OLED mobile phone screen linear defect detection method and system

Abstract

The invention relates to the technical field of mobile phone detection, in particular to an OLED mobile phone screen linear defect detection method and system. According to the method, the vibration time domain signal on the surface of the screen within the preset time period is obtained and does not depend on the display brightness of the screen, and the defect position is obtained through the abnormal signal interval, so that misjudgment caused by non-uniform brightness of the display image is avoided, the spectral data of each pixel is captured by using the spectral image, and the detection accuracy is improved. The influence of ambient light and display brightness can be effectively eliminated, defects are analyzed through the gradient magnitude of spectral data, misjudgment caused by light change and screen brightness difference can be reduced, abnormal pixels and normal pixels are distinguished by analyzing the gradient magnitude, subtle differences between the pixels can be recognized more accurately, and the accuracy of the detection result is improved. By acquiring the parameter features of each abnormal pixel, the characteristics of the defect can be described more comprehensively, so that the accuracy of defect identification is improved, the defect degree is evaluated through the parameter features, and more accurate defect degree evaluation can be obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of mobile phone detection, and particularly relates to a method and system for detecting linear defects of an OLED mobile phone screen. Background Art

[0002] OLED (Organic Light-Emitting Diode) is a display technology. Its main feature is that each pixel is composed of organic materials and can emit light independently. Different from traditional LCD screens, OLED does not require a backlight, and each pixel of the display can be independently controlled. Therefore, it has higher contrast, more vivid colors, deeper black display, and a thinner screen. Linear defects of the mobile phone screen refer to vertical or horizontal linear defects that appear in the display area of the OLED screen and do not conform to the normal display effect. These defects usually manifest as inconsistent brightness, color distortion, or completely black lines with no display.

[0003] In the prior art, the main method for detecting linear defects of the OLED mobile phone screen is to obtain the display image of the OLED screen and apply algorithms such as image filtering, edge detection, or line detection to identify lines or linear defects in the image. However, the image processing method usually depends on the display brightness of the screen, and any change in ambient light may affect the detection effect, making the detection algorithm unable to effectively identify, and further resulting in inaccurate detection defect degree results. Summary of the Invention

[0004] The main purpose of the present invention is to provide a method for detecting linear defects of an OLED mobile phone screen, aiming to solve the technical problems in the prior art.

[0005] The present invention proposes a method for detecting linear defects of an OLED mobile phone screen, including: Obtaining a plurality of vibration time-domain signals on the surface of the OLED mobile phone screen within a preset time period, and performing Fourier transform processing on each of the vibration time-domain signals to obtain corresponding vibration frequency-domain signals; Synchronously aligning and calibrating the plurality of vibration frequency-domain signals to obtain a plurality of synchronous vibration signals; Obtaining an abnormal signal interval according to the plurality of synchronous vibration signals, and obtaining a vibration abnormality degree according to the abnormal signal interval; Judging whether the vibration abnormality degree is greater than a preset abnormality degree; If the vibration abnormality degree is greater than the preset abnormality degree, it is determined that the OLED mobile phone screen has a linear defect, and the linear defect position of the OLED mobile phone screen is obtained according to the abnormal signal interval; Obtaining a spectral image of the linear defect position in the OLED mobile phone screen, and obtaining spectral data of each pixel point according to the spectral image; Obtain the corresponding gradient amplitude according to each of the spectral data, and determine abnormal pixel points and normal pixel points based on the gradient amplitude; Obtain the parameter features of each of the abnormal pixel points, and obtain the defect degree value based on the multiple parameter features.

[0006] Preferably, the step of synchronously aligning and calibrating the multiple vibration frequency domain signals to obtain multiple synchronous vibration signals includes: Use a high-pass filter to remove the DC component in each of the vibration frequency domain signals to obtain multiple dynamic vibration signals, and perform denoising and normalization processing on each of the dynamic vibration signals in sequence to obtain corresponding standard vibration signals; Sort the multiple standard vibration signals in chronological order, mark the standard vibration signal ranked first as the reference vibration signal, and mark the remaining standard vibration signals as calibration vibration signals; Obtain the first initial time point of the reference vibration signal and the second initial time point of each calibration vibration signal, and obtain the corresponding lag value based on the first initial time point and each second initial time point; Obtain the corresponding cross-correlation coefficient based on the reference vibration signal, each lag value, and the corresponding calibration vibration signal, and perform translation on the time axis of the corresponding calibration vibration signal according to each cross-correlation coefficient to obtain multiple synchronous vibration signals.

[0007] Preferably, the step of obtaining the abnormal signal interval according to the multiple synchronous vibration signals includes: Obtain the first amplitude of each of the synchronous vibration signals at each time point, and perform weighted average processing on the multiple first amplitudes at each time point to obtain a composite vibration signal; Obtain the first amplitude at each frequency point according to the composite vibration signal, and obtain the corresponding power spectral density according to each first amplitude; Establish a frequency-density coordinate axis with frequency as the X-axis and power spectral density as the Y-axis, and plot the power spectral density corresponding to each frequency as a connection point on the frequency-density coordinate axis; Connect the multiple connection points in sequence through a curve to obtain a power spectral waveform diagram, and obtain the peak value and valley value of the power spectral waveform diagram; Divide the power spectral waveform diagram into multiple curves according to adjacent two peak values and valley values, obtain the curvature of each curve, and determine whether the curvature is greater than a preset threshold; If the curvature is greater than the preset threshold, determine that the curve corresponding to the curvature is an abnormal curve, and obtain the abnormal signal interval according to the abnormal curve and the power spectral waveform diagram.

[0008] Preferably, the step of obtaining the vibration abnormality degree according to the abnormal signal interval includes: Obtaining the abnormal frequency components of each frequency point according to the abnormal signal interval, obtaining the historical frequency components of each same frequency point in the historical normal signal of the OLED mobile phone screen, and obtaining the frequency deviation degree according to the multiple historical frequency components and abnormal frequency components; Obtaining the total number of frequency bands, the maximum frequency and the minimum frequency according to the abnormal signal interval, and obtaining the center frequency of each frequency band according to the total number of frequency bands, the maximum frequency and the minimum frequency; Dividing the abnormal signal interval into multiple frequency interval segments according to each center frequency, obtaining the second amplitude of each frequency interval segment, obtaining the energy density of the corresponding frequency interval segment according to each second amplitude, and obtaining the abnormal frequency energy of the corresponding frequency according to each energy density; Obtaining multiple historical frequency energies in the historical normal signal of the OLED mobile phone screen, obtaining the frequency energy abnormality degree according to the multiple historical frequency energies and abnormal frequency energies, and obtaining the vibration abnormality degree according to the frequency energy abnormality degree and the frequency deviation degree.

[0009] Preferably, the step of obtaining the corresponding gradient amplitude according to each spectral data and determining the abnormal pixel points and normal pixel points according to the gradient amplitude includes: Obtaining the left pixel value and the right pixel value of the corresponding pixel point in the horizontal direction according to each spectral data, and obtaining the corresponding horizontal gradient according to each left pixel value and right pixel value; Obtaining the upper pixel value and the lower pixel value of the corresponding pixel point in the vertical direction according to each spectral data, and obtaining the corresponding vertical gradient according to each upper pixel value and lower pixel value; Obtaining the corresponding gradient amplitude according to each horizontal gradient and vertical gradient, and judging whether the gradient amplitude is greater than a preset gradient; If the gradient amplitude is greater than the preset gradient, determining that the pixel point corresponding to the gradient amplitude is an abnormal pixel point; If the gradient amplitude is not greater than the preset gradient, determining that the pixel point corresponding to the gradient amplitude is a normal pixel point.

