A colloidal visual inspection method, apparatus and computer readable storage medium
By reconstructing the three-dimensional morphology of transparent colloids through multi-angle image acquisition and phase unwrapping technology, the problem of internal defect detection in the visual inspection of transparent adhesive lines is solved, achieving high-precision and high-speed detection results. It is applicable to complex curved surfaces and reduces detection costs.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- YISHI ZHITONG TECH SHENZHEN CO LTD
- Filing Date
- 2026-03-25
- Publication Date
- 2026-06-23
AI Technical Summary
Existing technologies for visual inspection of defects in transparent adhesive lines or adhesive dots on transparent parts suffer from several problems, including a lack of internal defect detection capabilities, a contradiction between detection accuracy and speed, poor adaptability to complex curved surfaces, high material modification and cost, and poor environmental adaptability.
Transmitted and scattered images are acquired by multiple cameras at different angles. Phase unwrapping and contour reconstruction are performed. Combined with phase distortion and calibration parameters, three-dimensional topography data are extracted and multi-scale fusion and design model registration are carried out to determine the location and type of defects.
It enables faster and more efficient visual inspection of colloids, enhances applicability, reduces inspection costs, and improves production efficiency.
Smart Images
Figure CN122265221A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the fields of machine vision and industrial inspection technology, and in particular to a colloidal vision inspection method, equipment and computer-readable storage medium. Background Technology
[0002] Currently, dispensing technology is widely used in sealing, bonding, and encapsulation processes in industries such as 3C electronics manufacturing, automotive parts, and medical devices. Transparent adhesive lines (such as UV adhesives, silicone, and epoxy resins) are widely used in cover sealing, mid-frame bonding, and transparent component assembly due to their aesthetic and functional requirements.
[0003] However, the optical properties of transparent materials (e.g., high transmittance and low reflectance) make it difficult for traditional machine vision methods to obtain clear colloidal images, becoming a technical bottleneck in the industry.
[0004] Existing solutions include the following: First, there is the conventional optical imaging inspection. This method uses a high-resolution industrial camera with a ring light source for illumination. By adjusting the angle of the light source, the reflected image of the colloid surface is obtained. This method relies on the difference in surface reflectivity between the colloid and the substrate. It has no detection capability for internal defects of transparent colloids (such as bubbles, broken colloids, etc.), is sensitive to ambient light interference, and its detection accuracy is limited by optical contrast. Second, ultraviolet fluorescence detection; this method involves adding a fluorescent agent to the colloid and using an ultraviolet light source to excite the fluorescence and then imaging; this method requires modified colloidal materials, which increases costs, and the decay of fluorescence intensity affects long-term stability, and has limited ability to detect internal defects, making it unsuitable for high-speed online detection. Third, infrared thermal imaging detection; this method uses the difference in thermal radiation between the colloid and the substrate for imaging; this method is sensitive to the thermal properties of the material, is greatly affected by the ambient temperature, has low spatial resolution (for example, usually greater than 100μm), cannot meet the needs of micron-level detection, and has high equipment cost. Fourth, laser line scanning inspection; this method uses a laser line to project onto the colloidal surface and captures the contour information through a line scanning camera; this method has poor penetration of transparent materials, can only obtain the surface contour, cannot detect internal defects, has poor adaptability to curved surfaces, and the scanning speed is limited by mechanical movement. Fifth, optical coherence tomography (OCT); this method uses the principle of low-coherence optical interference to obtain the internal structure of transparent materials; however, the equipment for this method is complex and expensive, the imaging speed is slow, and the data processing volume is large, making it unsuitable for online detection on high-speed production lines, and mainly used for laboratory analysis.
