Three-dimensional structure defect detection device and method based on multi-point optical detection

Abstract

The invention discloses a three-dimensional structure defect detection device and method based on multi-point optical detection, and relates to the technical field of three-dimensional structure defect detection.The three-dimensional structure defect detection method based on multi-point optical detection comprises the specific steps that firstly, optical detection heads are arranged according to an actual detection site and a to-be-detected target, the method comprises the following steps of 1, setting module parameters, 2, driving optical path time-sharing switching and synchronous acquisition, obtaining a multi-point interference signal and recording parameters, and realizing multi-region data acquisition, and 3, performing preprocessing and coherent processing on the acquired signal, extracting depth information, splicing a tomographic image, constructing an image for defect analysis, and carrying out data acquisition. 4, defect detection is conducted on the tomographic image through an image processing algorithm, and a detection report containing defect information and qualification judgment is generated. Efficient imaging, quality analysis and defect detection of multiple surfaces and internal structures of a product are achieved through multi-channel switching and synchronous scanning probe configuration.

Description

Technical Field

[0001] This invention belongs to the field of three-dimensional structural defect detection technology, specifically relating to a three-dimensional structural defect detection device and method based on multi-point optical detection. Background Technology

[0002] 3D defect detection plays a crucial role in industrial automated manufacturing processes, improving product quality and safety, increasing production efficiency and reducing costs, and driving industrial automation upgrades. However, existing 3D defect detection technologies have the following drawbacks:

[0003] Vision-based inspection technologies can only image and identify defects on the product surface, unable to penetrate deep into the surface being tested. This prevents the imaging of internal defects below the material surface (such as incomplete welds in laser welding, internal bubbles in lenses or plastic seals, etc.), thus limiting the scope of application for such defects. Furthermore, changes in lighting conditions significantly impact the results of vision-based inspection technologies, easily leading to false positives or false negatives.

[0004] Ultrasonic testing equipment has limited resolution, complex calibration, high price, and relies on the operator's experience, making it unsuitable for full online inspection of product packaging areas in high-speed automated industrial manufacturing processes.

[0005] Because it's impossible to image different areas of a target collected by multiple optical lenses onto a single area array detector, industrial applications primarily use one optical lens per area array detector to detect defects in a single area. It's not possible to use a single system to simultaneously detect multiple areas of a target. Similarly, ultrasonic detection also cannot simultaneously detect targets at multiple locations.

[0006] Therefore, there is an urgent need for a three-dimensional structural defect detection device and method based on multi-point optical detection to solve the above problems. Summary of the Invention

[0007] The purpose of this invention is to provide a three-dimensional structural defect detection device and method based on multi-point optical detection, which solves the technical problem that the prior art cannot image internal defects below the surface of materials.

[0008] To achieve the above objectives, the present invention adopts the following technical solution:

[0009] A method for detecting three-dimensional structural defects based on multi-point optical inspection, comprising:

[0010] Step 1: Deploy the optical detector head and set the module parameters according to the actual testing site and the target to be inspected;

[0011] Step 2: Drive the optical path to switch in time and acquire synchronously, obtain multi-point interference signals and record parameters to achieve multi-region data acquisition;

[0012] Step 3: Preprocess and coherently process the acquired signals, extract depth information and stitch together tomographic images to construct images for defect analysis;

[0013] Step 4: Use image processing algorithms to perform defect detection on the tomographic image and generate an inspection report containing defect information and acceptance criteria.

[0014] Furthermore, the module parameters are set, specifically as follows:

[0015] The light source module is set to a broadband or frequency sweep light source, with an output optical power of a1-a2 and a center wavelength of a3.

[0016] The beam splitter is set to a fiber optic circulator to split the light into a reference optical path and a sample optical path.

[0017] The optical path distribution module is configured to use a fast reflector or galvanometer with a switching speed ≥ a4;

[0018] The optical path switching module is set as an optical fiber coupler with a coupling loss ≤ a5;

[0019] Set the XY scanning range of the galvanometer scanning module to A×A to B×B, with a step size of b1-b2 and a speed of b3-b4;

[0020] Set the preset light source timing, optical path switching window, and spectrum acquisition trigger frequency c of the control module;

[0021] Among them, a1, a2, a3, a4, a5, A, B, b1, b2, b3, b4 and c are all constant values, determined based on historical experimental data and actual needs.

