A self-adapting multi-angle variable light field surface defect detection composite light source system
Patent Information
- Application Number
- CN202610785235.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-02
- Publication Date
- 2026-08-18
AI Technical Summary
[0006]为解决上述背景技术中提出的问题,本发明提供一种自适应多角度变光场的表面缺陷检测复合光源系统以解决曲面光学镜片及类似高反射曲面件表面微小缺陷在现有照明条件下无法适配不同方向缺陷、难以适应曲面法向变化、检测效率低且存在定位误差的问题
[0019]1. Significantly improved detection stability: By using circumferentially multi-directional independent and controllable dark field illumination, combined with the scratch directional scattering evaluation algorithm, scratch defects in any direction can be effectively stimulated, improving the detection rate of scratches in unknown directions.
Smart Images

Figure CN122591676A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of optical detection technology for surface defects, and specifically to an adaptive multi-angle variable light field composite light source system for surface defect detection. Background Technology
[0002] Precision optical components such as high-end curved optical lenses, curved glass, and transparent covers are prone to minor defects such as scratches, pitting, shallow indentations, deposits, and microcracks during processing and assembly. These defects are small in size, have weak contrast, and exhibit significant direction dependence. At the same time, the continuous change of the normal to the curved surface leads to large differences in the incident conditions of fixed illumination in different areas, making it very easy for defects to be missed.
[0003] Current surface defect detection mainly adopts a technical approach of rapid detection in dark field and precise measurement in bright field: dark field imaging achieves bright defect imaging against a dark background through low-angle incident light, while bright field is used for defect boundary extraction and size measurement. However, most existing illumination systems are fixed optical path designs, including unidirectional dark field, full-ring dark field, and split bright-dark field switching structures, which are only suitable for planar workpieces or single defect types and cannot simultaneously meet the requirements of detection sensitivity, background suppression, and measurement consistency for curved surfaces.
[0004] Among them, the unidirectional dark field is only effective for scratches in a specific direction and is prone to failure when the curvature of the surface changes; the whole-ring dark field causes the background to rise due to the superposition of reflections in multiple directions, resulting in a decrease in the contrast of weak defects; the split bright and dark field requires switching stations or inspection heads, which has repeated positioning errors and low efficiency.
[0005] In summary, the existing technology has the following shortcomings: poor selectivity of illumination direction, unable to adapt to defects in different directions; difficulty in adapting to changes in the normal of curved surfaces; separate arrangement of bright and dark fields, resulting in low detection efficiency and positioning errors. Summary of the Invention
[0006] To address the problems mentioned in the background art, the present invention provides an adaptive multi-angle variable light field surface defect detection composite light source system to solve the problems that under existing lighting conditions, small defects on curved optical lenses and similar highly reflective curved surfaces cannot adapt to defects in different directions, are difficult to adapt to changes in the normal of the curved surface, have low detection efficiency, and have positioning errors.
[0007] To achieve the above objectives, the present invention provides the following technical solution:
[0008] An adaptive multi-angle variable light field surface defect detection composite light source system includes:
[0009] Imaging unit: The main optical axis of the imaging unit is set perpendicularly to the curved surface of the workpiece being measured;
[0010] Composite lighting structure: The light-emitting surface of the composite lighting structure is set to point towards the curved surface of the workpiece being measured. The composite lighting structure includes an annular bright field lighting component set with the main optical axis of the imaging unit as the center, and multiple dark field lighting units discretely set around the annular bright field lighting component in the circumferential direction. The dark field lighting units have the ability to adjust the incident tilt angle.
[0011] Control Module: The control module is electrically connected to the imaging unit, the annular bright field illumination assembly, and each dark field illumination unit. The control module controls image acquisition by the imaging unit within a given time sequence, and controls the on / off state and brightness of the annular bright field illumination assembly and each dark field illumination unit within a given time sequence. The control module is configured to execute the following steps in sequence:
[0012] S1: Obtain the detection task parameters, which include the curvature parameters and local surface normal data of the workpiece under test, as well as the image data of the workpiece under test.
[0013] S2: Based on the detection task parameters, calculate the actual incident angle of any detection point on the surface under test to each dark field illumination unit light source, calculate the defect imaging contrast under different actual incident angles and determine the optimal dark field illumination unit incident tilt angle of the current detection area, evaluate the excitation effect of dark field illumination unit light sources in different orientations on scratch defects, and calculate the comprehensive illumination adaptability or image quality evaluation value under different illumination configurations.
[0014] S3: Based on the actual incident angle, the incident tilt angle of the optimal dark field illumination unit, and the scratch excitation effect evaluation results obtained from S2, the optimal illumination scheme is selected by maximizing the overall illumination adaptability or image quality evaluation value, and the corresponding control command is generated.