[0010] Preferably, the step of obtaining the defect degree value according to the multiple parameter features includes: Obtaining the pixel area and multiple first abnormal principal component feature values of the corresponding abnormal pixel points according to each parameter feature; Obtaining the total defect area of the straight defect position in the OLED mobile phone screen, and obtaining the proportion of abnormal pixel defects according to the total defect area and multiple pixel areas; Obtaining the corresponding abnormal average feature value according to the multiple abnormal principal component feature values of each abnormal pixel point; Obtain multiple historical normal principal component eigenvalues corresponding to normal pixels of each same abnormal pixel in the historical normal spectral images of the OLED mobile phone screen, and obtain the corresponding normal average eigenvalue according to the multiple historical normal principal component eigenvalues of each normal pixel; Obtain the proportion of abnormal feature differences according to the multiple normal average eigenvalues and the abnormal average eigenvalue, and obtain the defect degree value according to the proportion of abnormal feature differences and the proportion of abnormal pixel defects.

[0011] This application also provides an OLED mobile phone screen linear defect detection system, including: A first acquisition module, configured to acquire multiple vibration time-domain signals within a preset time period on the surface of the OLED mobile phone screen, and perform Fourier transform processing on each of the vibration time-domain signals to obtain the corresponding vibration frequency-domain signal; A calibration module, configured to synchronously align and calibrate the multiple vibration frequency-domain signals to obtain multiple synchronous vibration signals; A second acquisition module, configured to obtain an abnormal signal interval according to the multiple synchronous vibration signals, and obtain the vibration abnormality degree according to the abnormal signal interval; A judgment module, configured to judge whether the vibration abnormality degree is greater than a preset abnormality degree; If the vibration abnormality degree is greater than the preset abnormality degree, it is determined that there is a linear defect in the OLED mobile phone screen, and the linear defect position of the OLED mobile phone screen is obtained according to the abnormal signal interval; A third acquisition module, configured to acquire a spectral image of the linear defect position in the OLED mobile phone screen, and acquire spectral data of each pixel point according to the spectral image; A determination module, configured to obtain the corresponding gradient amplitude according to each spectral data, and determine abnormal pixel points and normal pixel points according to the gradient amplitude; A fourth acquisition module, configured to acquire parameter characteristics of each abnormal pixel point, and obtain a defect degree value according to the multiple parameter characteristics.

[0012] Preferably, the fourth acquisition module includes: A first acquisition unit, configured to acquire the pixel area and multiple first abnormal principal component eigenvalues of the corresponding abnormal pixel point according to each parameter characteristic; A second acquisition unit, configured to acquire the total defect area of the linear defect position in the OLED mobile phone screen, and obtain the proportion of abnormal pixel defects according to the total defect area and multiple pixel areas; A third acquisition unit, configured to acquire the corresponding abnormal average eigenvalue according to the multiple abnormal principal component eigenvalues of each abnormal pixel point; A fourth acquisition unit, configured to acquire multiple historical normal principal component eigenvalues corresponding to normal pixel points of each same abnormal pixel point in the historical normal spectral images of the OLED mobile phone screen, and acquire corresponding normal average eigenvalues according to the multiple historical normal principal component eigenvalues of each normal pixel point; A fifth acquisition unit, configured to acquire an abnormal feature difference ratio according to the multiple normal average eigenvalues and the abnormal average eigenvalue, and acquire a defect degree value according to the abnormal feature difference ratio and the abnormal pixel defect ratio.

[0013] The present invention further provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned straight-line defect detection method for the OLED mobile phone screen are implemented.

[0014] The present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned straight-line defect detection method for the OLED mobile phone screen are implemented.

[0015] The beneficial effects of the present invention are as follows: By acquiring the vibration time-domain signal on the screen surface within a preset time period, the present invention does not depend on the display brightness of the screen, ensuring the stability and reliability of detection. By analyzing the vibration signal for detection, the acquisition of the vibration signal is not affected by the change in ambient light intensity. By performing Fourier transform on multiple vibration signals, high-frequency information can be obtained, accurately capturing the subtle anomalies in the vibration, improving the sensitivity of detection. By acquiring multiple synchronous vibration frequency-domain signals and calculating the vibration abnormality degree, a more objective and quantitative defect measurement standard can be provided. By obtaining the defect position through the abnormal signal interval, it does not depend on the display content and screen brightness, thus avoiding misjudgment caused by uneven display image brightness or ambient light change. Using spectral images to capture the spectral data of each pixel can effectively eliminate the influence of ambient light and display brightness. By analyzing the defect through the gradient amplitude of the spectral data, misjudgment caused by light change and screen brightness difference can be reduced. By analyzing the gradient amplitude to distinguish abnormal pixels from normal pixels, the subtle differences between pixels can be more accurately identified. By acquiring the parameter characteristics of each abnormal pixel, the characteristics of the defect can be more comprehensively described, thereby improving the accuracy of defect recognition. By evaluating the defect degree through multiple parameter characteristics, the defect characteristics in different dimensions can be comprehensively considered, thereby obtaining a more accurate and reliable defect degree evaluation. Description of the Drawings

[0016] Figure 1 It is a schematic flowchart of the method according to an embodiment of the present invention.

[0017] Figure 2 It is a schematic structural diagram of the system according to an embodiment of the present invention.

[0018] Figure 3 Schematic diagram of the internal structure of a computer device according to an embodiment of the present application.

[0019] The realization, functional characteristics and advantages of the present invention will be further described with reference to the embodiments and the accompanying drawings. Specific embodiments

[0020] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0021] As Figures 1 - 3 shown, the present application provides a method for detecting linear defects of an OLED mobile phone screen, including: S1. Obtain a plurality of vibration time-domain signals on the surface of the OLED mobile phone screen within a preset time period, and perform Fourier transform processing on each of the vibration time-domain signals to obtain corresponding vibration frequency-domain signals; S2. Synchronously align and calibrate the plurality of vibration frequency-domain signals to obtain a plurality of synchronous vibration signals; S3. Obtain an abnormal signal interval according to the plurality of synchronous vibration signals, and obtain a vibration abnormality degree according to the abnormal signal interval; S4. Determine whether the vibration abnormality degree is greater than a preset abnormality degree; If the vibration abnormality degree is greater than the preset abnormality degree, it is determined that there is a linear defect in the OLED mobile phone screen, and the linear defect position of the OLED mobile phone screen is obtained according to the abnormal signal interval; S5. Obtain a spectral image of the linear defect position in the OLED mobile phone screen, and obtain spectral data of each pixel point according to the spectral image; S6. Obtain a corresponding gradient amplitude according to each spectral data, and determine abnormal pixel points and normal pixel points according to the gradient amplitude; S7. Obtain parameter characteristics of each abnormal pixel point, and obtain a defect degree value according to the plurality of parameter characteristics.