[0005] Therefore, existing solutions for visual inspection of defects in transparent adhesive lines or adhesive dots on transparent parts suffer from technical shortcomings, including a lack of internal defect detection capabilities, a contradiction between detection accuracy and speed, poor adaptability to complex curved surfaces, high material modification and cost, and poor environmental adaptability. Summary of the Invention
[0006] In order to overcome the shortcomings of the prior art, the present invention aims to provide a colloidal visual inspection method, device and computer-readable storage medium to solve the problems existing in the current visual inspection of defects in transparent adhesive lines or transparent parts, such as lack of internal defect detection capability, contradiction between detection accuracy and speed, poor adaptability to complex curved surfaces, high material modification and cost, and poor environmental adaptability.
[0007] This invention proposes a visual inspection method for colloids, the method comprising: Multiple cameras were used to capture transmission and scattering images of pre-set structured light passing through the colloid at different angles; Phase unwrapping is performed on the transmitted image and the scattered image to obtain the absolute phase distribution of the colloid, and contour reconstruction is performed by combining the phase distortion and preset calibration parameters to obtain the three-dimensional morphology data of the colloid. Defect features from multiple angles are extracted from the three-dimensional topography data and fused at multiple scales. The data is then registered and compared with a preset design model to determine the location and type of defects in the colloid.
[0008] Optionally, the method further includes: The structured light is generated periodically using a preset digital light projection device or a preset grating projection module; The grating pattern is set as a combination of one or more of the following: sine stripes, Gray code patterns, and phase-shifted stripes.
[0009] Optionally, the method further includes: A preset pulsed laser or a preset high-speed LED is used as the lighting source; The pulse frequency of the illumination source is set to be synchronized with the exposure frequency of the camera and matched with the movement speed of the production line.
[0010] Optionally, the method further includes: The transmitted image is acquired by a camera arranged in a straight line with the structured light and the colloid; The scattered images are acquired by a camera positioned off-axis from the structured light beam.
[0011] Optionally, the method further includes: Acquire multiple phase-shifted images corresponding to the transmitted image and the scattered image; The phase unwrapping calculation of the multi-frame phase-shifted images is performed by one or more combinations of the preset four-step phase shift method, multi-frequency heterodyne method, and time phase expansion method to obtain the absolute phase distribution.
[0012] Optionally, the method further includes: Multiple defect types are preset, including glue breakage, glue overflow, air bubbles, and uneven glue width. Thresholds are set for the glue breakage feature, glue overflow feature, bubble feature, and glue width unevenness feature, respectively, and the defect features are extracted according to the thresholds.
[0013] Optionally, the method further includes: The defect features from multiple perspectives are fused at multiple scales using a preset weighted fusion or a preset Bayesian fusion algorithm. The fusion weights are dynamically adjusted based on the current camera viewpoint, lighting conditions, and defect type to eliminate occlusion and specular reflection interference.
[0014] Optionally, the method further includes: Detect the type of the colloid; When the colloid is a curved adhesive path, a non-uniform grating projection corresponding to the preset design model is used as the structured light, and a preset curved adaptive phase unwrapping algorithm is used to perform the phase unwrapping calculation.
[0015] The present invention also proposes a colloid visual inspection device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the colloid visual inspection method as described in any of the preceding claims.
[0016] The present invention also proposes a computer-readable storage medium storing a colloid vision inspection program, wherein when the colloid vision inspection program is executed by a processor, the steps of the colloid vision inspection method as described in any of the preceding claims are implemented.
