Thread detection method, system and equipment for high-temperature alloy fasteners

Abstract

The present application relates to a thread detection method, system and equipment for high-temperature alloy fasteners. First, images of the threads to be detected are collected from multiple observation angles to construct a first image set, and the first image set is aligned to form a second image set; then, a phase difference map is constructed based on the second image set, and the thread disturbance features are extracted in combination with a direction-sensitive filter group to generate a disturbance-enhanced image; finally, by analyzing the disturbance-enhanced image, it is identified whether there are micro-damages on the thread surface, and the identification results are output. Therefore, it is possible to achieve high-precision automatic identification of tiny surface defects without destroying the thread structure, solving the technical problems of poor stability, low efficiency and weak micro-damage identification ability of traditional visual inspection, and effectively improving the reliability and intelligence level of thread detection of high-temperature alloy fasteners.

Description

Technical Field

[0001] The present application relates to the technical field of image processing and quality inspection, and in particular to a thread inspection method, system and equipment for high-temperature alloy fasteners. Background Art

[0002] As high-end manufacturing sectors like aerospace and energy equipment continue to demand higher performance from components, high-temperature alloys are widely used in fastener manufacturing due to their superior high-temperature strength, oxidation resistance, and corrosion resistance. As critical connection and load-bearing components, the thread quality of these fasteners is directly related to the structural connection reliability and operational safety.

[0003] Currently, thread inspection primarily relies on manual visual inspection, performed with the aid of a magnifying glass. This traditional approach has significant shortcomings: First, the inspection process is highly dependent on the inspector's experience and subjective judgment, which can lead to unstable results. Second, it is inefficient, making it difficult to meet the demands of high-volume, high-consistency production. Third, its ability to identify small defects and microscopic damage is limited, leading to the risk of missed and false detections, making it difficult to ensure the comprehensiveness and reliability of fastener thread quality inspections.

[0004] That is, traditional thread detection methods have technical problems such as poor detection result stability, low detection efficiency and insufficient defect recognition accuracy. Summary of the Invention

[0005] Based on this, it is necessary to provide a thread detection method for high-temperature alloy fasteners, a thread detection system for high-temperature alloy fasteners and a computer device to address the above technical problems.

[0006] To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions:

[0007] In one aspect, a thread detection method for a high-temperature alloy fastener is provided, comprising:

[0008] Obtaining a first image set of the thread to be inspected; the thread to be inspected is a thread on a high-temperature alloy fastener, and the first image set is a collection of images obtained by imaging the thread to be inspected from different observation angles;

[0009] Registering the image area of ​​the thread to be detected in the first image set to obtain a second image set;

[0010] Based on the second image set, a phase difference map is constructed; the phase difference map is used to characterize the relative phase difference of the same thread area under different observation angles;

[0011] Based on the second image set and the phase difference image, a direction-sensitive filter bank is used to extract disturbance features of the thread to be detected, and a disturbance-enhanced image is obtained based on the disturbance features;

[0012] Based on the disturbance-enhanced image, the image recognition model identifies whether there is micro-damage on the surface of the thread to be inspected and outputs the recognition result; the image recognition model is constructed based on a convolutional neural network. Micro-damage refers to surface defects that are invisible under visual observation conditions. The types of micro-damage include thermal fatigue microcracks, oxidation pits, surface corrosion and stress corrosion cracks.

[0013] In one embodiment, in the thread detection method for high-temperature alloy fasteners, based on the disturbance-enhanced image, the image recognition model identifies whether there is micro-damage on the surface of the thread to be inspected, and outputs the recognition result, including:

[0014] The perturbation-enhanced image is input into an image recognition model; the image recognition model includes a feature adaptation module, a candidate box generation module, and a position regression module;

[0015] The feature adaptation module performs feature mapping on the disturbance enhanced image and outputs a feature map;

[0016] The candidate box generation module generates candidate detection areas based on the feature map;

[0017] The position regression module performs regression processing on the candidate detection area and outputs the position coordinates of the micro-damage; the recognition result includes the position coordinates.