[0022] Furthermore, the optical path is time-division switching and synchronous acquisition is performed using the following method:

[0023] The control module starts the light source at time t0. The light is split into a reference light path and a sample light path by the spectrometer. The reference light is reflected by the fixed mirror and then returns to the spectrum acquisition module.

[0024] During the t0 to t1 period, the control module outputs a high level, and the optical path distribution module switches to optical path branch 1, which is divided into two sub-channels by the optical path switching module. Optical probes 1 and 2 scan the Sample1 and Sample2 regions synchronously through the galvanometer scanning module. The sample light collected is transmitted to the spectrum acquisition module after interfering with the reference light.

[0025] During the t1 to t2 period, the control module outputs a low level, the optical path distribution module switches to optical path branch 2, which is divided into two sub-channels by the optical path switching module. Optical probes 3 and 4 scan the Sample3 and Sample4 regions simultaneously, and the collected interference signals are transmitted to the spectrum acquisition module.

[0026] After t2, the switching logic repeats until the scan is complete. The spectral acquisition module synchronously records the XY coordinates, optical path branch number, and timestamp corresponding to the signal.

[0027] Furthermore, the specific methods for preprocessing and coherent processing of the acquired signals are as follows:

[0028] The original signal is corrected in wavenumber space, background is removed and windowed, and then Fourier transformed from wavenumber space to depth space.

[0029] Based on the principle of low-coherence optical interference, the interference signal at each XY scanning position is converted from the wavelength domain to the depth domain by Fourier transform to obtain the intensity distribution curve of the depth direction at a single point.

[0030] Extract the peak position and intensity of the curve to determine the depth coordinates and structural characteristics of that point.

[0031] Furthermore, depth information is extracted and tomographic images are stitched together. The specific method is as follows:

[0032] All A-scan signals from optical path branch 1 during the time period t0 to t1 are arranged according to spatial coordinates and stitched together to form two-dimensional tomographic images of Sample1 and Sample2; similarly, signals from optical path branch 2 during the time period t1 to t2 are stitched together to form tomographic images of Sample3 and Sample4.

[0033] The detection area is divided according to the preset optical path difference. For example, the image of Sample1 in optical path branch 1 is located in the depth range of 0 to h1, and the image of Sample2 is located in the depth range of h1 to h2. h1 is the preset optical path difference. Optical path branch 2 is divided according to the same logic.

[0034] The present invention also provides a three-dimensional structural defect detection device based on multi-point optical detection, which is applied to a three-dimensional structural defect detection method based on multi-point optical detection. Specifically, it includes: a light source module, a beam splitting component, an optical path distribution module, an optical path switching module, an optical probe, a galvanometer scanning module, a control module, a spectrum acquisition module, and a data processing module.

[0035] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0036] 1. This invention achieves high-precision detection of surface and internal defects of products (micron-level lateral resolution, millimeter-level imaging depth) through the principle of optical interference. Combined with the time-division switching of multiple optical paths and the synchronous acquisition mode with preset optical path difference within the same branch, it supports at least 4-point parallel detection, which greatly improves detection efficiency. At the same time, it avoids the shortcomings of traditional visual inspection that cannot penetrate the interior and ultrasonic inspection that has low resolution and high cost. It is suitable for industrial automation online full inspection scenarios, which can effectively ensure product quality and reduce inspection costs. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 The diagram illustrates the steps of a three-dimensional structural defect detection method based on multi-point optical detection according to the present invention.

[0039] Figure 2 The hardware module composition and control flow of a three-dimensional structural defect detection device based on multi-point optical detection according to the present invention are shown. Detailed Implementation

[0040] 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.

[0041] Example 1, such as Figure 1 The method for detecting three-dimensional structural defects based on multi-point optical inspection, as shown, specifically includes the following steps:

[0042] Step 1: Arrange the optical detector head and set the module parameters according to the actual testing site and the target to be tested.

