Optimization method and system for curvature of circular arc-shaped diffuse reflection cover and circular arc-shaped diffuse reflection cover

Abstract

The application relates to a method and system for optimizing the curvature of a circular arc-shaped diffuse reflection cover and the circular arc-shaped diffuse reflection cover, the method comprising: determining geometric design variables that affect the curvature distribution of the inner surface of the circular arc-shaped diffuse reflection cover, and constructing a three-dimensional parameterized model of the circular arc-shaped diffuse reflection cover; establishing a geometric constraint range of the overall height according to the working conditions of a line scanning camera; generating a plurality of candidate models by traversing the design variables within the geometric constraint range, and performing light ray tracing simulation on each candidate model by using an optical simulation tool; obtaining the illumination distribution of each candidate model on the detection line of the to-be-tested surface, and eliminating the candidate models that do not satisfy a preset illumination uniformity threshold in the line scanning direction; for the retained candidate models, an illumination distribution perpendicular to the line scanning aperture direction is obtained by using a partitioned illumination strategy, and an effective light-emitting area is extracted; the illumination difference evaluation index of each candidate model in the effective light-emitting area is calculated, and the optimal candidate model is screened out.

Description

Technical Field

[0001] This application relates to the fields of optical and lighting technologies, and more specifically, to a method, system, and arc-shaped diffuse reflector for optimizing the curvature of an arc-shaped diffuse reflector. Background Technology

[0002] In the field of machine vision inspection, diffuse dome light sources (e.g., dome light sources) are widely used to inspect workpieces with high reflectivity or complex curved structures. The core principle is to uniformly project light onto the surface of the sample after multiple reflections through the inner surface of the diffuse reflector, thereby eliminating the light spots and shadows caused by specular reflection.

[0003] Early diffuse dome light sources primarily focused on basic light field uniformity, typically employing a single hemispherical structure with a simple diffuse coating. While technological advancements have enabled the achievement of higher light intensity uniformity, current diffuse dome light source designs have the following limitations when applied to online scanning imaging systems: First, existing diffuse dome light source designs only focus on the performance of the light source itself, neglecting the overall system characteristics of the machine vision system. Specifically, current technologies set the design goal of the light source singularly as "improving illumination uniformity," pursuing a highly uniform distribution of the light field. However, in actual line scan imaging systems, the camera, lens, light source, and algorithm together constitute a complete detection system. Simply pursuing light source uniformity cannot guarantee that the subsequent defect detection algorithm obtains the best image input.

[0004] Second, existing diffuse dome light sources are ineffective in detecting minute defects on highly reflective metal surfaces. For highly reflective metal surfaces (such as the edges of lithium batteries), most defect types (shallow scratches, micro-creases, micro-dimples, etc.) depend on the reflection direction or small height differences, thus producing visual features. Under uniform diffuse light, uniform illumination often suppresses the grayscale transition information required for the edges of micro-defects, causing the defects to appear as low-contrast, blurred features in the image, thereby affecting the recognition ability of subsequent detection algorithms. If attempts are made to solve the problem by improving the optical path and adjusting the light, current technologies can only rely on higher-performance cameras to slightly improve the clarity of defects, which is costly and has limited effectiveness.

[0005] Third, existing diffuse dome light sources do not take into account the directional characteristics of line scan imaging systems. Line scan imaging systems have a clear scanning direction (camera movement direction) and lateral dimension (perpendicular to the scanning direction). The angular distribution of light rays in different directions contributes differently to imaging. Therefore, existing arc-shaped diffuse reflector designs have not been optimized to address this characteristic. Summary of the Invention

[0006] The purpose of this invention is to provide a method for optimizing the curvature of an arc-shaped diffuse reflector. The arc-shaped diffuse reflector formed by this optimization method can enhance the intensity of changes in the angle of light in the vertical direction and improve the grayscale contrast between the defect and the background.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for optimizing the curvature of an arc-shaped diffuse reflector, comprising: S10: Determine the geometric design variables that affect the curvature distribution of the inner surface of the arc-shaped diffuse reflector, wherein the design variables include at least the bottom radius of the arc-shaped diffuse reflector. and overall height And construct a three-dimensional parametric model of the arc-shaped diffuse reflector; S20: Establish the overall height based on the working conditions of the line scan camera. The range of geometric constraints; S30: Within the geometric constraints, multiple candidate models are generated by traversing the design variables, and ray tracing simulation is performed on each candidate model using an optical simulation tool; S40: Obtain the illuminance distribution of each candidate model on the detection line of the surface to be tested, and eliminate candidate models that do not meet the preset illuminance uniformity threshold in the line scan direction; S50: For the retained candidate models, a zoned illumination strategy is used to obtain the illuminance distribution perpendicular to the direction of the line scan aperture, and the effective luminous area is extracted. S60: Calculate the illuminance difference evaluation index of each candidate model within the effective luminous area, and select the optimal candidate model with the goal of maximizing the illuminance difference evaluation index.

[0008] Furthermore, in step S20, the overall height The geometric constraint range must simultaneously satisfy the following conditions: Upper limit constraint: The overall height limit is determined based on the working distance of the line scan camera and the installation safety margin; Lower limit of field of view constraint: The lower limit of the height of the arc-shaped diffuser is determined based on the width of the top opening of the arc-shaped diffuser, the half angle of the field of view of the line scan camera perpendicular to the scanning direction, and the safety margin. Optical lower limit constraint: The minimum overall height limit determined through optical simulation. Further, the overall height includes the height of the arc-shaped diffuse reflector. and base thickness , The upper limit constraint must satisfy: , in, Indicates the working distance of the line scan camera. This indicates the safety margin during installation.

