Sheet surface quality detection system

Abstract

The invention relates to the technical field of thin plate surface quality detection, and particularly discloses a thin plate surface quality detection system, which comprises a laser scanning measurement module, a laser image processing and recess recognition module and a recess positioning and size analysis module, the laser scanning measurement module judges the depression of the surface of the thin plate according to the contact change of a laser beam and the surface of the thin plate; the laser image processing and recess recognition module captures an image of a laser beam on the surface of the thin plate, extracts feature points in the image and calculates the displacement of the laser beam; and the recess positioning and size analysis module is used for calculating a specific numerical value of the surface recess of the ink printing sheet based on the displacement of the laser beam. According to the invention, through combination of laser beam scanning, image processing, recess detection, recess positioning and size analysis, high-precision detection of the surface recess of the ink printing sheet is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of thin plate surface quality detection, and more specifically, to a thin plate surface quality detection system. Background Art

[0002] When printing with ink, depressions may appear on the surface of the sheet, which affects the final print quality and product performance.

[0003] The existing open document 1 (Study on the Appearance Inspection and Quality Assessment Method of Lap Laser Welds of Stainless Steel Sheets, 2019) proposes a fully automatic scanning inspection device, which controls the laser displacement sensor and the scanning driver, performs data processing and graphical output of the inspection results, and the scanning driver drives the sensor bracket to move, realizing the two-dimensional positioning scanning of the laser displacement sensor on the plane. The device can realize the rapid detection and automatic evaluation of the surface profile of the laser weld, with a built-in dedicated data processing algorithm, which can automatically calculate the weld size, measure the benchmark adaptively, store the inspection data and generate a three-dimensional image of the weld morphology.

[0004] The surface profile of laser welds of various sizes can be detected, and the welding process and weld quality can be evaluated based on the profile changes. However, the device is easily affected by ambient temperature and stray light, and the detection accuracy will also be affected. Existing public document 2 (Research on Eddy Current Detection Technology for Thin Plate Resistance Welding Weld Defects, 2016.) discloses a weld quality detection system, which includes six modules: weld detection, signal processing, data acquisition, data analysis, display and alarm. The weld detection module uses the eddy current detection principle to detect and receive signals; the signal processing module amplifies weak signals and eliminates interference; the data acquisition module converts analog signals into digital signals; the data analysis module processes data and transmits the results to the display and alarm module to achieve online evaluation of weld quality. However, the system lacks the collection of a large amount of experimental and simulation data, and cannot accurately judge some complex or special weld defects, which affects the accuracy and reliability of the detection results.

[0005] Therefore, how to detect the depressions that occur in the production process of ink-printed thin plates in real time and with high precision, and accurately calculate the size of the depressions, has become an urgent problem to be solved in the manufacturing industry. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art, the present invention provides a thin plate surface quality detection system, which obtains the surface data of the thin plate by multi-point scanning using a laser, extracts feature points using the Soble operator, calculates the displacement of the laser beam, and calculates the depth, width and area of ​​the depression on the surface of the thin plate based on the displacement, so as to solve the problems raised in the above-mentioned background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A thin plate surface quality detection system includes a laser scanning measurement module, a laser image processing and concave recognition module, and a concave positioning and size analysis module; the laser scanning measurement module determines the concave on the thin plate surface by the change of the contact between the laser beam and the thin plate surface; the laser image processing and concave recognition module captures the image of the laser beam on the thin plate surface, extracts the feature points in the image, and calculates the displacement of the laser beam; the concave positioning and size analysis module calculates the specific value of the concave on the surface of the ink-printed thin plate based on the displacement of the laser beam;

[0009] The laser image processing and depression recognition module calculates the curvature of the laser beam trajectory using the curvature formula: Where κ is the curvature, x' represents the tangential direction of the laser beam trajectory along the x-axis, y' represents the tangential direction of the laser beam trajectory along the y-axis, x" represents the acceleration direction of the laser beam trajectory along the x-axis, and y" represents the acceleration direction of the laser beam trajectory along the y-axis. Fourier transform is used to analyze the changes of the laser beam at different frequencies, and the calculation formula is: Among them, F(u) is the frequency domain signal, f(x) is the coordinate point of the laser beam trajectory, and u is the frequency domain variable; the displacement of the laser beam is calculated by combining the curvature of the laser beam and the frequency domain analysis, and the calculation formula is: Where D is the total displacement of the laser beam from time t to time t+1, is the velocity component of the laser beam trajectory in the x direction, is the velocity component of the laser beam trajectory in the y direction, λ1 is the curvature influence coefficient, and λ2 is the frequency influence coefficient.