[0022] As described in the above steps S1 - S7, the present invention obtains multiple vibration time - domain signals on the surface of the OLED mobile phone screen within a preset time period, performs Fourier transform processing on each vibration time - domain signal to obtain the corresponding vibration frequency - domain signal, synchronously aligns and calibrates the multiple vibration frequency - domain signals to obtain multiple synchronous vibration signals, that is, synchronously aligns and calibrates the vibration frequency - domain signals on the surface of the OLED mobile phone screen at different time points to the same time point, so as to eliminate the time deviation between signals and ensure the synchronism of vibration signals. An abnormal signal interval is obtained through the multiple synchronous vibration signals, and the vibration abnormality degree is obtained according to the abnormal signal interval. It is judged whether the vibration abnormality degree is greater than the preset abnormality degree. If the vibration abnormality degree is greater than the preset abnormality degree, it is determined that there is a linear defect in the OLED mobile phone screen. Traditional methods for detecting linear defects in OLED screens usually analyze the display image of the screen, which highly depends on the screen brightness and display content. If the display brightness is unstable, or due to changes in ambient light (such as light reflection, changes in ambient light intensity, etc.), it will interfere with the detection effect, resulting in inaccurate defect detection results. However, the present invention obtains the vibration time - domain signals on the screen surface within a preset time period. The vibration time - domain signal refers to the change process of the vibration signal recorded on the time axis. The vibration time - domain signal can reflect the intensity or amplitude of the object's vibration, helping to judge whether the object has excessive vibration and whether there are abnormal vibration behaviors (such as linear defects in the OLED mobile phone screen). Therefore, by obtaining the vibration time - domain signals, it does not depend on the screen display brightness, and the acquisition of the vibration time - domain signal has nothing to do with the screen display content. Therefore, the influence of external ambient light on the detection is avoided, ensuring the stability and reliability of the detection. By analyzing the vibration time - domain signals for detection, the acquisition of vibration signals is not affected by changes in ambient light intensity. Therefore, no matter how the ambient light conditions change, it will not affect the detection results. This enables the present invention to maintain high accuracy under different lighting conditions and has better adaptability. By performing Fourier transform processing on multiple vibration time - domain signals, the high - frequency information in the vibration frequency - domain signals can be obtained, accurately capturing the subtle abnormalities in the vibration. Since vibration abnormalities are directly related to physical defects (such as linear defects), this method can more accurately identify tiny defects, avoiding the limitations of traditional image methods and improving the detection sensitivity. By obtaining multiple synchronous vibration signals and calculating the vibration abnormality degree according to the abnormal signal interval obtained from the multiple synchronous vibration signals, this is because during the acquisition process of vibration signals, due to different positions of sensors, acquisition times, or other external factors (such as environmental noise, small errors of equipment, etc.), each vibration signal may have a time deviation. For example, signals from different sensors may be slightly advanced or delayed, resulting in inconsistent time alignment between them. If these signals are not synchronously aligned and directly combined for analysis, it may mislead the understanding of the vibration pattern and affect the accuracy of defect detection. When multiple vibration signals are synchronously aligned,It can more clearly reveal the vibration mode of the entire screen surface, thus helping to identify whether there are local defects or uneven vibration characteristics, and can provide a more objective and quantitative defect measurement standard. If the vibration abnormality exceeds a preset threshold, it can be very clearly determined that there is a linear defect in the OLED screen. This is because when multiple vibration signals are Fourier-transformed and synchronized and aligned, the changes in each vibration frequency-domain signal are analyzed to identify the parts different from the normal mode. This different part is the abnormal signal interval, which usually manifests as a frequency peak shift or peak increase, etc. When the vibration abnormality exceeds the preset threshold, it means that the degree of signal abnormality has reached a sufficient level, which usually corresponds to significant physical defects in the screen, especially linear defects. This is because linear defects will change the vibration mode, so the amplitude of vibration (vibration abnormality) will be significantly different from the vibration amplitude (vibration abnormality) of a normal screen. This quantitative abnormality detection method reduces the error of manual judgment, making the defect detection more accurate and reliable. By transforming the time-domain signal into a frequency-domain signal through Fourier transform, frequency-domain signal processing can effectively filter out noise and redundant information in the signal, improve the signal quality, making the subsequent abnormal signal analysis more accurate, and further improving the recognition accuracy of abnormal signals. Frequency-domain analysis can more clearly capture the frequency fluctuations caused by linear defects, further improving the detection accuracy and stability. Using automated vibration signal acquisition and frequency-domain analysis greatly improves the automation and intelligence level of the detection process. By setting a threshold to judge whether there are defects, the entire detection process can be completed efficiently and accurately, reducing human intervention and improving the detection efficiency of the production line, thus improving the quality control level of OLED mobile phone screens, and obtaining the linear defect position of the OLED mobile phone screen according to the abnormal signal interval. This is because by analyzing the frequency-domain characteristics within the abnormal signal interval, the position where the abnormal signal appears on the screen can be determined. Specifically, the abnormality of the vibration frequency-domain signal usually concentrates in the defect area. Therefore, by combining the frequency distribution of these abnormal signals, the defect position can be inferred. For example, linear defects usually appear as vibration abnormal sections parallel to the screen edge or center on the surface of the OLED mobile phone screen, and the volatility of the abnormal signal on the straight line may be relatively consistent. Therefore, the position of this linear defect can be calculated through the abnormal signal interval. By obtaining the spectral image of the linear defect position in the OLED mobile phone screen and obtaining the spectral data of each pixel point according to the spectral image, obtaining the corresponding gradient amplitude according to each spectral data, and determining abnormal pixel points and normal pixel points according to the gradient amplitude, obtaining the parameter characteristics of each abnormal pixel point, and obtaining the defect degree value according to multiple parameter characteristics, obtaining the defect position through the abnormal signal interval, independent of the display content and screen brightness, thus avoiding misjudgment caused by uneven display image brightness or ambient light changes. Capturing the spectral data of each pixel using the spectral image can effectively exclude the influence of ambient light and display brightness.Spectral images can provide more information than ordinary images, such as reflectivity and absorption characteristics in different bands, making detection more accurate and robust. By obtaining the spectral data of each pixel point, more information can be obtained from multiple wavelength ranges, enabling the identification of defects to be unaffected by changes in ambient light and the content displayed on the screen. The spectral data itself is not easily disturbed by changes in external conditions and can provide more stable and accurate detection results. Existing methods may rely on brightness gradients or color contrasts, but these data are greatly affected by the content displayed on the screen and its brightness changes, resulting in inaccuracies during the detection process. Analyzing defects through the gradient magnitude of spectral data can reduce false judgments caused by light changes and screen brightness differences. The gradient magnitude reflects the changes between pixels, rather than just the brightness or color itself, and can therefore maintain high accuracy under various lighting conditions. Vibration signal analysis can provide direct information about the physical structure of the screen, especially the dynamic response of the screen during operation. When a linear defect occurs in an OLED screen, the defective area may physically cause changes in the vibration mode. By performing a Fourier transform on the vibration signal of the screen and synchronously calibrating multiple signals, the location of the defect can be quickly determined by calculating the vibration abnormality degree in the abnormal signal interval. However, the processing of spectral images involves data from multiple bands, resulting in a high computational complexity. If spectral data is directly used for full-screen detection at the initial stage, the processing speed may be affected, and each pixel of the spectral data has information on multiple wavelengths. The extraction, analysis, and post-processing of this information require a large amount of computational resources. By first quickly determining the abnormal location through methods such as vibration time-domain signals and then using spectral images to accurately and stably detect the degree of abnormal defects, the linear defect detection of OLED mobile phone screens can be achieved more quickly and accurately. In traditional technologies, the determination of abnormal pixels usually relies on the brightness contrast and color differences of static images, but these may produce large deviations in different environments, resulting in the inability to accurately detect abnormal pixels. By analyzing the gradient magnitude to distinguish abnormal pixels from normal pixels, the subtle differences between pixels can be more accurately identified, avoiding errors based solely on brightness or color. Even in an environment with unstable lighting conditions, the analysis of the gradient magnitude can reliably identify the defective area. By obtaining the parameter characteristics of each abnormal pixel, the characteristics of the defect can be more comprehensively described, thereby improving the accuracy of defect identification. These parameter characteristics not only consider the brightness changes of the pixels but also cover more spectral and physical properties, enabling better capture and analysis of the abnormal area. By evaluating the degree of the defect through multiple parameter characteristics, the defect characteristics in different dimensions can be comprehensively considered, resulting in a more accurate and reliable assessment of the degree of the defect. This can not only determine the existence of the defect but also accurately reflect the severity of the defect, providing more comprehensive quality control.