[0009] The lower limit constraint of the field of view must satisfy: , in, This indicates the width of the opening at the top of the arc-shaped diffuser; This refers to the half-angle of the field of view of the line scan camera perpendicular to the scanning direction. This is for safety margin.

[0010] The optical lower limit constraint must satisfy: , in, This represents the minimum optical height required to meet basic diffuse reflection requirements.

[0011] Further, step S40 includes: The method for calculating the uniformity of illumination in the line scan direction is as follows: = ,in, This indicates the minimum illuminance on the testing line. Indicates maximum illuminance; Eliminate those that do not meet the requirements The surface shape, the preset line scan direction illuminance uniformity threshold .

[0012] Furthermore, in step S40, in addition to eliminating candidate models that do not meet the preset line scan direction illuminance uniformity threshold, an auxiliary screening step is also included: Obtain the overall light energy utilization rate of each candidate model on the test plane. The overall light energy utilization rate It is the ratio of the total luminous flux received by the detection surface to the total luminous flux of the light source; Excluding the overall light energy utilization rate The candidate reflector model is below the set lower limit value.

[0013] Further, in step S50, the method for extracting the effective luminous area using the zoned lighting strategy includes: Establish a coordinate system for the illuminance grid of a circular or hemispherical surface; The light source on the arc-shaped diffuser is divided into several independent lighting zones; Individual lighting zones are triggered and illuminated independently in sequence. The specific zone where the maximum illuminance value is located is extracted by ray tracing. This specific zone corresponds to the direction perpendicular to the line scan aperture. Within the specific partition, the illuminance gradient is calculated along the direction perpendicular to the line scan aperture. The position where the illuminance gradient first exceeds a preset threshold is taken as the boundary of the effective luminous area, and the area inside this boundary is the effective luminous area.

[0014] Further, in step S60, the illuminance difference evaluation index is: , in, This represents the maximum illuminance within the effective luminous area. This represents the minimum illuminance within the effective luminous area.

[0015] Furthermore, the optimization method also includes: The candidate models selected in step S60 are processed into physical objects and installed in the line scanning imaging system. The highly reflective metal sample was subjected to actual imaging, and the acquired image was input into a preset defect extraction algorithm for secondary verification. The model was then determined based on the verification results.

[0016] This application also provides a system for optimizing the curvature of an arc-shaped diffuse reflector, comprising: Parametric modeling module: used to model based on the base radius and overall height Construct a three-dimensional parametric model of an arc-shaped diffuse reflector; The geometric constraint module is used to establish the overall height based on the operating conditions of the line scan camera. The range of geometric constraints; The simulation iteration module is used to generate multiple candidate models within the geometric constraints and perform ray tracing simulation using optical simulation tools. The uniformity screening module is used to obtain the illuminance distribution of each candidate model on the detection line of the surface to be tested, and to remove candidate models that do not meet the preset line scan direction illuminance uniformity threshold. The partitioning and effective region extraction module is used to obtain the illuminance distribution perpendicular to the direction of the line scan aperture for the retained candidate models using a partitioned illumination strategy, and to extract the effective luminous region. The evaluation and selection module is used to calculate the illuminance difference evaluation index of each candidate model in the effective luminous area, and to select candidate models with the goal of maximizing the index.

[0017] This application also provides an arc-shaped diffuse reflector for use in a line scanning imaging system, wherein the arc-shaped diffuse reflector is obtained using the above-described optimization method.

[0018] The beneficial effects of this invention are as follows: The optimization method and system of this application selectively extract the effective light-emitting area and optimize the surface shape by maximizing the illuminance difference within this area (perpendicular to the direction of the line scan aperture). This significantly increases the intensity of the angle change of the emitted light rays in this direction (making its distribution steeper), allowing the small defects on the highly reflective metal surface to form a higher grayscale contrast during imaging. This greatly improves the accuracy and feature resolution of subsequent image processing algorithms such as defect detection from a physical optics perspective. The arc-shaped diffuse reflector formed by this optimization method enhances the intensity of the angle change of the light rays in the vertical direction, improving the grayscale contrast between the defects and the background.

[0019] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, the preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating a method for optimizing the curvature of an arc-shaped diffuse reflector according to an embodiment of this application; Figure 2 A comparison of illuminance distribution curves before and after optimization for an arc-shaped diffuse reflector; Figure 3a The original image acquired by a traditional hemispherical arc-shaped diffuse reflector; Figure 3b The original image acquired by the optimized non-hemispherical arc-shaped diffuse reflector according to an embodiment of this application; Figure 4a The defect detection results are shown under a traditional hemispherical arc-shaped diffuse reflector. Figure 4b The detection results are shown under an optimized arc-shaped diffuse reflector according to an embodiment of this application. Detailed Implementation

[0021] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] A "circular arc-shaped diffuse reflector" refers to a rotationally symmetric surface whose inner surface generatrix is ​​a circular arc curve. This includes, but is not limited to, hemispherical diffuse reflectors, spherical cap-shaped diffuse reflectors, and diffuse reflectors combining circular arcs and planes. This circular arc-shaped diffuse reflector is used in line-scan imaging systems. For ease of description, this embodiment uses a reflector with a dome-shaped inner surface (hereinafter referred to as a dome reflector) as an example. The dome structure enables sufficient diffuse reflection of light at multiple angles, which is beneficial for meeting the decoupling optimization requirements of line-scan imaging systems for one-dimensional uniformity and one-dimensional steepness.

