A laser-ultrasonic non-contact detection method and system based on polarity features

Abstract

The present application relates to the technical field of non-destructive testing of microelectronic devices, and particularly relates to a laser-ultrasonic non-contact detection method and system based on polarity characteristics, comprising: performing two-dimensional scanning on a sample surface and collecting time-domain diffraction signals of each scanning point; extracting the polarity of a first diffraction peak within a preset time window from the time-domain diffraction signals of each scanning point; calculating polarity spatial statistical characteristics based on the polarity of each scanning point; and generating a defect probability heat map based on the polarity spatial statistical characteristics. The present application extracts and utilizes the polarity fluctuation characteristics of the first diffraction peak in the laser-ultrasonic time-domain diffraction signals, converts the polarity fluctuation characteristics into polarity spatial statistical characteristics, and calculates and generates a defect probability heat map, thereby effectively overcoming the limitations of traditional detection methods that rely on signal amplitude and are susceptible to laser power fluctuations and mechanical vibration interference, and significantly improving the detection signal-to-noise ratio and positioning accuracy for subsurface nanoscale defects.

Description

Technical Field

[0001] This invention relates to the field of non-destructive testing technology for microelectronic devices, specifically to a laser-ultrasonic non-contact testing method and system based on polarity characteristics. Background Technology

[0002] With the rapid development of semiconductor manufacturing, advanced packaging, and micro / nano system integration technologies, device structures are increasingly becoming multilayered and miniaturized. In various microelectronic devices, defects beneath the metal layer or at thin film interfaces, such as particulate contamination, abrupt changes in interface roughness, or microcracks, have become key factors restricting device performance and long-term reliability. However, because these defects are located below the surface, traditional optical or electron microscopy methods struggle to achieve direct, non-destructive observation and localization. In recent years, ultrasonic testing technology based on ultrashort pulse lasers has gradually become an important tool for characterizing subsurface structures due to its non-contact nature and high spatiotemporal resolution. This technology uses a pump laser to excite ultrasonic waves in the material. When these waves encounter subsurface structures during propagation, they are reflected or scattered. The resulting surface optical response can be captured by the probe laser, thereby retrieving the morphology and physical state information of the subsurface.

[0003] However, existing laser ultrasonic testing methods primarily focus on morphological measurements of regular structures (such as periodic gratings), and their analysis largely relies on the time delay and signal amplitude information of the ultrasonic echo. In actual measurements, researchers have found that when samples are scanned in steps of tens of micrometers, the amplitude of the measured signal at different locations fluctuates significantly, and even frequent polarity reversals occur. These polarity reversals occur when the sample moves parallel or perpendicular to the structural direction, and are particularly pronounced in the presence of inclusions or abrupt interface changes. Although such polarity changes statistically exhibit local anomalies, existing techniques typically treat them as random noise caused by dust or optical interference, without in-depth analysis or utilization. Furthermore, because signal amplitude is easily affected by factors such as laser power fluctuations and mechanical vibrations, detection methods relying solely on amplitude information often exhibit limitations such as low signal-to-noise ratio and insufficient stability when dealing with nanoscale defects. They struggle to effectively identify local optical field disturbances caused by minute defects, thus limiting the potential of this technology for precise defect localization and classification. Summary of the Invention

[0004] To address the above technical problems, this invention provides a technical solution for a laser ultrasonic non-contact detection method and system based on polarity characteristics.

[0005] The technical problem solved by this invention can be achieved by the following technical solutions: A non-contact laser-ultrasound detection method based on polarity characteristics includes: Step S1: Perform a two-dimensional scan on the sample surface and acquire the time-domain diffraction signal at each scan point; Step S2: Extract the polarity of the first diffraction peak within a preset time window from the time-domain diffraction signal of each scanning point; Step S3: Calculate the polarity space statistical characteristics based on the polarity of each scan point; Step S4: Generate a defect probability heatmap based on the polarity space statistical features.

[0006] Preferably, the preset time window is set based on the appearance time of the first diffraction peak at the initial scan point, specifically expressed by the formula: , in, The time when the first diffraction peak appears at the initial scan point. For half the time.