[0015] S4: According to the control command, control the on / off state, brightness, incident tilt angle, and lighting combination mode of each dark field lighting unit, and control the mode switching between bright field lighting and dark field lighting;
[0016] S5: Drive the imaging unit to acquire detection images under the corresponding illumination mode; when acquiring images under dark illumination mode, perform preliminary defect detection to locate candidate defect regions; when acquiring images under bright illumination mode, perform boundary extraction, size measurement, and multifocal analysis on the located candidate defect regions.
[0017] S6: Combine the initial defect detection results in dark field mode and the analysis results in bright field mode to output the final defect detection results.
[0018] Compared with the prior art, the beneficial effects of the present invention are:
[0019] 1. Significantly improved detection stability: By using circumferentially multi-directional independent and controllable dark field illumination, combined with the scratch directional scattering evaluation algorithm, scratch defects in any direction can be effectively stimulated, improving the detection rate of scratches in unknown directions.
[0020] 2. Strong adaptability to curved surfaces: Through the adjustable incident tilt angle design of the dark field lighting unit, combined with the calculation of the actual incident angle of the curved surface and the algorithm for determining the optimal tilt angle, it can adapt to curved surface products with different radii of curvature, solve the problems of local background lifting and edge reflection residue caused by fixed tilt angle lighting, and reduce background noise.
[0021] 3. High detection efficiency: By coaxially integrating dark field illumination and bright field illumination into the same detection head, a seamless switch between rapid dark field detection and precise bright field retesting is achieved, eliminating the repetitive positioning errors of split-type arrangements and improving detection efficiency. Attached Figure Description
[0022] Figure 1 This is a schematic diagram of the structure from below in this application;
[0023] The diagram is marked as follows:
[0024] 1-Imaging unit; 2-Annular bright field illumination assembly; 3-Dark field illumination unit; 4-Annular base. Detailed Implementation
[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0026] like Figure 1 As shown, an adaptive multi-angle variable light field surface defect detection composite light source system includes an imaging unit 1, a composite illumination structure, and a control module. All components are integrated into the same detection head. Specifically:
[0027] The imaging unit 1 has its main optical axis vertically pointed to the surface being measured. It employs a combination of a high-resolution industrial camera and a telecentric lens, providing low-distortion imaging and ensuring consistent measurement accuracy of defect dimensions at different locations. The exposure time of the imaging unit is synchronized with the illumination sequence of the illumination unit.
[0028] The light-emitting surface of the composite illumination structure is oriented towards the surface being measured, and the entire structure is mounted on the front end of the lens of the imaging unit 1, with its relative position to the imaging unit 1 remaining fixed. The composite illumination structure includes an annular base 4, an annular bright field illumination component 2 horizontally positioned on the bottom surface of the annular base 4 with the main optical axis of the imaging unit 1 as the center, and eight dark field illumination units 3 discretely arranged horizontally along the outer periphery of the annular base 4. A through hole is provided on the central axis of the annular base for the imaging unit 1 to pass through.
[0029] The annular bright-field illumination component 2 adopts an annular segmented or continuous annular diffuse structure and is coaxially arranged with the imaging unit 1. The annular bright-field illumination component 2 is used for contour observation, boundary extraction, and multi-focal plane imaging of defect areas. In bright-field mode, light is incident on the measured surface along the near-normal direction. Specular reflections from intact surfaces return along the original path to the imaging unit 1, forming a bright background. Due to scattering, the intensity of reflected light changes in defect areas, thus forming dark defects in the image. This imaging mode can clearly present the boundaries and texture details of defects, making it suitable for dimensional measurement and multi-focal plane fusion imaging.
[0030] Eight dark field illumination units 3 are evenly distributed along the horizontal circumference, with an azimuth angle interval of 45° between adjacent units. The azimuth angles are 0°, 45°, 90°, 135°, 180°, 225°, 270°, and 315°, respectively. Each dark field illumination unit 3 can be independently controlled for on / off switching and brightness, with a brightness adjustment range of 0-100%.
[0031] Each dark field illumination unit 3 includes a light source module, a collimating lens, and a tilt adjustment mechanism:
[0032] Light source module: Employs narrowband LED light sources or collimated and shaped high-stability semiconductor light sources with a wavelength range of 450nm to 650nm. Single-wavelength, dual-wavelength, or multi-wavelength combinations can be used depending on the material being measured and the type of defect. For common optical glass and coated lenses, monochromatic or narrowband light sources in the visible light range are preferred; shorter wavelengths typically exhibit higher scattering sensitivity at minute edges and fine scratches.