[0023] Those skilled in the art will understand that the curvature optimization method proposed in this embodiment is not limited to dome reflectors; it can also be applied to rotationally symmetric diffuse reflectors with other generatrical shapes (such as elliptical arcs, parabolas, and free curves), as long as the curvature optimization is satisfied by the base radius. and overall height This is a prerequisite for parametric description. The arc-shaped diffuse reflector in this embodiment was optimized using the following method.

[0024] Figure 1 A flowchart illustrating a method for optimizing the curvature of an arc-shaped diffuse reflector according to one embodiment is shown. (Refer to...) Figure 1 As shown, the method for optimizing the curvature of the arc-shaped diffuse reflector includes: S10: Determine the geometric design variables that affect the curvature distribution of the inner surface of the arc-shaped diffuse reflector. The design variables shall include at least the bottom radius of the arc-shaped diffuse reflector. and overall height And construct a three-dimensional parametric model of the arc-shaped diffuse reflector; S20: Establish the overall height based on the working conditions of the line scan camera. The range of geometric constraints; S30: Within the geometric constraints, multiple candidate models are generated by traversing the design variables, and ray tracing simulation is performed on each candidate model using optical simulation tools. S40: Obtain the illuminance distribution of each candidate model on the detection line of the surface to be tested, and eliminate candidate models that do not meet the preset illuminance uniformity threshold in the line scan direction. S50: For the retained candidate models, a zoned illumination strategy is used to obtain the illuminance distribution perpendicular to the direction of the line scan aperture, and the effective luminous area is extracted. S60: Calculate the illuminance difference evaluation index of each candidate model within the effective luminous area, and select the optimal candidate model by maximizing the illuminance difference evaluation index.

[0025] In an alternative embodiment, in step S20, the overall height The geometric constraints must simultaneously meet the following conditions to ensure that the light source, reflector, and line scan camera have no mechanical interference and do not obstruct the imaging field of view. Specifically, the upper limit constraint is the overall height upper limit determined based on the working distance of the line scan camera and the installation safety margin, to avoid the top of the arc-shaped diffuse reflector colliding with the camera lens; the lower limit constraint is the height of the arc-shaped diffuse reflector determined based on the width of the opening at the top of the arc-shaped diffuse reflector, the half-angle of the field of view of the line scan camera perpendicular to the scanning direction, and the safety margin, to avoid the edge of the opening at the top of the arc-shaped diffuse reflector entering the camera's field of view and causing image obstruction; the lower optical limit constraint is the minimum overall height lower limit determined through optical simulation to ensure basic diffuse reflection effects.

[0026] The overall height includes the height of the arc-shaped diffuser. and base thickness In one application, Overall height of the light source It must be less than the working distance of the line scan camera. Therefore, the upper limit constraint must satisfy: , in, Indicates the working distance of the line scan camera. This indicates a safety margin for installation; the thickness of the base is [not specified]. For quantitative purposes.

[0027] In the lower limit constraint of the field of view, if the overall height of the light source is... If the size is too small, the edge of the opening will fall within the camera's field of view, causing image occlusion. Therefore, the lower limit constraint of the field of view must satisfy: , in, This indicates the width of the opening at the top of the arc-shaped diffuser; This refers to the half-angle of the field of view of the line scan camera perpendicular to the scanning direction. This is for safety margin.

[0028] In the optical lower limit constraint, the overall height of the light source If the reflector is too small, it will be too flat, and the light cannot be diffused sufficiently. Therefore, the optical lower limit constraint must be met: , in, This represents the minimum optical height required to meet basic diffuse reflection requirements.

[0029] It should be noted that in practical applications, a safety margin should be considered during installation. The setting is primarily used to prevent physical collisions or mechanical interference between the top of the diffuser and the lens of the line scan camera or its associated structure. The specific selection of this value needs to consider the manufacturing tolerances, assembly precision, and vibration amplitude of each component of the line scan imaging system during operation. Generally, The setting can be within the range of 2mm to 10mm. For example, in a precision laboratory, when the repeatability of the mechanical structure is high, A thickness of 2mm is acceptable; however, in industrial environments such as production lines where there is some mechanical vibration, to ensure absolute safety and ease of disassembly and assembly for later maintenance, We recommend using 5mm or 10mm to ensure the overall height of the light source. Always less than the camera's working distance .

[0030] Safety margin The purpose of this setting is to ensure that the physical edge of the opening at the top of the diffuser does not enter the camera's imaging field of view, thereby avoiding image edge occlusion due to installation misalignment. This margin is the half-angle of the camera's field of view perpendicular to the scanning direction. Redundancy space has been reserved. In engineering practice, The value of this margin is usually determined based on the camera's mounting alignment accuracy and lens distortion characteristics, and is generally set between 1mm and 5mm. By setting this margin, even with a slight offset between the reflector and the camera's optical axis, the camera's imaging field of view is within the reflector height. The cross-sectional radius at that point is still smaller than the opening radius. This ensures the integrity and brightness uniformity of the acquired images from a physical structure perspective.