[0043] Based on the number, spatial distribution, and detection accuracy requirements of the targets to be detected, multiple optical probes are deployed (the number is determined by the number of branches in the optical path distribution module and the number of sub-channels in each branch, with a minimum of 4). Optical probes within the same optical path branch (such as probe 1 and probe 2 in optical path branch 1) require a pre-set fixed optical path difference, specifically:

[0044] The optical path difference needs to be more than twice the system's depth resolution to ensure that the signals collected by different probes do not overlap in the depth domain and to avoid signal interference.

[0045] The optical probe is pointed directly at the area to be tested. Its spatial position needs to be adjusted according to the shape of the sample (such as a plane or a curved surface) to ensure that the detection direction is perpendicular to the surface to be tested (or tilted at a preset angle, with the angle error compensated by an algorithm).

[0046] Set up the light source module and select a broadband light source (suitable for static detection) or a frequency sweep light source (suitable for dynamic high-speed detection) according to the detection scenario. Set the output optical power (e.g., 5-20mW) and center wavelength (e.g., 850nm or 1310nm, to match the optical characteristics of the material being tested).

[0047] A beam splitting assembly is set up, using a fiber optic circulator as the beam splitting device to divide the light output from the light source into a reference light path (accounting for 30%-50%) and a sample light path (accounting for 50%-70%), ensuring that the intensity of the two light paths is matched.

[0048] An optical path distribution module is set up, using a fast reflector or galvanometer as the core switching device, and the optical path branch switching speed is preset (e.g., ≥1kHz) to ensure the continuity of time-division detection.

[0049] Set up an optical path switching module and use fiber optic couplers (such as 1×2 or 1×4 couplers) to divide each optical path branch into two or more sub-channels, with coupling loss controlled to ≤3dB;

[0050] Set up the galvanometer scanning module and preset the XY plane scanning range (e.g., 10mm×10mm to 100mm×100mm), scanning step size (e.g., 5-20μm, adjusted according to the horizontal resolution requirements) and scanning speed (e.g., 100-500mm / s, matching the production line speed).

[0051] Configure the control module and preset various timing parameters, including the light source start / stop time, optical path branch switching time window (e.g., t0~t1 is branch 1, t1~t2 is branch 2), and spectrum acquisition trigger frequency (synchronized with the scanning step size, e.g., 10kHz).

[0052] Step 2: Drive the optical path to switch in time and acquire synchronously, obtain multi-point interference signals and record parameters to achieve multi-region data acquisition.

[0053] The control module sends a start command at time t0, and the light source module emits an optical signal, which is split into a reference optical path and a sample optical path by the beam splitter (fiber optic circulator).

[0054] The optical signal in the reference optical path is reflected by the fixed mirror and then returns to the spectrum acquisition module along the original path;

[0055] The optical signal in the sample optical path enters the optical path distribution module and waits for branch switching.

[0056] During the time interval t0 to t1, the control module outputs a high level, and the optical path distribution module (galvanometer / fast reflector) switches the sample optical path to optical path branch 1. At this time, optical path branch 1 is divided into two sub-channels by the optical path switching module (fiber optic coupler), corresponding to optical probe 1 and probe 2 entering the working state.

[0057] The galvanometer scanning module drives probe 1 to scan the Sample1 region in the XY plane, while probe 2 simultaneously scans the Sample2 region.

[0058] The sample light signals collected by the two probes are transmitted to the spectral acquisition module via sub-channels, where they interfere with the reference light signal.

[0059] During the time period t1 to t2, the control module outputs a low level, and the optical path distribution module switches to optical path branch 2. Optical path branch 2 is divided into two sub-channels through another optical path switching module, and optical probes 3 and 4 start working:

[0060] The galvanometer scanning module drives probe 3 to scan the Sample 3 region, while probe 4 simultaneously scans the Sample 4 region;

[0061] The collected sample optical signal is transmitted to the spectral acquisition module after interfering with the reference optical signal.