[0018] In one embodiment, in the above-mentioned thread detection method for high-temperature alloy fasteners, the image recognition model also includes a type discrimination module, which is used to identify the damage type of micro-damage based on the position coordinates and the corresponding candidate detection area, and output the position coordinates and the corresponding damage type; the recognition result includes the position coordinates and the corresponding damage type.

[0019] In one embodiment, in the above-mentioned thread detection method for high-temperature alloy fasteners, in the first image set, each observation angle includes a corresponding visible light image and infrared image;

[0020] Registering the image region of the thread to be detected in the first image set to obtain a second image set includes:

[0021] For each observation angle, the corresponding visible light image and infrared image are fused to obtain a fused image at each observation angle;

[0022] The image regions of the threads to be detected in each fused image are registered to obtain a second image set.

[0023] In one embodiment, the thread detection method for high-temperature alloy fasteners further includes, after outputting the recognition result, visually measuring the micro-damage to obtain geometric size parameters of the micro-damage.

[0024] In one embodiment, in the thread inspection method for high-temperature alloy fasteners, visual measurement of micro-damage to obtain geometric size parameters of the micro-damage includes:

[0025] Locating a corresponding visual detection area in the second image set according to the position coordinates;

[0026] For the visual inspection area, extract the first boundary contour of the micro-damage;

[0027] The geometric dimension parameters of the micro-damage are calculated according to the first boundary contour, the image acquisition resolution and the calibration model; the calibration model is used to characterize the mapping relationship between the pixel coordinates in the image acquisition module and the actual space coordinates.

[0028] In one embodiment, in the thread inspection method for high-temperature alloy fasteners, visual measurement of micro-damage to obtain geometric size parameters of the micro-damage includes:

[0029] According to the position coordinates, the second boundary contour of the micro-damage is extracted from the disturbance-enhanced image;

[0030] An image transformation model is constructed based on the second image set, the phase difference map, and the disturbance-enhanced image to obtain a pixel position transformation correction factor during the image enhancement process; the image transformation model is used to characterize the pixel position change between the disturbance-enhanced image and the second image set;

[0031] The geometric size parameters of the micro-damage are calculated according to the second boundary contour, the image acquisition resolution and the correction factor.

[0032] On the other hand, a thread detection system for high-temperature alloy fasteners is also provided, comprising:

[0033] An image acquisition module is configured to obtain a first image set of the thread to be inspected; the thread to be inspected is a thread on a high-temperature alloy fastener, and the first image set is a collection of images obtained by imaging the thread to be inspected from different observation angles;

[0034] A registration module, configured to register the image region of the to-be-detected thread in the first image set to obtain a second image set;

[0035] A construction module is used to construct a phase difference map based on the second image set; the phase difference map is used to represent the relative phase difference of the same thread area under different observation angles;

[0036] A generation module is used to extract disturbance features of the thread to be detected by using a direction-sensitive filter bank based on the second image set and the phase difference map, and obtain a disturbance-enhanced image based on the disturbance features;

[0037] The image recognition model is used to identify whether there is micro-damage on the surface of the thread to be inspected based on the disturbance-enhanced image and output the recognition result. The image recognition model is constructed based on a convolutional neural network. Micro-damage refers to surface defects that are invisible under visual observation conditions. The types of micro-damage include thermal fatigue microcracks, oxidation pits, surface corrosion and stress corrosion cracks.

[0038] In one embodiment, the thread detection system for high-temperature alloy fasteners further includes: a measurement module for visually measuring micro-damages to obtain geometric size parameters of the micro-damages.

[0039] On the other hand, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned thread detection method for high-temperature alloy fasteners when executing the computer program.

