Prediction method of interference image reflected by illumination source for surface defect detection of highly reflective materials

Abstract

The present invention discloses a method for predicting the reflection interference image of an illumination light source for surface defect detection of a highly reflective material, and belongs to the field of surface defect detection of curved surfaces of highly reflective metals, glass and other materials. The actual optical system is calibrated by the Zhang Zhengyou calibration method, an equivalent pinhole imaging model of the actual optical system is established according to the calibration parameters, an ideal single lens model is established on the basis of the pinhole imaging model, the entrance pupil diameter is calculated according to the depth of field size of the actual optical system, and then a single lens finite aperture ray tracing model with an entrance pupil size is established in the COMSOL Multiphysics software geometric optics module or similar functional software to simulate the imaging characteristics of the reflection interference image of the illumination light source when detecting surface defects of highly reflective materials under the actual optical system. The present invention can quickly obtain the imaging characteristics of the reflection interference image of the illumination light source when detecting defects at different curvature positions on a highly reflective surface, and then provide a theoretical basis for the design of the optical system for surface defect detection and the suppression of the reflection interference image of the light source.

Description

Technical Field

[0001] The present invention belongs to the field of curved surface defect detection of highly reflective metal, glass and other materials, and more specifically relates to a method for predicting interference images reflected by an illumination light source for surface defect detection of highly reflective materials. Background Art

[0002] Machine vision surface defect detection collects scattered light from surface defects and determines the presence of defects based on the scattered image formed by the collected defects on the image plane. However, when performing machine vision defect detection on curved surfaces of some highly reflective materials, due to the influence of the surface curvature, it is easy for part of the surface illumination light source reflected light to enter the imaging system for imaging. The reflected image area of ​​the illumination light source will cover the scattered image position of the defect, resulting in a blind spot in defect detection. In addition, in some light fields, such as dark fields, the reflected image of the illumination light source and the scattered image of the defect have similar image grayscale characteristics, which can easily lead to misjudgment of defects in defect detection. The interference of the reflected image of the surface illumination light source on surface defect detection directly affects the efficiency and accuracy of surface defect detection. Therefore, it is necessary to predict the interference image of the illumination light source reflection when detecting surface defects of highly reflective materials on curved surfaces, and then adjust the illumination numerical aperture angle, incident light angle and other parameters according to the position where the interference image of the illumination light source appears on the surface to suppress the interference image.

[0003] During the inspection of curved surface defects of highly reflective materials (such as metal, glass, etc.), the reflected image area of ​​the illumination light source may cover the scattered image position of the defect. As a result, there may be blind spots in defect detection, or in light fields such as dark fields, the reflected image and the scattered image have similar image grayscale characteristics, further leading to misjudgment results. These will have a negative impact on the efficiency and accuracy of surface defect detection and lead to erroneous results. Summary of the invention

[0004] To this end, the present invention proposes a method for predicting the reflected interference image of the illumination light source for surface defect detection of highly reflective materials based on finite aperture imaging modeling. Specific technical steps include: using Zhang Zhengyou calibration method to calibrate the actual optical system parameters, constructing a pinhole imaging model of the actual optical system, and establishing an equivalent single-lens imaging model; calculating the entrance pupil diameter according to the depth of field of the actual imaging system and establishing a finite aperture imaging model; inputting the spatial posture parameters of the camera and the light source relative to the surface during sub-aperture scanning, and predicting the imaging results of the illumination light source reflection image at each scanning position. The goal is to achieve the prediction of the reflected interference image of the illumination light source for surface defect detection of highly reflective materials, and to improve the accuracy of defect detection.

[0005] In order to achieve the above object, the present invention is implemented by adopting the following technical scheme: the method comprises:

[0006] The optical parameters of the actual optical system are calibrated by Zhang Zhengyou calibration method;

[0007] Establish a pinhole imaging model of the actual optical system based on the optical parameters obtained by calibration;

[0008] An equivalent ideal single-lens imaging model is established based on the pinhole imaging model;

[0009] Calculate the entrance pupil diameter of the actual system according to the depth of field of the actual imaging system;

[0010] A single-lens finite aperture imaging model with an actual entrance pupil size was established in the COMSOL Multiphysics software geometric optics module.

