Titanium rod coating damage visual detection method and device based on optical imaging

Abstract

The invention discloses a titanium rod coating damage visual detection method and device based on optical imaging, and relates to the technical field of titanium rod coating detection. The method comprises the following steps: acquiring processing information of a titanium rod coating, performing two-dimensional focal plane deployment and over-focus trajectory planning based on processing damage risk, and determining an optical scanning scheme; connecting the optical detection equipment with the precise displacement table, carrying out incident scanning and CCD target surface imaging based on scattered light on the titanium rod according to an optical scanning scheme, and determining imaging information; a dual-filtering detection module is developed in an imaging system, dual-channel imaging and mutual inspection based on high-pass filtering and low-pass filtering are carried out on imaging information, target imaging is determined, three-dimensional space stacking reconstruction is executed, and coating damage distribution is determined. The technical problems that in the prior art, the titanium rod coating damage detection precision is low, and tiny losses are difficult to recognize are solved, and the technical effects that the detection precision is improved, and coating damage visualization is achieved are achieved.

Description

Technical Field

[0001] The invention relates to the technical field of titanium rod coating detection, and in particular to a titanium rod coating damage visualization detection method and device based on optical imaging. Background Art

[0002] As a high-performance metal material, titanium rods are widely used in aerospace, medical equipment, chemical equipment and other key fields due to their high strength, low density, good corrosion resistance and biocompatibility. However, in practical applications, the surface of titanium rods often needs to be coated with a layer of specific materials to improve their wear resistance, corrosion resistance or other specific properties. The quality of this coating is directly related to the service life and overall performance of the titanium rod. However, during the processing and coating of titanium rods, due to various factors such as improper control of process parameters, differences in material properties, operational errors, etc., various types of damage may appear on the coating surface, such as scratches, cracks, peeling, etc. These damages will not only reduce the aesthetics and performance of titanium rods, but may also become potential failure points, affecting the reliability and safety of products. Existing coating damage detection technologies mainly rely on manual visual inspection, ultrasonic testing or conventional optical testing, but these methods often have problems such as low detection accuracy, inability to effectively identify tiny damage or sub-surface defects, and large interference from environmental factors. Summary of the invention

[0003] The present application provides a method and device for visually detecting titanium rod coating damage based on optical imaging, which solves the technical problems in the prior art of low accuracy in detecting titanium rod coating damage and difficulty in identifying subtle losses.

[0004] In a first aspect of the present application, a method for visually detecting titanium rod coating damage based on optical imaging is provided, the method comprising:

[0005] The processing information of the titanium rod coating is obtained, and two-dimensional focal plane deployment and over-focus trajectory planning are performed based on the processing damage risk, and the optical scanning scheme is determined, wherein the processing damage risk includes surface damage and sub-surface damage; the optical detection equipment is connected to the precision translation stage, and according to the optical scanning scheme, the titanium rod is incident scanned and CCD target surface imaging based on scattered light is performed to determine the imaging information, wherein at least two groups of parallel detections are included; a dual-filter detection module is developed in the imaging system, and the imaging information is subjected to dual-channel imaging and mutual verification based on high-pass filtering and low-pass filtering to determine the target imaging and perform three-dimensional space stacking reconstruction to determine the coating damage distribution, wherein the coating damage distribution has a priori identification based on damage and contaminants, and the target imaging is effective information under dual-filter mutual verification and parallel detection mutual verification.

[0006] The second aspect of the present application provides a titanium rod coating damage visualization detection device based on optical imaging, the device comprising:

[0007] A scanning scheme determination module is used to obtain processing information of titanium rod coating, perform two-dimensional focal plane deployment and over-focus trajectory planning based on processing damage risks, and determine an optical scanning scheme, wherein processing damage risks include surface damage and sub-surface damage; an imaging information acquisition module is used to connect optical detection equipment and a precision translation stage, and perform incident scanning and CCD target imaging based on scattered light on the titanium rod according to the optical scanning scheme to determine imaging information, wherein at least two sets of parallel detections are included; a loss determination module is used to develop a dual-filter detection module in the imaging system, perform dual-channel imaging and mutual verification based on high-pass filtering and low-pass filtering on the imaging information, determine target imaging and perform three-dimensional spatial stacking reconstruction, and determine coating damage distribution, wherein the coating damage distribution has a priori identification based on damage and contaminants, and the target imaging is effective information under dual-filter mutual verification and parallel detection mutual verification.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] First, the processing information of the titanium rod coating is obtained, and the two-dimensional focal plane deployment and over-focus trajectory planning based on the processing damage risk are carried out to determine the optical scanning scheme, where the processing damage risk includes surface damage and sub-surface damage. Then, the optical detection equipment is connected to the precision translation stage, and according to the optical scanning scheme, the titanium rod is incident scanned and the CCD target surface imaging based on scattered light is performed to determine the imaging information, which includes at least two sets of parallel detection. Finally, a dual-filter detection module is developed in the imaging system, and the imaging information is subjected to dual-channel imaging and mutual verification based on high-pass filtering and low-pass filtering, the target imaging is determined, and three-dimensional space stacking reconstruction is performed to determine the coating damage distribution, where the coating damage distribution has a priori identification based on damage and contaminants, and the target imaging is effective information under dual-filter mutual verification and parallel detection mutual verification. The technical problem of low detection accuracy of titanium rod coating damage and difficulty in identifying subtle losses in the prior art is solved, and the technical effect of improving detection accuracy and realizing visualization of coating damage is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0011] Figure 1 A schematic diagram of a process flow of a titanium rod coating damage visualization detection method based on optical imaging provided in an embodiment of the present application;

[0012] Figure 2 A schematic diagram of the structure of a titanium rod coating damage visualization detection device based on optical imaging provided in an embodiment of the present application.