[0040] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0041] The aforementioned thread inspection method, system, and equipment for high-temperature alloy fasteners first capture images of the threads to be inspected from multiple observation angles to construct a first image set, which is then registered to form a second image set. A phase difference map is then constructed based on the second image set, and a direction-sensitive filter bank is used to extract thread perturbation features, generating a perturbation-enhanced image. Finally, the perturbation-enhanced image is analyzed to identify the presence of micro-damage on the thread surface and output the identification results. This allows for high-precision, automated identification of tiny surface defects without damaging the thread structure, resolving the technical issues of traditional visual inspection, which suffer from poor stability, low efficiency, and weak micro-damage recognition capabilities. This effectively enhances the reliability and intelligence of thread inspection for high-temperature alloy fasteners. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0043] Figure 1 Schematic diagram of a process for detecting threads of high-temperature alloy fasteners in one embodiment.

[0044] Figure 2 Schematic diagram of the process of thread micro-damage identification based on disturbance-enhanced images in one embodiment.

[0045] Figure 3 Schematic diagram of the module structure of a thread detection system for high-temperature alloy fasteners in one embodiment Figure 1 .

[0046] Figure 4 Schematic diagram of the module structure of a thread detection system for high-temperature alloy fasteners in one embodiment Figure 2 . DETAILED DESCRIPTION

[0047] Specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the described embodiments are merely some, and not all, of the embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the description of the present invention without inventive effort are intended to fall within the scope of protection of the present invention.

[0048] In the description of the present invention, unless otherwise specified or limited, the terms "disposed," "installed," and "connected" should be understood broadly. For example, they may refer to fixed, detachable, or integral connections; mechanical or electrical connections; and direct or indirect connections through an intermediary. Those skilled in the art will understand the specific meanings of these terms based on the specific circumstances.

[0049] The directions or positional relationships indicated by terms such as "upper", "lower", "left", "right", "front", "back", "top", "bottom", "inside" and "outside" are based on the directions or positional relationships shown in the accompanying drawings, or are the directions or positional relationships in which the inventive product is usually placed when in use. They are only for the convenience and simplification of description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operate in a specific direction. Therefore, they should not be understood as limiting the present invention.

[0050] The terms "first," "second," "third," etc. are merely used to distinguish between elements of similar nature and do not indicate or imply relative importance or a particular order.

[0051] The terms "comprises," "comprising," or any other variations thereof, are intended to cover a non-exclusive inclusion of elements other than the listed elements and may also include additional elements not specifically listed.

[0052] The following describes the embodiments of the present invention in detail with reference to the accompanying drawings.

[0053] In one embodiment, Figure 1As shown, the embodiment of the present application provides a thread detection method for high-temperature alloy fasteners, including the following processing steps S11-S15:

[0054] S11, obtaining a first image set of the thread to be inspected; the thread to be inspected is a thread on a high-temperature alloy fastener, and the first image set is a collection of images obtained by imaging the thread to be inspected from different observation angles.

[0055] It can be understood that the thread structure is a three-dimensional spiral structure and there may be tiny defects on the surface. A multi-angle imaging method (such as rotational shooting or multi-camera arrangement) is used to shoot the thread surface from multiple different perspectives to obtain a first image set.

[0056] In some embodiments, a multi-angle imaging device can be used to capture images of threads on high-temperature alloy fasteners. This device can include a rotation control platform, an industrial camera, and a light source system. First, the high-temperature alloy fastener to be inspected is secured to the rotation control platform. The platform's step-by-step rotation function rotates the threads at preset angle intervals (e.g., every 20 degrees). At each angle, the industrial camera captures the thread surface, producing a clear image.

[0057] Before shooting, the light source can be controlled to perform adaptive adjustments to ensure image contrast and edge clarity. The images obtained at multiple angles (for example, 18 viewing angles) are numbered in order of angle to form a first image set of the thread to be inspected.

[0058] In other embodiments, a multi-camera arrangement is used to capture images of threads on high-temperature alloy fasteners. For example, multiple (e.g., 4-8) industrial cameras can be evenly spaced around the thread to be inspected, with each camera simultaneously capturing images of the thread surface from a different direction. Each camera is positioned at a specific angle relative to the thread axis to cover the visible area of ​​the thread surface.