[0010] As a further solution of the present invention, the laser scanning measurement module determines the depression of the thin plate surface by the change of the contact between the laser beam and the thin plate surface, including the following specific contents: the laser scanning measurement module projects a high-precision laser beam on the surface of the ink-printed thin plate, and uses the change of the laser beam after reflection on the surface of the ink-printed thin plate to detect the slight deformation of the surface. During the detection process, a high-stability laser emits a finely collimated laser beam, and the wavelength of the laser beam is selected within the visible light or near-infrared light range to ensure that a clear reflection signal can be formed on the surface of the ink-printed thin plate of different materials. The laser beam is modulated by a group of high-precision optical lenses so that it can evenly cover the entire detection area and ensure that the size of the laser spot is suitable for high-precision measurement.

[0011] In order to improve the detection accuracy and reduce noise interference, the laser scanning measurement module adopts a multi-point scanning method, that is, the laser beam is scanned along the surface of the thin plate multiple times to obtain the surface data of the ink-printed thin plate in multiple directions, and construct a three-dimensional model of the depression on the surface of the ink-printed thin plate: the surface of the ink-printed thin plate is represented by a continuous surface function Z(x,y), where x is the horizontal coordinate on the surface of the ink-printed thin plate, and y is the vertical coordinate on the surface of the ink-printed thin plate. Z(x,y) represents the height of a certain point on the surface of the ink-printed thin plate. Ideally, if the surface of the thin plate is completely flat, then Z(x,y) is a plane function that satisfies Z0(x,y)=A0, where Z0 is the height distribution of the ideal plane ink-printed thin plate surface, and A0 is the average height of the ideal plane. However, due to the influence of the depression on the surface of the ink-printed thin plate, the actual surface will have deviations: Z(x,y)=Z0(x,y)+ΔZ(x,y), where ΔZ(x,y) represents the height change caused by defects, satisfying Where M is the order of Fourier expansion in the x direction, which indicates the complexity of the change of the ink printed thin plate surface in the horizontal direction, N is the order of Fourier expansion in the y direction, which indicates the complexity of the change of the ink printed thin plate surface in the vertical direction, and C mn represents the amplitude value of the depression, m represents the Fourier mode index in the x direction, which represents the undulation frequency of the surface along the x axis, n represents the Fourier mode index in the y direction, which represents the undulation frequency of the surface along the y axis, L x Indicates the length of the ink-printed sheet, L y Indicates the width of the ink-printed sheet, φ mn Indicates the phase shift of different notch waveforms.

[0012] In order to ensure the accuracy of the data, the laser scanning measurement module will also preprocess the collected laser reflection data. The preprocessing includes noise filtering, background light compensation, data smoothing and other operations to eliminate errors that may be caused by external light source interference or sensor noise.

[0013] As a further solution of the present invention, the laser image processing and depression recognition module captures the image of the laser beam on the surface of the thin plate, extracts the feature points in the image, and calculates the displacement of the laser beam, including the following specific contents: the laser image processing and depression recognition module uses an RGB camera to obtain the image of the laser beam on the surface of the ink-printed thin plate, and the RGB camera can capture the projection of the laser beam on the surface of the ink-printed thin plate with high resolution.

[0014] The image processing unit in the laser image processing and concave recognition module extracts the features of the laser beam from the image acquired by the RGB camera, and the specific steps are as follows:

[0015] Step Z1, using Gaussian filtering to reduce the noise in the image to avoid high-frequency noise interfering with edge detection, the calculation formula is: Among them, σ is the standard deviation of the Gaussian kernel, x is the horizontal coordinate of the current pixel relative to the center of the Gaussian kernel, and y is the vertical coordinate of the current pixel relative to the center of the Gaussian kernel.

[0016] Step Z2, use the Sobel operator to calculate the gradient of the image I in the x and y directions, and obtain the gradient magnitude and direction: Among them, G is the gradient image, G x is the horizontal gradient image, G y is the vertical gradient image, is the rate of change of the pixel value in image I in the x-axis direction, is the rate of change of the pixel value in image I in the y-axis direction.

[0017] In step Z3, by checking the adjacent pixel values ​​in the pixel gradient direction, only the local maximum is retained to refine the edge, and morphological processing is performed to further clean up the noise in the image.

[0018] Step Z4, according to the brightness and color information of the laser beam, the image is segmented into different areas using threshold segmentation, and the area of ​​the laser beam in the image is extracted.