[0013] Explanation of the reference numerals: scanning scheme determination module 11 , imaging information acquisition module 12 , loss determination module 13 . DETAILED DESCRIPTION

[0014] The present application solves the technical problems in the prior art of low accuracy in detecting titanium rod coating damage and difficulty in identifying subtle losses by providing a method and device for visually detecting titanium rod coating damage based on optical imaging.

[0015] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0016] It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or are inherent to these processes, methods, products or devices.

[0017] Embodiment 1, as Figure 1 As shown, the present application provides a method for visually detecting titanium rod coating damage based on optical imaging, wherein the method comprises:

[0018] The processing information of the titanium rod coating is obtained, and two-dimensional focal plane deployment and over-focus trajectory planning are performed based on the processing damage risk, and the optical scanning plan is determined. The processing damage risk includes surface damage and sub-surface damage.

[0019] The processing information of the titanium rod coating is obtained through manufacturing process records, material characterization tests (such as X-ray fluorescence spectroscopy analysis, scanning electron microscopy observation, etc.) or direct measurement (such as white light interferometry, profilometer measurement). The processing information includes but is not limited to the coating material type, coating thickness, substrate material characteristics, coating process parameters (such as deposition method, temperature, atmosphere conditions, etc.), surface roughness, stress distribution, etc.

[0020] Based on the processing information, the possible processing damage risks are analyzed and the damage types are classified, where the processing damage risks include surface damage and sub-surface damage. The surface damage category includes but is not limited to scratches, cracks, peeling, pores and other defects on the coating surface, and the sub-surface damage category includes but is not limited to hidden defects such as microcracks, voids, and delamination inside the coating or near the interface.

[0021] After determining the risk of processing damage, a coating space model is constructed based on the coating thickness and the two-dimensional coordinates of the surface, and a two-dimensional focal plane deployment is performed in the coating space. Specifically, the coating thickness is used as the longitudinal depth parameter, and the three-dimensional coating space model is constructed with the two-dimensional coordinate system of the titanium rod surface. The axial direction of the titanium rod is used as the X-axis, and the expansion angle or linear expansion length formed after the titanium rod is expanded in the circumferential direction is used as the Y-axis to establish a two-dimensional coordinate system on the titanium rod surface, thereby achieving accurate calibration of the titanium rod surface position. Secondly, based on historical defect data, the damage risk points in the two-dimensional coordinate system on the titanium rod surface are identified and located. According to the distribution of the damage risk points, the key detection areas are set, and different detection focal planes are set for surface damage and sub-surface damage, respectively, to complete the two-dimensional focal plane deployment. The surface damage detection focal plane is set at a value close to zero on the Z axis (i.e., the outermost layer of the coating), while the sub-surface damage detection focal plane is deployed in layers according to the set layer depth.

[0022] Based on the damage risk points and the coating space, the over-focus trajectory planning is carried out to optimize the optical detection scheme. Specifically, for surface damage, a fixed focal plane, that is, the coating surface layer (Z-axis zero layer) is used for full-area scanning detection to cover all surface damage risk areas; for sub-surface damage, based on the layer depth positioning of the damage risk point, the key focus is set, and imaging information of different depths is obtained by layer-by-layer scanning; at the same time, according to the spatial distribution of the damage risk point, the over-focus scanning trajectory is formulated to optimize the detection path to ensure complete coverage of possible damage areas, so as to finally determine the optical scanning scheme. Among them, the setting of the key focus includes: according to the coordinates of each sub-surface damage risk point in the three-dimensional coating space model, its Z-axis depth information, that is, its corresponding layer depth position, is extracted; all sub-surface damage risk points are layered and classified according to the Z axis to form multiple depth layer groups; in each depth layer group, the geometric center or weighted center of all risk points in the layer in the XY plane is calculated (the weight can be based on the risk level or distribution density), and the center point is set as the key focus of the layer.

[0023] Furthermore, two-dimensional focal plane deployment and over-focus trajectory planning are performed, including:

[0024] Based on the processing information, damage risk points are located, wherein the damage risk points include surface damage risk points and sub-surface damage risk points; the two-dimensional space is determined by the coating surface, and the third spatial dimension is determined by the coating thickness to construct the coating space; based on the damage risk points and the coating space, two-dimensional focal plane deployment and over-focus trajectory planning are performed.