[0062] After time t2, the control module repeats the high-low level switching logic until all areas to be tested are scanned (the number of scans is determined according to the sample size and scanning range).

[0063] The spectral acquisition module is set up to synchronously acquire interference light signals in real time (including interference information between sample light and reference light), receive synchronization markers sent by the control module, and record the XY scanning position coordinates corresponding to each signal (feedback in real time by the galvanometer scanning module, with an accuracy of ±1μm), optical path branch number (1 or 2, used for subsequent image stitching) and timestamp (accurate to the μs level to ensure timing consistency).

[0064] For time-division switching, if it is a two-way switching, it can be achieved by the simplest high-low level switching;

[0065] If the system is expanded to multiple channels, other methods need to be added, including but not limited to: limit switches, photoelectric sensors, encoders, or position sensors.

[0066] Step 3: Preprocess and coherently process the acquired signals, extract depth information, and stitch together tomographic images to construct images for defect analysis.

[0067] The data processing module preprocesses the raw signal output by the spectral acquisition module to remove noise such as dark current and ambient light (using a background subtraction algorithm) and normalizes the signal (to eliminate the influence of light intensity fluctuations).

[0068] Based on the principle of low-coherence optical interference, the interference signal at each XY scan position is coherently processed:

[0069] The interference signal is converted from the wavelength domain to the depth domain by Fourier transform, and the intensity distribution curve (A-scan signal) of a single point in the depth direction is obtained.

[0070] Extract the peak position and intensity from the curve, and determine the depth coordinates (accuracy ±5μm) and structural characteristics (such as density and refractive index changes) of that point.

[0071] All A-scan signals at XY positions acquired by optical path branch 1 during the time period t0 to t1 are arranged according to spatial coordinates and stitched together to form two-dimensional tomographic images (B-scan images) of Sample 1 and Sample 2. Similarly, the signals of optical path branch 2 during the time period t1 to t2 are stitched together to form tomographic images of Sample 3 and Sample 4.

[0072] Multi-probe region differentiation: Based on preset optical path difference information, the detection regions of different probes are divided in the tomographic image. For example, in optical path branch 1, the tomographic image of Sample 1 is located in the depth range of 0 to h1, and the image of Sample 2 is located in the depth range of h1 to h2 (h1 is the preset optical path difference);

[0073] The region division logic for optical path branch 2 is the same as above.

[0074] Step 4: Use image processing algorithms to perform defect detection on the tomographic image and generate a detection report.

[0075] Set a defect judgment threshold (e.g., use twice the background gray value as the boundary, and mark areas exceeding this value as suspected defects);

[0076] The shape (e.g., pores are circular / elliptical, cracks are linear, and cold welds are irregular areas), size (area, depth, length), and spatial distribution of defects are identified through edge detection and region growing algorithms.

[0077] Statistical analysis of the number, density, and maximum size of defects (e.g., in plastic sealant testing, the number of bright spots is linearly negatively correlated with the edge sealing strength, and can be directly used for strength assessment).

[0078] Output a detection report containing the following information:

[0079] Three-dimensional tomographic images of each sample to be tested (with defect locations and types labeled);

[0080] Defect parameter table (including defect ID, type, coordinates, size, severity level);

[0081] The result of the conformity assessment (based on a preset defect threshold, such as "conforming", "unconforming" or "rework recommendation");

[0082] Applicable scenarios (e.g., detection of incomplete welds in laser welding, detection of air bubbles inside lenses, and detection of the strength of adhesive sealant boundaries).

[0083] Example 2, as follows Figure 2 The three-dimensional structural defect detection device based on multi-point optical detection shown includes: a light source module, a beam splitting component, an optical path distribution module, an optical path switching module, an optical probe, a galvanometer scanning module, a control module, a spectrum acquisition module, and a data processing module.

[0084] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

[0085] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to specific implementations. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for detecting three-dimensional structural defects based on multi-point optical inspection, characterized in that, include: Step 1: Deploy the optical detector head and set the module parameters according to the actual testing site and the target to be inspected; Step 2: Drive the optical path to switch in time and acquire synchronously, obtain multi-point interference signals and record parameters to achieve multi-region data acquisition; Step 3: Preprocess and coherently process the acquired signals, extract depth information and stitch together tomographic images to construct images for defect analysis; Step 4: Use image processing algorithms to perform defect detection on the tomographic image and generate an inspection report containing defect information and acceptance criteria.