[0019] The formula for calculating the image gradient by the Sobel operator is expressed as: x =g x *I,I y =g y *I, I represents the input image, I y Represents the gradient component of the image in the x direction, I y Represents the gradient component of image I in the y direction, g x is the first convolution kernel of the Sobel operator, which is used to calculate the horizontal gradient and satisfies: g y is the second convolution kernel of the Sobel operator, which is used to calculate the vertical gradient and satisfies

[0020] As a further solution of the present invention, the depression positioning and size analysis module calculates the specific value of the depression on the surface of the ink-printed thin plate based on the displacement of the laser beam, including the following specific contents: the depression positioning and size analysis module compares the laser beam scanning trajectory under normal conditions with the actual scanning trajectory, and finds that the area deviating from the normal trajectory is the location where the depression exists. The depression positioning and size analysis module establishes a depression positioning model, calculates the depth value, width value and area of ​​the depression on the surface of the ink-printed thin plate, and determines the center point position of the depression area by continuously scanning and comparing multiple laser beam positions. The specific steps are as follows:

[0021] Step W1, comparing the normal scanning trajectory of the laser beam with the actual scanning trajectory, and finding that the area deviating from the normal trajectory is the location where the depression exists.

[0022] Step W2, set a displacement threshold T d , used to determine whether a certain area is a concave area. If the laser beam displacement in a certain area is greater than T d , then it is considered that there is a depression in the area.

[0023] Step W3, by continuously scanning and comparing multiple laser beam positions, determine the center point position P of the concave area center =(X center ,Y center ), P center is the coordinate of the center point of the concave area, X center is the horizontal coordinate of the center point of the concave area, Y center is the vertical coordinate of the center point of the concave area.

[0024] After determining the center position and boundary of the depression on the thin plate surface, the size of the depression on the thin plate surface can be further calculated. The size of the depression on the thin plate surface includes the depth value of the depression, the width value of the depression and the area of ​​the depression. The depth value of the depression refers to the vertical distance between the deepest point of the depression area and the surface of the normal non-depression ink printed thin plate. The calculation formula of the depth value of the depression is: d depth (X center ,Y cemter )=|Z defect (X center ,Y center )-Z surface (X center ,Y center )|, where d depth is the depth of the depression, Z defect is the laser beam reflection height at the center of the concave area, Z surface It is the height of the surface of the normal non-depressed ink printed thin plate.

[0025] The calculation formula for the width of the depression is: defect =max(|X left -X right |,|Y top -Y bottom |), where W defect is the width of the concave, X left is the leftmost position of the concave area along the x-axis direction of the concave center point, X right Y is the rightmost position of the concave area along the x-axis direction of the concave center point. topY is the top position of the concave area along the y-axis direction of the concave center point. bottom It is the lowest position of the concave area along the y-axis direction of the concave center point.

[0026] The area of ​​the depression is obtained by calculating the total surface area of ​​the depression region, and the number of pixels of the depression region on the scanning path is N defect , then the calculation formula for the concave area is: Among them, A defect is the area of ​​the depression, i is the index of the pixel, Δx is the horizontal size of the pixel, and Δy is the vertical size of the pixel.

[0027] After the calculation of the size of the depression is completed, the depression positioning and size analysis module can intuitively display the location and size of the depression by superimposing a heat map on the surface image of the ink-printed thin plate, making it easier for operators to quickly take corrective measures.

[0028] The technical effects and advantages of a thin plate surface quality detection system of the present invention: The present invention realizes high-precision detection of depressions on the surface of ink-printed thin plates by combining laser beam scanning, image processing and depression detection, and depression positioning and size analysis modules. The laser scanning measurement module adopts a high-stability laser and uses a multi-point scanning method to obtain accurate data on the surface of the thin plate and construct a three-dimensional model, which can effectively reduce the impact of external interference on the detection accuracy. The laser image processing and depression recognition module captures the image of the laser beam on the surface through a high-resolution RGB camera, and uses an algorithm to extract feature points to achieve accurate displacement calculation. The depression positioning and size analysis module can accurately calculate the depth, width and area of ​​the depression based on the displacement of the laser beam, and intuitively display it in the form of a thermal map, which is convenient for operators to make corrections. The detection system of the present invention has the advantages of high detection accuracy, strong anti-interference ability, and high degree of automation. It can effectively improve the efficiency and accuracy of thin plate surface quality detection and solve the problems of insufficient accuracy and complex operation in traditional methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 The figure is a schematic structural diagram of a thin plate surface quality detection system of the present invention.