[0025] First, based on the processing information, the processing characteristics of the coating are determined, and risk assessment is performed on potential damage areas to locate damage risk points. Damage risk points include surface damage risk points and sub-surface damage risk points. Surface damage risk points refer to risk areas where scratches, cracks, peeling or pores may appear on the coating surface, while sub-surface damage risk points refer to areas where hidden defects such as microcracks, voids, and delamination may exist inside the coating or near the interface. Then, a coating space model is constructed based on the coating thickness and the two-dimensional surface coordinates, in which the coating thickness is the longitudinal dimension and the two-dimensional surface coordinates are the lateral dimension to form a three-dimensional data structure. In the coating space, the detection area is divided according to the distribution of damage risk points, and the detection strategies corresponding to different damage types are determined.

[0026] Furthermore, based on the damage risk points and coating space, two-dimensional focal plane deployment and over-focus trajectory planning are performed. For surface damage risk points, a fixed focal plane is set on the coating surface, and a full-area scanning strategy is adopted to cover the entire surface area; for sub-surface damage risk points, the focal plane depth is adjusted layer by layer in combination with the coating thickness information, and detection focal planes of different depths are set to obtain complete sub-surface damage information. At the same time, based on the spatial distribution of damage risk points, an over-focus scanning trajectory is formulated to ensure that the optical detection equipment can pass through the set key focal points in sequence during the scanning process, so as to optimize the scanning path and improve detection efficiency and accuracy.

[0027] Furthermore, based on the damage risk point and the coating space, two-dimensional focal plane deployment and over-focus trajectory planning are performed, including:

[0028] For the sub-surface damage risk point, the layer depth is located based on the coating thickness to determine the two-dimensional focal plane; the sub-surface damage risk point is traversed to determine the key focus at the position located at the corresponding two-dimensional focal plane; the focus trajectory is planned with the key focus as the necessary position of the scanning trajectory to determine the sub-surface scanning plan; the surface scanning plan and the sub-surface scanning plan are integrated to determine the optical scanning plan, wherein the surface damage plan is a full-area scan with a focus on the surface damage risk point.

[0029] Specifically, based on the coating space model, the relative depth of each sub-surface damage risk point in the coating thickness direction is calculated, and the detection focal plane is set at the corresponding depth to ensure that the optical detection system can accurately focus on the target layer and realize effective detection of sub-surface damage. Then, the sub-surface damage risk points are traversed, and the key focus is determined at the corresponding two-dimensional focal plane position. The key focus refers to the high-priority detection area corresponding to the known damage risk point on a specific focal plane. Next, according to the spatial distribution of the key focus, the scanning trajectory is designed so that the optical detection equipment can adjust the focus position according to the set trajectory during the scanning process, and obtain sub-surface damage information of different depths layer by layer. Finally, the surface scanning scheme and the sub-surface scanning scheme are fused to determine the final optical scanning scheme. Among them, the surface scanning scheme focuses on the surface damage risk point and adopts a full-domain scanning strategy to cover the entire coating surface to ensure comprehensive detection of surface damage; the sub-surface scanning scheme performs layered scanning for sub-surface damage of different depths according to the path planned by the over-focus trajectory. Through the fusion of surface and sub-surface scanning, a complete optical scanning scheme is finally formed to achieve efficient and accurate detection of titanium rod coating damage.

[0030] The optical detection device is connected to the precision translation stage, and according to the optical scanning scheme, the titanium rod is incident scanned and the CCD target surface is imaged based on the scattered light to determine the imaging information, which includes at least two sets of parallel detection.

[0031] First, the optical detection equipment is connected to the precision displacement stage, wherein the optical detection equipment includes a light source, a lens system, a CCD imaging component, etc., and the precision displacement stage is used to carry and accurately control the position and scanning trajectory of the titanium rod. The coordinated operation of the optical detection equipment and the precision displacement stage ensures high stability and high precision in the optical scanning process. Then, according to the optical scanning scheme, the titanium rod coating is incident scanned. Specifically, the light source is controlled to emit incident light at a specific angle and wavelength range so that the light beam is irradiated to the surface of the titanium rod, and the position of the titanium rod is adjusted using the precision displacement stage according to the set scanning trajectory to ensure that the light beam covers all target detection areas during the scanning process. Then, imaging information is obtained based on CCD target surface imaging of scattered light, wherein scattered light refers to backscattering, side scattering or transmission signals generated after the incident light interacts with the surface or sub-surface damage characteristics of the coating. The CCD imaging component receives the scattered light and projects it onto the CCD target surface through a high-precision optical system to form optical imaging information, which can not only reflect the surface damage characteristics, but also image the sub-surface damage under the appropriate optical parameter configuration. In addition, in order to improve imaging reliability and information integrity, at least two sets of parallel detection are performed during the optical scanning process. Specifically, two different detection viewing angles are set respectively, so that the optical detection device obtains independent imaging information at two different angles.

[0032] Furthermore, at least two parallel tests are included, including:

[0033] By coordinating the optical detection device and the precision displacement stage, a first detection viewing angle and a second detection viewing angle are determined, wherein the titanium rod is placed on the precision displacement stage; and according to the first detection viewing angle and the second detection viewing angle, parallel detection based on the optical scanning scheme is performed.

