Optical part surface detection method and optical part surface detection system

Abstract

The invention relates to the technical field of optics, and provides an optical part surface detection method and an optical part surface detection system. The method comprises the steps of determining a shooting angle of a shooting device relative to a to-be-detected surface of a to-be-detected optical part, and determining an incident angle of a light source relative to the to-be-detected surface based on the shooting angle; performing light source irradiation on the to-be-detected surface based on the incident angle, and acquiring dark field image data of the to-be-detected surface in light source irradiation based on a shooting device; according to the method, the incident angle of the light source is dynamically determined by taking the shooting angle as a reference, so that the mirror reflection direction of the to-be-detected surface avoids the shooting view angle of the shooting device during subsequent light source irradiation, and the defect detection result of the to-be-detected surface is obtained. Therefore, an illumination environment meeting dark field imaging conditions is constructed, dark field image data with low background brightness are formed in the obtained image, and detection of tiny defects on the to-be-detected surface is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of optical technology, and particularly to a detection method for a surface of an optical part and a detection system for the surface of the optical part. BACKGROUND

[0002] In the prior art, in the field of industrial production quality inspection, a quality inspector usually performs manual visual inspection, that is, the quality inspector directly observes the surface of a workpiece by naked eye to determine whether defects exist.

[0003] With the increasing refinement of industrial products, the limitations of manual visual inspection are increasingly highlighted, resulting in poor quality inspection effect. SUMMARY

[0004] Therefore, the embodiments of the present application provide a detection method for a surface of an optical part and a detection system for the surface of the optical part to solve the problem of poor quality inspection effect caused by manual quality inspection of the surface of the optical part in the prior art.

[0005] In a first aspect, the embodiments of the present application provide a detection method for a surface of an optical part, which comprises: determining a shooting angle of a shooting device relative to a to-be-detected surface of a to-be-detected optical part, and determining an incident angle of a light source relative to the to-be-detected surface based on the shooting angle; performing light source irradiation on the to-be-detected surface based on the incident angle, and acquiring dark-field image data of the to-be-detected surface in the light source irradiation based on the shooting device; and performing defect detection on the to-be-detected surface based on the dark-field image data to obtain a defect detection result of the to-be-detected surface.

[0006] In a second aspect, the embodiments of the present application provide a detection system for a surface of an optical part, which comprises: a shooting device, an illumination device, and an image processing device. The illumination device is configured to perform light source irradiation on a to-be-detected surface of a to-be-detected optical part based on an incident angle, the incident angle being determined based on a shooting angle of the shooting device relative to the to-be-detected surface of the to-be-detected optical part, and the shooting device is configured to acquire dark-field image data of the to-be-detected surface in the light source irradiation. The image processing device is configured to perform defect detection on the to-be-detected surface based on the dark-field image data to obtain a defect detection result of the to-be-detected surface.

[0007] The beneficial effects of the embodiment of the present application compared with the prior art are: the detection method of the surface of the optical part in the embodiment of the present application determines the shooting angle of the shooting device relative to the surface to be detected of the optical part to be detected, and determines the incident angle of the light source relative to the surface to be detected based on the shooting angle; the surface to be detected is irradiated by the light source based on the incident angle, and the dark field image data of the surface to be detected in the light source irradiation is acquired based on the shooting device; the surface to be detected is detected based on the dark field image data, and the defect detection result of the surface to be detected is obtained, the present application dynamically determines the incident angle of the light source with the shooting angle as the reference, so that the mirror reflection direction of the surface to be detected avoids the shooting visual angle of the shooting device in subsequent light source irradiation, thereby constructing a light irradiation environment meeting the dark field imaging condition, so that the dark field image data with low background brightness and high defect contrast is formed in the acquired image, and then the detection of the micro defect on the surface to be detected is realized by image processing on the above-mentioned dark field image data, thereby improving the quality inspection effect, and avoiding the problem that the quality inspection effect is poor in the prior art by manually inspecting the surface of the optical part. BRIEF DESCRIPTION OF DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0009] Figure 1 is a flowchart of a detection method of a surface of an optical part provided by the embodiment of the present application. DETAILED DESCRIPTION

[0010] In the following description, specific details such as specific system structures, techniques, etc. are presented in order to thoroughly understand the embodiments of the present application. However, it should be clear to those skilled in the art that the present application can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits and methods are omitted to avoid unnecessary details that hinder the description of the present application.