[0059] During the capture process, all cameras are controlled and synchronized to capture images, minimizing errors caused by workpiece vibration or environmental changes. Furthermore, independent or ring-shaped adjustable light sources can be configured based on camera position to provide sufficient and uniform illumination. Polarizers or diffuse reflectors can also be incorporated to reduce surface reflections. The multiple viewpoint images captured are organized by camera number to form the first image set.

[0060] S12, registering the image area of ​​the thread to be detected in the first image set to obtain a second image set.

[0061] As you can understand, registration refers to aligning two or more images taken from different angles to the same reference coordinate system. In other words, regions at the same physical location in thread images acquired from different perspectives are aligned in image space, allowing them to be compared or fused at the pixel level.

[0062] Specifically, the image registration process mainly includes steps such as geometric pre-registration and feature refinement alignment.

[0063] In some embodiments, image acquisition is performed using a rotational capture method. During the geometric pre-registration phase, a unified 3D observation reference coordinate system is constructed based on the platform's rotation angle (e.g., 20-degree rotations), combined with the known rotation axis orientation and the industrial camera's pose. Based on the rotation angle information for each frame, the pose transformation relationship of the industrial camera relative to the fastener at the time of image acquisition can be inferred, allowing each image to be spatially mapped into a unified reference coordinate system. Specifically, an affine transformation or a projective transformation can be used to initially geometrically map the images, ensuring that the threaded areas at different angles are roughly aligned with a unified reference frame in the image plane.

[0064] In other embodiments, a multi-camera arrangement is used for image acquisition. In the geometric pre-registration stage, the images acquired by each camera are uniformly projected into a common reference coordinate system according to the position and orientation parameters of each camera in three-dimensional space, and a preliminary geometric correspondence between the images is established.

[0065] After completing geometric pre-registration, the feature refinement and alignment phase begins. This phase primarily targets the threaded area, further improving alignment accuracy between different images. Specifically, it automatically identifies recognizable structural features within the image and finds corresponding relationships between these features across multiple viewpoints. A consistency check mechanism is then introduced to eliminate potential mismatches and retain reliable feature pairs. This feature alignment process further improves the consistency of the threaded area in the image across different viewpoints, providing a more accurate image foundation for subsequent thread disturbance analysis, defect identification, and other tasks.

[0066] To further improve registration accuracy and robustness, auxiliary geometric reference objects can be introduced in some implementations. For example, in rotating capture scenarios, reference markers can be placed on the capture platform or background plate to assist in identifying the platform's rotation center and compensating for accumulated rotation errors. In multi-camera capture scenarios, reference markers can also be set as common features between multiple viewpoints to improve the accuracy of relative pose estimation between cameras. This is particularly useful for workpiece surfaces with repetitive textures or weak patterns.

[0067] S13, constructing a phase difference map based on the second image set; the phase difference map is used to characterize the relative phase difference of the same thread area under different observation angles.

[0068] It can be understood that, given that image registration has been completed, a phase difference map is constructed to quantify local phase changes by analyzing the relative changes in grayscale or texture features at the same physical location in images from different observation angles. This phase difference map effectively reflects the response consistency of the thread surface under multi-view imaging, and its phase differences can reveal potential micro-disturbance areas on the thread surface, such as phase shifts caused by micro-damage such as microcracks and pits.

[0069] Specifically, a sequence of registered images covering the same thread region is first extracted from the second image set. Sequential modeling is then performed on the grayscale or texture features at each image location. For example, within a local image window, a sliding window approach is used to obtain the grayscale variation sequence of each pixel across multiple images. Fourier analysis or other phase estimation methods are then used to calculate the relative phase value of that location at different viewing angles.