[0011] Input the sub-aperture scanning spatial pose parameters of the actual system for surface defect detection of highly reflective materials;

[0012] Output the imaging result of the illumination light source reflection image at the scanning position of the surface.

[0013] In one embodiment, the modeling optical parameters include: object distance d, focal length , Actual physical size of the image plane a× b .

[0014] In one embodiment, the image distance of the lens in the ideal single-lens imaging model is equal to the focal length of the pinhole imaging model, and the focal length of the lens is calculated by the following formula:

[0015] (1)

[0016] In the formula is the focal length of the lens, is the focal length of the pinhole imaging model, and d is the object distance.

[0017] In one embodiment, the entrance pupil diameter is determined by solving the following equation:

[0018] (2)

[0019] In the formula is the depth of field of the actual system, is the entrance pupil diameter to be solved, is the minimum value of the diffuse spot diameter that can be resolved by the actual system, d The object distance.

[0020] In one embodiment, the steps of establishing the single-lens finite aperture imaging model are as follows:

[0021] (1) Geometric model setting;

[0022] (2) Material property setting: assigning actual refractive index properties to single lenses and highly reflective surface samples to be tested;

[0023] (3) Setting of physical fields and boundary conditions;

[0024] (4) Mesh division settings.

[0025] In one embodiment, the geometric model is set as follows: d , the focal length of the lens obtained by formula (1) , respectively set the position of the object plane and the image plane, where the actual physical size of the image plane is a × b ; Establish an aperture stop close to the lens in the object space, and the entrance pupil diameter According to formula (2), the radius of curvature of the front and rear surfaces of the lens is and Satisfies the following equation,

[0026] (3)

[0027] In the formula is the radius of curvature of the front surface of the lens, is the radius of curvature of the rear surface of the lens, n is the lens refractive index, h is the lens thickness.

[0028] In one scheme, the physical field and boundary conditions are set as follows: for the area outside the entrance pupil diameter where the entrance pupil plane is located, an absorption boundary is set; for the image plane area, a freezing boundary condition is set; for the highly reflective surface, a mirror reflection boundary condition is set; and the light source is set according to the spatial position and incident angle of the light source in the actual system.

[0029] In one embodiment, the meshing is configured as follows: the geometric model is meshed, wherein the image plane and the entrance pupil plane are meshed using free triangle meshes, and the object to be measured is meshed using free tetrahedral meshes.

[0030] Beneficial effects of the present invention:

[0031] 1. It can effectively predict the reflection interference image of the lighting source during surface defect detection of highly reflective materials, and help the actual optical inspection system to handle the reflection interference image in advance, so as to avoid or reduce the detection misjudgment and detection blind spots caused by the reflection interference image of the lighting source during defect detection.

[0032] 2. This method comprehensively utilizes technologies such as optical parameter calibration, pinhole imaging model, single lens imaging model and finite aperture imaging model, and comprehensively considers the optimization of detection effect from both theoretical and practical application perspectives, so it can significantly improve the efficiency and accuracy of surface defect detection of highly reflective materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 is a flow chart of the method of the present invention;

[0034] Figure 2 This is a flow chart of calibrating the optical parameters of an actual optical system by the Zhang Zhengyou calibration method of the present invention;

[0035] Figure 3 This is a flow chart of the single-lens finite aperture imaging model of the present invention. DETAILED DESCRIPTION

[0036] In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. Typical embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described in the present invention. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

[0037] Unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as those understood by those skilled in the art of the present invention. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0038] Embodiment 1:

[0039] like Figure 1 As shown, a method for predicting the interference image reflected by the illumination light source for surface defect detection of highly reflective materials based on finite aperture imaging modeling comprises the following steps:

[0040] like Figure 2 As shown in S1, the optical parameters of the actual optical system are calibrated using MATLAB software through the Zhang Zhengyou calibration method.

[0041] S101, prepare a plane grid calibration plate: used to take photos for subsequent analysis. The calibration plate is usually a black and white checkered pattern with no pattern inside the grid.