[0011] The detection method of the surface of the optical part and the detection system of the surface of the optical part according to the embodiment of the present application will be described in detail below with reference to the drawings.

[0012] Figure 1 is a flowchart of a detection method of a surface of an optical part provided by the embodiment of the present application, as shown in Figure 1 The detection method of the surface of the optical part comprises:

[0013] S101, determine the shooting angle of the shooting device relative to the surface to be detected of the optical part to be detected, and determine the incident angle of the light source relative to the surface to be detected based on the shooting angle;

[0014] S102, irradiate the surface to be detected with the light source based on the incident angle, and acquire dark field image data of the surface to be detected in the irradiation with the light source based on the shooting device;

[0015] S103, detect defects on the surface to be detected based on the dark field image data, and obtain the defect detection result of the surface to be detected.

[0016] The optical part to be detected is an optical workpiece, an optical product or an optical material having one or more surfaces to be detected, which can be a consumer electronic component (such as a mobile phone glass cover plate or a lens piece), an optical element (such as a lens, a mirror or a filter), a precision mechanical component (such as a metal processing surface or an injection molding part), a coating, a film layer or a transparent material surface, and the surface to be detected is a surface area of the optical part to be detected that needs to be detected.

[0017] The shooting device is a device for collecting images of the surface to be detected of the optical part to be detected, which can be an industrial high-resolution area array camera or a line array scanning camera, and is configured with an appropriate lens module.

[0018] The light source can be a ring-shaped light source, a strip-shaped light source or a coaxial oblique light source, which can be emitted by LED, laser or near-infrared light source equipment, and can be used in combination with diffusion sheets, collimating sheets and polarizing sheets to form good dark field illumination.

[0019] In some examples, the shooting angle of the shooting device relative to the surface to be detected of the optical part to be detected can be set to an angle of 0° between the camera optical axis and the normal line of the surface to be detected, i.e. the camera is perpendicular to the surface to be detected for vertical shooting; of course, it can also be set to a slight inclination angle according to the surface material or process requirements of the surface to be detected of the optical part to be detected.

[0020] Then, the application determines the incident angle of the light source relative to the surface to be detected based on the shooting angle of the shooting device relative to the surface to be detected of the optical component to be detected. Preferably, the incident angle of the light source relative to the surface to be detected and the shooting angle form a complementary angle, avoiding the mirror reflection of the surface to be detected into the shooting device, thereby forming a dark field imaging condition. For example, when the camera is vertically shot, the incident angle of the light source can be set to an angle of inclination of 25° to 35° relative to the normal of the surface to be detected, to form a clear bright / dark contrast. The angle can be obtained by table lookup, formula calculation or device calibration. For another example, when the shooting angle of the shooting device relative to the surface to be detected of the optical component to be detected is 25°, the camera is obliquely shot at an angle of 25° relative to the surface to be detected, and the light source is obliquely irradiated in a direction with an angle of -25° (or 25° relative to the other side) relative to the normal, so that the mirror reflection light is far away from the camera lens, thereby reserving a dark background and highlighting the contrast of defect scattering light in the subsequent dark field image data.

[0021] After determining the incident angle of the light source relative to the surface to be detected, the light source emits a directional illumination beam to irradiate the surface to be detected. Preferably, an LED oblique light source, a ring-shaped dark field light source or a strip-shaped directional light source is used, and a diffuser or a collimator is further used to control the irradiation uniformity and directionality.

[0022] Subsequently, the shooting device collects image data of the surface to be detected while the light source is irradiating, to obtain dark field image data. Under the dark field imaging condition, the ideal smooth area is deviated from the shooting visual angle (lens of the shooting device) of the shooting device due to mirror reflection, and the imaging area appears as a dark background. The position with surface defects (such as scratches, pits, dirty spots, etc.) will reflect part of the light to the shooting visual angle (shooting lens of the shooting device) of the shooting device due to scattering effect, thereby forming a bright spot, a bright line or a high-contrast area in the dark field image data.