[0031] In a specific example, taking a certain type of lithium battery edge detection system, a line scan camera is used with a lens focal length of 50mm and a working distance (from the front of the lens to the surface of the workpiece being measured) of 250mm. The calculated half-angle of the camera's field of view perpendicular to the scanning direction... The angle is approximately 5°. The top opening of the diffuser is designed as an elliptical hole with a major axis (scanning direction) of 70 mm and a minor axis (perpendicular to the scanning direction) of 10 mm. The theoretical alignment deviation between the center of the opening and the camera optical axis is controlled within ±0.3 mm through mechanical positioning. The edge field of view offset introduced by lens distortion is approximately 0.2 mm.

[0032] reflector height .exist At height, the theoretical value of the camera's field of view radius (perpendicular to the scanning direction) is... In actual installation, considering optical axis eccentricity and distortion, the required maximum field-of-view section radius is approximately 10.94 + 0.3 + 0.2 = 11.44 mm.

[0033] The minor axis of the elliptical aperture is 5mm (half of the 10mm minor axis). The physical semi-aperture radius in this direction is only 5mm, smaller than the required field-of-view radius of 11.44mm. This indicates that without a safety margin, the aperture edge will severely obstruct the imaging field of view. Therefore, by adjusting the installation height of the dome light source or increasing the aperture size, the effective aperture radius in the minor axis direction can be made to meet the requirements. ≥Required field of view radius+ In this embodiment, the minor axis of the elliptical opening is increased to 26mm (minor semi-axis 13mm), and a safety margin is set. =1.5mm, then 13mm ≥ 11.44mm + 1.5mm = 12.94mm, the condition is met. The selected value ensures that even with an installation deviation of ±0.3mm and a distortion of 0.2mm, the edge of the opening still maintains a physical gap of more than 1.5mm with the boundary of the field of view, thereby ensuring that the camera captures complete and unobstructed images, while avoiding the impact of an excessively large opening on the reflection efficiency of the dome light source.

[0034] In an alternative embodiment, step S30 includes: The overall height determined in step S20 Within geometric constraints (e.g.) (Take a radius of 200mm to 300mm), and a preset bottom radius. Within a value range (e.g., 100mm to 150mm), iterate through all values ​​with appropriate step sizes (e.g., equal intervals of 5mm). Combining and building a series of candidate models in 3D modeling software; For each candidate model, perform the following optical simulation operations: Import the candidate model into the optical simulation software, set the light source to a Lambertian LED array (the actual light source model used), and apply the optical properties (reflectivity) of a white diffuse reflection coating to the inner surface of the reflector. (and the Lambertian scattering model), perform Monte Carlo ray tracing, and the number of rays traced must meet the statistical convergence requirement.

[0035] In a specific example, consider a candidate model with H=140 mm and R=130 mm. A hemispherical dome reflector is constructed in 3D modeling software, with an observation hole (aperture radius...) at the top. =25 mm, safety margin =3 mm). Import the model into optical simulation software, set the light source to a Lambertian LED array, and apply a white diffuse reflection coating to the inner surface of the reflector (reflectivity set to 0.9, using the Lambertian scattering model). Perform Monte Carlo ray tracing to ensure that the number of rays meets the convergence requirement (i.e., the illuminance distribution no longer changes significantly after increasing the number of rays). After tracing is completed, extract the illuminance distribution curve along the perpendicular scanning direction on the workpiece surface (5 mm below the bottom opening of the reflector), and calculate the evaluation index of the drastic change in illuminance on this curve. Record the values ​​under this (H,R) combination. Values. Iterate through all (H, R) combinations (H = 130~170 mm, R = 100-160 mm, step size 5 mm), comparing the values ​​of each combination. Value, will The top 5% combinations with the largest size are selected as the preferred design domain.

[0036] In an alternative embodiment, to ensure the most basic imaging quality of the line scan imaging system, to ensure the consistency of image grayscale during the scanning process of the line scan camera, and to avoid misjudgment or missed judgment by the algorithm due to uneven lighting of the light source itself, step S40 specifically includes: The calculation method for the uniformity of illuminance in the line scan direction is as follows: = ,in, This indicates the minimum illuminance on the testing line. Indicates maximum illuminance; Eliminate those that do not meet the requirements The surface shape, preset line scan direction illuminance uniformity threshold .

[0037] In a specific example, taking a candidate model (H=130mm, R=140mm) as an example, after completing the ray tracing simulation, a detection line along the scanning direction on the workpiece surface is extracted, and illuminance values ​​are sampled at 1.8mm intervals. The minimum illuminance is measured. =44810 lx, maximum illuminance =48089lx, according to the formula = Calculated ≈0.932. Preset threshold for uniformity of illumination along the line scan direction. =0.9, the model satisfies Values ​​≥0.9 are retained.

[0038] Another candidate model (H=160mm, R=140mm) measured the minimum illuminance on the same detection line with the same sampling interval. =43000lx, maximum illuminance =48089lx, calculated to The value is approximately 0.894, which is below the threshold of 0.9, therefore the model does not meet the requirements. ≥ Those that are removed are excluded. Through the above calculations and filtering, all are retained. Face shapes with a value ≥0.9 will be further optimized.

[0039] Furthermore, in step S40, in addition to eliminating candidate models that do not meet the preset line scan direction illuminance uniformity threshold, an auxiliary screening step is also included: Obtain the overall light energy utilization rate of each candidate model on the test plane. Overall light energy utilization rate It is the ratio of the total luminous flux received by the detection surface to the total luminous flux of the light source; Excluding overall light energy utilization Below the set lower limit (for example, Candidate reflector models.