[0007] Preferably, step S2 includes: Step S21, the time-domain diffraction signal is subjected to DC component removal and filtering, specifically expressed by the following formula: , , in, The time-domain diffraction signal, The time-domain diffraction signal after DC removal. This is the normalized time-domain diffraction signal; Step S22: Extract a signal segment within the preset time window from the preprocessed time-domain diffraction signal, and identify the time point corresponding to the extreme point of the first diffraction peak within the signal segment. The formula is as follows: , in, The time point corresponding to the extreme point; Step S23: Determine and output the polarity based on the peak intensity corresponding to the extreme point, specifically expressed by the formula: , Wherein, s represents the polarity, and A represents the peak intensity.

[0008] Preferably, in step S3, the polarity space statistical features include polarity reversal rate, symbol entropy, and amplitude variation coefficient.

[0009] Preferably, the formula for calculating the polarity reversal rate is: , in, Center scan point polarity reversal rate, Center scan point polarity, neighboring points The polarity of N, where N is a neighboring point. The total number.

[0010] Preferably, the formula for calculating the symbol entropy is: , , in, Center scan point symbolic entropy, The proportion of positive polarity within the neighborhood. This represents the proportion of negative polarity within the neighborhood.

[0011] Preferably, the formula for calculating the amplitude variation coefficient is: , in, Center scan point The amplitude variation coefficient, where A is the peak intensity. The standard deviation of the peak intensity within the neighborhood. It represents the average value of the peak intensity within the neighborhood.

[0012] Preferably, in step S4, the generation model for the defect probability heatmap is: , in, Center scan point Defect probability heatmap Center scan point polarity reversal rate, Center scan point symbolic entropy, The maximum value of symbolic entropy. Center scan point The amplitude variation coefficient, , , All are weighting coefficients, and .

[0013] Preferably, in step S1, the step distance of the two-dimensional scan is not less than twice the spatial coherence length Lc.

[0014] It also includes a laser-ultrasonic non-contact detection system based on polarity characteristics, which implements the laser-ultrasonic non-contact detection method based on polarity characteristics as described above, including: Laser emitting module, used to generate pump light and probe light; The scanning module, connected to the laser emitting module, is used to perform two-dimensional scanning of the sample based on a preset scanning step and path; The signal acquisition and processing module is connected to the scanning module and is used to acquire the time-domain diffraction signal of each scanning point, extract the polarity of the first diffraction peak in the time-domain diffraction signal, and calculate the polarity spatial statistical characteristics. The heat map generation module, connected to the signal acquisition and processing module, is used to generate a defect probability heat map based on the polarity space statistical characteristics. The display module, connected to the heatmap generation module, is used to visualize the detection results.

[0015] Beneficial effects: This invention extracts and utilizes the polarity fluctuation characteristics of the first diffraction peak in the laser ultrasonic time-domain diffraction signal, transforms it into polarity spatial statistical characteristics, and calculates and generates a defect probability heatmap. This effectively overcomes the limitations of traditional signal amplitude-dependent detection, which is susceptible to interference from laser power fluctuations and mechanical vibrations. It significantly improves the detection signal-to-noise ratio and positioning accuracy for subsurface nanoscale defects, providing a reliable means for non-destructive, high-sensitivity identification and classification of hidden defects in microelectronic devices. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the time-domain diffraction signal acquisition process of the present invention; Figure 3 This is a flowchart illustrating step S2 of the present invention; Figure 4 This is a system block diagram of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0019] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the scope of the invention.

[0020] Reference Figure 1This invention provides a laser-ultrasound non-contact detection method based on polarity characteristics, comprising: Step S1: Perform a two-dimensional scan on the sample surface and acquire the time-domain diffraction signal at each scan point; Step S2: Extract the polarity of the first diffraction peak within a preset time window from the time-domain diffraction signal of each scanning point; Step S3: Calculate the polarity space statistical characteristics based on the polarity of each scan point; Step S4: Generate a defect probability heatmap based on the polarity space statistical features.

[0021] Specifically, in this embodiment of the invention, in response to the technical problem that existing laser ultrasonic testing technology suffers from low signal-to-noise ratio and insufficient stability due to its reliance on signal amplitude, making it difficult to effectively identify and locate subsurface nanoscale defects, the invention extracts and analyzes signal polarity characteristics that are more sensitive to minute defects and less susceptible to common noise interference. This avoids the inherent limitations of traditional amplitude analysis methods, which are easily affected by laser power fluctuations and mechanical vibrations, and achieves high sensitivity, high signal-to-noise ratio, and non-contact location and detection of subsurface minute defects.