[0034] By coordinating the optical detection device and the precision displacement stage, the first detection angle of view and the second detection angle of view are determined. The titanium rod is fixed on the precision displacement stage, and the precision displacement stage can accurately adjust the position and angle of the titanium rod to ensure that the optical detection device performs imaging at the set detection angle of view. The first detection angle of view and the second detection angle of view refer to the optical configuration for detecting the titanium rod coating from different incident angles or receiving angles, respectively, in order to optimize the appearance effect of damage characteristics. According to the first detection angle of view and the second detection angle of view, parallel detection based on the optical scanning scheme is performed. Specifically, at the first detection angle of view, the light source irradiates the surface of the titanium rod according to the set incident angle, and the scattered light is projected onto the CCD target surface through the optical system to form a first set of imaging information; at the second detection angle of view, the same scanning detection is performed at different incident angles or receiving angles to obtain the second set of imaging information. Through two sets of parallel detection, multi-angle observation of coating damage can be achieved, the visibility of damage characteristics can be improved, and the recognition ability of complex damage morphology (such as microcracks, spalling, sub-surface voids, etc.) can be enhanced. In addition, parallel detection results can be used for cross-validation to reduce detection errors that may be caused by a single perspective, improve detection accuracy and stability, and provide more reliable imaging data for subsequent damage analysis and three-dimensional reconstruction.

[0035] A dual-filter detection module is developed in the imaging system, and the imaging information is subjected to dual-channel imaging and mutual verification based on high-pass filtering and low-pass filtering, the target imaging is determined, and three-dimensional spatial stacking reconstruction is performed to determine the coating damage distribution, wherein the coating damage distribution has a priori identification based on damage and contaminants, and the target imaging is effective information under dual-filter mutual verification and parallel detection mutual verification.

[0036] A dual-filter detection module is built in the imaging system. The module includes a high-pass filter channel and a low-pass filter channel, which are used to extract information of different spatial frequency characteristics. The high-pass filter channel is used to enhance the detailed features of the coating damage area, such as microcracks and small peeling, so that it is clearer in the imaging; the low-pass filter channel is used to smooth the background and highlight the large-scale damage area, such as large-area coating peeling or contaminant attachment.

[0037] Based on the dual-channel imaging information, a mutual verification process is performed. Specifically, in the first channel (high-pass filtering), imaging information of bright damage with dark background is generated, that is, the damaged area appears as a high-brightness signal; in the second channel (low-pass filtering), imaging information of dark damage with bright background is generated, that is, the damaged area appears as a low-brightness signal. By mutual verification of the two imaging results, that is, comparing the imaging consistency of the same area under the two channels, the true coating damage characteristics are screened out, and noise or false detection information is eliminated. Furthermore, the target imaging information screened out by the dual-filter mutual verification is combined with the imaging data of parallel detection to perform mutual verification again to ensure that the identified coating damage information can be stably presented at different detection angles, thereby further improving the accuracy and reliability of the detection. Based on the final target imaging, three-dimensional space stacking reconstruction is performed. Specifically, based on the imaging data on different focal planes during the scanning process, stacking processing is performed according to the coating space model to reconstruct the three-dimensional distribution structure of the coating damage.

[0038] In addition, in the coating damage distribution information, a priori identification based on damage characteristics and contaminant characteristics is preset to achieve automatic classification of damage categories. By comparing the differences in optical imaging characteristics between damage and contaminants, such as scattering patterns, morphological characteristics, spectral response, etc., it is possible to effectively distinguish coating damage (such as cracks, peeling, wear) from contaminant attachment (such as particle deposition, chemical contamination, etc.), ensuring the accuracy of the test results.

[0039] Furthermore, the imaging information is subjected to dual-channel imaging and mutual verification based on high-pass filtering and low-pass filtering, including:

[0040] For the imaging information of the first parallel detection group, combined with the dual-filter detection module, dark background bright damage imaging under high-pass filter imaging processing and bright background dark damage imaging under low-pass filter imaging processing are performed to determine the first imaging and the second imaging; the first imaging and the second imaging are mapped and cross-verified to determine the first target imaging; wherein, if the cross-verification is successful, the target imaging is the first imaging or the second imaging.

[0041] First, for the imaging information obtained by the first parallel detection group, combined with the dual-filter detection module, high-pass filter imaging processing and low-pass filter imaging processing are performed respectively. In the high-pass filter imaging processing, the high-frequency features of the coating damage area are mainly extracted, including microcracks, small peeling, surface microscopic defects, etc., so that the damaged area appears as a highlight area in the imaging, while the background part is weakened, thereby forming imaging information of dark background and bright damage, that is, the first imaging. In the low-pass filter imaging processing, the high-frequency noise is mainly suppressed, while the smoothness of the coating background area is enhanced, making larger-scale damage (such as wide-area peeling, contaminant attachment) more significant. This imaging method makes the damaged area appear as a low-brightness signal, while the background area maintains a higher brightness, thereby forming imaging information of bright background and dark damage, that is, the second imaging.