[0023] Finally, the surface of the optical component to be detected is detected based on the dark field image data, which can specifically include the processes of identifying the bright abnormal area, extracting the defect morphological feature, judging the defect type, etc., to obtain the complete detection result of the surface of the optical component, including the spatial position, geometric boundary and defect category of the defect, etc.

[0024] According to the technical scheme provided in the embodiment of the present application, the shooting angle of the shooting device relative to the surface to be detected of the optical part to be detected is determined, and the incident angle of the light source relative to the surface to be detected is determined based on the shooting angle; the surface to be detected is irradiated by the light source based on the incident angle, and the dark field image data of the surface to be detected in the irradiation by the light source is acquired based on the shooting device; the surface to be detected is detected for defects based on the dark field image data, and the defect detection result of the surface to be detected is obtained. The present application dynamically determines the incident angle of the light source with the shooting angle as a reference, so that when the surface to be detected is irradiated by the light source subsequently, the mirror reflection direction of the surface to be detected avoids the shooting visual angle of the shooting device, thereby constructing a light environment meeting the dark field imaging condition, so that a dark field image data with low background brightness and high defect contrast is formed in the acquired image. Then, the detection of the micro defects on the surface to be detected is realized by image processing on the dark field image data, thereby improving the quality inspection effect and avoiding the problem of poor quality inspection effect caused by manual quality inspection in the prior art.

[0025] In some examples, after the defect detection result of the surface to be detected is obtained based on the dark field image data, the method further includes: generating a defect prompt image based on the defect detection result, the defect prompt image being used to indicate the defect area on the surface to be detected; acquiring the pose information of the optical part to be detected, and displaying the defect prompt image on the augmented reality device based on the pose information.

[0026] In the embodiment, after the illumination image is acquired and the defect detection is completed, the present application further generates a defect prompt image and prepares for augmented reality display.

[0027] Firstly, the present application structures the information such as defect position, contour, type, severity, etc. in the defect detection result and outputs the information as the basic data of the defect prompt image. The prompt image can be in the form of a two-dimensional layer, a three-dimensional grid or a semantic annotation image, and the data structure thereof contains the coordinate point set, category identifier, color coding and display rule (such as whether to flash, whether to dynamically scale, etc.) of multiple defect areas.

[0028] For example, if a long strip scratch is detected, the present application generates a red graphic object with the corresponding boundary contour and attaches a “severe” level label for AR highlight display. At the same time, auxiliary information such as defect ID, timestamp, image number and the like metadata can also be added.

[0029] Subsequently, this application acquires the pose information (including 3D position and attitude information) of the optical component under inspection in real space through a spatial positioning module, i.e., its displacement vector and translation / rotation matrix in the world coordinate system. This module can be implemented using vision-based SLAM (Simultaneous Localization and Mapping), structured light, QR code positioning, inertial navigation unit (IMU), etc. The spatial pose information is used to establish the transformation relationship between the defect image coordinate system and the augmented reality device coordinate system.

[0030] In some examples, displaying defect warning images on an augmented reality device based on pose information includes: registering the defect warning image with a view of the optical component to be inspected in the augmented reality device based on the pose information; and displaying the registered defect warning image overlaid on the corresponding position of the optical component to be inspected by the augmented reality device.

[0031] In this embodiment, after receiving the defect warning image and the spatial pose information of the optical component to be inspected, registration and image overlay operations are performed to achieve accurate visualization of the defect on the surface of the real object.

[0032] Specifically, this application spatially aligns the defect warning image with the real-time camera footage in an AR device based on spatial pose information. This process includes the following steps:

[0033] Coordinate system transformation: This application establishes a mapping relationship between the defect indication image coordinate system, the optical component under inspection coordinate system, and the AR device's own world coordinate system. For example, a homogeneous transformation matrix is ​​used to map the defect indication image coordinates from the surface coordinate system of the optical component under inspection to the AR device's view coordinate system.

[0034] Perspective transformation and projection correction: Based on the intrinsic parameters (such as focal length, principal point position, distortion coefficient, etc.) and extrinsic parameters (such as the current orientation of the device) of the AR device, perspective transformation and image distortion correction are performed on the defect indication image to ensure that the defect indication image can be accurately projected onto the surface of the optical part to be inspected as seen by the user.