[0042] S102, taking photos: placing the calibration plate at different positions and angles, and taking at least three photos at different viewing angles with a camera.

[0043] S103, extracting feature points: extracting grid corner points from the photo, that is, points where black and white grids intersect, a total of four feature points.

[0044] S104, establishing a camera imaging model: When the camera is imaging, it can be described by a pinhole model. After the three-dimensional points in the physical world are transformed by rotation, translation, lens distortion, projection, etc., they are mapped to two-dimensional coordinates on the image plane.

[0045] S105, solving parameters: By optimizing the method, the distance between the image feature points and the projected points is minimized, so that the camera's internal parameters, including focal length, optical axis center coordinates, etc., and external parameters, including rotation matrix and translation matrix, can be solved.

[0046] S2. Establish a pinhole imaging model of the actual optical system based on the optical parameters obtained by calibration.

[0047] The pinhole imaging model, also known as the pinhole model, is the most basic model in camera modeling. Its main assumption is that the camera can be regarded as a small hole in optics, through which the three-dimensional world is projected onto a two-dimensional plane.

[0048] Specifically, according to the pinhole imaging model, the relationship between the object point P and the point p on the image plane is as follows:

[0049] The optical parameters obtained by calibration can be used to establish a pinhole imaging model of the actual optical system. These calibration parameters include object distance d, focal length f and actual physical size of the image plane. a×b .

[0050] When building a model, the first thing to consider is the geometry of the imaging system. The pinhole imaging model (also known as the pinhole camera model) is a simplified camera model that simulates the working principle of a real camera to a certain extent. In this model, light passes through a small hole and creates an inverted image of the scene on the plane behind it (called the imaging plane).

[0051] Assume that a point P(X,Y,Z) on the object is projected onto a point p(x,y) on the imaging plane through the pinhole O. Considering the distance between the pinhole O and the imaging plane as the focal length f, the following equation can be obtained based on the properties of similar triangles:

[0052]

[0053]

[0054] This completes the establishment of the pinhole imaging model of the actual optical system.

[0055] S3. Based on the pinhole imaging model, an equivalent ideal single-lens imaging model is established.

[0056] The image distance of the lens in S3 is equal to the focal length of the pinhole imaging model , the focal length of the lens Calculated by the following formula

[0057] (1)

[0058] In the formula is the focal length of the lens, is the focal length of the pinhole imaging model, d The object distance.

[0059] S4. Calculate the entrance pupil diameter of the actual system according to the depth of field of the actual imaging system.

[0060] The entrance pupil diameter in S4 is realized by solving the following equation:

[0061] (2)

[0062] In the formula is the depth of field of the actual system, is the entrance pupil diameter to be solved, is the minimum value of the diffuse spot diameter that can be resolved by the actual system, d The object distance.

[0063] S5. A single-lens finite aperture imaging model with an actual entrance pupil size is established in the geometric optics module of COMSOL Multiphysics software.

[0064] like Figure 3 As shown, the S5 includes the following sub-steps:

[0065] S501, geometric model setting: according to the object distance obtained in S1 d , the focal length of the lens obtained by formula (1) , respectively set the position of the object plane and the image plane, where the actual physical size of the image plane is a × b . An aperture stop is established close to the lens in the object space, and the entrance pupil diameter According to formula (2), the lens is a spherical lens, the lens thickness is set to 1 mm, and the lens diameter is set to mm, radius of curvature of the front and rear surfaces of the lens and The following equation needs to be satisfied,

[0066] (3)

[0067] In the formula is the radius of curvature of the front surface of the lens, is the radius of curvature of the rear surface of the lens, n is the lens refractive index, h is the lens thickness.

[0068] S502, material property setting: assigning actual refractive index properties to the single lens and the high-reflective surface sample to be measured;

[0069] S503, physical field and boundary condition setting: for the area outside the entrance pupil diameter where the entrance pupil plane is located, set the absorption boundary. For the image plane area, set the freezing boundary condition. For the highly reflective surface, set the mirror reflection boundary condition. Set the light source position and light direction vector according to the spatial position and incident angle of the light source in the actual system;

[0070] S504, mesh generation setting: mesh the geometric model, wherein the maximum mesh size of the image surface is in the order of 0.1 mm; the maximum mesh size of the high-reflection surface is set in the order of 1 mm.