[0040] In the process of simply pursuing high contrast (illuminance difference), some extremely distorted surface shapes may be screened out, resulting in serious waste of light energy. By adopting this auxiliary screening step, the luminous efficiency of the light source can be ensured to be within a reasonable range, avoiding the need to blindly increase the power of the equipment due to excessively low light energy utilization, which may lead to engineering problems such as poor heat dissipation or excessive energy consumption.

[0041] In a practical application, due to structural obstruction or limitations in LED arrangement, the light distribution at the edge of the light source is often irregular and unusable in actual lighting. To eliminate interference from ineffective light, step S50 includes the following methods for extracting the effective light-emitting area: Establish a coordinate system for the illuminance grid of the arc or hemispherical surface: with the central axis of the dome reflector as the polar axis, define... The coordinates represent the angular or radial distance from the detection point to the top of the reflector. The coordinates are the azimuth angles of the detection point on the horizontal plane (0°~360°, along the circumference).

[0042] The light source on the arc-shaped diffuser is divided into several independent illumination zones (e.g., along...). The direction is divided into four equal sectors, that is, the light source is divided into four 90° sectors in the horizontal plane according to the azimuth angle. Individual lighting zones are triggered and illuminated independently in sequence. Each time, only one area (e.g., a quarter of the area) is triggered to light up, while the rest of the area is turned off. The specific zone where the maximum illuminance value is located is extracted by ray tracing. This specific zone corresponds to the direction perpendicular to the line scan aperture. Within a specific zone, the illuminance gradient is calculated along the direction perpendicular to the line scan aperture. The point where the illuminance gradient first exceeds a preset threshold is taken as the boundary of the effective luminous area. The area inside this boundary is the effective luminous area, excluding interference edges caused by structural obstruction or LED arrangement. It should be noted that the area outside the boundary is not used in actual lighting due to structural obstruction or LED arrangement limitations. Therefore, in the design, this boundary is used as a limit, and only the inner area is included in the calculation range of the illuminance difference evaluation index.

[0043] In one example, taking a candidate dome reflector (H=130mm, R=140mm) as an example, a spherical illuminance grid coordinate system is established in the optical simulation software: with the central axis of the reflector as the polar axis, a... The coordinates are the angular distance between the detection point and the top of the reflector (i.e., the angle with the axis). The coordinates are the azimuth angle (0°~360°) of the detection point on the horizontal plane. The light source on the arc-shaped diffuser is positioned along... The direction is divided into four equal parts, that is, divided into four independent lighting zones: Zone 1 ( =0°~90°), Zone 2 ( =90°~180°), Zone 3 ( =180°~270°), Zone 4 ( =270°~360°).

[0044] Individual zones were lit sequentially and independently: first, only zone 1 was lit, the remaining zones were turned off, ray tracing was performed, and the illuminance values ​​at each point on the workpiece surface detection line were counted, recording the maximum illuminance value in that zone; similarly, zones 2, 3, and 4 were tested sequentially. The test results showed that zone 2 ( The maximum illuminance is generated at 90° to 180°, indicating that this zone corresponds to the direction perpendicular to the line scan aperture (i.e., the direction perpendicular to the scanning direction of the line scan camera), which is the key lighting area to be utilized in this design.

[0045] Within zone 2, the illuminance gradient is calculated point-by-point along the direction perpendicular to the line-scanned aperture (i.e., radial, corresponding to the direction of increasing angle θ). A gradient threshold of T = 10000 lx / mm (a design empirical value) is set. Starting from a position near the central axis within the zone (θ = 0°), moving outwards, when the illuminance gradient at a point relative to the adjacent inner point first exceeds 10000 lx / mm, that point is marked as the boundary of the effective luminous area. The area outside the boundary experiences a sharp drop in illuminance and extremely irregular distribution due to structural obstruction or LED arrangement limitations, and is therefore not used in actual lighting. Therefore, subsequent illuminance difference evaluation indicators... The calculations are performed only within the effective luminous area, thus eliminating interference from ineffective light and ensuring that design optimization is only for the actual usable lighting area.

[0046] In one specific embodiment, the defect resolution depends on the actual difference in image grayscale, which manifests as an illuminance difference within the effective luminous area at the light source level. Therefore, in step S60, the illuminance difference evaluation index... for: , in, This represents the maximum illuminance within the effective luminous area. This represents the minimum illuminance within the effective luminous area. This illuminance difference evaluation index... The larger the value, the more drastic the change in illuminance, which is more beneficial for subsequent algorithms to identify and process defects.

[0047] In one example, a candidate model (H=130mm, R=140mm) after the effective luminous area is extracted in step S50 is used. Within the effective luminous area of ​​this model, illuminance values ​​are collected point by point along the scanning direction perpendicular to the scanning direction, and the maximum illuminance is measured. =3416672 lx, minimum illuminance =504413lx, according to the formula Calculated =2912259lx.

[0048] In contrast, another candidate model (H=140mm, R=130mm) measured the maximum illuminance within the effective luminous area. =3296917 lx, minimum illuminance =492162lx, calculated to =2804755 lx. The illuminance difference value of this model is slightly smaller than that of the previous model, indicating that its illuminance change is relatively less drastic.

[0049] Illuminance Difference Evaluation Index It directly reflects the range of illuminance variation within the effective luminous area. The larger the value, the steeper the illuminance gradient of the emitted light perpendicular to the scanning direction. For small defects on highly reflective metal surfaces, such as shallow scratches and micro-pits that rely on grayscale jumps, a clearer image contrast can be formed, thereby reducing the algorithm's false negative rate. In comparison, the surface shape in this embodiment with H=130mm and R=140mm... The design with the higher value was selected as the preferred design.