[0017] The colloidal visual inspection method, device, and computer-readable storage medium of this invention acquire transmission and scattering images of a colloid through multiple cameras at different angles when pre-set structured light passes through it. The transmission and scattering images are then unwrapped to obtain the absolute phase distribution of the colloid. This data is combined with phase distortion and pre-set calibration parameters to reconstruct the colloid's three-dimensional morphology. Defect features from multiple angles are extracted from the 3D morphology data and fused at multiple scales. These features are then registered and compared with a pre-set design model to determine the location and type of defects in the colloid. This invention provides a faster and more effective colloidal visual inspection solution, enhancing its applicability, saving inspection costs, and improving production efficiency. Attached Figure Description
[0018] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings: Figure 1 This is a flowchart of the first embodiment of the colloidal visual inspection method of the present invention; Figure 2 This is a flowchart of the second embodiment of the colloidal visual inspection method of the present invention; Figure 3 This is a flowchart of the third embodiment of the colloidal visual inspection method of the present invention; Figure 4 This is a flowchart of the fourth embodiment of the colloidal visual inspection method of the present invention; Figure 5 This is a flowchart of the fifth embodiment of the colloidal visual inspection method of the present invention; Figure 6 This is a flowchart of the sixth embodiment of the colloidal visual inspection method of the present invention; Figure 7 This is a flowchart of the seventh embodiment of the colloidal visual inspection method of the present invention; Figure 8 This is a flowchart of the eighth embodiment of the colloidal visual inspection method of the present invention; Figure 9 This is a system structure diagram of the colloid visual inspection method of the present invention; Figure 10 This is a schematic diagram of structured light for the colloidal visual inspection method of the present invention; Figure 11 This is a pulse-synchronized illumination timing diagram of the colloidal vision detection method of the present invention; Figure 12 This is a multi-angle image acquisition optical path diagram of the colloidal visual detection method of the present invention; Figure 13 This is a flowchart of the phase unwrapping and three-dimensional reconstruction of the colloidal visual detection method of the present invention; Figure 14 This is a schematic diagram of defect feature extraction in the colloidal visual inspection method of the present invention; Figure 15 This is a schematic diagram illustrating the multi-scale fusion discrimination principle of the colloidal visual detection method of the present invention; Figure 16 This is a diagram of the online detection system architecture for the colloid visual inspection method of the present invention. Detailed Implementation
[0019] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.
[0020] In the following description, the use of suffixes such as "module," "part," or "unit" to denote elements is solely for the purpose of illustrative purposes and has no specific meaning in itself. Therefore, "module," "part," or "unit" may be used interchangeably.
[0021] Example 1 Figure 1 This is a flowchart of the first embodiment of the colloidal visual inspection method of the present invention. A colloidal visual inspection method, the method comprising: S1. Multiple cameras are used to capture transmission and scattering images of pre-set structured light passing through the colloid at different angles; S2. Perform phase unwrapping on the transmission image and the scattering image to obtain the absolute phase distribution of the colloid, and combine the phase distortion and preset calibration parameters to reconstruct the contour to obtain the three-dimensional morphology data of the colloid. S3. Extract defect features from multiple angles in the three-dimensional topography data and perform multi-scale fusion, then register and compare them with the preset design model to determine the location and type of defects in the colloid.
[0022] like Figure 9 As shown, the system structure applying this method specifically includes a pulsed laser source, a structured light projection module, a multi-angle camera array, and a synchronization controller. The pulsed laser source provides a general light source to the workpiece under test on the conveyor belt. The structured light projection module transmits structured light to the workpiece under test. The multi-angle camera array provides the image processing unit with the transmitted and scattered images of the workpiece under test. The synchronization controller synchronously controls the pulsed laser source, the structured light projection module, the multi-angle camera array, and the production line's cycle time, ensuring they operate synchronously.
[0023] In this embodiment, the colloidal vision inspection is applicable to the three-dimensional vision inspection of defects in transparent adhesive lines or transparent parts, and is mainly used in quality inspection scenarios of transparent colloidal materials such as cover sealant, mid-frame sealant, and transparent UV adhesive lines in 3C electronic products. Compared with the existing technology, which has technical problems such as low imaging contrast, inability to detect internal defects, and poor adaptability to complex curved surfaces when inspecting transparent material adhesive lines, this embodiment projects a periodic structured grating onto the area of the transparent adhesive to be inspected, uses pulse illumination in conjunction with a high frame rate camera to capture the structured light pattern distortion caused by internal defects in the transparent adhesive, and uses an algorithm combining phase unwrapping and triangulation to reconstruct the three-dimensional morphology of the adhesive, realizing micron-level online detection of defects such as adhesive breakage, adhesive overflow, bubbles, and uneven width. It is easy to see that the above-mentioned detection method in this embodiment does not rely on surface optical contrast, can detect internal defects in transparent adhesives, and has good adaptability to adhesive lines on complex curved surfaces. The detection accuracy can reach the micron level, meeting the online inspection requirements of high-speed production lines.