[0023] In one embodiment, step S2 of synchronously aligning and calibrating the multiple vibration frequency domain signals to obtain multiple synchronous vibration signals includes: S21. Using a high-pass filter to remove the DC component in each vibration frequency domain signal to obtain multiple dynamic vibration signals, and sequentially performing denoising and normalization processing on each dynamic vibration signal to obtain corresponding standard vibration signals; S22. Sorting the multiple standard vibration signals in chronological order, marking the standard vibration signal ranked first as the reference vibration signal, and marking the remaining standard vibration signals as calibration vibration signals; S23. Obtaining a first initial time point of the reference vibration signal and a second initial time point of each calibration vibration signal, and obtaining corresponding lag values according to the first initial time point and each second initial time point; S24. Obtaining corresponding cross-correlation coefficients according to the reference vibration signal, each lag value, and the corresponding calibration vibration signal, and performing translation on the corresponding calibration vibration signal on the time axis according to each cross-correlation coefficient to obtain multiple synchronous vibration signals.

[0024] As described in the above steps S21 - S24, the present invention obtains a plurality of dynamic vibration signals by removing the DC components in each vibration frequency domain signal using a high - pass filter. By sequentially denoising and normalizing each dynamic vibration signal, corresponding standard vibration signals are obtained. By sorting the plurality of standard vibration signals in chronological order, the standard vibration signal ranked first is marked as the reference vibration signal, and the remaining standard vibration signals are marked as calibration vibration signals. By obtaining the first initial time point of the reference vibration signal and the second initial time point of each calibration vibration signal, and obtaining the corresponding lag value according to the first initial time point and each second initial time point. By using the reference vibration signal, each lag value and the corresponding calibration vibration signal to obtain the corresponding cross - correlation coefficient, and translating each calibration vibration signal on the time axis according to each cross - correlation coefficient to obtain a plurality of synchronous vibration signals. Removing the DC component through a high - pass filter can filter out low - frequency noise, making the dynamic part of the vibration signal clearer. This method makes the detection no longer rely on the display image itself, so it is not affected by the display brightness and ambient light, and can effectively avoid the decrease in detection accuracy caused by changes in environmental factors. By denoising and normalizing the vibration signal, the influence of noise can be effectively reduced and the amplitude of the signal can be unified, making the characteristics of each vibration signal more obvious and easy to analyze. The standardized signal can reduce the differences between the environment and the equipment, enhancing the stability and robustness of the detection algorithm. By sorting the vibration signals in time and determining the reference vibration signal and the calibration vibration signals, the time alignment of all signals in subsequent analysis can be ensured. This process can eliminate the time deviation between signals, ensuring the synchronization of vibration signals and improving the accuracy and reliability of defect detection. By accurately calculating the lag value and using it for time alignment, it can be ensured that all vibration signals are compared and analyzed in the same time frame. This precise time synchronization can eliminate the influence caused by signal delay or inconsistent transmission, thereby improving the accuracy of detection. By calculating the cross - correlation coefficient, the similarity between signals can be quantitatively measured. This method can effectively identify the time correlation and similarity between different vibration signals, making defect detection more accurate and avoiding misjudgment caused by changes in ambient light or image noise in traditional image detection. By translating the signals on the time axis, ensuring that all signals are synchronized to a common time standard, the comparability between signals is further improved. This synchronization method can accurately capture the vibration patterns related to defects, avoiding misjudgment caused by different time delays and improving the accuracy and reliability of defect identification.

[0025] In one embodiment, step S3 of obtaining the abnormal signal interval according to the plurality of synchronous vibration signals includes: S31. Obtain the first amplitude of each of the synchronous vibration signals at each time point, and perform weighted average processing on the multiple first amplitudes at each time point to obtain a composite vibration signal; S32. Obtain the first amplitude at each frequency point according to the composite vibration signal, and calculate the corresponding power spectral density according to each of the first amplitudes, where the calculation formula is: ; where G(PM) represents the power spectral density, F(DZ) represents the first amplitude, and C(XM) represents the number of frequency points; S33. Establish a frequency-density coordinate axis with frequency as the X-axis and power spectral density as the Y-axis, and plot the power spectral density corresponding to each frequency as a connection point on the frequency-density coordinate axis; S34. Connect the multiple connection points in sequence through a curve to obtain a power spectral waveform diagram, and obtain the peak value and valley value of the power spectral waveform diagram; S35. Divide the power spectral waveform diagram into multiple curves according to adjacent peaks and valleys, and obtain the curvature of each curve; S36. Determine whether the curvature is greater than a preset threshold; If the curvature is greater than the preset threshold, determine that the curve corresponding to the curvature is an abnormal curve, and obtain an abnormal signal interval according to the abnormal curve and the power spectral waveform diagram.