[0030] Figure 2 The present invention is a structural schematic diagram of a fully automatic scanning and detection device in the prior art.

[0031] Figure 3 The present invention is a structural schematic diagram of a weld quality detection system in the prior art.

[0032] Figure 4 The present invention is a flow chart of an image processing unit extracting laser beam features from an image acquired by an RGB camera.

[0033] Figure 5 It is a curve diagram of Fourier transform of the surface depression of the ink printed thin plate of the present invention.

[0034] Figure 6 It is a thermal diagram of the depression on the surface of the ink printed thin plate of the present invention. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] Example 1

[0037] See also Figure 1 As shown in the structural schematic diagram, an embodiment of the present invention provides a thin plate surface quality detection system, which includes a laser scanning measurement module, a laser image processing and depression recognition module and a depression positioning and size analysis module; the laser scanning measurement module determines the depression on the surface of the thin plate by the change of the contact between the laser beam and the thin plate surface; the laser image processing and depression recognition module captures the image of the laser beam on the surface of the thin plate, extracts the feature points in the image, and calculates the displacement of the laser beam; the depression positioning and size analysis module calculates the specific value of the depression on the surface of the ink-printed thin plate based on the displacement of the laser beam.

[0038] In this embodiment, see Figure 2 The structural diagram shown is a fully automatic scanning detection device proposed in the prior art. The device controls the laser displacement sensor and the scanning driver to process the detection results and output them graphically. The scanning driver drives the sensor bracket to move to achieve two-dimensional positioning scanning of the laser displacement sensor on the plane. The device can realize rapid detection and automatic evaluation of the surface profile of the laser weld. It has a built-in dedicated data processing algorithm that can automatically calculate the weld size, measure the base adaptively, store the detection data and generate a three-dimensional image of the weld morphology. Figure 3The structural schematic diagram shown is a weld quality detection system proposed in the prior art. The system includes six modules: weld detection, signal processing, data acquisition, data analysis, display and alarm. The weld detection module uses the eddy current detection principle to detect and receive signals; the signal processing module amplifies weak signals and eliminates interference; the data acquisition module converts analog signals into digital signals; the data analysis module processes data and transmits the results to the display and alarm module to achieve online evaluation of weld quality. However, the system lacks the collection of a large amount of experimental and simulation data, and cannot accurately judge some complex or special weld defects, which affects the accuracy and reliability of the detection results.

[0039] Furthermore, the laser scanning measurement module determines the depression of the thin plate surface by the change of the contact between the laser beam and the thin plate surface, including: the laser scanning measurement module projects a high-precision laser beam on the surface of the ink-printed thin plate, and uses the change of the laser beam after reflection on the surface of the ink-printed thin plate to detect the slight deformation of the surface. During the detection process, a high-stability laser emits a finely collimated laser beam, and the wavelength of the laser beam is selected within the visible light or near-infrared light range to ensure that a clear reflection signal can be formed on the surface of the ink-printed thin plate of different materials. The laser beam is modulated by a group of high-precision optical lenses so that it can evenly cover the entire detection area and ensure that the size of the laser spot is suitable for high-precision measurement.

[0040] In order to improve the detection accuracy and reduce noise interference, the laser scanning measurement module adopts a multi-point scanning method, that is, the laser beam is scanned along the surface of the thin plate multiple times to obtain the surface data of the ink-printed thin plate in multiple directions, and construct a three-dimensional model of the depression on the surface of the ink-printed thin plate: the surface of the ink-printed thin plate is represented by a continuous surface function Z(x,y), where x is the horizontal coordinate on the surface of the ink-printed thin plate, and y is the vertical coordinate on the surface of the ink-printed thin plate. Z(x,y) represents the height of a certain point on the surface of the ink-printed thin plate. Ideally, if the surface of the thin plate is completely flat, then Z(x,y) is a plane function that satisfies Z0(x,y)=A0, where Z0 is the height distribution of the ideal plane ink-printed thin plate surface, and A0 is the average height of the ideal plane. However, due to the influence of the depression on the surface of the ink-printed thin plate, the actual surface will have deviations: Z(x,y)=Z0(x,y)+ΔZ(x,y), where ΔZ(x,y) represents the height change caused by defects, satisfying Where M is the order of Fourier expansion in the x direction, which indicates the complexity of the change of the ink printed thin plate surface in the horizontal direction, N is the order of Fourier expansion in the y direction, which indicates the complexity of the change of the ink printed thin plate surface in the vertical direction, and C mnrepresents the amplitude value of the depression, m represents the Fourier mode index in the x direction, which represents the undulation frequency of the surface along the x axis, n represents the Fourier mode index in the y direction, which represents the undulation frequency of the surface along the y axis, L x Indicates the length of the ink-printed sheet, L y Indicates the width of the ink-printed sheet, φ mn Indicates the phase shift of different notch waveforms.