[0024] The beneficial effects of this embodiment are as follows: Multiple cameras acquire transmission and scattering images of structured light passing through the colloid at different angles; phase unwrapping is performed on the transmission and scattering images to obtain the absolute phase distribution of the colloid; and contour reconstruction is performed by combining phase distortion and preset calibration parameters to obtain the three-dimensional morphology data of the colloid. Defect features from multiple angles are extracted from the three-dimensional morphology data and fused at multiple scales, then registered and compared with a preset design model to determine the location and type of defects in the colloid. This achieves a faster and more effective colloid visual inspection scheme, enhances its applicability, saves inspection costs, and improves production efficiency.
[0025] Example 2 Figure 2 This is a flowchart of a second embodiment of the colloid visual inspection method of the present invention. Based on the above embodiment, the method further includes: S01. The structured light is generated periodically using a preset digital light projection device or a preset grating projection module; S02. Set the grating pattern to one or more combinations of sine stripes, Gray code patterns, and phase-shifted stripes.
[0026] In this embodiment, a digital light projection device or a grating projection module is used to generate a periodic structured grating pattern, with the pattern period P ranging from 0.5mm to 5mm. Optionally, the pattern period P can be adaptively adjusted according to the colloid size and defect accuracy requirements.
[0027] In this embodiment, please refer to Figure 10 The schematic diagram of structured light shown includes sinusoidal fringes, Gray code patterns, and phase-shifting fringes.
[0028] In this embodiment, the modulation effect of transparent colloid on structured light patterns (e.g., refraction, scattering, absorption, etc.) is used to detect internal defects, thereby solving the contrast barrier in imaging of transparent materials.
[0029] The beneficial effect of this embodiment is that the penetration effect is achieved by the reverse application of structured light. Unlike the existing solution that uses structured light to measure the surface of opaque objects, this embodiment applies structured light in reverse to transparent materials and uses the penetration effect to perform three-dimensional imaging of internal defects, thereby improving the detection accuracy and detection effect.
[0030] Example 3 Figure 3 This is a flowchart of the third embodiment of the colloid visual inspection method of the present invention. Based on the above embodiment, the method further includes: S03. Use a preset pulsed laser or a preset high-speed LED as the lighting source; S04. Set the pulse frequency of the lighting source to be synchronized with the exposure frequency of the camera and matched with the movement speed of the production line.
[0031] In this embodiment, a nanosecond-level pulsed laser or a high-speed LED is used as the illumination source; optionally, the pulse width τ ranges from 10ns to 1μs, and the illumination source is synchronized with the camera exposure; optionally, the pulse frequency f matches the production line movement speed v and the detection cycle; further, f ≥ v / P to ensure that at least one sampling is performed in each grating cycle.
[0032] In this embodiment, please refer to Figure 11 The pulsed synchronous lighting timing diagram shown specifically includes the emission timing of the pulsed laser / LED, the camera exposure timing, the workpiece movement timing, and the synchronization relationship between them.
[0033] The beneficial effect of this embodiment is that, unlike the existing solutions that use pulse illumination or phase unwrapping alone, this embodiment applies pulse illumination and phase unwrapping in combination. On the one hand, nanosecond-level pulse illumination eliminates motion blur, and on the other hand, phase unwrapping extracts sub-pixel-level information, thereby achieving high-speed and high-precision detection. The combined application of the two solves the contradiction between detection accuracy and speed.