[0022] Specifically, in practical applications, the zero reference point (x0, y0) is used as the initial scanning point. The waveform signal of this scanning point is acquired, and the first diffraction peak is locked as the reference reference. Subsequently, the two-dimensional precision displacement stage performs two-dimensional scanning motion in the horizontal (e.g., (x1, y0), (x2, y0), ...) and vertical directions (e.g., (x0, y1), (x0, y1), ...) with set steps, systematically recording the polarity data of the first diffraction peak at each scanning point. Finally, by statistically analyzing the spatial distribution characteristics of these polarity data, a two-dimensional heat map that intuitively reflects the probability distribution of defects is generated.

[0023] More specifically, in the embodiments of the present invention, reference is made to... Figure 2The time-domain diffraction signal acquisition system uses a Ti:sapphire laser as its core, generating laser pulses with a repetition frequency on the order of MHz and a pulse width of tens of femtoseconds. This pulse is split into two paths by a beam splitter: one path serves as the pump light, which, after modulation by an optical modulator, is focused into a spot with a diameter of approximately 50 micrometers and acts on the sample surface. Its pulse energy is controlled at the nanojoule level and can be precisely adjusted according to experimental requirements. The other path serves as the probe light, which, after precise time delay control by an optical delay line, is focused into a spot with a diameter of approximately 40 micrometers and coincides with the pump light spot on the sample surface through a lens. When the ultrasonic waves excited by the pump light propagate in the material and interact with subsurface defects, scattering or diffraction occurs at the defect interface, causing transient changes in local refractive index or displacement. After sensing this change, the probe light's reflected light interferes with the reference light. The interference signal is received by a photodetector and converted into an electrical signal, ultimately forming a voltage-time curve, i.e., the time-domain diffraction signal S(t). This method overcomes the limitations of traditional amplitude analysis in detecting nanoscale defects by directly capturing the polarity reversal characteristics of diffraction signals induced by the interaction between ultrasound and defects, and significantly improves the sensitivity and reliability of detection.

[0024] In a preferred embodiment of the present invention, in step S1, the step distance of the two-dimensional scan is not less than twice the spatial coherence length Lc.

[0025] Specifically, since an excessively small scanning step can cause the detection areas of adjacent measurement points to overlap, resulting in a strong correlation of signal polarity characteristics, thereby obscuring the accurate location of the defect and reducing the detection resolution, in this embodiment of the invention, the scanning step is set to be no less than twice the spatial coherence length Lc to ensure the independence and statistical validity of the signals at each scanning point, with a preferred range of 30-70 μm.

[0026] Furthermore, in this embodiment of the invention, a scanning step of 50 μm is preferred. This setting can effectively avoid crosstalk between signals and achieve efficient two-dimensional scanning while ensuring detection accuracy, thereby laying the foundation for the accurate extraction of polarity features and the reliable generation of defect probability heatmaps.

[0027] In a preferred embodiment of the present invention, the preset time window is set based on the appearance time of the first diffraction peak at the initial scan point, specifically expressed by the formula: , in, The time when the first diffraction peak appears at the initial scan point. For half the time.

[0028] Specifically, considering that the propagation speed of ultrasound in materials is relatively stable, and that the signal time shift caused by subsurface defects is usually within a certain range, in this embodiment of the invention, before extracting the polarity of the first diffraction peak at each scanning point, the polarity measured at the initial scanning point (x0, y0) is pre-set. Centered on approximately 285 ps Peak search can be performed within a typical range of 10-20 ps, ​​depending on actual testing needs. This setting effectively eliminates interference from other reflected waves or noise signals in subsequent time windows, ensuring that the extracted polarity characteristics of the first diffraction peak accurately reflect the initial response of the target subsurface structure, thereby improving the accuracy and reliability of defect identification and location.

[0029] It is evident that by limiting the time range of peak search, not only is the efficiency of signal processing improved, but the correspondence between polarity features and subsurface defects is also enhanced, providing a reliable data foundation for the subsequent generation of high-resolution defect probability heatmaps.

[0030] As a preferred embodiment of the present invention, refer to Figure 3 Step S2 includes: Step S21, the time-domain diffraction signal is subjected to DC component removal and filtering, specifically expressed by the following formula: , , in, The time-domain diffraction signal, The time-domain diffraction signal after DC removal. This is the normalized time-domain diffraction signal; Step S22: Extract a signal segment within the preset time window from the preprocessed time-domain diffraction signal, and identify the time point corresponding to the extreme point of the first diffraction peak within the signal segment. The formula is as follows: , in, The time point corresponding to the extreme point; Step S23: Determine and output the polarity based on the peak intensity corresponding to the extreme point, specifically expressed by the formula: , Wherein, s represents the polarity, and A represents the peak intensity.