[0026] As described in the above steps S31 - S36, the calculation formulas of the power spectral density all perform normalization processing on the parameters of the first amplitude in advance to eliminate the dimensional differences between different variables. The purpose is to ensure that all variables are on the same order of magnitude, so as to make the calculation more stable and effective. The present invention obtains the first amplitude of each synchronous vibration signal at each time point, and performs weighted average processing on the multiple first amplitudes at each time point to obtain a composite vibration signal. The first amplitude at each frequency point is obtained through the composite vibration signal, and the corresponding power spectral density is calculated according to each first amplitude. By establishing a frequency - density coordinate axis with frequency as the X - axis and power spectral density as the Y - axis, and plotting the power spectral density corresponding to each frequency as a connection point on the frequency - density coordinate axis. By connecting multiple connection points in sequence through a curve to obtain a power spectral waveform diagram, and obtaining the peak value and valley value of the power spectral waveform diagram. The power spectral waveform diagram is divided into multiple curves by adjacent two peak values and valley values, and the curvature of each curve is obtained. By judging whether the curvature is greater than a preset threshold, if the curvature is greater than the preset threshold, it is determined that the curve corresponding to the curvature is an abnormal curve, and an abnormal signal interval is obtained according to the abnormal curve and the power spectral waveform diagram. Compared with the traditional image - based detection method, the vibration signal can be collected through sensors or devices, avoiding dependence on the brightness of the display screen. This means that no matter how the brightness of the displayed image changes, it will not affect the detection process, thus overcoming the interference problem of environmental light changes on the results in the traditional method. By weighted - averaging the amplitudes of multiple signals, the errors and noises caused by the fluctuations of a single signal can be reduced, making the final composite vibration signal more stable and reliable. This processing method reduces the fluctuations caused by external factors compared with the traditional method, making the detection result more accurate. By analyzing the frequency characteristics of the vibration signal, the state of the screen can be understood more comprehensively. Frequency analysis can more effectively identify potential defects than only relying on image brightness. Especially when dealing with environmental light changes, the stability of frequency characteristics can improve the robustness of detection. By establishing the relationship between frequency and power spectral density, the change trend and frequency components of the vibration signal can be clearly observed. This enables the position and nature of the defect to be more accurately located, rather than relying on the deviation that may occur in the image information due to environmental light or display brightness. By using the detection of peak values and valley values, the periodic changes of the signal can be further analyzed to discover potential abnormal fluctuations. Compared with the traditional image detection method, this frequency - domain analysis is more reliable because it is not directly affected by the displayed image, avoiding the errors generated in image analysis due to environmental light or display setting changes. By performing curvature analysis on the power spectral waveform diagram, the complexity of signal changes can be revealed, and normal signals and abnormal signals can be distinguished. This method can perform more detailed analysis on complex vibration signals, thus effectively distinguishing different types of defects and improving the accuracy of defect detection. By setting a curvature threshold, abnormal vibration characteristics can be automatically detected and a quick response can be made.This algorithm can effectively identify subtle defects that are difficult to capture in traditional image detection, such as minor display unevenness or structural damage, avoiding false detections and missed detections. By accurately locating the signal interval where the abnormal curve is located, the position and degree of the defect can be precisely identified. This makes the detection result not only stay at the level of whether there is a defect, but further provides detailed information about the defect, which helps subsequent analysis and processing.

[0027] In one embodiment, step S3 of obtaining the vibration abnormality degree according to the abnormal signal interval includes: S37. Obtain the abnormal frequency components of each frequency point according to the abnormal signal interval, and obtain the historical frequency components of each same frequency point in the historical normal signal of the OLED mobile phone screen; S38. Calculate the frequency deviation degree according to the multiple historical frequency components and abnormal frequency components, where the calculation formula is: ; Where P(PY) represents the frequency deviation degree, N represents the number of abnormal frequency components, i represents the serial number of the abnormal frequency component, Y(PC) i represents the i-th abnormal frequency component, and L(PC) i represents the i-th historical frequency component; S39. Obtain the total number of frequency bands, the maximum frequency, and the minimum frequency according to the abnormal signal interval, and obtain the center frequency of each frequency band according to the total number of frequency bands, the maximum frequency, and the minimum frequency; S310. Divide the abnormal signal interval into multiple frequency interval segments according to each center frequency, and obtain the second amplitude of each frequency interval segment; S311. Obtain the energy density of the corresponding frequency interval segment according to each second amplitude, and obtain the abnormal frequency energy corresponding to each energy density; S312. Obtain multiple historical frequency energies in the historical normal signal of the OLED mobile phone screen, and calculate the frequency energy abnormality degree according to the multiple historical frequency energies and abnormal frequency energies, where the calculation formula is: ; Where P(NY) represents the frequency energy abnormality degree, M represents the number of abnormal frequency energies, m represents the serial number of the abnormal frequency energy, Y(PN) m represents the m-th abnormal frequency energy, and L(PN) m represents the m-th historical frequency energy; S313. Obtain the vibration abnormality degree according to the frequency energy abnormality degree and the frequency deviation degree.

[0028] As described in the above steps S37 - S313, the calculation formulas of the frequency energy abnormality degree perform normalization processing on the two parameters of the historical frequency energy and the abnormal frequency energy in advance to eliminate the dimensional differences between different variables. The purpose is to ensure that all variables are on the same order of magnitude, so as to make the calculation more stable and effective. Here, the abnormal frequency component refers to the change or abnormality of the signal at a specific frequency point, and its unit is Hertz (Hz), and the unit of the frequency deviation degree is also Hertz (Hz). For example, if the historical frequency components are (10, 20, 30) and the abnormal frequency components are (12, 19, 28), substituting them into the above calculation formula of the frequency deviation degree, it can be known that , in the prior art, the calculation of the frequency offset degree is usually the difference between the actual frequency and the reference frequency, that is, the frequency offset degree = actual frequency - reference frequency. However, only considering a single frequency offset will bring errors to the detection result. In the present invention, the overall average frequency offset degree calculated from the historical frequency components and abnormal frequency components corresponding to multiple frequency points is mainly calculated by the mean square deviation of the frequency components to calculate the offset degree compared with the calculation method in the prior art. It can be averaged at multiple frequency points, thereby providing an overall and comprehensive error assessment. In a multi-frequency signal, the calculated frequency offset degree is more accurate because it takes into account the deviation of each frequency component, rather than just a single frequency deviation. The present invention obtains the abnormal frequency component of each frequency point through the abnormal signal interval, and obtains the historical frequency component of each same frequency point in the historical normal signal of the OLED mobile phone screen. The frequency offset degree is calculated through multiple historical frequency components and abnormal frequency components. The total frequency band number, the maximum frequency, and the minimum frequency are obtained through the abnormal signal interval, and the center frequency of each frequency band is obtained according to the total frequency band number, the maximum frequency, and the minimum frequency. The abnormal signal interval is divided into multiple frequency interval segments through each center frequency, and the second amplitude of each frequency interval segment is obtained. The energy density of the corresponding frequency interval segment is obtained through each second amplitude, and the abnormal frequency energy corresponding to each energy density is obtained. By obtaining multiple historical frequency energies in the historical normal signal of the OLED mobile phone screen, and calculating the frequency energy abnormality degree according to the multiple historical frequency energies and the abnormal frequency energy. The vibration abnormality degree is obtained through the frequency energy abnormality degree and the frequency offset degree. By analyzing based on the frequency components of the signal, it is possible to bypass the influence of the display brightness and external light sources, and directly extract abnormal information from the signal characteristics. In this way, the abnormal analysis of the signal is more accurate, and the interference of external factors such as ambient light can be eliminated. By combining with the historical normal signal for comparison, a benchmark for signal change can be established, with higher recognition accuracy. This approach can identify long-term offset trends through comparison of historical data, further reducing the probabilities of false positives (false alarms) and false negatives (missed reports). Through the calculation of the frequency offset degree, the deviation between the signal and the historical normal signal can be accurately measured. This method can reveal the law of frequency change, help identify even subtle signal changes or local abnormalities, and provide a more comprehensive and accurate defect detection ability. By dividing and analyzing the signal frequency band, the frequency distribution of the signal can be more clearly understood, and the frequency interval that may cause defects can be identified. This frequency domain analysis can capture problems more precisely, especially those that may not be detected by image detection. By calculating the center frequency of each frequency band, the characteristics of each frequency band can be clarified, and precise signal segmentation can be provided for subsequent analysis. This method makes the abnormal points of the signal easier to identify and reduces the interference of noise signals. By dividing the frequency interval segments, the characteristics and changes of each frequency segment can be finely analyzed, which makes the detection more refined.And it can locate the problem areas in specific frequency bands, providing higher detection accuracy. By analyzing the energy density of each frequency interval, the intensity distribution of abnormal signals can be deeply understood. Energy density analysis provides a quantitative standard to help detect whether there is abnormal frequency energy aggregation, further improving the accuracy of defect identification. Energy density is one of the important indicators of frequency signal abnormality. Through energy calculation, the abnormal frequency energy of the signal can be more directly captured, thus identifying subtle abnormal signals. The advantage of this method lies in its high degree of quantification, which can provide precise abnormal measurement. By comparing the historical frequency energy with the current abnormal frequency energy, the trend and law of signal change can be revealed, helping to identify long-term abnormalities. By calculating the frequency energy abnormality degree, the abnormal degree of the signal can be quantitatively evaluated, thus not only considering the deviation of a single frequency, but also comprehensively evaluating the change of the overall frequency energy, enhancing the reliability and comprehensiveness of the detection result. Combining the frequency deviation degree and the frequency energy abnormality degree can comprehensively evaluate the overall abnormal condition of the signal, accurately detecting potential vibration abnormalities. This multi-dimensional abnormal measurement method can effectively improve the detection accuracy and reduce false alarms and missed detections.