[0035] Registration accuracy correction: If this application uses IMU or environmental anchor points (such as QR codes, positioning point clouds) to assist in positioning, the real-time registration results can be dynamically fine-tuned to improve the fit and stability between the defect indication image and the optical part to be inspected.

[0036] After registration, the augmented reality device presents the defect warning image in the user's field of vision as an overlay, accurately covering the actual defect location of the optical component to be inspected.

[0037] In addition, the application can also control the display state of the prompt image according to the operation instruction of the user (for example, through gesture or voice interaction), such as hiding a certain type of defect, enlarging a certain area, switching the defect information layer, and the like, so as to realize the interactive intelligent quality inspection auxiliary function.

[0038] Finally, the registered defect prompt image is superimposed and displayed on the actual position of the optical part to be detected in the user's field of view through the AR glasses, AR tablet or other augmented reality display device, so as to realize the "spatial intuitive magnification display" of the defect. The operator can clearly see the position, shape and category information of each defect without affecting the original surface of the object, thereby significantly improving the efficiency and accuracy of defect review and processing.

[0039] For example, in the process of detecting a piece of mobile phone glass cover plate, a small scratch about 5mm long is detected in the upper left corner, and the type is "light linear scratch". The application generates a yellow dashed outline frame according to the defect, and obtains the spatial pose of the glass cover plate in combination with the VSLAM technology, so as to accurately project the prompt frame on the actual scratch position of the upper left corner of the glass in the field of view of the AR glasses for intuitive confirmation and labeling by the quality inspector.

[0040] According to the scheme provided in the embodiment, the defect prompt image is generated based on the defect detection result, and the defect prompt image is used to indicate the defect area on the surface to be detected. The pose information of the optical part to be detected is obtained, and the defect prompt image is displayed on the augmented reality device based on the pose information. By obtaining the pose information of the workpiece, the defect information can be accurately superimposed on the surface of the optical part to be detected. Without paper marking, labeling or manual measurement, the position of the defect on the actual workpiece can be "one-to-one" visually presented. The operator can intuitively view the real position and shape of the defect without the aid of professional defect image analysis software. The dependence on human experience is greatly reduced, and the missed detection and misjudgment are reduced.

[0041] In some examples, determining the incident angle of the light source relative to the surface to be detected based on the shooting angle includes: matching the shooting angle with a pre-set corresponding relationship between shooting angle and incident angle; and determining the incident angle of the light source relative to the surface to be detected according to the matching result.

[0042] Specifically, the application pre-sets the corresponding relationship between the shooting angle and the incident angle. The corresponding relationship between the shooting angle and the incident angle can be stored in any form such as key-value pair, hash, database or table. The corresponding relationship between the shooting angle and the incident angle can be obtained based on theoretical optical calculation, experimental test or device calibration data.

[0043] For example, the corresponding relationship between the shooting angle and the incident angle is in the form of a table as shown in Table 1.

[0044] Table 1: Correspondence table between shooting angle and incident angle

[0045]

[0046]

[0047] The above correspondence reflects the recommended light source incident angle configuration for obtaining an ideal dark field image under different shooting angles. The system can take the currently set or measured shooting angle as an input parameter, and look up or interpolate the optimal incident angle value corresponding thereto.

[0048] For example, when the shooting device is set to shoot obliquely at an angle of 10° relative to the normal of the surface to be detected, the table lookup gives a recommended light source incident angle of 25°, so that the light source device is controlled to adjust to a direction that is 25° away from the normal of the surface to be detected. This incident angle configuration makes the specular reflection direction away from the camera field of view, and at the same time enhances the scattered light from the defect into the shooting device, thereby improving the contrast of the defect in the image.

[0049] This method does not need to perform complex real-time optical modeling calculation, and only needs to establish the angle correspondence table once, so as to realize fast and efficient light angle configuration.

[0050] In some examples, determining the incident angle of the light source relative to the surface to be detected based on the shooting angle includes: inputting the shooting angle into a pre-set angle calculation formula to obtain the incident angle of the light source relative to the surface to be detected.

[0051] First, the present application determines the shooting angle of the shooting device relative to the surface of the optical part to be detected. In the present embodiment, the included angle between the optical axis of the shooting device and the normal of the surface to be detected is set as θ p .