[0050] In an alternative embodiment, to ensure the absolute reliability of the optimization results in industrial applications, the optimization method further includes: The candidate models selected in step S60 are processed into physical objects and installed in the line scanning imaging system. The highly reflective metal sample was subjected to actual imaging, and the acquired image was input into a preset defect extraction algorithm for secondary verification. The model was then determined based on the verification results.

[0051] In one example, the preferred candidate model selected in step S60 is used. A dome-shaped light source reflector is fabricated according to the model's geometric parameters. The inner wall of the reflector is coated with a barium sulfate diffuse reflection coating, and an elliptical observation hole with a major axis of 70 mm and a minor axis of 26 mm is opened at the top. The fabricated reflector is assembled with a Lambertian LED array to form a complete dome light source, which is then installed in a line scanning imaging system. System configuration: Line scanning camera resolution of 4k, pixel size of 7 μm, lens focal length of 50 mm, working distance of 250 mm, and motion platform scanning speed of 50 mm / s.

[0052] A highly reflective metal sample (such as the edge of a lithium battery aluminum shell) with pre-existing shallow scratches and micro-pits on its surface was selected. The sample was placed on a motion platform, and a dome light source was illuminated at its rated power while a camera simultaneously acquired images. The acquired raw images were transmitted to an industrial control computer via an image acquisition card and input into a preset defect extraction algorithm. The algorithm identified and labeled the scratches and pits in the images.

[0053] The secondary verification results show that the grayscale gradient of the defect reaches 135 grayscale differences, which is higher than the grayscale difference detected by the algorithm before optimization. The verification results meet the requirements of industrial field inspection. Based on this, the surface shape (H=130 mm, R=140 mm) is finalized as the mass production solution. If the verification results do not meet the standards (e.g., grayscale difference less than 81 levels), return to step S60 to select a suboptimal candidate model (e.g., H=140 mm, R=130 mm) and repeat the above measurement and verification process until a finalized solution that meets the reliability requirements is obtained.

[0054] This application also provides a system for optimizing the curvature of an arc-shaped diffuse reflector. This system can execute optimization methods for the curvature of an arc-shaped diffuse reflector and can be mounted on a computer workstation, a dedicated optical design software platform, or a computer medium. The system may include a parametric modeling module, a geometric constraint module, a simulation iteration module, a uniformity screening module, a partitioning and effective region extraction module, and an evaluation and optimization module. The parametric modeling module is used to optimize the curvature of the diffuse reflector based on the base radius. and overall height A 3D parametric model of the arc-shaped diffuse reflector is constructed. The geometric constraint module is used to establish the overall height based on the operating conditions of the line scan camera. The geometric constraint range is defined. The simulation iteration module generates multiple candidate models within this range and performs ray tracing simulation using optical simulation tools. The uniformity screening module obtains the illuminance distribution of each candidate model on the detection line of the surface under test, eliminating candidate models that do not meet the preset line scan direction illuminance uniformity threshold (e.g., 0.9). The partitioning and effective region extraction module uses a partitioned illumination strategy to obtain the illuminance distribution perpendicular to the line scan opening direction for the retained candidate models and extracts the effective luminous region. The evaluation and optimization module calculates the illuminance difference evaluation index of each candidate model within the effective luminous region and selects candidate models with maximizing this index as the optimization objective.

[0055] In a specific example, the method for optimizing the curvature of the arc-shaped diffuse reflector includes: Step S10: Determine the initial geometric constraints For lithium battery edge detection scenarios, the overall height H of the dome reflector is set to range from 130mm to 170mm, and the base radius is... The value range is 130mm to 160mm. Meanwhile, based on the line scan camera parameters (working distance 250mm, field of view half-angle)... ) and safety margin =1.5 mm, and the minor axis of the top elliptical observation hole needs to be ≥26 mm, and the major axis is 70 mm.

[0056] Step S20: Establish the parameterized model and traversal With a step size of 5 mm and Perform a mesh traversal to generate a set of candidate models. Then, take one of the candidate models ( =130mm, Taking a 140mm diameter dome as an example, a hemispherical dome reflector is constructed in 3D modeling software, with an observation hole at the top that has a major axis of 70mm and a minor axis of 26mm.

[0057] Step S30, Optical Simulation Import the model into the optical simulation software, set the light source to a Lambertian LED array, and apply a white diffuse reflection coating (reflectivity 0.92, Lambertian scattering model) to the inner surface of the reflector. Perform Monte Carlo ray tracing (ensuring the number of rays meets the convergence requirement) to extract the illuminance values ​​of the detection lines along the scanning direction on the workpiece surface.

[0058] Step S40, Screening of Illumination Uniformity in Line Scan Direction The model measures samples at 1.8 mm intervals along the scanning direction. =44810lx, =48089lx, calculate =0.932, preset threshold =0.9, which meets the requirements and is retained. The overall light energy utilization rate η meets the requirements for simultaneous auxiliary screening.

[0059] Step S50: Extraction of effective luminescent region A zoned lighting strategy is adopted, with the reflector positioned along the azimuth angle. Divide the space into four equal parts, and light up each section independently in turn. Result display The illuminance maximum is generated in the 90° to 180° zone, which corresponds to the perpendicularity of the scanning direction. The illuminance gradient is calculated radially within this zone, and a gradient threshold of 10000 lx / mm is set. The position where the threshold is first exceeded is recorded, and the area inside this position is defined as the effective luminous region.