[0029] In one embodiment, step S6 of obtaining the corresponding gradient magnitude according to each of the spectral data and determining the abnormal pixel points and normal pixel points according to the gradient magnitude includes: S61. Obtain the left pixel value and the right pixel value of the corresponding pixel point in the horizontal direction according to each of the spectral data, and obtain the corresponding horizontal gradient according to each of the left pixel value and the right pixel value; S62. Obtain the upper pixel value and the lower pixel value of the corresponding pixel point in the vertical direction according to each of the spectral data, and obtain the corresponding vertical gradient according to each of the upper pixel value and the lower pixel value; S63. Obtain the corresponding gradient magnitude according to each of the horizontal gradient and the vertical gradient; S64. Determine whether the gradient magnitude is greater than a preset gradient; If the gradient magnitude is greater than the preset gradient, determine that the pixel point corresponding to the gradient magnitude is an abnormal pixel point; If the gradient magnitude is not greater than the preset gradient, determine that the pixel point corresponding to the gradient magnitude is a normal pixel point.

[0030] As described in the above steps S61 - S64, the present invention obtains the left pixel value and the right pixel value of the corresponding pixel point in the horizontal direction through each spectral data, and obtains the corresponding horizontal gradient according to each left pixel value and right pixel value. It obtains the upper pixel value and the lower pixel value of the corresponding pixel point in the vertical direction through each spectral data, and obtains the corresponding vertical gradient according to each upper pixel value and lower pixel value. It obtains the corresponding gradient magnitude through each horizontal gradient and vertical gradient, and determines whether the gradient magnitude is greater than a preset gradient. If the gradient magnitude is greater than the preset gradient, it is determined that the pixel point corresponding to the gradient magnitude is an abnormal pixel point. If the gradient magnitude is not greater than the preset gradient, it is determined that the pixel point corresponding to the gradient magnitude is a normal pixel point. In the existing technology, the detection of straight line defects on the OLED screen usually depends on the displayed image content, such as test patterns or specific images, which makes the results affected by factors such as ambient light and brightness, resulting in unstable detection. By using spectral data to obtain pixel values and calculate gradients, the dependence on the screen display content can be reduced, and the interference caused by changes in image brightness or content can be avoided. The gradient calculation in the vertical direction also avoids the influence of the screen content or display brightness. The gradient calculations in the vertical and horizontal directions can comprehensively consider the local changes of pixels without depending on the specific display content, which provides more stable and consistent features for detection. Regardless of how the image content changes, the algorithm can maintain a high accuracy under different ambient lights. The calculation of the gradient magnitude combines the information in the horizontal and vertical directions, providing a more comprehensive pixel change situation. The existing technology may only rely on gradients in a single direction or simple image content detection, ignoring the information in other directions. By combining horizontal and vertical gradients, abnormal changes in pixel points can be better captured, improving the robustness and accuracy of detection, especially being able to more precisely identify abnormal points when dealing with straight line defects. By setting the preset gradient value, the sensitivity of detection can be flexibly controlled and the detection results can be optimized. This method can adjust the preset threshold according to the needs of actual applications, so as to adapt to different defect detection standards. However, the existing technology depends on the display content and brightness changes of the image and cannot flexibly adjust the detection standards. Therefore, under different environmental conditions, the accuracy and consistency of detection may be affected. By using the gradient magnitude as the determination basis, the detection process is more objective and stable, reducing the influence of human intervention or image brightness changes. The magnitude of the gradient magnitude can effectively distinguish normal pixels from abnormal pixels. Abnormal pixels usually show a large gradient difference from the surrounding pixels. Therefore, identifying abnormal pixels through the gradient magnitude can more accurately identify straight line defects. The advantage of this method is that through digital gradient calculation, the dependence on the display content is avoided, and defects can be stably identified without being interfered by changes in ambient light or screen content. Traditional methods often rely on brightness differences or image content, which will be affected by factors such as changes in ambient light and screen brightness adjustment, thus affecting the accuracy of detection.By relying not on image content and brightness but instead on spectral data and pixel gradient information, the detection process becomes more stable and consistent, capable of overcoming interference from ambient light, screen brightness, or content changes. This means that regardless of how the ambient light changes or what the image content displayed on the screen is, the effect of defect detection remains reliable, thereby improving the accuracy and practicality of defect detection.

[0031] In one embodiment, step S7 of obtaining a defect degree value according to the plurality of parameter features includes: S71. Obtain the pixel area of the corresponding abnormal pixel point and a plurality of first abnormal principal component feature values according to each of the parameter features; S72. Obtain the total defect area at the position of the linear defect in the OLED mobile phone screen, and obtain the proportion of abnormal pixel defects according to the total defect area and the plurality of pixel areas; S73. Obtain the corresponding abnormal average feature value according to the plurality of abnormal principal component feature values of each abnormal pixel point; S74. Obtain the plurality of historical normal principal component feature values of the corresponding normal pixel points of each same abnormal pixel point in the historical normal spectral image of the OLED mobile phone screen, and obtain the corresponding normal average feature value according to the plurality of historical normal principal component feature values of each normal pixel point; S75. Obtain the proportion of abnormal feature differences according to the plurality of normal average feature values and abnormal average feature values, and obtain the defect degree value according to the proportion of abnormal feature differences and the proportion of abnormal pixel defects.