[0052] Next, the present application inputs the shooting angle θ p into a pre-set angle calculation formula to determine the light source incident angle θ1 corresponding to the shooting angle. The calculation formula used in the present embodiment is as follows:

[0053] θ1=2×θ p +Δ;

[0054] Wherein, θ p is the shooting angle (i.e. the included angle between the camera optical axis and the normal of the surface to be detected); θ1 is the light source incident angle (i.e. the included angle between the light and the normal of the surface to be detected); Δ is the offset compensation angle, which is used to adjust the actual deviation caused by factors such as light source size, target surface micro-curvature or spatial layout, and is usually an empirical value, for example, 5° or 10°.

[0055] The above calculation model aims to ensure that the specular reflection direction of the light source is away from the imaging field of view of the camera, thereby enhancing the scattering contrast of the defect area under dark field imaging conditions.

[0056] After obtaining the incident angle θ1, the light source rotating mechanism is controlled to adjust the light source to the angle position, and subsequent illumination, image acquisition, defect detection and result processing procedures are performed.

[0057] Compared with the table lookup matching method, the present embodiment has higher flexibility, is especially suitable for dynamic detection of continuous variable angles, can realize real-time automatic calculation and rapid adjustment of the light source angle, and improves the automation degree and detection efficiency of the present application.

[0058] In some examples, defect detection is performed on the surface to be detected based on the dark field image data to obtain a defect detection result of the surface to be detected, including: identifying a brightness abnormal area from the dark field image data, and taking the identified brightness abnormal area as a defect area; determining a defect category of the defect area based on a feature of the defect area, and taking the defect area and the category of the defect area as the defect detection result.

[0059] In this embodiment, the present application performs defect detection on the surface to be detected based on the dark field image data collected by the camera to obtain a complete defect detection result. Specifically, the dark field image data is a reflection image obtained under illumination at a specific incident angle, wherein the ideal smooth area deviates from the camera view due to specular reflection, and appears as a dark overall background. The area with surface defects (such as scratches, pits, dents, etc.) will cause light scattering due to surface structure damage, and part of the scattered light enters the camera imaging, which appears as a bright spot or bright line in the image. Therefore, there is a significant brightness contrast between the defect area and the background area.

[0060] The present application first pre-processes the dark field image data, including gray scale normalization, noise filtering, brightness enhancement and other operations, to improve the clarity and robustness of the defect signal. Subsequently, the present application uses a threshold segmentation algorithm or a local brightness contrast algorithm to identify the brightness abnormal area in the image. The specific method can include: comparing the brightness value of each pixel point in the image with the average brightness value of its surrounding neighborhood, and if it exceeds the set threshold, it is determined that the pixel point belongs to the brightness abnormal area.

[0061] The identified brightness abnormal area is taken as a candidate defect area, and the present application further extracts its feature information, including but not limited to: area, aspect ratio, contour shape, edge strength, brightness distribution pattern, etc. Based on these features, the present application uses a pre-trained classification model or a rule-based classification logic to identify the defect type of each defect area, such as determining whether it belongs to a linear scratch, a point-like pit, a stain or a depression, etc.

[0062] Finally, the application outputs the position, contour boundary and corresponding defect type of each defect area as the defect detection result for subsequent AR display module or quality inspection application. The detection method relies on the high contrast characteristics of the dark field image, so that small defects can also be quickly and accurately identified in the form of a significant bright spot.

[0063] It can be understood that, in order to improve the accuracy of defect detection and robustness in complex scenes, the embodiment can also use a deep learning-based image segmentation model to analyze and process the dark field image. Specifically, the application can call a pre-trained convolutional neural network (CNN) model, such as a fully convolutional network with a U-Net structure, to perform pixel-level semantic segmentation on the dark field image. The model takes the dark field image as input and outputs a probability map (Mask map) of each pixel belonging to "defect" or "non-defect", thereby directly obtaining the accurate contour of the defect area. In the model training stage, a training set can be constructed based on a large number of labeled defect image samples, supervised learning can be performed using a cross-entropy loss function, and the generalization ability can be improved through data augmentation (such as rotation, scaling, brightness disturbance). In the inference stage, the model can complete global analysis of large-size images in a short time, which is suitable for high-speed industrial scenes. Compared with the traditional brightness threshold method, this method can identify weaker, edge blurred or irregular shaped defects, while reducing the risk of false positives caused by noise and uneven lighting.