2. The method for detecting three-dimensional structural defects based on multi-point optical detection according to claim 1, characterized in that, To set module parameters, the specific method is as follows: The light source module is set to a broadband or frequency sweep light source, with an output optical power of a1-a2 and a center wavelength of a3. The beam splitter is set to a fiber optic circulator to split the light into a reference optical path and a sample optical path. The optical path distribution module is configured to use a fast reflector or galvanometer with a switching speed ≥ a4; The optical path switching module is set as an optical fiber coupler with a coupling loss ≤ a5; Set the XY scanning range of the galvanometer scanning module to A×A to B×B, with a step size of b1-b2 and a speed of b3-b4; Set the preset light source timing, optical path switching window, and spectrum acquisition trigger frequency c of the control module; Among them, a1, a2, a3, a4, a5, A, B, b1, b2, b3, b4 and c are all constant values, determined based on historical experimental data and actual needs.

3. The method for detecting three-dimensional structural defects based on multi-point optical detection according to claim 1, characterized in that, The specific method for driving time-division switching and synchronous acquisition of the optical path is as follows: The control module starts the light source at time t0. The light is split into a reference light path and a sample light path by the spectrometer. The reference light is reflected by the fixed mirror and then returns to the spectrum acquisition module. During the t0 to t1 period, the control module outputs a high level, and the optical path distribution module switches to optical path branch 1, which is divided into two sub-channels by the optical path switching module. Optical probes 1 and 2 scan the Sample1 and Sample2 regions synchronously through the galvanometer scanning module. The sample light collected is transmitted to the spectrum acquisition module after interfering with the reference light. During the t1 to t2 period, the control module outputs a low level, the optical path distribution module switches to optical path branch 2, which is divided into two sub-channels by the optical path switching module. Optical probes 3 and 4 scan the Sample3 and Sample4 regions simultaneously, and the collected interference signals are transmitted to the spectrum acquisition module. After t2, the switching logic repeats until the scan is complete. The spectral acquisition module synchronously records the XY coordinates, optical path branch number, and timestamp corresponding to the signal.

4. The method for detecting three-dimensional structural defects based on multi-point optical detection according to claim 1, characterized in that, The specific methods for preprocessing and coherent processing of the acquired signals are as follows: The original signal is corrected in wavenumber space, background is removed and windowed, and then Fourier transformed from wavenumber space to depth space. Based on the principle of low-coherence optical interference, the interference signal at each XY scanning position is converted from the wavelength domain to the depth domain by Fourier transform to obtain the intensity distribution curve of the depth direction at a single point. Extract the peak position and intensity of the curve to determine the depth coordinates and structural characteristics of that point.

5. The method for detecting three-dimensional structural defects based on multi-point optical detection according to claim 1, characterized in that, The specific method for extracting depth information and stitching together tomographic images is as follows: All A-scan signals from optical path branch 1 during the time period t0 to t1 are arranged according to spatial coordinates and stitched together to form two-dimensional tomographic images of Sample1 and Sample2; similarly, signals from optical path branch 2 during the time period t1 to t2 are stitched together to form tomographic images of Sample3 and Sample4. The detection area is divided according to the preset optical path difference. For example, the image of Sample1 in optical path branch 1 is located in the depth range of 0 to h1, and the image of Sample2 is located in the depth range of h1 to h2. h1 is the preset optical path difference. Optical path branch 2 is divided according to the same logic.

6. A three-dimensional structural defect detection device based on multi-point optical detection, applied to the three-dimensional structural defect detection method based on multi-point optical detection as described in any one of claims 1-5, characterized in that, Specifically, it includes: The system includes a light source module, a beam splitter, an optical path distribution module, an optical path switching module, an optical probe, a galvanometer scanning module, a control module, a spectrum acquisition module, and a data processing module.

Images