[0060] Step S60, Evaluation of Illuminance Difference Within the effective luminous area, the maximum illuminance was measured. =3416672 lx, minimum illuminance =504413 lx, calculate the illuminance difference evaluation index. = - =2912259 lx. This value is significantly higher than that of another candidate model (H=140 mm, R=130 mm). =2804755lx), indicating that the former has a more drastic change in illuminance, which is more conducive to the grayscale contrast of micro-defects.

[0061] Step S70, Physical verification and finalization A model (H=130mm, R=140mm) was fabricated into a physical dome light source and installed in a line-scan imaging system. Pre-existing shallow scratches on a lithium-ion battery aluminum shell sample were used for actual imaging. The acquired images were validated using a defect extraction algorithm: before optimization, the edge defect grayscale difference was 81 levels; after optimization, the edge defect grayscale difference was 135 levels, representing a resolution improvement of 1.67 times. This surface was ultimately determined as the mass production solution, and the curvature parameters of the optimized arc-shaped diffuse reflector were recorded.

[0062] Figure 2A comparison of illuminance distribution curves before and after optimization of the arc-shaped diffuse reflector is shown. Specifically, ray tracing simulations were performed on the dome reflector before optimization and the dome reflector after optimization using this embodiment, and the relative light intensity of each angular zone in the direction perpendicular to the line scan aperture was statistically analyzed. In the figure, the horizontal axis represents the illuminance region (corresponding to different spatial positions or emission angle zones in the direction perpendicular to the line scan aperture), and the vertical axis represents the normalized illuminance value. As can be seen from the figure, the illuminance distribution curve of the dome reflector before optimization is relatively flat, and the gradient of illuminance value with position is small, indicating that the angular distribution of emitted light in the direction perpendicular to the line scan aperture is relatively uniform and the changes are not drastic. The illuminance distribution curve of the dome reflector after optimization shows a significantly steeper rising and falling edge in the effective luminous area, and the difference between the illuminance peak and valley values ​​is significantly increased, indicating that the angular distribution of emitted light in this direction is more concentrated and the changes are more drastic. This comparison shows that the optimization method of this embodiment effectively increases the drastic degree of illuminance change in the direction perpendicular to the line scan aperture, and the quantitative evaluation index The value is significantly higher than before optimization, providing an optical basis for achieving high-contrast imaging of tiny defects on highly reflective metal surfaces.

[0063] Figure 3a , Figure 3b The comparison of the original images before and after optimization is shown. Figure 3a This illustrates the use of a traditional hemispherical arc-shaped diffuse reflector to acquire images of the lithium battery's edge. Figure 3b The image shows an image of the same location acquired using a non-hemispherical arc-shaped diffuse reflector optimized according to this embodiment. A comparison reveals that the optimized original image exhibits more pronounced grayscale contrast at the defect location, and the defect edges are more clearly discernible. This demonstrates that the optimization method of this embodiment can significantly improve the visualization of minute defects on highly reflective metal surfaces in the original image.

[0064] Figure 4a , Figure 4b The comparison of the algorithm processing results before and after optimization is shown. Figure 4a These are the defect detection results under a traditional hemispherical arc-shaped diffuse reflector. Figure 4b The detection results under the optimized arc-shaped diffuse reflector of this invention, after quantitative analysis, show that in the dimension perpendicular to the scanning direction, the grayscale difference between the defect and the background produced by the traditional hemispherical reflector is 81 levels, while the optimized reflector reaches 135 levels, an improvement of approximately 1.67 times. These results demonstrate that the method for optimizing the curvature of the arc-shaped diffuse reflector proposed in this embodiment can significantly enhance the imaging contrast of small defects on highly reflective metal surfaces, which is beneficial for improving the recognition accuracy and stability of subsequent defect detection algorithms.

[0065] The key performance indicators before and after optimization are shown in the table below:

[0066] As shown in the table above, the reflector optimized by this method can significantly enhance the drastic change in the angle of light in the sensitive direction of the algorithm, improve the detection performance of highly reflective metal surface defects, and verify the effectiveness of this method.

[0067] The comparative results above demonstrate that the proposed method for optimizing the curvature of the arc-shaped diffuse reflector significantly outperforms traditional methods in three aspects: steepness of illumination distribution, visibility of defects in the original image, and grayscale contrast detected by the algorithm. This verifies the effectiveness of the optimization method in enhancing the detection capability of minute defects on highly reflective metal surfaces. The optimization method and system selectively extract the effective luminescent area and optimize the surface shape by maximizing the illumination difference within this area (perpendicular to the direction of the line scan aperture). This significantly increases the drastic change in the angle of emitted light rays in this direction (making its distribution steeper), allowing minute defects on highly reflective metal surfaces to form higher grayscale contrast during imaging. This substantially improves the accuracy and feature resolution of subsequent image processing algorithms, such as defect detection, from a physical optics perspective. The arc-shaped diffuse reflector formed by this optimization method enhances the drastic change in the angle of light rays in the vertical direction, improving the grayscale contrast between defects and the background.