[0064] In another embodiment, to improve the adaptability of the detection algorithm to the brightness distribution under different lighting conditions, the application can use a brightness gradient change detection algorithm instead of using a fixed global threshold.

[0065] Specifically, the application calculates the brightness gradient value of each pixel in the image based on the difference between the gray value of the pixel and the average gray value of its surrounding neighborhood; when the brightness gradient of a certain pixel exceeds a dynamically set local threshold, the point is determined to belong to a possible defect edge region.

[0066] Further, the application can combine an edge detection algorithm (such as Sobel, Laplacian or Canny operator) to identify the defect contour, and fuse the brightness gradient and region connectivity information to refine the defect area, and eliminate invalid areas such as isolated noise points and false alarm responses.

[0067] Compared with the traditional static threshold method, this method is more suitable for complex reflection, shadowing, or local uneven lighting image environments, and improves the completeness and accuracy of the defect area.

[0068] In some examples, the defect detection result further includes a defect evaluation result of evaluating the defect area in terms of defect degree, and the defect area is evaluated in terms of defect degree based on an area of the defect area to obtain the defect evaluation result of the defect area.

[0069] Specifically, in order to provide a basis for subsequent quality judgment, AR labeling and operation decision, the application can further comprehensively evaluate the severity of each defect after completing defect detection and classification.

[0070] The evaluation process can be quantitatively calculated based on the following two main factors: Intensity Mean: used to measure the scattering intensity of the defect area. The brighter it is, the more intense the light scattering is, which may correspond to a deeper or rougher defect; Pixel Area: used to reflect the actual size of the defect in the image, which can be converted into physical units (such as mm 2 ) according to the number of pixels.

[0071] The application can combine the two factors by using linear weighting or classification rule tree to generate a defect severity score.

[0072] In some examples, after defect detection is performed on the to-be-detected surface based on the dark-field image data to obtain a defect detection result of the to-be-detected surface, the method further includes: generating a voice alarm prompt and prompting a user through the voice alarm prompt.

[0073] In some examples, determining the shooting angle of the shooting device relative to the to-be-detected surface of the to-be-detected optical part includes: rotating the shooting device through the translation table so that the shooting device shoots the to-be-detected surface of the to-be-detected optical part at the shooting angle. It can be understood that for larger or high-precision positioning elements, the application can also integrate an X-Y translation table or a rotary table to move the shooting device or the to-be-detected optical part to realize full-surface scanning of the to-be-detected optical part.

[0074] In some examples, the application further provides a detection system for a surface of an optical part, which includes: a shooting device, an illumination device, and an image processing device. The illumination device is used to irradiate a to-be-detected surface of a to-be-detected optical part with a light source based on an incident angle, the incident angle being determined based on a shooting angle of the shooting device relative to the to-be-detected surface of the to-be-detected optical part, and the shooting device is used to obtain dark-field image data of the to-be-detected surface in the light source irradiation. The image processing device is used to perform defect detection on the to-be-detected surface based on the dark-field image data to obtain a defect detection result of the to-be-detected surface.

[0075] In some examples, the optical surface detection system further comprises an augmented reality device, the image processing device is further configured to generate a defect prompt image based on the defect detection result, the defect prompt image is used to indicate a defect area on the surface to be detected, and the augmented reality device is configured to acquire pose information of the optical part to be detected and display the defect prompt image based on the pose information.

[0076] Specifically, the camera can use a face array CCD or CMOS camera, and sometimes a line array camera (for a scanning system). The resolution needs to be high enough to capture tiny defects (such as microns).

[0077] The illumination device (light source) adopts a dark field illumination mode. The illumination device (usually an LED array) irradiates the surface to be detected of the optical part to be detected at a specific, large angle (usually much larger than the receiving angle of the camera lens).

[0078] Imaging lens: high-quality lens to ensure imaging clarity and distortion control.