[0070] It should be understood that, considering that the thread to be inspected is a three-dimensional entity with a 360-degree annular structure, while the phase difference map is a two-dimensional image (a 2D grayscale image or heat map), the registered image can be divided into several viewing areas (for example, four viewing areas). For each viewing area, a phase difference map is constructed to characterize the relative phase changes in that viewing area at different observation angles. Multiple phase difference maps can be combined to represent the phase response distribution of the entire thread surface.

[0071] S14, based on the second image set and the phase difference image, using a direction-sensitive filter bank to extract disturbance features of the thread to be detected, and obtaining a disturbance-enhanced image based on the disturbance features.

[0072] It can be understood that, based on the second image set and the phase difference map, a direction-sensitive filter bank is used to respond to image changes in multiple directions in the threaded region to extract possible disturbance features. A direction-sensitive filter can significantly respond to features in specific directions within the image (such as edges, textures, and disturbances). The direction-sensitive filter bank includes multiple direction-sensitive filters with different directional response characteristics. Disturbance features refer to areas that still exhibit local discontinuities, sudden changes, or asymmetric features after multi-view fusion and phase analysis.

[0073] Specifically, a Gabor filter bank or a directional derivative filter bank can be used to jointly process the second image set and the phase difference map. Within each local area window of the image, the filter responses are applied along multiple directions to generate a directional response map. This directional response map is then fused based on the amplitude, directional consistency, or local perturbation index of the filter responses to produce a disturbance-enhanced image. This disturbance-enhanced image enhances the slightly perturbed regions while preserving the original structural features, providing a more significant input for micro-damage characterization in subsequent image recognition models.

[0074] S15, based on the disturbance-enhanced image, the image recognition model identifies whether there is micro-damage on the surface of the thread to be inspected and outputs the recognition result; the image recognition model is constructed based on a convolutional neural network. Micro-damage refers to surface defects that are invisible under visual observation conditions. The types of micro-damage include thermal fatigue microcracks, oxidation pits, surface corrosion and stress corrosion cracks.

[0075] It can be understood that after obtaining the disturbance-enhanced image, the image recognition model based on the convolutional neural network is used to automatically identify and distinguish the thread surface in the disturbance-enhanced image to determine whether there is micro-damage.

[0076] Microdamage refers to surface defects that are invisible to visual inspection, but they can potentially impact the service performance of fasteners. High-temperature alloy fasteners are often subjected to harsh service conditions such as high temperatures, high stresses, and severe corrosion. Surface integrity is crucial to the long-term reliability of these fasteners. The threaded area, as a stress concentration point, is more susceptible to early damage, and small surface disturbances have a high potential to develop into fatigue cracks or corrosion sources.

[0077] The types of micro-damage on the thread surface of high-temperature alloy fasteners mainly include thermal fatigue micro-cracks, oxidation pits, surface corrosion and stress corrosion cracks.

[0078] For high-temperature alloy fasteners, the manufacturing process goes through warm heading, grinding and thread rolling, solution treatment, aging heat treatment and rolling process. During warm heading and grinding and thread rolling, the material is subjected to high temperature and high strain rate conditions, which can easily introduce residual tensile stress and micro-plastic deformation in stress concentration areas such as the root of the thread. If the local deformation is uneven or the surface quality control is not strict, thermal fatigue microcracks or surface corrosion may be induced. Solution treatment and aging heat treatment involve high temperature and long-term treatment processes, and high-temperature oxidation reactions are active. If surface protection measures are not in place, local oxidation pits are easily formed on the thread surface, becoming the starting source of later crack propagation. In addition, if the residual stress is too large or there is residual corrosive medium on the surface during the manufacturing process, stress corrosion cracking is likely to occur.

[0079] Specifically, the constructed image recognition model can be used to detect micro-damage on the threaded surfaces of high-temperature alloy fasteners based on a disturbance-enhanced image of the threaded surface to be inspected. The image recognition model uses a convolutional neural network architecture and can include a module for extracting image features and a module for outputting recognition results.