[0031] Specifically, since the original time-domain diffraction signal usually contains interference components such as DC offset, high-frequency noise, and low-frequency drift, it will seriously affect the accuracy of the polarity determination of the first diffraction peak. Based on this, in this embodiment of the invention, the signal baseline offset is first eliminated by DC removal processing, and then high-frequency noise and low-frequency drift are effectively suppressed by bandpass filtering (e.g., 0.1-3 GHz). Subsequently, the influence of signal amplitude fluctuation is eliminated by normalization processing. On this basis, the interval where the first diffraction peak is located is accurately located by a preset time window, the peak position is determined by an extreme value search algorithm, and finally the polarity determination result is automatically output based on the positive or negative value of the peak intensity.

[0032] This "filtering → limiting time window → finding maximum value → determining sign" processing flow effectively improves the stability and repeatability of polarity features, making them more sensitive to signal changes caused by minor defects. At the same time, it significantly reduces the interference of environmental noise and system fluctuations on the detection results, laying a reliable signal foundation for the subsequent accurate generation of defect probability heatmaps.

[0033] In a preferred embodiment of the present invention, in step S3, the polarity space statistical features include polarity reversal rate, symbol entropy, and amplitude variation coefficient.

[0034] Specifically, considering that subsurface defects can disrupt the propagation consistency of ultrasonic signals, leading to polarity anomalies and amplitude fluctuations in local areas, this embodiment of the invention quantifies such anomalies by calculating multi-dimensional statistical characteristics within the neighborhood of each scanning point. In the detection system, regions with stable signs and gradual amplitude changes are defined as normal regions, while regions with defects such as inclusions or abrupt changes in interface roughness are defined as abnormal regions. These regions typically exhibit characteristics of random polarity reversals and strong amplitude jitter.

[0035] Furthermore, for any scan point (i,j) in the scan grid, its neighborhood (e.g., a 3×3 pixel region with N=8 neighborhood points; a 5×5 pixel region with N=24 neighborhood points, etc.) is used as the statistical unit, and the following key features are calculated: 1) Polarity reversal rate The calculation formula is as follows: , in, Center scan point The polarity reversal rate is used to reflect the frequency of polarity changes between adjacent points. Center scan point The polarity of (take +1 or -1). neighboring points The polarity of N, where N is a neighboring point. The number (excluding the center scan point).

[0036] This polarity reversal rate The polarity stability of a local region is quantified by statistically analyzing the proportion of neighborhoods with polarities inconsistent with the center point; the value ranges from 0 to 1. In normal regions, this value approaches 0, indicating a stable polarity distribution; while in abnormal regions, due to signal disturbances caused by defects, this value increases significantly, reflecting an unstable state with frequent polarity reversals.

[0037] 2) Symbolic entropy The calculation formula is as follows: , , in, Center scan point symbolic entropy, The proportion of positive polarity (s=+1) within the neighborhood. The proportion of negative polarity (s=-1) in the neighborhood.

[0038] The symbolic entropy The degree of disorder in the polarity distribution within a neighborhood is assessed based on information entropy theory. Normal regions exhibit lower entropy values ​​due to their uniform and orderly polarity distribution; while abnormal regions show significantly higher entropy values ​​due to their random polarity distribution, indicating enhanced disorder in local signal patterns.

[0039] 3) Amplitude variation coefficient The calculation formula is as follows: , in, Center scan point The amplitude variation coefficient, where A is the peak intensity. The standard deviation of the peak intensity within the neighborhood. It represents the average value of the peak intensity within the neighborhood.

[0040] The amplitude variation coefficient The coefficient of variation characterizes the relative fluctuation of signal intensity within a neighborhood; a larger value indicates a more non-uniform signal. In normal regions, the coefficient of variation is small due to stable signal propagation; however, in abnormal regions, the coefficient of variation increases significantly due to strong scattering and signal jitter caused by defects, indicating potential subsurface structural anomalies.

[0041] In addition, embodiments of the present invention can introduce other spatial statistical features, such as local polarity autocorrelation functions or gradient distribution features, according to specific detection needs, to construct a more complete defect identification and classification system. By integrating the above multidimensional features, the system can effectively distinguish signal patterns in normal and abnormal regions, providing a reliable quantitative basis for the accurate generation of subsequent defect probability heatmaps.