[0034] Example 4 Figure 4 This is a flowchart of the fourth embodiment of the colloid visual inspection method of the present invention. Based on the above embodiment, the method further includes: S11. Acquire the transmitted image using a camera arranged in a straight line with the structured light and the colloid; S12 acquires the scattering image using a camera positioned off-axis from the structured light beam.
[0035] In this embodiment, a high frame rate industrial camera (e.g., frame rate greater than 1000fps) is used to acquire modulated images of structured light patterns passing through a transparent colloid from multiple angles (e.g., at least two viewing angles), including transmission and scattering images, to capture optical path distortion caused by defects inside the colloid.
[0036] In this embodiment, please refer to Figure 12 The diagram shows the optical path for multi-angle image acquisition. It illustrates the optical path for both transmission and scattering image acquisition. In the former, the light source-colloid-camera are arranged in a straight line, while in the latter, the camera is arranged off-axis.
[0037] The beneficial effect of this embodiment is that it captures structured light modulation information from different perspectives through multi-angle image acquisition, thereby providing a data foundation for the subsequent weighted fusion algorithm to eliminate specular reflection and occlusion interference.
[0038] Example 5 Figure 5 This is a flowchart of the fifth embodiment of the colloid visual inspection method of the present invention. Based on the above embodiment, the method further includes: S21. Acquire multiple phase-shifted images corresponding to the transmitted image and the scattered image; S22. The phase unwrapping calculation of the multi-frame phase-shifted image is performed by one or more combinations of the preset four-step phase shift method, multi-frequency heterodyne method and time phase expansion method to obtain the absolute phase distribution.
[0039] Please refer to Figure 13 The flowchart shown illustrates the phase unwrapping and 3D reconstruction process, illustrating the calculation flow from multiple phase-shifted images to the absolute phase distribution φ(x, y), and then to the 3D coordinates (x, y, z). Specifically, it includes: performing phase calculation on the original images (the transmission image and the scattering image), then performing phase unwrapping to generate a point cloud and obtain a 3D model; in this embodiment, phase unwrapping calculation is performed on the acquired multiple phase-shifted images to obtain the absolute phase distribution φ(x, y) of the colloidal region; further, combining the triangulation principle, based on the phase distortion Δφ and system calibration parameters, the 3D coordinates (x, y, z) of the colloidal surface and internal defects are calculated to reconstruct the 3D morphology of the colloidal body.
[0040] The beneficial effect of this embodiment is that by combining multi-angle image acquisition with the principle of triangulation, it is possible to reconstruct the three-dimensional morphology of colloidal material on complex curved surfaces (e.g., 3D cover plates, curved mid-frames, etc.); and by eliminating the phase ambiguity of the curved surface through phase unwrapping, it does not require preset curved surface parameters and has good surface self-adaptation capability.
[0041] Example 6 Figure 6This is a flowchart of the sixth embodiment of the colloid visual inspection method of the present invention. Based on the above embodiment, the method further includes: S31. Preset multiple defect types, including glue breakage, glue overflow, air bubble, and glue width unevenness. S32. Set thresholds corresponding to the glue breakage feature, glue overflow feature, bubble feature, and glue width unevenness feature respectively, and extract the defect features according to the thresholds.
[0042] Please refer to Figure 14 The diagram illustrates defect feature extraction, showcasing the principles of weighted fusion or Bayesian fusion algorithms for multi-angle defect features, and the registration and comparison process with a design model (e.g., CAD). Here, a represents the original image, b represents the phase map, c represents the depth map, d represents the gradient feature, e represents the texture feature, and f represents the fusion result. In this embodiment, based on the reconstructed 3D topography data, a voxel segmentation algorithm is used to extract defect features, specifically including the following four types: First, the characteristics of gel breakage: continuity loss, phase abrupt change > threshold T1; Second, the characteristics of glue overflow: the height / width of the glue exceeds the design range [H_min, H_max] or [W_min, W_max]; Third, bubble characteristics: internal voids, phase distortion gradient > threshold T2; Fourth, the characteristic of uneven width: the rate of change of width ΔW / W > the threshold T3.