[0080] The image recognition model can adopt different structural configurations: one approach includes only an image feature extraction module and a judgment module, which outputs a result to determine whether microdamage is present on the thread surface. Another approach is for the image recognition model to further include a position regression module, which outputs the location coordinates upon identifying the presence of microdamage. Preferably, the image recognition model also integrates a type discrimination module to identify the type of detected microdamage. The output recognition results include the location coordinates and damage type, thereby enabling accurate identification and classification of thermal fatigue microcracks, oxidation pits, surface corrosion, and stress corrosion cracks.

[0081] In the above-mentioned thread detection method for high-temperature alloy fasteners, by acquiring thread images from multiple observation angles and aligning them, a phase difference map is constructed, which effectively enhances the characterization capability of micro-damages on complex curved surfaces and improves the stability of the detection results. Combining a direction-sensitive filter group to extract disturbance features and generate disturbance-enhanced images makes small disturbances easier to identify, significantly improving the accuracy of defect recognition. Finally, an image recognition model based on a convolutional neural network is used to automatically identify the disturbance-enhanced image, which has good generalization performance and efficient processing capabilities. In summary, the key technical problems of low detection efficiency, poor stability of detection results, and insufficient micro-damage recognition accuracy in the existing technology are solved, and it is suitable for the high-reliability detection needs of micro-damages on the surface of high-temperature alloy fasteners.

[0082] In some embodiments, the thread detection method for high temperature alloy fasteners is as follows: Figure 2 As shown, step S15: based on the disturbance enhanced image, the image recognition model identifies whether there is micro-damage on the thread surface to be inspected, and outputs the recognition result, which specifically includes:

[0083] S151, input the disturbance-enhanced image into an image recognition model; the image recognition model includes a feature adaptation module, a candidate box generation module and a position regression module.

[0084] It is understood that the image recognition model can be trained offline on a perturbation-enhanced image dataset containing a large number of thread micro-damage samples to enhance its sensitivity to micro-damage and recognition accuracy. The overall image recognition model adopts a multi-stage structure. The feature adaptation module can be constructed based on a convolutional neural network to perform feature mapping on the perturbation-enhanced image. The candidate box generation module can integrate a region proposal network to generate candidate micro-damage detection areas. The position regression module can use a lightweight fully connected layer or a small convolutional sub-network to accurately regress the position of the candidate detection areas.

[0085] S152: The feature adaptation module performs feature mapping processing on the disturbance enhanced image and outputs a feature map.

[0086] Understandably, since the input to the image recognition model is a perturbation-enhanced image with more pronounced surface detail, there's no need to perform deep semantic feature extraction. The feature adaptation module primarily performs spatial mapping of the effective information in the perturbation-enhanced image while preserving the image's structural features for use by subsequent modules. The feature adaptation module can be built on shallow network structures such as ResNet (residual network) or MobileNet (lightweight convolutional neural network), retaining only the first few convolutional and normalization layers to achieve effective feature mapping of the perturbation-enhanced image.

[0087] S153, the candidate frame generation module generates a candidate detection area according to the feature map.

[0088] As you can understand, the candidate box generation module is used to propose a series of regions on the feature map that may contain microdamages, called candidate detection regions. Specifically, the candidate box generation module can integrate a region proposal network (RPN) to generate multiple anchor boxes by sliding them at different locations in the feature map. Based on these anchor boxes, it predicts the probability of the presence of microdamages and the rough bounding box coordinates. Subsequently, by sorting the prediction scores and combining them with a non-maximum suppression (NMS) strategy, the highest-scoring candidate boxes are retained from overlapping boxes, and redundant boxes are suppressed. This results in a set of candidate detection regions with high confidence for subsequent positional fine-grained regression processing.

[0089] S154, the position regression module performs regression processing on the candidate detection area and outputs the position coordinates of the micro-damage; the recognition result includes the position coordinates.