[0033] Collimating lens: Located in front of the light source module, it is used to shape the divergent light emitted by the light source into an illumination beam with a low divergence angle, ensuring that the illumination beam can illuminate the designated area of the surface under test at a stable incident angle.
[0034] The tilt adjustment mechanism includes a miniature stepper motor and a hinge. The light source module is hinged to the outer periphery of the annular base. The miniature stepper motor is fixedly mounted on the annular base, and its output shaft is directly and coaxially fixedly connected to the hinge shaft of the rotating bracket. The incident tilt angle adjustment range of the tilt adjustment mechanism is 5° to 40°, preferably 10° to 30° (relative to the local tangent plane of the surface being measured). By adjusting the incident tilt angle, the angle between the illumination beam and the surface being measured can be changed, thereby optimizing the defect scattering intensity and background suppression effect.
[0035] By controlling the on / off state, brightness, and tilt angle of different dark field lighting units 3, the system can form the following 5 lighting modes:
[0036] Unidirectional dark field mode: Only one dark field illumination unit is lit, which is used to detect scratch defects in a specific direction;
[0037] Bidirectional combined dark field mode: Illuminates two dark field illumination units in opposite or adjacent directions to detect cross scratches or complex shape defects;
[0038] Sequential scanning dark field mode: The dark field illumination units in all directions are lit in sequence to acquire multiple images of the same area, which are used to detect scratches and defects in unknown directions.
[0039] Full-circle composite dark field mode: Simultaneously illuminates dark field illumination units in all directions for rapid scanning of large areas and initial screening of defects;
[0040] Bright field precision measurement mode: Turn off all dark field illumination units and turn on the ring bright field illumination component for precise measurement and morphology analysis of the located defects.
[0041] The control module is electrically connected to the imaging unit 1, the annular bright field illumination component 2, and each dark field illumination unit 3, respectively, and can realize synchronous control and data interaction of all components. The control module includes a processor, a signal interface, a storage unit, and a power management unit.
[0042] The control module is configured to execute the complete detection process in the following order:
[0043] S1: Obtain detection task parameters; the control module first obtains the detection task parameters, including:
[0044] Workpiece parameters: radius of curvature, surface equation, external dimensions, coating type;
[0045] Local surface normal data: pre-obtained normal vectors at each point on the measured surface;
[0046] Image feedback data: Real-time captured preview images used for online adjustment of lighting parameters.
[0047] S2: Multi-dimensional lighting condition calculation; Based on the acquired detection task parameters, the control module performs the following core calculations:
[0048] Calculation of the actual incident angle of the surface; assuming the unit normal of the measured surface at any detection point P is... The incident direction unit vector of the k-th dark field light source is Then the actual angle of incidence at that point is:
[0049] ;
[0050] By calculating the actual incident angle of each detection point for all dark field light sources, the illumination effectiveness of different light sources at that point can be evaluated.
[0051] Calculation of defect scattering intensity; The scattering intensity of surface defects is related to the incident light parameters, defect geometric parameters, and material optical parameters, and its general expression is:
[0052] ;
[0053] in, scattering direction The intensity of scattered light on the surface, The intensity of the incident light. This is the bidirectional scattering distribution function, which is related to the light source wavelength λ, defect depth h, defect width w, material refractive index n, and incident tilt angle. and incident azimuth Related, and correspond.
[0054] Defect imaging contrast calculation and optimal tilt angle determination; Let α be the angle between the dark field beam and the local tangent plane of the measured surface. The defect imaging contrast is defined as:
[0055] ;
[0056] Among them, I d (α) represents the imaging intensity of the defect region at tilt angle α, I b (α) represents the imaging intensity of the background region under the same conditions, and ε is a regularization term introduced to prevent the denominator from being too small. The optimal dark field incident tilt angle is the tilt angle corresponding to the maximum contrast:
[0057] ;
[0058] By calculating the contrast at different tilt angles, the optimal illumination tilt angle for the current detection area is determined.
[0059] Evaluation of the directional scattering effect of scratches; scratches can usually be approximated as linear microstructures with a length much greater than their width, and their scattering intensity is closely related to the angle between the incident direction and the scratch direction. Let t be the unit vector of the principal direction of the scratch in the tangent plane of the surface, and pi be the projection direction of the incident light in the same plane, with the angle between them being... The scratch scattering response intensity is:
[0060] ;
[0061] Where m is an empirical index characterizing directional selectivity. The modulation function is related to the incident tilt angle, scratch depth h, characteristic width w, and light source wavelength λ. The excitation effect is evaluated by calculating the scattering response intensity of the light source to the scratch from different orientations.