[0079] Precise motion platform (optional): for larger or high-precision positioning of the surface to be detected of the optical part to be detected, the optical surface detection system further comprises an X-Y translation stage or a rotary stage to move the camera / lens or the optical part to be detected, thereby realizing full-surface scanning of the optical part to be detected.

[0080] The image processing device is responsible for controlling the hardware, acquiring images, applying algorithms for real-time detection, identification, classification of defects, and measurement of their size (length, width, area), position, and finally output of a report. The efficiency of the software directly determines the overall efficiency of the system.

[0081] Computer: running control software and image processing algorithms.

[0082] The optical surface detection system provided by the present application has the advantages of non-contact, fast imaging, automation, objectivity, consistency, etc.

[0083] Non-contact: avoids scratching the precise optical surface of the optical part to be detected.

[0084] Fast imaging: short exposure time of the camera, combined with the motion platform to enable high-speed scanning of large-area surfaces.

[0085] Automation: the entire optical surface detection process (scanning, image acquisition, analysis, determination, report generation) is fully automated, requiring no human intervention or only a small amount of operation (such as loading / unloading).

[0086] Objectivity and consistency: the software algorithm makes judgments based on pre-set standards, the results are objective and repeatable, avoiding the subjectivity and fatigue errors of manual detection.

[0087] High throughput: Especially suitable for online or offline batch detection on production lines.

[0088] High sensitivity: Can reliably detect tiny (even sub-micron) scratches and pits.

[0089] The physical basis of the detection system for the surface of the optical part provided in this application is the "dark field imaging" and "scattered light" principle:

[0090] (1) Ideal smooth surface: When light is incident on an ideal smooth optical surface at a certain angle (dark field angle), the light will reflect according to the law of specular reflection. Since the camera lens is directly opposite the surface (or receives light within a very small angle range), the specularly reflected light will completely avoid the lens of the camera. Therefore, the camera "sees" a dark background (dark field).

[0091] (2) Presence of surface defects: scratches, pits, dents, and other surface defects destroy the local smoothness. When the illumination light shines on these defect areas, scattering occurs. Light no longer propagates only in the direction of specular reflection, but diverges in all directions (including towards the camera lens). These scattered lights can enter the camera lens, forming bright spots or lines on the originally dark background.

[0092] (3) Image formation and detection:

[0093] The image captured by the camera lens has a dark background (no specularly reflected light enters) and bright defect areas (scattered light enters).

[0094] This high-contrast image makes it very easy, fast and accurate for software algorithms to: identify the presence of bright spots (detect defects). Measure the size and shape of bright spots (distinguish between scratches, which are linear bright lines, and pits, which are point-like bright spots). Locate the position of bright spots.

[0095] Data acquisition: The system uses a high-sensitivity line / surface array scanner to scan the workpiece surface at high speed, quickly acquiring high-precision image data (dark field image data) of the workpiece surface. The scanner has high sensitivity characteristics, which can capture tiny defect details and provide a basis for accurate defect detection.

[0096] Defect recognition: The AI processing module receives image data transmitted by the scanner and uses built-in AI algorithms to analyze and process images in real time. The algorithm can accurately identify various defects such as scratches and pits, and determine the outline, type and severity of the defects. For example, through machine learning training algorithms, different shapes and sizes of scratches can be accurately classified, and the severity of the scratches can be evaluated according to parameters such as length and depth.

[0097] Spatial positioning: the spatial positioning module (such as VSLAM, visual simultaneous localization and mapping) obtains the spatial position and attitude information of the workpiece in real time, provides accurate spatial coordinate reference for AR, and ensures that the position corresponds to the actual defect position accurately.

[0098] Defects: according to the defect information output by the AI processing module and the coordinate information provided by the spatial positioning module, the high-precision AR instrument identifies the defect contour, type and severity, and marks it in the form of 5-10 times magnification through different colors, flashing and other ways, and accurately covers the actual defect position of the workpiece. For example, for defects with high severity, red flashing is used for identification; for slight defects, yellow static identification is used.

[0099] Inspection operation: the inspector only needs to confirm, classify or review the defects according to the clear "magnification prompt" in front of his eyes. The operation control module can control and adjust the entire inspection process, such as setting scanning parameters, modes, etc.

[0100] All the optional technical solutions described above can be combined to form optional embodiments of the present application, and will not be repeated here.