[0090] It can be understood that the position regression module is used to fine-tune the positioning of the candidate detection areas selected in the previous stage. Specifically, the position regression module can perform coordinate regression on each candidate detection area based on a lightweight fully connected layer or a small convolutional subnetwork, and output more accurate bounding box coordinates, that is, the position coordinates of the micro-damage. It should be understood that the candidate box generation module outputs a rough target position, while the position regression module further calibrates these positions and outputs the final recognition result. In addition, the position regression module can also learn the prediction function of coordinate offset based on the damage annotation information during the training process, thereby improving the positioning accuracy of micro-damage recognition.

[0091] If a candidate detection area is judged to have no valid micro-damage after regression during the detection process (for example, the classification confidence is low, or the regression result does not have obvious damage characteristics), then the candidate detection area will be excluded and will not be included in the final recognition result.

[0092] The thread inspection method for high-temperature alloy fasteners described above significantly improves the accuracy of micro-damage detection on the threaded surfaces of high-temperature alloy fasteners by combining perturbation-enhanced images with a multi-stage image recognition model. The feature adaptation module, based on a lightweight convolutional structure, efficiently extracts significant texture details from perturbation-enhanced images. The candidate box generation module selects high-confidence candidate detection areas. The position regression module further fine-tunes the detection box position and learns to predict offsets based on damage annotation information, resulting in more accurate micro-damage location coordinates.

[0093] In some embodiments, in the above-mentioned thread detection method for high-temperature alloy fasteners, the image recognition model also includes a type discrimination module, which is used to identify the damage type of micro-damage based on the position coordinates and the corresponding candidate detection area, and output the position coordinates and the corresponding damage type; the recognition result includes the position coordinates and the corresponding damage type.

[0094] It is understood that the type discrimination module can be constructed using a lightweight classification neural network or a convolutional neural network. By performing offline training on an image dataset containing multiple micro-damage types, the image recognition model is able to distinguish between multiple damage types, such as thermal fatigue microcracks, oxidation pits, surface corrosion, and stress corrosion cracks. The type discrimination module receives the position coordinates output by the position regression module and, based on these position coordinates, extracts the image subregion corresponding to the candidate detection area, identifies the damage type of this image subregion, and ultimately outputs the position coordinates and corresponding damage type of each micro-damage.

[0095] The thread inspection method for high-temperature alloy fasteners described above incorporates a type discrimination module into the image recognition model, enabling precise identification of micro-damage on the thread surface of high-temperature alloy fasteners. This method not only pinpoints the location of micro-damage but also identifies the specific damage type. This significantly improves the integrity and practicality of the inspection results, making them particularly useful for damage analysis, lifespan assessment, and maintenance decision-making.

[0096] In some embodiments, in the thread inspection method for high-temperature alloy fasteners, each observation angle in the first image set includes a corresponding visible light image and infrared image. Registering the image regions of the threads to be inspected in the first image set to obtain the second image set includes: fusing the corresponding visible light image and infrared image for each observation angle to obtain a fused image at each observation angle; and registering the image regions of the threads to be inspected in each fused image to obtain the second image set.

[0097] It is understood that each observation angle in the first image set includes a visible light image and an infrared image. To improve the availability and information integrity of the detection image, synchronized visible light and infrared images are acquired for each observation angle and image fusion processing is performed on the two to obtain a fused image at each observation angle.

[0098] Visible light images provide clear information about surface structure and texture, while infrared images reveal uneven heat conduction caused by micro-damage by displaying thermal radiation distribution and temperature differences. Infrared images provide crucial information, particularly for deep cracks and micro-cracks that are difficult to detect with the naked eye. By fusing these two images, the image recognition model can more comprehensively assess micro-damage on the thread surface, significantly improving its accuracy.

[0099] In the actual acquisition process, a dual-camera system can be used to synchronously capture visible light images and infrared images. The cameras can be a color camera and an infrared thermal imager respectively. The two achieve consistency of image data under the same observation angle through perspective correction and synchronous triggering mechanism.