[0032] As described in the above steps S71 - S75, the present invention obtains the pixel area and multiple first abnormal principal component eigenvalues of the corresponding abnormal pixel points through each parameter feature, obtains the total defect area of the straight - line defect positions in the OLED mobile phone screen, and obtains the proportion of abnormal pixel defects according to the total defect area and multiple pixel areas. The corresponding abnormal average eigenvalue is obtained through the multiple abnormal principal component eigenvalues of each abnormal pixel point. The multiple historical normal principal component eigenvalues of the corresponding normal pixel points of each same abnormal pixel point in the historical normal spectral image of the OLED mobile phone screen are obtained, and the corresponding normal average eigenvalue is obtained according to the multiple historical normal principal component eigenvalues of each normal pixel point. The proportion of abnormal feature differences is obtained through the multiple normal average eigenvalues and abnormal average eigenvalues, and the defect degree value is obtained according to the proportion of abnormal feature differences and the proportion of abnormal pixel defects. By extracting the features of each pixel point and analyzing its principal component eigenvalues, the over - reliance on screen brightness can be avoided, and the detection accuracy can still be maintained under different ambient light conditions. In this way, the abnormal degree of each abnormal pixel point can be judged more precisely, and the stability of detection can be improved. By measuring the total area of the straight - line defect, the distribution of the defect can be comprehensively understood. For high - resolution display screens such as OLED screens, the comprehensiveness of the detection method can improve the accuracy of defect recognition. Especially in the case of large - area defects or multiple defects coexisting, the overall situation can be better judged. By calculating the proportion of abnormal pixels, the severity of the defect can be quantitatively evaluated, rather than relying solely on the recognition of the defect position. Compared with traditional methods, it can provide a more quantitative result, which can be flexibly adjusted and evaluated in different situations, avoiding the deviation caused by simply relying on manual experience. By extracting multiple principal components of abnormal pixels and calculating their average eigenvalues, the characteristics of abnormal pixels can be analyzed more comprehensively. Especially for defects with irregular shapes or difficult to judge by traditional brightness, it can provide more accurate abnormal recognition ability. Introducing the comparative analysis of historical normal spectral images greatly improves the accuracy of abnormal detection. Traditional methods usually cannot compare with the normal state, but only judge defects based on real - time display images. With the support of historical normal data, the currently detected abnormal pixels can be compared with the historical normal state, further improving the recognition accuracy of abnormal patterns and reducing the risk of misjudgment or missed judgment. It can not only extract detailed feature information from the historical data of normal pixel points, but also, by calculating the average eigenvalue in the normal state, enable the detection algorithm to still accurately grasp the normal state in the face of ambient light changes. Compared with existing methods, this comparison based on historical data makes the detection system more sensitive and has higher robustness. By calculating the proportion of differences between normal and abnormal features, the nature and severity of the defect can be evaluated more carefully. This method can more accurately reflect the health status of the screen, reduce errors, and provide a more effective basis for subsequent fault diagnosis.By comprehensively considering the proportion of abnormal feature differences and the proportion of pixel defects, the finally obtained defect degree value is a comprehensive quantitative result. Compared with traditional methods, this detection method that comprehensively considers multiple dimensions can comprehensively evaluate the actual impact of defects. Especially in complex situations where multiple factors interact, this method can provide a more accurate defect assessment, providing a scientific basis for subsequent quality control and maintenance.

[0033] This application also provides an OLED mobile phone screen linear defect detection system, including: A first acquisition module, configured to acquire multiple vibration time-domain signals on the surface of the OLED mobile phone screen within a preset time period, and perform Fourier transform processing on each of the vibration time-domain signals to obtain corresponding vibration frequency-domain signals; A calibration module, configured to synchronously align and calibrate the multiple vibration frequency-domain signals to obtain multiple synchronous vibration signals; A second acquisition module, configured to obtain an abnormal signal interval according to the multiple synchronous vibration signals, and obtain a vibration abnormality degree according to the abnormal signal interval; A judgment module, configured to judge whether the vibration abnormality degree is greater than a preset abnormality degree; If the vibration abnormality degree is greater than the preset abnormality degree, it is determined that there is a linear defect in the OLED mobile phone screen, and the linear defect position of the OLED mobile phone screen is obtained according to the abnormal signal interval; A third acquisition module, configured to acquire a spectral image of the linear defect position in the OLED mobile phone screen, and acquire spectral data of each pixel point according to the spectral image; A determination module, configured to obtain a corresponding gradient amplitude according to each spectral data, and determine abnormal pixel points and normal pixel points according to the gradient amplitude; A fourth acquisition module, configured to acquire parameter features of each abnormal pixel point, and acquire a defect degree value according to the multiple parameter features.

[0034] In one embodiment, the fourth acquisition module includes: A first acquisition unit, configured to acquire the pixel area of the corresponding abnormal pixel point and multiple first abnormal principal component feature values according to each parameter feature; A second acquisition unit, configured to acquire the total defect area of the linear defect position in the OLED mobile phone screen, and acquire the abnormal pixel defect proportion according to the total defect area and multiple pixel areas; A third acquisition unit, configured to acquire a corresponding abnormal average feature value according to the multiple abnormal principal component feature values of each abnormal pixel point; A fourth acquisition unit, configured to acquire multiple historical normal principal component eigenvalues of normal pixels corresponding to each same abnormal pixel in the historical normal spectral image of the OLED mobile phone screen, and acquire a corresponding normal average eigenvalue according to the multiple historical normal principal component eigenvalues of each normal pixel; A fifth acquisition unit, configured to acquire a proportion of abnormal feature difference according to the multiple normal average eigenvalues and the abnormal average eigenvalue, and acquire a defect degree value according to the proportion of abnormal feature difference and the proportion of abnormal pixel defects.

[0035] It should be noted that each module and unit in the straight-line defect detection system of the OLED mobile phone screen corresponds one by one to the steps in the straight-line defect detection method of the OLED mobile phone screen.

[0036] As Figure 3 shown, the present application further provides a computer device, which may be a server, and its internal structure may be as Figure 3 shown. The computer device includes a processor, a memory, a network interface, and a database connected through a system bus. Among them, the processor of the computer design is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store all data required for the process of the straight-line defect detection method of the OLED mobile phone screen. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements the straight-line defect detection method of the OLED mobile phone screen.

[0037] Those skilled in the art can understand that Figure 3 the structure shown in

[0038] is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied.

[0039] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium provided in this application and used in the embodiments can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0040] It should be noted that in this article, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements not only includes those elements but also includes other elements not explicitly listed, or further includes elements inherent to such process, device, article or method. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of another identical element in the process, device, article or method including that element.

[0041] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structure or equivalent process transformation made by using the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. A method for detecting linear defects in an OLED mobile phone screen, characterized in that, Including: Obtain a plurality of vibration time-domain signals on the surface of the OLED mobile phone screen within a preset time period, and perform Fourier transform processing on each of the vibration time-domain signals to obtain corresponding vibration frequency-domain signals; Synchronously align and calibrate the plurality of vibration frequency-domain signals to obtain a plurality of synchronous vibration signals; Obtain an abnormal signal interval according to the plurality of synchronous vibration signals, and obtain a vibration abnormality degree according to the abnormal signal interval; Judge whether the vibration abnormality degree is greater than a preset abnormality degree; If the vibration abnormality degree is greater than the preset abnormality degree, it is determined that there is a linear defect in the OLED mobile phone screen, and the linear defect position of the OLED mobile phone screen is obtained according to the abnormal signal interval; Obtain a spectral image of the linear defect position in the OLED mobile phone screen, and obtain spectral data of each pixel point according to the spectral image; Obtain a corresponding gradient amplitude according to each spectral data, and determine abnormal pixel points and normal pixel points according to the gradient amplitude; Obtain parameter characteristics of each abnormal pixel point, and obtain a defect degree value according to the plurality of parameter characteristics.

2. The method for detecting linear defects of an OLED mobile phone screen according to claim 1, wherein The step of synchronously aligning and calibrating the plurality of vibration frequency-domain signals to obtain a plurality of synchronous vibration signals includes: Use a high-pass filter to remove the DC component in each vibration frequency-domain signal to obtain a plurality of dynamic vibration signals, and perform denoising and normalization processing on each dynamic vibration signal in turn to obtain corresponding standard vibration signals; Sort the plurality of standard vibration signals in chronological order, mark the first standard vibration signal as a reference vibration signal, and mark the remaining standard vibration signals as calibration vibration signals; Obtain a first initial time point of the reference vibration signal and a second initial time point of each calibration vibration signal, and obtain corresponding lag values according to the first initial time point and each second initial time point; Obtain corresponding cross-correlation coefficients according to the reference vibration signal, each lag value and the corresponding calibration vibration signal, and perform time-axis translation on the corresponding calibration vibration signal according to each cross-correlation coefficient to obtain a plurality of synchronous vibration signals.