[0071] S6. Input the sub-aperture scanning spatial pose parameters of the actual system for surface defect detection of highly reflective materials.

[0072] Input the sub-aperture scanning spatial pose parameters of the actual system for surface defect detection of highly reflective materials, including the position and direction of the illumination light source and the camera, as well as the size and position of the sub-aperture.

[0073] This step is mainly to simulate actual lighting and imaging conditions. The position and direction of the illumination source will affect the way the light hits the object, and thus affect the observation of the object surface; while the size and position of the subaperture will directly determine the imaging quality of the imaging system.

[0074] The specific input parameters are as follows:

[0075] The position information of the lighting source and the camera is generally expressed in three-dimensional coordinates, such as (x, y, z).

[0076] The direction of the lighting source and the camera is usually expressed as a direction vector, which points to the direction of the light source's ray and the camera's optical axis.

[0077] The size of the subaperture is usually expressed in terms of diameter.

[0078] The position information of the sub-aperture is represented by the three-dimensional coordinates of its center point, such as (u, v, w).

[0079] S7. Output the imaging result of the reflection image of the illumination light source at a given scanning position or different curvature positions on the surface, and obtain the imaging law of the interference image reflected by the illumination light source at different curvature positions on the surface.

[0080] First, the model will simulate the light emitted from the illumination light source, passing through the aperture, hitting the surface of the detection object, and then reflected by the surface of the object into the imaging system to form an image on the image plane according to the sub-aperture scanning spatial pose parameters given in S6.

[0081] Then, by analyzing the generated image, the light source interference image generated by the reflection of the illumination light source can be identified. During the analysis process, a common method is to perform threshold segmentation on the image, and the area that is obviously brighter than the background can be regarded as the light source interference image.

[0082] After this step, the imaging law of the interference image reflected by the illumination light source at the curvature position of the surface of different objects can be obtained. This law can be used as the basis for the design and adjustment of the actual light source, so that the light source can better adapt to the surface of different objects, reduce reflection interference, and improve the accuracy of detection.

[0083] Embodiment 2:

[0084] First, perform step S1 and use MATLAB software to calibrate the optical parameters of the actual optical system using the Zhang Zhengyou calibration method. Assume that the object distance d is 303 mm, the focal length of the pinhole imaging model is 293 mm, and the actual physical size of the image plane is a×b It is 10mm×15mm.

[0085] Then, steps S2 and S3 are performed to establish a pinhole imaging model and an equivalent ideal single lens imaging model according to the above parameters. According to the object distance and focal length in the pinhole imaging model, the focal length of the ideal lens can be obtained to be 150 mm. Assuming the lens thickness is 1 mm and the refractive index is 1.52, according to formula (3) combined with the focal length of the lens, the front surface curvature radius of the lens can be obtained to be 127 mm and the rear surface curvature radius is -200 mm.

[0086] Then, step S4 is performed. Assuming that the depth of field of the actual imaging system is 2.5 mm, the entrance pupil diameter of the actual system is calculated to be approximately 21 mm using formula (2).

[0087] Then, step S5 is performed. In the geometric optics module of COMSOL Multiphysics software, a single-lens finite aperture imaging model with an actual entrance pupil size is established based on the above data and parameters such as the curvature radius of the front and rear surfaces of the lens and the lens refractive index.

[0088] Then comes step S6, where the spatial pose parameters of the sub-aperture scanning for surface defect detection of highly reflective materials in the actual system are input. For example, the spatial position of the light source is (0, -200, 200) mm, the spatial position of the camera is (0, 0, 303), the sub-aperture center point is (0, 0, 0), and the unit vector of the camera optical axis direction is set to (0, 0, -1). The unit vector of the incident direction of the light source is .