[0068] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0069] The above embodiments merely illustrate several implementation methods of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A method for optimizing the curvature of an arc-shaped diffuse reflector, characterized in that, The optimization method includes: S10: Determine the geometric design variables that affect the curvature distribution of the inner surface of the arc-shaped diffuse reflector, wherein the design variables include at least the bottom radius of the arc-shaped diffuse reflector. and overall height And construct a three-dimensional parametric model of the arc-shaped diffuse reflector; S20: Establish the overall height based on the working conditions of the line scan camera. The range of geometric constraints; S30: Within the geometric constraints, multiple candidate models are generated by traversing the design variables, and ray tracing simulation is performed on each candidate model using an optical simulation tool; S40: Obtain the illuminance distribution of each candidate model on the detection line of the surface to be tested, and eliminate candidate models that do not meet the preset illuminance uniformity threshold in the line scan direction; S50: For the retained candidate models, a zoned illumination strategy is used to obtain the illuminance distribution perpendicular to the direction of the line scan aperture, and the effective luminous area is extracted. S60: Calculate the illuminance difference evaluation index of each candidate model within the effective luminous area, and select the optimal candidate model with the goal of maximizing the illuminance difference evaluation index.

2. The method for optimizing the curvature of an arc-shaped diffuse reflector as described in claim 1, characterized in that, In step S20, the overall height The geometric constraint range must simultaneously satisfy the following conditions: Upper limit constraint: The overall height limit is determined based on the working distance of the line scan camera and the installation safety margin; Lower limit of field of view constraint: The lower limit of the height of the arc-shaped diffuser is determined based on the width of the top opening of the arc-shaped diffuser, the half angle of the field of view of the line scan camera perpendicular to the scanning direction, and the safety margin. Optical lower limit constraint: The minimum overall height limit determined through optical simulation.

3. The method for optimizing the curvature of an arc-shaped diffuse reflector as described in claim 2, characterized in that, The overall height includes the height of the arc-shaped diffuse reflector. and base thickness , The upper limit constraint must satisfy: ， in, Indicates the working distance of the line scan camera. This indicates a safety margin for installation. The lower limit constraint of the field of view must satisfy: , in, This indicates the width of the opening at the top of the arc-shaped diffuser; This refers to the half-angle of the field of view of the line scan camera perpendicular to the scanning direction. For safety margin, the optical lower limit constraint must satisfy: ,in, This represents the minimum optical height required to meet basic diffuse reflection requirements.

4. The method for optimizing the curvature of an arc-shaped diffuse reflector as described in claim 1, characterized in that, Step S40 includes: the calculation method for the uniformity of illumination in the line scan direction is as follows: = ,in, This indicates the minimum illuminance on the testing line. Indicates maximum illuminance; excludes those that do not meet the requirements. The surface shape, the preset line scan direction illuminance uniformity threshold .

5. The method for optimizing the curvature of an arc-shaped diffuse reflector as described in claim 4, characterized in that, In step S40, after eliminating candidate models that do not meet the preset line scan direction illuminance uniformity threshold, an auxiliary screening step is also included: Obtain the overall light energy utilization rate of each candidate model on the test plane. The overall light energy utilization rate It is the ratio of the total luminous flux received by the detection surface to the total luminous flux of the light source; Excluding the overall light energy utilization rate The candidate reflector model is below the set lower limit value.

6. The method for optimizing the curvature of an arc-shaped diffuse reflector as described in claim 1, characterized in that, In step S50, the partitioned lighting strategy and the method for extracting the effective luminous area include: Establish a coordinate system for the illuminance grid of a circular or hemispherical surface; The light source on the arc-shaped diffuser is divided into several independent lighting zones; Individual lighting zones are triggered and illuminated independently in sequence. The specific zone where the maximum illuminance value is located is extracted by ray tracing. This specific zone corresponds to the direction perpendicular to the line scan aperture. Within the specific partition, the illuminance gradient is calculated along the direction perpendicular to the line scan aperture. The position where the illuminance gradient first exceeds a preset threshold is taken as the boundary of the effective luminous area, and the area inside this boundary is the effective luminous area.

7. The method for optimizing the curvature of an arc-shaped diffuse reflector as described in claim 1, characterized in that, In step S60, the illuminance difference evaluation index is: , in, This represents the maximum illuminance within the effective luminous area. This represents the minimum illuminance within the effective luminous area.

8. The method for optimizing the curvature of an arc-shaped diffuse reflector as described in claim 1, characterized in that, The optimization method further includes: The candidate models selected in step S60 are processed into physical objects and installed in the line scanning imaging system. The highly reflective metal sample was subjected to actual imaging, and the acquired image was input into a preset defect extraction algorithm for secondary verification. The model was then determined based on the verification results.

9. A system for optimizing the curvature of an arc-shaped diffuse reflector, characterized in that, include: Parametric modeling module: used to model based on the base radius and overall height Construct a three-dimensional parametric model of an arc-shaped diffuse reflector; The geometric constraint module is used to establish the overall height based on the operating conditions of the line scan camera. The range of geometric constraints; The simulation iteration module is used to generate multiple candidate models within the geometric constraints and perform ray tracing simulation using optical simulation tools. The uniformity screening module is used to obtain the illuminance distribution of each candidate model on the detection line of the surface to be tested, and to remove candidate models that do not meet the preset line scan direction illuminance uniformity threshold. The partitioning and effective region extraction module is used to obtain the illuminance distribution perpendicular to the direction of the line scan aperture for the retained candidate models using a partitioned illumination strategy, and to extract the effective luminous region. The evaluation and selection module is used to calculate the illuminance difference evaluation index of each candidate model in the effective luminous area, and to select candidate models with the goal of maximizing the index.

10. An arc-shaped diffuse reflector, used in a line scanning imaging system, characterized in that, The arc-shaped diffuse reflector is obtained using the optimization method described in any one of claims 1-9.

Images