[0101] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above division of functional units and modules is taken as an example for illustration. In actual application, the above functions can be completed by different functional units or modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit or module in the embodiment can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or software.

[0102] The integrated modules / units, if implemented in the form of software functional units and sold or used as independent products, can be stored in a computer readable storage medium. Based on such understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. The computer program can be executed by a processor to implement the steps of the above-mentioned various method embodiments. The computer program can include computer program code, which can be in the form of source code, object code, executable files or some intermediate forms. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer readable medium does not include electric carrier signals and telecommunication signals.

[0103] The above embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method of inspecting an optical part surface, characterized by, The method comprises: determining a shooting angle of a shooting device relative to a to-be-detected surface of a to-be-detected optical component, and determining an incident angle of a light source relative to the to-be-detected surface based on the shooting angle; performing light source irradiation on the to-be-detected surface based on the incident angle, and acquiring dark field image data of the to-be-detected surface in the light source irradiation based on the shooting device; performing defect detection on the to-be-detected surface based on the dark field image data to obtain a defect detection result of the to-be-detected surface.

2. The method of claim 1, wherein, After performing defect detection on the to-be-detected surface based on the dark field image data to obtain a defect detection result of the to-be-detected surface, the method further comprises: generating a defect prompt image based on the defect detection result, the defect prompt image being used to indicate a defect area on the to-be-detected surface; acquiring pose information of the to-be-detected optical component, and displaying the defect prompt image on an augmented reality device based on the pose information.

3. The method of claim 2, wherein, Displaying the defect prompt image on the augmented reality device based on the pose information comprises: registering the defect prompt image with a view of the to-be-detected optical component in the augmented reality device based on the pose information; and superimposing and displaying the registered defect prompt image on a corresponding position of the to-be-detected optical component through the augmented reality device.

4. The method of claim 1, wherein, Determining the incident angle of the light source relative to the to-be-detected surface based on the shooting angle comprises: matching the shooting angle with a pre-set corresponding relationship between shooting angles and incident angles; and determining the incident angle of the light source relative to the to-be-detected surface according to a matching result.

5. The method of claim 1, wherein, Determining the incident angle of the light source relative to the to-be-detected surface based on the shooting angle comprises: inputting the shooting angle into a pre-set angle calculation formula to obtain the incident angle of the light source relative to the to-be-detected surface.

6. The method of claim 1, wherein, Performing defect detection on the to-be-detected surface based on the dark field image data to obtain a defect detection result of the to-be-detected surface comprises: identifying a brightness abnormal area from the dark field image data, taking the identified brightness abnormal area as a defect area; and determining a defect category of the defect area based on a feature of the defect area, and taking the defect area and the category of the defect area as the defect detection result.

7. The method of claim 6, wherein, The defect detection result further comprises a defect evaluation result of defect degree evaluation on the defect area, and the defect degree evaluation on the defect area comprises: performing defect degree evaluation on the defect area based on an area of the defect area to obtain a defect evaluation result of the defect area.

8. The method of claim 1, wherein, After performing defect detection on the to-be-detected surface based on the dark field image data to obtain a defect detection result of the to-be-detected surface, the method further comprises: generating a voice alarm prompt, and prompting a user through the voice alarm prompt.

9. A system for inspecting a surface of an optical part, characterized in that, The system comprises a shooting device, an illumination device, and an image processing device. The illumination device is configured to irradiate a to-be-detected surface of a to-be-detected optical component with a light source based on an incident angle, the incident angle being determined based on a shooting angle of the shooting device relative to the to-be-detected surface of the to-be-detected optical component. The shooting device is configured to acquire dark-field image data of the to-be-detected surface in the irradiation with the light source. The image processing device is configured to perform defect detection on the to-be-detected surface based on the dark-field image data, and obtain a defect detection result of the to-be-detected surface.

10. The system of claim 9, wherein, The system further comprises an augmented reality device. The image processing device is further configured to generate a defect prompt image based on the defect detection result, the defect prompt image being configured to indicate a defect region on the to-be-detected surface. The augmented reality device is configured to acquire pose information of the to-be-detected optical component, and display the defect prompt image based on the pose information.

Images