[0089] Finally, step S7 is performed, and the software model simulates and calculates the output of the imaging result of the illumination light source reflection image at the surface scanning position. The result can show the imaging law of the interference image reflected by the illumination light source at different curvature positions on the surface, so that the illumination method with a specific illumination angle and illumination numerical aperture at a given curvature position can be designed before detection to suppress the appearance of the interference image of the illumination light source.

[0090] Through this method, the interference image reflected by the lighting source in the surface defect detection of curved and highly reflective materials can be predicted and processed more accurately, thereby improving the efficiency and accuracy of defect detection.

[0091] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM).

[0092] It should be understood that the detailed description of the technical solutions of the present invention by means of the preferred embodiments is illustrative rather than restrictive. A person skilled in the art may modify the technical solutions described in the embodiments, or replace some of the technical features by equivalents, based on reading the specification of the present invention; and these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting interference images of illumination light sources reflected from surface defects of highly reflective materials, characterized by: The method includes: The optical parameters of the actual optical system are calibrated by Zhang Zhengyou calibration method; Establish a pinhole imaging model of the actual optical system based on the optical parameters obtained by calibration; An equivalent ideal single-lens imaging model is established based on the pinhole imaging model; The image distance of the lens in the ideal single-lens imaging model is equal to the focal length of the pinhole imaging model. The focal length of the lens is calculated by the following formula: (1) In the formula is the focal length of the lens, is the focal length of the pinhole imaging model, d is the object distance; Calculate the entrance pupil diameter of the actual system according to the depth of field of the actual imaging system; A single-lens finite aperture imaging model with an actual entrance pupil size was established in the COMSOL Multiphysics software geometric optics module. Input the spatial pose parameters of the camera and light source relative to the surface during sub-aperture scanning of the actual system for surface defect detection of highly reflective materials; Output the imaging result of the illumination light source reflection image at the surface sub-aperture scanning position.

2. The method for predicting interference images reflected by illumination sources for surface defect detection of highly reflective materials according to claim 1, characterized in that: Modeling optical parameters include: object distance d ,focal length , Actual physical size of the image plane a × b .

3. The method for predicting interference images reflected by illumination sources for surface defect detection of highly reflective materials according to claim 1, characterized in that: The entrance pupil diameter is determined by solving the following equation: (2) In the formula is the depth of field of the actual system, is the entrance pupil diameter to be solved, is the minimum value of the diffuse spot diameter that can be resolved by the actual system, d The object distance.

4. The method for predicting interference images reflected by illumination sources for surface defect detection of highly reflective materials according to claim 1, characterized in that: The steps for establishing the single-lens finite aperture imaging model are as follows: (1) Geometric model setting; (2) Material property setting: assigning actual refractive index properties to single lenses and highly reflective surface samples to be tested; (3) Setting of physical fields and boundary conditions; (4) Mesh division settings.

5. The method for predicting interference images reflected by illumination sources for surface defect detection of highly reflective materials according to claim 4, characterized in that: The geometric model is set as follows: according to the object distance obtained d , the focal length of the lens obtained by formula (1) , respectively set the position of the object plane and the image plane, where the actual physical size of the image plane is a × b ; Establish an aperture stop close to the lens in the object space, and the entrance pupil diameter According to formula (2), the radius of curvature of the front and rear surfaces of the lens is and Satisfies the following equation, (3) In the formula is the radius of curvature of the front surface of the lens, is the radius of curvature of the rear surface of the lens, n is the lens refractive index, h is the lens thickness.

6. The method for predicting interference images reflected by illumination sources for surface defect detection of highly reflective materials according to claim 4, characterized in that: The physical field and boundary conditions are set as follows: for the area outside the entrance pupil diameter where the entrance pupil plane is located, an absorption boundary is set; for the image plane area, a freezing boundary condition is set; for the highly reflective surface, a mirror reflection boundary condition is set; and the light source is set according to the spatial posture and incident angle of the light source in the actual system.

7. The method for predicting interference images reflected by illumination sources for surface defect detection of highly reflective materials according to claim 4, characterized in that: The mesh generation is set as follows: the geometric model is meshed, wherein the image plane and the entrance pupil plane are meshed using free triangle meshes, and the object to be measured is meshed using free tetrahedron meshes.

Images