[0041] In order to ensure the accuracy of the data, the laser scanning measurement module will also preprocess the collected laser reflection data. The preprocessing includes noise filtering, background light compensation, data smoothing and other operations to eliminate errors that may be caused by external light source interference or sensor noise.

[0042] Furthermore, the laser image processing and depression recognition module captures the image of the laser beam on the surface of the thin plate, extracts feature points in the image, and calculates the displacement of the laser beam, including: the laser image processing and depression recognition module uses an RGB camera to obtain the image of the laser beam on the surface of the ink-printed thin plate, and the RGB camera can capture the projection of the laser beam on the surface of the ink-printed thin plate with high resolution.

[0043] See also Figure 4 As shown in the flowchart, the image processing unit in the laser image processing and concave recognition module extracts the features of the laser beam from the image acquired by the RGB camera, and the specific steps are as follows:

[0044] Step Z1, using Gaussian filtering to reduce the noise in the image to avoid high-frequency noise interfering with edge detection, the calculation formula is: Among them, σ is the standard deviation of the Gaussian kernel, x is the horizontal coordinate of the current pixel relative to the center of the Gaussian kernel, and y is the vertical coordinate of the current pixel relative to the center of the Gaussian kernel.

[0045] Step Z2, use the Sobel operator to calculate the gradient of the image I in the x and y directions, and obtain the gradient magnitude and direction: Among them, G is the gradient image, G x is the horizontal gradient image, G y is the vertical gradient image, is the rate of change of the pixel value in image I in the x-axis direction, is the rate of change of the pixel value in image I in the y-axis direction.

[0046] In step Z3, by checking the adjacent pixel values ​​in the pixel gradient direction, only the local maximum is retained to refine the edge, and morphological processing is performed to further clean up the noise in the image.

[0047] Step Z4, according to the brightness and color information of the laser beam, the image is segmented into different areas using threshold segmentation, and the area of ​​the laser beam in the image is extracted.

[0048] The following table shows the use of the Sobel operator to process a 600×200 pixel laser beam image and successfully extract the edge points of the laser beam:

[0049] Table 1 Part of the feature point data

[0050] Y(row) X(column) 1 102 2 105 3 108 4 110 5 112 6 115 7 117 8 120 9 123 10 125

[0051] The data in Table 1 show that the laser beam is displaced in the Y-axis direction as the line number increases, which is caused by the depression on the surface of the ink-printed thin plate.

[0052] The formula for calculating the image gradient by the Sobel operator is expressed as: x =g x *I,I y =g y *I, I represents the input image, I x Represents the gradient component of the image in the x direction, I y Represents the gradient component of image I in the y direction, g x is the first convolution kernel of the Sobel operator, which is used to calculate the horizontal gradient and satisfies: g y is the second convolution kernel of the Sobel operator, which is used to calculate the vertical gradient and satisfies

[0053] In order to quantify the degree of deformation of the laser beam, the curvature formula is used to calculate the curvature of the laser beam trajectory. The calculation formula is: Among them, κ is the curvature, which indicates the curvature of the laser beam trajectory. x' and y' are first-order derivatives, x' indicates the tangential direction of the laser beam trajectory along the x-axis, y' indicates the tangential direction of the laser beam trajectory along the y-axis. x" and y" are second-order derivatives, x" indicates the acceleration direction of the laser beam trajectory along the x-axis, and y" indicates the acceleration direction of the laser beam trajectory along the y-axis.

[0054] See also Figure 5 The curve graph shown uses Fourier transform to analyze the changes of the laser beam at different frequencies. The calculation formula is: Among them, F(u) is the frequency domain signal, f(x) is the coordinate point of the laser beam trajectory, and u is the frequency domain variable.

[0055] The displacement of the laser beam is calculated by combining the curvature of the laser beam and the frequency domain analysis. The calculation formula is: Where D is the total displacement of the laser beam between time t and time t+1; is the velocity component of the laser beam trajectory in the x direction; is the velocity component of the laser beam trajectory in the y direction; λ1 is the curvature influence coefficient, which is used to adjust the weight of the curvature on the displacement calculation; λ2 is the frequency influence coefficient, which is used to adjust the influence of Fourier frequency domain information on the displacement calculation.