[0062] Calculation of integrated lighting adaptability and image quality evaluation value; to comprehensively evaluate the suitability of different lighting configurations, a comprehensive lighting adaptability function is defined:
[0063] ;
[0064] in, Let be the defect imaging contrast value under the k-th illumination configuration, and let be the scratch scattering response intensity value under the k-th illumination configuration. As a penalty term for background reflection or false brightness response, , , These are the weighting coefficients. Simultaneously, a light field pattern image quality evaluation value is defined:
[0065] ;
[0066] in, As a defect significance indicator, This is a local contrast index. For background noise or stray reflection intensity, , , These are the weighting coefficients.
[0067] S3: Selection of the optimal lighting scheme;
[0068] Based on the actual incident angle, optimal dark field incident tilt angle, and scratch excitation effect evaluation results obtained from S2, the control module maximizes the overall lighting adaptability. or image quality rating The optimal lighting configuration is selected from all possible lighting configurations using the following method:
[0069] or ;
[0070] in, Let k be the number of the optimal lighting configuration, and k is the set of all possible lighting configurations.
[0071] The control module selects the optimal lighting configuration. Generate corresponding control commands, including the on / off status of each lighting unit, brightness value, tilt angle value, lighting combination mode, image acquisition sequence, and bright / dark field mode switching timing.
[0072] S4: Lighting and data acquisition control;
[0073] The control module sends control signals to the annular bright field illumination component 2 and each dark field illumination unit 3 according to the generated control commands, adjusting their on / off state, brightness and tilt angle; at the same time, it sends trigger signals to the imaging unit 1 to synchronously control the image acquisition timing.
[0074] S5: Image Acquisition and Processing;
[0075] The control module drives the imaging unit 1 to acquire detection images under the corresponding illumination mode and processes the acquired images in real time:
[0076] Dark field mode image processing: Perform initial defect detection, extract features such as the location and area of defects, and locate candidate defect regions;
[0077] Bright field mode image processing: perform boundary extraction, size measurement and multi-focal plane fine analysis on the located defect candidate regions.
[0078] S6: Output the result;
[0079] The control module integrates the initial defect detection results in dark field mode and the detailed analysis results in bright field mode to generate a final defect detection report, including information such as the location, size, type and grade of the defect, and outputs it to an external device for display and storage.
[0080] For curved products with different radii of curvature, coating conditions, or dimensions, the system can pre-establish corresponding light field configuration templates in the control module. These templates include recommended dark field orientation combinations, brightness levels for each channel, tilt angle settings, and bright field switching rules. When inspecting a specific product, the control module can automatically call the corresponding preset light field template based on the product model, eliminating the need for manual adjustments to the light source structure. If necessary, the control module can also fine-tune the template based on real-time image feedback to further improve the visibility of defects in localized areas.
[0081] The following are specific application examples:
[0082] Example 1: Used for detecting scratches of unknown direction on the surface of curved optical lenses;
[0083] When detecting scratches in unknown directions, the system first illuminates each dark field direction sequentially, acquiring multiple dark field images of the same area. Then, based on the saliency of the defects in the images or by directly fusing multiple images, it obtains defect detection results that respond well to scratches in different directions. For the located scratch area, the dark field light source is turned off, and the annular bright field illumination is turned on for high-resolution observation and multi-focal plane acquisition, used for subsequent boundary extraction and dimensional measurement. In practical use, compared to fixed single-direction dark field illumination, the 8-direction adjustable dark field structure in this embodiment significantly improves the detection stability of scratches in unknown directions.
[0084] Example 2: Used for edge area inspection of curved coated glass;
[0085] For the edge region of curved coated glass, due to the significant change in edge curvature and the tendency to produce specular reflection residue, simultaneous illumination with all dark-field light sources can easily raise the background, affecting the visibility of subtle defects. In this embodiment, the system activates only 1 to 2 dark-field illumination units that better match the local edge normal, based on the spatial location of the edge region, and adjusts their tilt angle to be closer to grazing incidence. This reduces reflections from defect-free areas into the imaging optical path while enhancing the scattering response of edge scratches and minor edge chips. This implementation method can achieve locally optimized illumination for edge regions through selective illumination and angle adjustment without changing the overall hardware structure.