[0042] Subsequently, the first image and the second image are mapped and the mutual verification process is performed. By comparing the consistency of the damage information under the two imaging channels, the real coating damage characteristics are screened out, and the false detection information that may be caused by noise or optical errors is excluded. Specifically, the mutual verification includes: calculating the matching degree of the two imaging results in the damaged area, that is, the damage information that can be displayed under both high-pass and low-pass filtering is considered to be credible damage information; using morphological analysis or pixel-level matching to ensure that the damage in the same area has obvious characteristics under both imaging modes; based on the mutual verification results, if the two imaging results match in the damage characteristics, the mutual verification is considered to be successful, and the first target imaging is determined. If the mutual verification is successful, the target imaging is finally determined to be the first imaging or the second imaging to ensure the authenticity of the damage information and the imaging stability. Through this dual-channel imaging and mutual verification mechanism, the detection accuracy of coating damage can be effectively improved, the false detection rate can be reduced, and high-quality input data can be provided for subsequent coating damage analysis and three-dimensional reconstruction.

[0043] Furthermore, the second target imaging of the second parallel detection group is determined, and the first target imaging and the second target imaging are cross-verified to determine the target imaging.

[0044] In the second parallel detection group, the acquired imaging information is first processed to generate the second target imaging. Similar to the first parallel detection group, the second parallel detection group also uses high-pass and low-pass filtering to extract information of different frequencies. Specifically, the imaging process of the second parallel detection group is also processed based on high-pass filtering and low-pass filtering to generate the second imaging (bright background dark damage or dark background bright damage) respectively. In this step, the imaging information of the second parallel detection group is processed in a consistent manner with the imaging information of the first parallel detection group to ensure the consistency and integrity of the detection results. Next, the first target imaging and the second target imaging are mutually verified. The purpose of this step is to verify the accuracy and consistency of the first target imaging and the second target imaging in the same or overlapping area. By comparing the results of the two target imaging, it can be determined which areas of damage show similar characteristics under two different viewing angles, thereby verifying the reliability of the imaging. Through the mutual verification process, if the first target imaging and the second target imaging are highly consistent in the damage area, it is considered that the mutual verification is successful, and the final target imaging is determined.

[0045] Furthermore, the target imaging is two-dimensional, including a two-dimensional focal plane and a coating surface; for the target imaging, a three-dimensional space stacking reconstruction based on the coating space is performed to determine the coating damage distribution.

[0046] After double filtering and mutual verification processing, the target imaging obtained is two-dimensional imaging data, which contains the imaging information of the coating surface and different focal planes.

[0047] In order to further improve the accurate identification and positioning of coating damage, a three-dimensional space stacking reconstruction based on the coating space is performed for the obtained two-dimensional target imaging. Specifically, a coating space model is established based on the thickness information of the titanium rod coating and the three-dimensional geometric shape of the coating. The model can describe the geometric distribution of the coating in space, ensuring that the subsequent stacking reconstruction process can accurately reflect the actual structure of the coating. Each two-dimensional image in the target imaging represents the imaging information of the coating on different focal planes. By scanning the data on different focal planes, the collected imaging information contains the damage information of different depth levels of the coating, where each focal plane corresponds to a specific depth or position of the coating. The two-dimensional imaging information from different focal planes is aligned, merged and stacked according to the coating space model. Through the three-dimensional reconstruction algorithm, the imaging information of each level is reconstructed into a complete three-dimensional coating damage distribution model. The three-dimensional model can show the damage distribution inside and outside the coating, such as surface cracks, sub-surface voids, peeling, etc.

[0048] Further, after determining the coating damage distribution, it includes:

[0049] The damage detection records of titanium rod coating are retrieved, and adversarial training is performed with the goal of determining damage characteristics and contaminant characteristics to determine a damage verification branch. The damage verification branch is used to perform a binary classification verification on the coating damage distribution to divide coating damage and coating contaminants. Special identifiers are introduced to classify and mark coating damage and coating contaminants within the coating damage distribution.

[0050] After determining the distribution of coating damage, the historical damage detection records of titanium rod coating are retrieved. These records include the specific type, location and severity of coating damage, and also contain known contaminant characteristics. Based on the historical damage detection records, combined with the existing damage characteristics and contaminant characteristics database, a damage and contaminant labeled sample set is constructed as training data. Based on the training data, a generative adversarial network (GAN) structure is constructed, and adversarial training is performed. The GAN structure contains two sub-networks, the generator and the discriminator. The generator is used to generate pseudo-damage images or contaminant images to improve the generalization ability of the model for complex or edge features; the discriminator is used to determine whether the input image is a real sample or a generated sample, and at the same time determine whether the image is coating damage or contaminants. During the training process, in order to enhance the recognition accuracy and robustness of the model in actual detection, the joint loss function is used to optimize the model parameters. The optimization of model parameters uses the backpropagation algorithm combined with the Adam optimizer, and dynamically adjusts the learning rate to improve the convergence efficiency. During the training iteration process, the classification accuracy and damage recognition ability on the validation set are continuously evaluated, and the training strategy is dynamically adjusted based on the indicator feedback until the model converges stably. After the training is completed, a set of model branches with fixed weight parameters are output as damage verification branches, which are used to distinguish the distribution of coating damage in real time during the detection process, distinguish coating damage from contaminants, and provide accurate basis for subsequent marking and quality assessment.