[0056] In this embodiment, the following is a Python language code example for realizing the integrated curvature and Fourier transform of the laser beam and calculating the displacement of the laser beam:

[0057] import numpy as np

[0058] import matplotlib.pyplot as plt

[0059] from scipy.fftpack import fft,ifft

[0060] y=a*x^2+b*x+c

[0061] a,b,c=0.1,0,0#parabola coefficient

[0062] x = np.linspace(-5,5,100) #x coordinate (simulated time series)

[0063] y=a*x**2+b*x+c

[0064] dy_dx = np.gradient(y,x) #Calculate the first-order derivative of y with respect to x (i.e. gradient)

[0065] d2y_dx2 = np.gradient(dy_dx,x) # Calculate the second-order derivative of y with respect to x (i.e. acceleration)

[0066] curvature=np.abs(d2y_dx2) / (1+dy_dx**2)**(3 / 2)

[0067] plt.figure(figsize=(10,6))

[0068] plt.subplot(2,1,1)

[0069] plt.plot(x,y,label='Laser Beam Trajectory(y=ax^2+bx+c)')

[0070] plt.title('Laser Beam Trajectory')

[0071] plt.xlabel('x')

[0072] plt.ylabel('y')

[0073] plt.grid(True)

[0074] plt.subplot(2, 1, 2)

[0075] plt.plot(x, curvature, label='Curvature(κ)', color='r')

[0076] plt.title('Curvature of Laser Beam')

[0077] plt.xlabel('x')

[0078] plt.ylabel('Curvature(κ)')

[0079] plt.grid(True)

[0080] plt.tight_layout()

[0081] plt.show()

[0082] frequency_domain = fft(y) # Fourier transform, calculate frequency domain

[0083] frequency_magnitude = np.abs(frequency_domain)

[0084] frequency_phase = np.angle(frequency_domain)

[0085] plt.figure(figsize=(10, 6))

[0086] plt.plot(frequency_magnitude, label='Frequency Magnitude')

[0087] plt.title('Fourier Transform of Laser Beam Trajectory')

[0088] plt.xlabel('Frequency(Hz)')

[0089] plt.ylabel('Magnitude')

[0090] plt.grid(True)

[0091] plt.legend()

[0092] plt.show()

[0093] dx_dt = np.gradient(x) #Assume that x changes over time

[0094] dy_dt = np.gradient(y) #how fast y changes over time

[0095] D=integral of sqrt((dx / dt)^2+(dy / dt)^2)

[0096] displacement=np.sqrt(dx_dt**2+dy_dt**2)

[0097] lambda1=0.1#curvature influence coefficient

[0098] curvature_impact=1+lambda1*np.abs(curvature)

[0099] lambda2=0.05#Fourier transform influence coefficient

[0100] frequency_impact=1+lambda2*frequency_magnitude

[0101] final_displacement=np.trapz(displacement*curvature_impact*frequency_impact,x)

[0102] print(f"Final Displacement(integrated over trajectory):{final_displacement:.4f}")

[0103] Furthermore, the depression positioning and size analysis module calculates the specific value of the depression on the surface of the ink-printed thin plate based on the displacement of the laser beam, including: the depression positioning and size analysis module compares the laser beam scanning trajectory under normal conditions with the actual scanning trajectory, and finds that the area deviating from the normal trajectory is the location where the depression exists. The depression positioning and size analysis module establishes a depression positioning model, calculates the depth value, width value and area of ​​the depression on the surface of the ink-printed thin plate, and determines the center point position of the depression area by continuously scanning and comparing multiple laser beam positions. The specific steps are as follows:

[0104] Step W1, comparing the normal scanning trajectory of the laser beam with the actual scanning trajectory, and finding that the area deviating from the normal trajectory is the location where the depression exists.

[0105] Step W2, set a displacement threshold T d , used to determine whether a certain area is a concave area. If the laser beam displacement in a certain area is greater than T d , then it is considered that there is a depression in the area.

[0106] Step W3, by continuously scanning and comparing multiple laser beam positions, determine the center point position P of the concave area center =(X center ,Y center ), P center is the coordinate of the center point of the concave area, X center is the horizontal coordinate of the center point of the concave area, Y center is the vertical coordinate of the center point of the concave area.