[0086] Example 3: Full-circle composite dark field mode;
[0087] In certain inspection tasks, the system can first use a multi-directional dark field to rapidly scan a large curved surface to obtain candidate defect regions. Then, keeping the relative position of the workpiece and the inspection head unchanged, the system switches the illumination mode to a ring-shaped bright field only through the control module, and performs high-resolution magnified imaging and multi-focal plane acquisition in the same area. If the system further integrates a confocal depth sensor, the bright field image and depth measurement results can also share the same spatial positioning relationship. This implementation demonstrates the structural advantages of this patent in the integration of detection and precise measurement.
[0088] 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 within the protection scope of the present invention.
Claims
1. An adaptive multi-angle variable light field composite light source system for surface defect detection, characterized in that, include: Imaging unit; The main optical axis of the imaging unit is set perpendicularly to the curved surface of the workpiece being measured; Composite lighting structure; The light-emitting surface of the composite lighting structure is set to point towards the curved surface of the workpiece being measured. The composite lighting structure includes an annular bright field lighting component set with the main optical axis of the imaging unit as the center, and multiple dark field lighting units set circumferentially discretely around the outer periphery of the annular bright field lighting component. The dark field lighting units have the ability to adjust the incident tilt angle. The control module is electrically connected to the imaging unit, the annular bright field illumination assembly, and each dark field illumination unit. The control module controls image acquisition by the imaging unit within a given time sequence, and controls the on / off state and brightness of the annular bright field illumination assembly and each dark field illumination unit within a given time sequence. The control module is configured to execute the following steps in sequence: S1: Obtain the detection task parameters, which include the curvature parameters and local surface normal data of the workpiece under test, as well as the image data of the workpiece under test. S2: Based on the detection task parameters, calculate the actual incident angle of any detection point on the surface under test to each dark field illumination unit light source, calculate the defect imaging contrast under different actual incident angles and determine the optimal dark field illumination unit incident tilt angle of the current detection area, evaluate the excitation effect of dark field illumination unit light sources in different orientations on scratch defects, and calculate the comprehensive illumination adaptability or image quality evaluation value under different illumination configurations. S3: Based on the actual incident angle, the incident tilt angle of the optimal dark field illumination unit, and the scratch excitation effect evaluation results obtained from S2, the optimal illumination scheme is selected by maximizing the overall illumination adaptability or image quality evaluation value, and the corresponding control command is generated. S4: According to the control command, control the on / off state, brightness, incident tilt angle, and lighting combination mode of each dark field lighting unit, and control the mode switching between bright field lighting and dark field lighting; S5: Drive the imaging unit to acquire detection images under the corresponding illumination mode; when acquiring images under dark illumination mode, perform preliminary defect detection to locate candidate defect regions; When acquiring images under bright field illumination, boundary extraction, size measurement, and multifocal plane analysis are performed on the located defect candidate regions. S6: Combine the initial defect detection results in dark field mode and the analysis results in bright field mode to output the final defect detection results.
2. The adaptive multi-angle variable light field surface defect detection composite light source system according to claim 1, characterized in that, The azimuth interval between two adjacent dark field illumination units is between 30° and 90°.
3. The adaptive multi-angle variable light field surface defect detection composite light source system according to claim 1, characterized in that, The incident tilt angle of the dark field lighting unit can be adjusted from 5° to 40°.
4. The adaptive multi-angle variable light field surface defect detection composite light source system according to claim 1, characterized in that, Both the dark field illumination unit and the ring bright field illumination component use narrow-band LED light sources or collimated and shaped semiconductor light sources. The light source wavelength range of the dark field illumination unit is 450nm to 650nm.
5. The adaptive multi-angle variable light field surface defect detection composite light source system according to claim 1, characterized in that, The incident tilt angle adjustment capability of the dark field lighting unit can be achieved through manual adjustment, arc-shaped slide lock, screw drive, micro motor drive, or piezoelectric actuation.
6. The adaptive multi-angle variable light field surface defect detection composite light source system according to claim 1, characterized in that, The ring-shaped bright field lighting component adopts a ring-shaped segmented or continuous ring-shaped diffused structure.
7. The adaptive multi-angle variable light field surface defect detection composite light source system according to claim 1, characterized in that, The relative position of the workpiece and the imaging unit remains unchanged during the acquisition of dark field mode images and bright field mode images.
8. The adaptive multi-angle variable light field surface defect detection composite light source system according to claim 1, characterized in that, The control module has established light field configuration templates for curved surface products with different radii of curvature, different coating states, or different external dimensions. The light field configuration template also includes corresponding control instructions. The light field configuration template is used to call the corresponding instructions according to the product model. When using the light field configuration template, the control module is configured to directly execute S4-S6.
9. The adaptive multi-angle variable light field surface defect detection composite light source system according to claim 1, characterized in that, Each dark field illumination unit is paired with a collimating lens to form a low-divergence illumination beam.