[0043] The beneficial effect of this embodiment is that by combining CAD model registration and multi-view fusion algorithm, accurate detection of adhesive paths on complex curved surfaces can be achieved.
[0044] Example 7 Figure 7 This is a flowchart of the seventh embodiment of the colloid visual inspection method of the present invention. Based on the above embodiment, the method further includes: S33. Use a preset weighted fusion or a preset Bayesian fusion algorithm to perform multi-scale fusion of the defect features from multiple angles; S34. Dynamically adjust the fusion weights based on the current camera viewpoint, lighting conditions, and defect type to eliminate occlusion and specular reflection interference.
[0045] Please refer to Figure 15The diagram illustrates the principle of multi-scale fusion discrimination, showcasing the weighted fusion or Bayesian fusion algorithm for multi-angle defect features and the registration and comparison process with the CAD model. It utilizes optical, geometric, and texture features for weight allocation, combined with feature alignment and multi-dimensional fusion to achieve defect discrimination. In this embodiment, defect features acquired from multiple angles are fused using a weighted fusion or Bayesian fusion algorithm to eliminate occlusion and specular reflection interference, improving defect recognition accuracy. Furthermore, in this embodiment, registration and comparison are performed with the design CAD model to label the defect location and type.
[0046] The beneficial effect of this embodiment is that, compared with existing two-dimensional image processing, this embodiment improves the accuracy and robustness of defect identification by using a voxel segmentation algorithm based on three-dimensional topography data.
[0047] Example 8 Figure 8 This is a flowchart of the eighth embodiment of the colloid visual inspection method of the present invention. Based on the above embodiment, the method further includes: S41. Detect the type of the colloid; S42. When the colloid is a curved adhesive path, a non-uniform grating projection corresponding to the preset design model is used as the structured light, and a preset curved adaptive phase unwrapping algorithm is used to perform the phase unwrapping calculation.
[0048] In this embodiment, the system detects and outputs results online and in real time, including pass / fail determination, defect location coordinates, defect type classification, and defect size measurement. Optionally, this embodiment also supports integration with the production line MES (Manufacturing Execution System) to achieve quality traceability and process feedback.
[0049] Please refer to Figure 16 The diagram illustrates the architecture of the online inspection system. It showcases the software architecture for the inspection algorithm, result output, and MES system integration, as well as the data flow for quality traceability and process feedback. The application layer includes production management, quality traceability, and statistical analysis; the service layer includes image processing, AI inference, and data storage; the control layer includes equipment control, synchronization triggering, and parameter configuration; and the perception layer includes camera arrays, light source modules, and sensors. In this embodiment, the inspection results are output in real time and integrated with the MES system to achieve quality traceability and process feedback, supporting closed-loop quality control of the production process and improving the intelligence level of the production line.
[0050] The beneficial effects of this embodiment are that by providing a complete online inspection and MES system integration solution, quality traceability and process feedback can be achieved; through the combined application of the above embodiments, key indicators such as internal defect detection capability, detection accuracy, detection speed, and surface adaptability can be enhanced.
[0051] Example 9 Based on the above embodiments, the present invention also proposes a colloid visual inspection device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, it implements the steps of the colloid visual inspection method as described in any of the above embodiments.
[0052] It should be noted that the above-described device embodiments and method embodiments belong to the same concept. The specific implementation process can be found in the method embodiments, and the technical features in the method embodiments are also applicable to the device embodiments, which will not be repeated here.
[0053] Example 10 Based on the above embodiments, the present invention also proposes a computer-readable storage medium storing a colloid vision inspection program, wherein when the colloid vision inspection program is executed by a processor, the steps of the colloid vision inspection method as described in any of the above claims are implemented.