[0107] After determining the center position and boundary of the depression on the thin plate surface, the size of the depression on the thin plate surface can be further calculated. The size of the depression on the thin plate surface includes the depth value of the depression, the width value of the depression and the area of ​​the depression. The depth value of the depression refers to the vertical distance between the deepest point of the depression area and the surface of the normal non-depression ink printed thin plate. The calculation formula of the depth value of the depression is: d depth (X center ,Y center )=|Z defect (X center ,Y center )-Z surface (X center ,Y center )|, where d depth is the depth of the depression, Z defect is the laser beam reflection height at the center of the concave area, Z surface It is the height of the surface of the normal non-depressed ink printed thin plate.

[0108] The calculation formula for the width of the depression is: defect=max(|X left -X right |,|Y top -Y bottom |), where W defect is the width of the concave, X left is the leftmost position of the concave area along the x-axis direction of the concave center point, X right Y is the rightmost position of the concave area along the x-axis direction of the concave center point. top Y is the top position of the concave area along the y-axis direction of the concave center point. bottom It is the lowest position of the concave area along the y-axis direction of the concave center point.

[0109] The area of ​​the depression is obtained by calculating the total surface area of ​​the depression region, and the number of pixels of the depression region on the scanning path is N defect , then the calculation formula for the area of ​​the depression is: Among them, A defect is the area of ​​the depression, i is the index of the pixel, Δx is the horizontal size of the pixel, and Δy is the vertical size of the pixel.

[0110] See also Figure 6 As shown in the schematic diagram, after the calculation of the size of the depression is completed, the depression positioning and size analysis module can intuitively display the location and size of the depression by superimposing a heat map on the surface image of the ink-printed thin plate, making it easy for operators to quickly take corrective measures.

[0111] The present invention realizes high-precision detection of depressions on the surface of ink-printed thin plates by combining laser beam scanning, image processing and depression detection, as well as depression positioning and size analysis modules. The laser scanning measurement module adopts a high-stability laser and uses a multi-point scanning method to obtain accurate data on the surface of the thin plate and construct a three-dimensional model, which can effectively reduce the impact of external interference on the detection accuracy. The laser image processing and depression recognition module captures the image of the laser beam on the surface through a high-resolution RGB camera, and uses an algorithm to extract feature points to achieve accurate displacement calculation. The depression positioning and size analysis module can accurately calculate the depth, width and area of ​​the depression based on the displacement of the laser beam, and intuitively display it in the form of a thermal map, which is convenient for operators to make corrections. The detection system of the present invention has the advantages of high detection accuracy, strong anti-interference ability, and high degree of automation. It can effectively improve the efficiency and accuracy of thin plate surface quality detection and solve the problems of insufficient accuracy and complex operation in traditional methods.

[0112] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0113] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A thin plate surface quality detection system, characterized in that: It includes a laser scanning measurement module, a laser image processing and concave recognition module, and a concave positioning and size analysis module; the laser scanning measurement module determines the concave on the surface of the thin plate by the change in the contact between the laser beam and the surface of the thin plate; the laser image processing and concave recognition module captures the image of the laser beam on the surface of the thin plate, extracts the feature points in the image, and calculates the displacement of the laser beam; The depression positioning and size analysis module calculates the specific value of the depression on the surface of the ink printed sheet based on the displacement of the laser beam; The laser image processing and concave recognition module uses the curvature formula to calculate the curvature of the laser beam trajectory: Where κ is the curvature, x' represents the tangential direction of the laser beam trajectory along the x-axis, y' represents the tangential direction of the laser beam trajectory along the y-axis, x" represents the acceleration direction of the laser beam trajectory along the x-axis, and y" represents the acceleration direction of the laser beam trajectory along the y-axis. Fourier transform is used to analyze the changes of the laser beam at different frequencies, and the calculation formula is: Among them, F(u) is the frequency domain signal, f(x) is the coordinate point of the laser beam trajectory, and u is the frequency domain variable; the displacement of the laser beam is calculated by combining the curvature of the laser beam and the frequency domain analysis, and the calculation formula is: Where D is the total displacement of the laser beam from time t to time t+1, is the velocity component of the laser beam trajectory in the x direction, is the velocity component of the laser beam trajectory in the y direction, λ1 is the curvature influence coefficient, and λ2 is the frequency influence coefficient.