[0051] Joint loss function: L total =λ 1 ·L adv +λ 2 ·L cls +λ 3 ·L rec +λ 4 ·L per , where L adv To counter the loss, it is used to optimize the game ability between the generator and the discriminator and improve the authenticity of the generated image; L cls is the classification loss, which is used to optimize the discriminator’s accuracy in classifying coating damage and contaminants; L rec L is the reconstruction loss, which is used to maintain structural consistency in the encoding-decoding structure; perTo perceive the loss, a pre-trained convolutional neural network is used to extract high-level semantic features of the image, thus improving the model’s ability to identify subtle damage; 1 ~λ 4 Divided into corresponding loss weights, which are set according to experimental experience.

[0052] After obtaining the damage verification branch, the model is used to perform a binary verification on the generated coating damage distribution, that is, the areas in the coating damage distribution are classified according to the characteristics of the damage and the characteristics of the contaminants, and the coating damage and coating contaminants are distinguished. Specifically, damage may appear as cracks, peeling, pits, etc., while contaminants may appear as particle attachment, stains, chemical deposition, etc. Through verification, the two can be accurately distinguished and marked separately in the damage distribution map.

[0053] For the classification results of coating damage and contaminants, special identifiers are introduced for further marking and classification. Each damaged area and contaminated area will be assigned a unique identifier based on its type, location, size and other characteristics. For example, the damaged area may be marked with a specific color, shape or symbol, while the contaminant area uses another color or shape identifier. Through this marking, the distribution of different types of damage and contaminants in the coating can be intuitively displayed, providing a clear basis for subsequent repair decisions, quality assessment and risk management.

[0054] Further, after determining the coating damage distribution, it includes:

[0055] The first-order weighting is performed based on the damage type, and the second-order weighting is performed based on the damage level. The coating damage distribution is traversed to perform a global assessment of the titanium rod damage and determine the damage coefficient. The damage coefficient is judged to be qualified. If the qualified threshold is not met, a titanium rod coating quality warning is issued.

[0056] After determining the coating damage distribution, each type of damage is assigned a first-order weight according to different damage types. Damage types include surface damage, subsurface damage, cracks, spalling, corrosion, etc. Each type of damage has a different impact weight. By assigning a corresponding weight value to each damage type, the degree of influence of the damage type on the overall titanium rod coating performance is obtained.

[0057] Next, a second-order weighting is performed based on the damage level (i.e., the severity of the damage). The damage level is usually assessed based on factors such as the size, depth, and affected area of ​​the damage. For example, a deep crack or extensive spalling damage level is higher, whereas a minor surface scratch or contaminant may have a lower damage level. The level of each damage point is assigned a corresponding weight value to reflect the impact of the damage on the overall performance of the titanium rod.

[0058] By traversing the damage distribution of the entire coating, each type of damage is weighted by the first and second order to obtain the damage coefficient. The damage coefficient is a comprehensive indicator that represents the degree of influence of coating damage on the overall performance of the titanium rod. It is usually the weighted sum of the weighted types and levels of each damage. The size of the damage coefficient can reflect the overall damage degree of the titanium rod coating.

[0059] After obtaining the damage coefficient, the next step is to make a qualified judgment. Set a qualified threshold, which represents the maximum allowable damage degree of the titanium rod coating in quality control. If the calculated damage coefficient is greater than the threshold, it means that the damage to the coating has exceeded the predetermined allowable range, and the overall performance of the titanium rod may be affected and cannot meet the use requirements. If the damage coefficient does not meet the qualified threshold, the titanium rod coating quality warning is triggered. This warning signal reminds relevant personnel that the titanium rod coating has significant damage or potential failure risks, and further inspection, repair or replacement is required.

[0060] In summary, the embodiments of the present application have at least the following technical effects:

[0061] First, the processing information of the titanium rod coating is obtained, and the two-dimensional focal plane deployment and over-focus trajectory planning based on the processing damage risk are carried out to determine the optical scanning scheme, where the processing damage risk includes surface damage and sub-surface damage. Then, the optical detection equipment is connected to the precision translation stage, and according to the optical scanning scheme, the titanium rod is incident scanned and the CCD target surface imaging based on scattered light is performed to determine the imaging information, which includes at least two sets of parallel detection. Finally, a dual-filter detection module is developed in the imaging system, and the imaging information is subjected to dual-channel imaging and mutual verification based on high-pass filtering and low-pass filtering, the target imaging is determined, and three-dimensional space stacking reconstruction is performed to determine the coating damage distribution, where the coating damage distribution has a priori identification based on damage and contaminants, and the target imaging is effective information under dual-filter mutual verification and parallel detection mutual verification. The technical problem of low detection accuracy of titanium rod coating damage and difficulty in identifying subtle losses in the prior art is solved, and the technical effect of improving detection accuracy and realizing visualization of coating damage is achieved.