[0042] In a preferred embodiment of the present invention, in step S4, the generation model of the defect probability heatmap is as follows: , in, Center scan point Defect probability heatmap The maximum value of symbolic entropy. , , All are weighting coefficients, and .

[0043] Specifically, in order to effectively integrate the statistical features of multidimensional polarity space into a unified defect probability index, this embodiment of the invention employs a weighted linear combination model to construct a defect probability heatmap. That is, by using the polarity reversal rate... Normalized symbolic entropy and amplitude variation coefficient The three feature quantities are linearly superimposed according to preset weights, wherein the weight coefficients are preferably... =0.5:0.3:0.2 to highlight the dominant role of polarity reversal rate in defect identification; the calculated PFI values ​​are normalized and mapped to the [0,1] interval, and finally an intuitive defect probability heatmap is generated through pseudo-color coding.

[0044] This weighted linear combination model fully utilizes the sensitivity of polarity features to minute defects, while improving the reliability of detection results through multi-feature fusion, thus achieving accurate localization and visualization of subsurface contamination and defects.

[0045] Reference Figure 4 The present invention also includes a laser-ultrasonic non-contact detection system based on polarity characteristics, which implements the laser-ultrasonic non-contact detection method based on polarity characteristics as described above, including: Laser emitting module 100 is used to generate pump light and probe light; The scanning module 200 is connected to the laser emitting module 100 and is used to perform two-dimensional scanning of the sample based on a preset scanning step and path; The signal acquisition and processing module 300 is connected to the scanning module 200 and is used to acquire the time-domain diffraction signal of each scanning point, extract the polarity of the first diffraction peak in the time-domain diffraction signal, and calculate the polarity spatial statistical characteristics. The heat map generation module 400 is connected to the signal acquisition and processing module 300 and is used to generate a defect probability heat map based on the polarity space statistical features. The display module 500 is connected to the heat map generation module 400 and is used to visualize the detection results.

[0046] Specifically, in this embodiment of the invention, the laser emitting module 100 includes a Ti:sapphire laser, a beam splitter, an optical modulator, and an optical delay line. The Ti:sapphire laser generates femtosecond laser pulses, the beam splitter divides the pulses into pump and probe beams, the optical modulator modulates the intensity and phase of the pump beam to ensure the stability and signal-to-noise ratio of the ultrasonic excitation, and the optical delay line precisely controls the time delay of the probe beam relative to the pump beam. The scanning module 200 includes a lens and a two-dimensional precision displacement stage. The lens focuses the pump beam and probe beam onto the same point on the sample surface, and the two-dimensional precision displacement stage carries the sample and performs two-dimensional scanning according to preset parameters. The signal acquisition and processing module 300 includes a photodetector and a data acquisition system. The photodetector receives the reflected signal of the probe beam and converts it into an electrical signal, and the data acquisition system acquires the time-domain diffraction signal and executes a signal processing algorithm. The heat map generation module 400 and the display module 500 are integrated into a computer system, and generate and visualize a defect probability heat map by executing a preset algorithm.

[0047] In summary, this invention extracts and analyzes the polarity fluctuation characteristics of the first diffraction peak in the laser ultrasonic time-domain diffraction signal, and constructs a multi-dimensional spatial statistical model based on polarity reversal rate, symbol entropy, and amplitude variation coefficient, thereby achieving accurate localization and visual detection of subsurface defects.

[0048] Compared with the prior art, the present invention has the following outstanding advantages: Significantly improved detection sensitivity: By analyzing polar characteristics that are more sensitive to minute defects, replacing the traditional amplitude analysis method which is susceptible to interference, it can effectively identify subsurface defects at the nanoscale.

[0049] Strong anti-interference capability: The polarity characteristic is not sensitive to common noise sources such as laser power fluctuations and mechanical vibrations, which greatly improves the detection stability and reliability under complex working conditions.

[0050] Spatial resolution optimization: By setting a scanning step size of not less than twice the spatial coherence length, the independence of signals from adjacent measurement points is ensured, effectively improving the accuracy of defect location.

[0051] High detection efficiency: It adopts an automatic peak recognition algorithm with a preset time window, combined with an efficient two-dimensional scanning strategy, to achieve fast and automated full-field defect detection.

[0052] Visual and intuitive: By generating a defect probability heatmap through multi-feature weighted fusion, the abstract ultrasonic signal is transformed into an intuitive pseudo-color image, which greatly improves the efficiency of interpreting the detection results.