[0054] It should be noted that the above-described medium embodiments and method embodiments belong to the same concept. The specific implementation process can be found in the method embodiments, and the technical features in the method embodiments are also applicable to the medium embodiments, which will not be repeated here.
[0055] The colloidal visual inspection method, device, and computer-readable storage medium of this invention acquire transmission and scattering images of a colloid through multiple cameras at different angles when pre-set structured light passes through it. The transmission and scattering images are then unwrapped to obtain the absolute phase distribution of the colloid. This data is combined with phase distortion and pre-set calibration parameters to reconstruct the colloid's three-dimensional morphology. Defect features from multiple angles are extracted from the 3D morphology data and fused at multiple scales. These features are then registered and compared with a pre-set design model to determine the location and type of defects in the colloid. This invention provides a faster and more effective colloidal visual inspection solution, enhancing its applicability, saving inspection costs, and improving production efficiency.
[0056] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.
[0057] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.
[0058] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.
[0059] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.
Claims
1. A visual inspection method for colloids, characterized in that, The method includes: Multiple cameras were used to capture transmission and scattering images of pre-set structured light passing through the colloid at different angles; Phase unwrapping is performed on the transmitted image and the scattered image to obtain the absolute phase distribution of the colloid, and contour reconstruction is performed by combining the phase distortion and preset calibration parameters to obtain the three-dimensional morphology data of the colloid. Defect features from multiple angles are extracted from the three-dimensional topography data and fused at multiple scales. The data is then registered and compared with a preset design model to determine the location and type of defects in the colloid.
2. The colloid visual inspection method according to claim 1, characterized in that, The method further includes: The structured light is generated periodically using a preset digital light projection device or a preset grating projection module; The grating pattern is set as a combination of one or more of the following: sine stripes, Gray code patterns, and phase-shifted stripes.
3. The colloid visual inspection method according to claim 1, characterized in that, The method further includes: A preset pulsed laser or a preset high-speed LED is used as the lighting source; The pulse frequency of the illumination source is set to be synchronized with the exposure frequency of the camera and matched with the movement speed of the production line.
4. The colloid visual inspection method according to claim 1, characterized in that, The method further includes: The transmitted image is acquired by a camera arranged in a straight line with the structured light and the colloid; The scattered images are acquired by a camera positioned off-axis from the structured light beam.
5. The colloid visual inspection method according to claim 1, characterized in that, The method further includes: Acquire multiple phase-shifted images corresponding to the transmitted image and the scattered image; The phase unwrapping calculation of the multi-frame phase-shifted images is performed by one or more combinations of the preset four-step phase shift method, multi-frequency heterodyne method, and time phase expansion method to obtain the absolute phase distribution.
6. The colloid visual inspection method according to claim 1, characterized in that, The method further includes: Multiple defect types are preset, including glue breakage, glue overflow, air bubbles, and uneven glue width. Thresholds are set for the glue breakage feature, glue overflow feature, bubble feature, and glue width unevenness feature, respectively, and the defect features are extracted according to the thresholds.
7. The colloid visual inspection method according to claim 6, characterized in that, The method further includes: The defect features from multiple perspectives are fused at multiple scales using a preset weighted fusion or a preset Bayesian fusion algorithm. The fusion weights are dynamically adjusted based on the current camera viewpoint, lighting conditions, and defect type to eliminate occlusion and specular reflection interference.
8. The colloid visual inspection method according to claim 7, characterized in that, The method further includes: Detect the type of the colloid; When the colloid is a curved adhesive path, a non-uniform grating projection corresponding to the preset design model is used as the structured light, and a preset curved adaptive phase unwrapping algorithm is used to perform the phase unwrapping calculation.
9. A colloidal visual inspection device, characterized in that, The device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the colloid visual inspection method as described in any one of claims 1 to 8.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an application colloid vision inspection program, which, when executed by a processor, implements the steps of the colloid vision inspection method as described in any one of claims 1 to 8.