[0062] Embodiment 2, based on the same inventive concept as the titanium rod coating damage visualization detection method based on optical imaging in the previous embodiment, Figure 2 As shown, the present application provides a titanium rod coating damage visualization detection device based on optical imaging, wherein the device comprises:

[0063] The scanning scheme determination module 11 is used to obtain the processing information of the titanium rod coating, perform two-dimensional focal plane deployment and over-focus trajectory planning based on the processing damage risk, and determine the optical scanning scheme, wherein the processing damage risk includes surface damage and sub-surface damage; the imaging information acquisition module 12 is used to connect the optical detection equipment and the precision translation stage, and perform incident scanning and CCD target imaging based on scattered light on the titanium rod according to the optical scanning scheme to determine the imaging information, wherein at least two sets of parallel detection are included; the loss determination module 13 is used to develop a dual-filter detection module in the imaging system, perform dual-channel imaging and mutual verification based on high-pass filtering and low-pass filtering on the imaging information, determine the target imaging and perform three-dimensional space stacking reconstruction, and determine the coating damage distribution, wherein the coating damage distribution has a priori identification based on damage and contaminants, and the target imaging is effective information under dual-filter mutual verification and parallel detection mutual verification.

[0064] Furthermore, the scanning scheme determination module 11 is used to perform the following method:

[0065] Based on the processing information, damage risk points are located, wherein the damage risk points include surface damage risk points and sub-surface damage risk points; the two-dimensional space is determined by the coating surface, and the third spatial dimension is determined by the coating thickness to construct the coating space; based on the damage risk points and the coating space, two-dimensional focal plane deployment and over-focus trajectory planning are performed.

[0066] Furthermore, the scanning scheme determination module 11 is used to perform the following method:

[0067] For the sub-surface damage risk point, the layer depth is located based on the coating thickness to determine the two-dimensional focal plane; the sub-surface damage risk point is traversed to determine the key focus at the position located at the corresponding two-dimensional focal plane; the focus trajectory is planned with the key focus as the necessary position of the scanning trajectory to determine the sub-surface scanning plan; the surface scanning plan and the sub-surface scanning plan are integrated to determine the optical scanning plan, wherein the surface damage plan is a full-area scan with a focus on the surface damage risk point.

[0068] Furthermore, the loss determination module 13 is used to perform the following method:

[0069] For the imaging information of the first parallel detection group, combined with the dual-filter detection module, dark background bright damage imaging under high-pass filter imaging processing and bright background dark damage imaging under low-pass filter imaging processing are performed to determine the first imaging and the second imaging; the first imaging and the second imaging are mapped and cross-verified to determine the first target imaging; wherein, if the cross-verification is successful, the target imaging is the first imaging or the second imaging.

[0070] Furthermore, the loss determination module 13 is used to perform the following method:

[0071] Determine a second target imaging of a second parallel detection group, perform mutual verification on the first target imaging and the second target imaging, and determine the target imaging.

[0072] Furthermore, the loss determination module 13 is used to perform the following method:

[0073] The target imaging is two-dimensional, including a two-dimensional focal plane and a coating surface; for the target imaging, a three-dimensional space stacking reconstruction based on the coating space is performed to determine the coating damage distribution.

[0074] Furthermore, the loss determination module 13 is used to perform the following method:

[0075] The damage detection records of titanium rod coating are retrieved, and adversarial training is performed with the goal of determining damage characteristics and contaminant characteristics to determine a damage verification branch. The damage verification branch is used to perform a binary classification verification on the coating damage distribution to divide coating damage and coating contaminants. Special identifiers are introduced to classify and mark coating damage and coating contaminants within the coating damage distribution.

[0076] Furthermore, the imaging information acquisition module 12 is used to perform the following method:

[0077] By coordinating the optical detection device and the precision displacement stage, a first detection viewing angle and a second detection viewing angle are determined, wherein the titanium rod is placed on the precision displacement stage; and according to the first detection viewing angle and the second detection viewing angle, parallel detection based on the optical scanning scheme is performed.

[0078] Furthermore, the loss determination module 13 is used to perform the following method:

[0079] The first-order weighting is performed based on the damage type, and the second-order weighting is performed based on the damage level. The coating damage distribution is traversed to perform a global assessment of the titanium rod damage and determine the damage coefficient. The damage coefficient is judged to be qualified. If the qualified threshold is not met, a titanium rod coating quality warning is issued.

[0080] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0081] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0082] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A visual detection method for titanium rod coating damage based on optical imaging, characterized in that: The method comprises: Obtain the processing information of the titanium rod coating, perform two-dimensional focal plane deployment and over-focus trajectory planning based on the processing damage risk, and determine the optical scanning plan. The processing damage risk includes surface damage and sub-surface damage. Connecting the optical detection device to the precision translation stage, and performing incident scanning on the titanium rod and CCD target surface imaging based on scattered light according to the optical scanning scheme to determine imaging information, wherein at least two sets of parallel detection are included; A dual-filter detection module is developed in the imaging system, and the imaging information is subjected to dual-channel imaging and mutual verification based on high-pass filtering and low-pass filtering, the target imaging is determined, and three-dimensional spatial stacking reconstruction is performed to determine the coating damage distribution, wherein the coating damage distribution has a priori identification based on damage and contaminants, and the target imaging is effective information under dual-filter mutual verification and parallel detection mutual verification.