[0053] High system integration: It organically integrates optical, mechanical and signal processing modules to build a complete non-contact detection system, providing a feasible technical solution for industrial field applications.

[0054] The above description is merely a preferred embodiment of the present invention and does not limit the implementation and protection scope of the present invention. Those skilled in the art should realize that any equivalent substitutions and obvious changes made based on the description and illustrations of the present invention should be included within the protection scope of the present invention.

Claims

1. A laser-ultrasonic non-contact detection method based on polarity characteristics, characterized in that, include: Step S1: Perform a two-dimensional scan on the sample surface and acquire the time-domain diffraction signal at each scan point; Step S2: Extract the polarity of the first diffraction peak within a preset time window from the time-domain diffraction signal of each scanning point; Step S3: Calculate the polarity space statistical characteristics based on the polarity of each scan point; The polarity space statistical features include polarity reversal rate, symbol entropy, and amplitude variation coefficient; Step S4: Generate a defect probability heatmap based on the polarity space statistical features.

2. The laser-ultrasonic non-contact detection method based on polarity characteristics according to claim 1, characterized in that, The preset time window is set based on the appearance time of the first diffraction peak at the initial scan point, and is specifically expressed by the formula: ， in, The time when the first diffraction peak appears at the initial scan point. For half the time.

3. The laser-ultrasonic non-contact detection method based on polarity characteristics according to claim 2, characterized in that, Step S2 includes: Step S21, the time-domain diffraction signal is subjected to DC component removal and filtering, specifically expressed by the following formula: ， ， in, The time-domain diffraction signal, The time-domain diffraction signal after DC removal. This is the normalized time-domain diffraction signal; Step S22: Extract a signal segment within the preset time window from the preprocessed time-domain diffraction signal, and identify the time point corresponding to the extreme point of the first diffraction peak within the signal segment. The formula is as follows: ， in, The time point corresponding to the extreme point; Step S23: Determine and output the polarity based on the peak intensity corresponding to the extreme point, specifically expressed by the formula: ， in, For the aforementioned polarity, The peak intensity is denoted as .

4. The laser-ultrasonic non-contact detection method based on polarity characteristics according to claim 1, characterized in that, The formula for calculating the polarity reversal rate is: ， in, Center scan point polarity reversal rate, Center scan point polarity, neighboring points polarity, neighboring points The quantity.

5. The laser-ultrasonic non-contact detection method based on polarity characteristics according to claim 1, characterized in that, The formula for calculating the symbol entropy is: ， in, Center scan point symbolic entropy, The proportion of positive polarity within the neighborhood. This represents the proportion of negative polarity within the neighborhood.

6. The laser-ultrasonic non-contact detection method based on polarity characteristics according to claim 1, characterized in that, The formula for calculating the amplitude variation coefficient is: ， in, Center scan point The amplitude variation coefficient, Peak intensity The standard deviation of the peak intensity within the neighborhood. It represents the average value of the peak intensity within the neighborhood.

7. The laser-ultrasonic non-contact detection method based on polarity characteristics according to claim 1, characterized in that, In step S4, the generation model for the defect probability heatmap is: ， in, Center scan point Defect probability heatmap Center scan point polarity reversal rate, Center scan point symbolic entropy, The maximum value of symbolic entropy. Center scan point The amplitude variation coefficient, , , All are weighting coefficients, and .

8. The laser-ultrasonic non-contact detection method based on polarity characteristics according to claim 1, characterized in that, In step S1, the step distance of the two-dimensional scan is not less than twice the spatial coherence length Lc.

9. A laser-ultrasonic non-contact detection system based on polarity characteristics, characterized in that, The application implements a laser-ultrasonic non-contact detection method based on polarity characteristics as described in any one of claims 1-8, comprising: Laser emitting module, used to generate pump light and probe light; The scanning module, connected to the laser emitting module, is used to perform two-dimensional scanning of the sample based on a preset scanning step and path; The signal acquisition and processing module, connected to the scanning module, is used to acquire the time-domain diffraction signal at each scanning point, extract the polarity of the first diffraction peak in the time-domain diffraction signal, and calculate the polarity spatial statistical characteristics. A heat map generation module, connected to the signal acquisition and processing module, is used to generate a defect probability heat map based on the polarity space statistical characteristics. The display module, connected to the heatmap generation module, is used to visualize the detection results.

Images