2. A thin plate surface quality detection system according to claim 1, characterized in that ,The image processing unit in the laser image processing and concave recognition module extracts the features of the laser beam from the image acquired by the RGB camera. The specific steps are as follows: Step Z1, using Gaussian filtering to reduce the noise in the image to avoid high-frequency noise interfering with edge detection, the calculation formula is: Among them, σ is the standard deviation of the Gaussian kernel, x is the horizontal coordinate of the current pixel relative to the center of the Gaussian kernel, and y is the vertical coordinate of the current pixel relative to the center of the Gaussian kernel; Step Z2, use the Sobel operator to calculate the gradient of the image I in the x and y directions, and obtain the gradient magnitude and direction: Among them, G is the gradient image, G x is the horizontal gradient image, G y is the vertical gradient image, is the rate of change of the pixel value in image I in the x-axis direction, is the rate of change of the pixel value in the image I in the y-axis direction; Step Z3, by checking the adjacent pixel values ​​in the pixel gradient direction, retaining the local maximum to refine the edge, and performing morphological processing to further clean up the noise in the image; Step Z4, according to the brightness and color information of the laser beam, the image is segmented into different areas using threshold segmentation, and the area of ​​the laser beam in the image is extracted.

3. A thin plate surface quality detection system according to claim 1, characterized in that , the laser scanning measurement module constructs a three-dimensional model of the depression on the surface of the ink-printed thin plate: the surface of the ink-printed thin plate is represented by a continuous surface function Z(x,y), x is the horizontal coordinate on the surface of the ink-printed thin plate, and y is the vertical coordinate on the surface of the ink-printed thin plate; ideally, if the surface of the thin plate is completely flat, then Z(x,y) is a plane function that satisfies Z0(x,y)=A0, Z0 is the height distribution of the ideal plane ink-printed thin plate surface, and A0 is the average height of the ideal plane; but due to the depression on the surface of the ink-printed thin plate, the actual surface will have deviations: Z(x,y)=Z0(x,y)+ΔZ(x,y), ΔZ(x,y) represents the height change caused by the defect, satisfying Where M is the order of Fourier expansion in the x direction, N is the order of Fourier expansion in the y direction, and C mn represents the amplitude value of the depression, m represents the Fourier mode index in the x direction, which represents the undulation frequency of the surface along the x axis, n represents the Fourier mode index in the y direction, which represents the undulation frequency of the surface along the y axis, L x Indicates the length of the ink-printed sheet, l y Indicates the width of the ink-printed sheet, φ mn Indicates the phase shift of different notch waveforms.

4. A thin plate surface quality detection system according to claim 1, characterized in that: The formula for calculating the image gradient by the Sobel operator is expressed as: x =g x *I,I y =g y *I, I represents the input image, I x Represents the gradient component of the image in the x direction, I y Represents the gradient component of image I in the y direction, g x is the first convolution kernel of the Sobel operator, used to calculate the horizontal gradient, satisfying: g y is the second convolution kernel of the Sobel operator, which is used to calculate the vertical gradient and satisfies 5. A thin plate surface quality detection system according to claim 1, characterized in that: The depression positioning and size analysis module determines the center point position of the depression area through continuous scanning, that is: center =(X center ,Y center ), P center is the coordinate of the center point of the concave area, X center is the horizontal coordinate of the center point of the concave area, Y center is the vertical coordinate of the center point of the concave area.

6. A thin plate surface quality detection system according to claim 5, characterized in that: After the center position and boundary of the depression on the thin plate surface are determined, the size of the depression on the thin plate surface is further calculated, and the size of the depression on the thin plate surface includes the depth value of the depression, the width value of the depression and the area of ​​the depression.

7. A thin plate surface quality detection system according to claim 6, characterized in that: The depth calculation formula of the depression is: depth (X center ,Y center )=|Z defect (X center ,Y center )-Z surface (X center ,Y center )|, where d depth is the depth of the depression, Z defect is the laser beam reflection height at the center of the concave area, Z surface It is the height of the surface of the normal non-depressed ink printed thin plate.

8. A thin plate surface quality detection system according to claim 6, characterized in that: The calculation formula for the width of the depression is: defect =max(|X left -X right |,|Y top -Y bottom |), where W defect is the width of the concave, X left is the leftmost position of the concave area along the x-axis direction of the concave center point, X right Y is the rightmost position of the concave area along the x-axis direction of the concave center point. top Y is the top position of the concave area along the y-axis direction of the concave center point. bottom It is the lowest position of the concave area along the y-axis direction of the concave center point.

9. A thin plate surface quality detection system according to claim 6, characterized in that: The area of ​​the depression is obtained by calculating the total surface area of ​​the depression region, and the number of pixels of the depression region on the scanning path is N defect , then the calculation formula for the area of ​​the depression is: Among them, A defect is the area of ​​the depression, i is the index of the pixel, Δx is the horizontal size of the pixel, and Δy is the vertical size of the pixel.

Images