2. The titanium rod coating damage visualization detection method based on optical imaging according to claim 1, characterized in that: Perform two-dimensional focal plane deployment and over-focus trajectory planning, including: Locating damage risk points according to the processing information, wherein the damage risk points include surface damage risk points and sub-surface damage risk points; The coating space is constructed by determining the two-dimensional space with the coating surface and determining the third spatial dimension with the coating thickness; Based on the damage risk points and the coating space, two-dimensional focal plane deployment and over-focus trajectory planning are performed.

3. The titanium rod coating damage visualization detection method based on optical imaging as claimed in claim 2, characterized in that: Based on the damage risk point and the coating space, two-dimensional focal plane deployment and over-focus trajectory planning are performed, including: For the sub-surface damage risk point, the layer depth is located based on the coating thickness to determine the two-dimensional focal plane; Traversing the sub-surface damage risk points to determine the critical focus at the positions located in the corresponding two-dimensional focal planes; Taking the key focus as the necessary position of the scanning trajectory, performing focus trajectory planning, and determining the sub-surface scanning plan; The surface scanning scheme and the sub-surface scanning scheme are integrated to determine the optical scanning scheme, wherein the surface damage scheme is a full-area scan focusing on the surface damage risk points.

4. The titanium rod coating damage visualization detection method based on optical imaging according to claim 1, characterized in that: Performing dual-channel imaging and mutual verification based on high-pass filtering and low-pass filtering on the imaging information includes: Based on the imaging information of the first parallel detection group, the dual-filter detection module is combined to perform dark background bright damage imaging under high-pass filter imaging processing and bright background dark damage imaging under low-pass filter imaging processing to determine the first imaging and the second imaging; Mapping the first imaging and the second imaging, performing mutual verification, and determining a first target imaging; If the mutual verification is successful, the target imaging is the first imaging or the second imaging.

5. The titanium rod coating damage visualization detection method based on optical imaging as claimed in claim 4, characterized in that: Determine a second target imaging of a second parallel detection group, perform mutual verification on the first target imaging and the second target imaging, and determine the target imaging.

6. The titanium rod coating damage visualization detection method based on optical imaging as claimed in claim 5, characterized in that: The target imaging is two-dimensional, including a two-dimensional focal plane and a coating surface; For the target imaging, three-dimensional space stacking reconstruction based on the coating space is performed to determine the coating damage distribution.

7. The titanium rod coating damage visualization detection method based on optical imaging according to claim 6, characterized in that: After determining the coating damage distribution, including: Retrieve the damage detection records of titanium rod coating, conduct adversarial training with the goal of determining damage characteristics and contaminant characteristics, and determine the damage verification branch; The damage verification branch performs a binary verification on the coating damage distribution to divide coating damage and coating contaminants; Special identifiers are introduced to classify and mark the coating damage and coating contaminants within the coating damage distribution.

8. The titanium rod coating damage visualization detection method based on optical imaging according to claim 1, characterized in that: At least two parallel tests are included, including: Determine a first detection viewing angle and a second detection viewing angle by coordinating the optical detection device and the precision displacement stage, wherein the titanium rod is placed on the precision displacement stage; According to the first detection viewing angle and the second detection viewing angle, parallel detection based on the optical scanning scheme is performed.

9. The titanium rod coating damage visualization detection method based on optical imaging according to claim 1, characterized in that: After determining the coating damage distribution, including: Perform first-order weighting based on damage type, perform second-order weighting based on damage level, traverse the coating damage distribution to perform global assessment of titanium rod damage, and determine the damage coefficient; The damage coefficient is judged to be qualified. If the qualified threshold is not met, a quality warning of the titanium rod coating is issued.

10. A visual detection device for titanium rod coating damage based on optical imaging, characterized in that: The device is used to implement the titanium rod coating damage visualization detection method based on optical imaging according to any one of claims 1 to 9, and comprises: The scanning scheme determination module is used to obtain the processing information of the titanium rod coating, perform two-dimensional focal plane deployment and over-focus trajectory planning based on the processing damage risk, and determine the optical scanning scheme, where the processing damage risk includes surface damage and sub-surface damage; An imaging information acquisition module is used to connect the optical detection equipment and the precision translation stage, and to perform incident scanning on the titanium rod and CCD target surface imaging based on scattered light according to the optical scanning scheme to determine imaging information, wherein at least two sets of parallel detection are included; The loss determination module is used to develop a dual-filter detection module in the imaging system, perform dual-channel imaging and mutual verification based on high-pass filtering and low-pass filtering on the imaging information, determine the target imaging and perform three-dimensional space stacking reconstruction, and determine the coating damage distribution, wherein the coating damage distribution has a priori identification based on damage and contaminants, and the target imaging is effective information under dual-filter mutual verification and parallel detection mutual verification.

Images