3. The method for detecting linear defects of an OLED mobile phone screen according to claim 1, characterized in that, The step of obtaining an abnormal signal interval according to the plurality of synchronous vibration signals includes: Obtain a first amplitude of each synchronous vibration signal at each time point, and perform weighted average processing on the plurality of first amplitudes at each time point to obtain a composite vibration signal; Obtain a first amplitude at each frequency point according to the composite vibration signal, and obtain a corresponding power spectral density according to each first amplitude; Establish a frequency-density coordinate axis with frequency as the X axis and power spectral density as the Y axis, and plot the power spectral density corresponding to each frequency as a connection point on the frequency-density coordinate axis; Connect the plurality of connection points in sequence through a curve to obtain a power spectral waveform diagram, and obtain the peak value and valley value of the power spectral waveform diagram; Divide the power spectral waveform diagram into multiple curves according to adjacent two peak values and valley values, obtain the curvature of each curve, and judge whether the curvature is greater than a preset threshold; If the curvature is greater than a preset threshold, it is determined that the curve corresponding to the curvature is an abnormal curve, and an abnormal signal interval is obtained based on the abnormal curve and the power spectrum waveform diagram.

4. The method for detecting linear defects of an OLED mobile phone screen according to claim 1, wherein The step of obtaining the vibration abnormality degree according to the abnormal signal interval includes: Obtaining the abnormal frequency component of each frequency point according to the abnormal signal interval, obtaining the historical frequency component of each same frequency point in the historical normal signal of the OLED mobile phone screen, and obtaining the frequency deviation degree according to the multiple historical frequency components and the abnormal frequency components; Obtaining the total number of frequency bands, the maximum frequency and the minimum frequency according to the abnormal signal interval, and obtaining the center frequency of each frequency band according to the total number of frequency bands, the maximum frequency and the minimum frequency; Dividing the abnormal signal interval into multiple frequency interval segments according to each center frequency, obtaining the second amplitude of each frequency interval segment, obtaining the energy density of the corresponding frequency interval segment according to each second amplitude, and obtaining the abnormal frequency energy corresponding to each energy density; Obtaining multiple historical frequency energies in the historical normal signal of the OLED mobile phone screen, obtaining the frequency energy abnormality degree according to the multiple historical frequency energies and the abnormal frequency energy, and obtaining the vibration abnormality degree according to the frequency energy abnormality degree and the frequency deviation degree.

5. The method for detecting linear defects of an OLED mobile phone screen according to claim 1, wherein The step of obtaining the corresponding gradient amplitude according to each spectral data and determining the abnormal pixel points and normal pixel points according to the gradient amplitude includes: Obtaining the left pixel value and the right pixel value of the corresponding pixel point in the horizontal direction according to each spectral data, and obtaining the corresponding horizontal gradient according to each left pixel value and right pixel value; Obtaining the upper pixel value and the lower pixel value of the corresponding pixel point in the vertical direction according to each spectral data, and obtaining the corresponding vertical gradient according to each upper pixel value and lower pixel value; Obtaining the corresponding gradient amplitude according to each horizontal gradient and vertical gradient, and determining whether the gradient amplitude is greater than a preset gradient; If the gradient amplitude is greater than the preset gradient, it is determined that the pixel point corresponding to the gradient amplitude is an abnormal pixel point; If the gradient amplitude is not greater than the preset gradient, it is determined that the pixel point corresponding to the gradient amplitude is a normal pixel point.

6. The method for detecting linear defects of an OLED mobile phone screen according to claim 1, wherein, The step of obtaining the defect degree value according to the multiple parameter features includes: Obtaining the pixel area and multiple first abnormal principal component feature values of the corresponding abnormal pixel points according to each parameter feature; Obtaining the total defect area of the straight defect position in the OLED mobile phone screen, and obtaining the abnormal pixel defect ratio according to the total defect area and multiple pixel areas; Obtaining the corresponding abnormal average feature value according to the multiple abnormal principal component feature values of each abnormal pixel point; Obtaining multiple historical normal principal component feature values of the corresponding normal pixel points of each same abnormal pixel point in the historical normal spectral image of the OLED mobile phone screen, and obtaining the corresponding normal average feature value according to the multiple historical normal principal component feature values of each normal pixel point; Obtaining the abnormal feature difference ratio according to the multiple normal average feature values and the abnormal average feature values, and obtaining the defect degree value according to the abnormal feature difference ratio and the abnormal pixel defect ratio.

7. An OLED mobile phone screen linear defect detection system, characterized in that, including: A first acquisition module, configured to acquire a plurality of vibration time-domain signals on the surface of the OLED mobile phone screen within a preset time period, and perform Fourier transform processing on each of the vibration time-domain signals to obtain corresponding vibration frequency-domain signals; A calibration module, configured to synchronously align and calibrate the plurality of vibration frequency-domain signals to obtain a plurality of synchronous vibration signals; A second acquisition module, configured to acquire an abnormal signal interval according to the plurality of synchronous vibration signals, and acquire a vibration abnormality degree according to the abnormal signal interval; A judgment module, configured to judge whether the vibration abnormality degree is greater than a preset abnormality degree; If the vibration abnormality degree is greater than the preset abnormality degree, it is determined that there is a linear defect in the OLED mobile phone screen, and the linear defect position of the OLED mobile phone screen is acquired according to the abnormal signal interval; A third acquisition module, configured to acquire a spectral image of the linear defect position in the OLED mobile phone screen, and acquire spectral data of each pixel point according to the spectral image; A determination module, configured to acquire a corresponding gradient amplitude according to each of the spectral data, and determine abnormal pixel points and normal pixel points according to the gradient amplitude; A fourth acquisition module, configured to acquire parameter characteristics of each of the abnormal pixel points, and acquire a defect degree value according to the plurality of parameter characteristics; 8. The OLED mobile phone screen linear defect detection system according to claim 7, characterized in that The fourth acquisition module includes: A first acquisition unit, configured to acquire a pixel area and a plurality of first abnormal principal component feature values of a corresponding abnormal pixel point according to each of the parameter characteristics; A second acquisition unit, configured to acquire a total defect area of the linear defect position in the OLED mobile phone screen, and acquire an abnormal pixel defect ratio according to the total defect area and the plurality of pixel areas; A third acquisition unit, configured to acquire a corresponding abnormal average feature value according to the plurality of abnormal principal component feature values of each abnormal pixel point; A fourth acquisition unit, configured to acquire a plurality of historical normal principal component feature values of the corresponding normal pixel points of each same abnormal pixel point in the historical normal spectral image of the OLED mobile phone screen, and acquire a corresponding normal average feature value according to the plurality of historical normal principal component feature values of each normal pixel point; A fifth acquisition unit, configured to acquire an abnormal feature difference ratio according to the plurality of normal average feature values and the abnormal average feature values, and acquire a defect degree value according to the abnormal feature difference ratio and the abnormal pixel defect ratio; 9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images