determined in the power spectral density using the 3σ criterion, and the brightness oscillation period T was calculated. Based on the brightness time series St and the oscillation period T, the instantaneous phase φ(t) is calculated. According to the principle of light interference, the phase change is converted into the change of optical path difference. Generate an optical path difference distribution map containing the maximum optical path difference change and its location, as the phase distortion field.

5. The optical thin film quality detection method based on image processing according to claim 4, characterized in that, The dominant frequency component is determined in the power spectral density using the 3σ criterion: By identifying the main peak boundary, noise interference is eliminated, ensuring the accuracy of the dominant frequency component extraction. The specific calculation formula is as follows: f main ∈[μ f -3s f ,m f +3s f ] Among them, f main Represents the dominant frequency component, μ f σ represents the mean of the peak frequency of the power spectrum. f The standard deviation of the peak frequency of the power spectrum is used to calculate the brightness oscillation period based on the dominant frequency component. The specific formula for calculating the brightness oscillation period is as follows: Where T represents the brightness oscillation period, and the abnormal period value is interpolated and corrected using the local neighborhood weighted average method.

6. The optical thin film quality detection method based on image processing according to claim 4, characterized in that, The spatial gradient of the optical path difference variation distribution map is calculated using the Sobel operator. The specific calculation formula for locating the phase abrupt change region is as follows: in, The partial derivative operator with respect to the x and y directions represents the instantaneous rate of change of optical path difference in the x and y directions. By comparing the optical path difference change trends of adjacent pixels, outliers that do not conform to physical continuity are marked. The calculated optical path difference change ΔOPD map is then subjected to Laplacian sharpening to enhance the phase change features of the defect edges. An adaptive threshold segmentation algorithm is applied to enhance high gradient regions and preliminarily locate potential defects.

7. The optical thin film quality inspection method based on image processing according to claim 1, characterized in that, Using the photoelastic effect, the phase distortion field data is converted into stress distribution data. The conversion formula is as follows: Where C represents the photoelastic coefficient, d represents the thickness of the optical film, and σ represents the stress per unit area inside the optical film. Combining the preset elastic modulus model E(x,y)=E0·[1+δ(x,y)], where E0 represents the reference value of the elastic modulus and δ(x,y) represents the perturbation function of the local modulus anomaly, the strain of each region of the film is calculated by Hooke's law, and the correspondence between stress and strain is established. The calculated strain field is compared with the measured deformation field, and the elastic modulus distribution is iteratively optimized by the least squares method to form the final relationship.

8. The optical thin film quality detection method based on image processing according to claim 7, characterized in that, It also includes calculating the spatial gradient of the elastic modulus field, with the specific calculation formula as follows: in, This represents the spatial gradient of the elastic modulus field. represents the partial derivative of the elastic modulus in the x-direction, and represents the instantaneous rate of change of the elastic modulus in the x-direction. Let represent the partial derivative of the elastic modulus in the y-direction, and represent the instantaneous rate of change of the elastic modulus in the y-direction. The calculated gradient of the elastic modulus is then compared with a preset threshold T. E In contrast, when When this occurs, the region is determined to be a region of modulus abrupt change.

9. The optical thin film quality detection method based on image processing according to claim 8, characterized in that, Defect localization in regions of modulus abrupt change includes: Extract the boundaries of regions with abrupt changes in modulus, perform polynomial fitting on the boundaries, calculate the boundary curvature K, and then select regions with curvatures exceeding a preset curvature threshold T. K The area is identified as the core defect area.

Images