[0100] For each observation angle, the corresponding visible light image and infrared image are fused to obtain a fused image at each observation angle. Specifically, the visible light image and infrared image can be registered first, and then an image fusion algorithm can be used to integrate the feature information of the two images.

[0101] For example, using the weighted averaging method, by assigning a weight to each pixel point, the pixel values ​​of the visible light image and the infrared image are weighted and fused according to a certain weight, and finally a fused image is obtained.

[0102] For example, the principal component analysis (PCA) algorithm can be used to extract the key features of the visible light image and infrared image and combine them in a new space to obtain a fused image. Specifically, the visible light image and infrared image are first converted to grayscale images and subjected to principal component analysis to obtain their respective principal component matrices. The principal component matrices of the two images are then combined, and appropriate principal components are selected for synthesis. Finally, the fused principal component image is converted back into a spatial image through an inverse transformation.

[0103] The image regions of the threads to be detected in each fused image are registered to obtain a second image set. Specifically, a fused image with a clear thread structure is first selected as a reference view. The thread region is manually or automatically annotated to serve as a template image. Edge detection combined with Fourier transform is used to extract the periodic features of the thread and determine the axial direction and pitch of the thread. Template image matching is then used to preliminarily locate the thread region in the remaining fused images.

[0104] For each thread region in the fused image, the grayscale of the thread region in the template image and the current image is first normalized to reduce the impact of brightness differences on registration accuracy. Subsequently, an iterative optimization method based on image gradients is used to estimate the affine transformation parameters, including geometric relationships such as rotation, scaling, translation, and shearing. During this optimization process, a residual weighting mechanism can be introduced to weightedly suppress or eliminate pixels with large errors, thereby improving the robustness and accuracy of the registration process. Ultimately, the affine transformation matrix of the thread region in the current image relative to the template image is obtained.

[0105] After obtaining the affine transformation matrix, bicubic interpolation is used to transform and align the current thread image region to the coordinate system of the template image, ensuring pixel-level continuity and smoothness. After registration, the entire transformed thread image region is retained. The thread images of all registered fused images are then assembled into a second image set, ordered by observation angle. This second image set has a highly consistent spatial structure, significantly improving the training and detection accuracy of the image recognition model.

[0106] The thread inspection method for high-temperature alloy fasteners described above fully integrates surface texture and thermal distribution information by fusing visible light and infrared images, improving the detectability and accuracy of micro-damage identification. Furthermore, through precise registration of the fused images, a second image set with consistent structure is constructed, effectively eliminating spatial deviations between multi-angle images. This improves the training effectiveness and detection stability of the subsequent image recognition model, significantly enhancing the ability to accurately identify micro-damage in high-temperature alloy fastener threads.

[0107] In some embodiments, the above-mentioned thread detection method for high-temperature alloy fasteners, after outputting the recognition result, further includes: visually measuring the micro-damage to obtain geometric size parameters of the micro-damage.

[0108] It is understood that after outputting the identification results, the thread inspection method for high-temperature alloy fasteners further performs visual measurement of the micro-damage area to obtain the geometric parameters of the micro-damage. Geometric parameters can include but are not limited to the length, width, and strike angle of microcracks, or the area and depth of micro-pits and surface erosion areas, which can be used for structural integrity assessment and failure risk analysis.

[0109] The present application provides two methods for visual measurement. The first method involves performing recognition based on the disturbance-enhanced image and performing visual measurement on a second image set. This means that the recognition results based on the disturbance-enhanced image output can be used to locate the corresponding locations of microdamages on the second image set, and visual measurement can then be performed on the second image set. The second method involves performing visual measurement directly on the disturbance-enhanced image and then applying a correction factor to correct for deviations in the measurement results.

[0110] In the aforementioned thread inspection method for high-temperature alloy fasteners, visual measurement, after identifying micro-damage, can obtain the specific geometric parameters of the micro-damage, helping to quantitatively assess the extent of thread damage, thereby providing a more accurate basis for subsequent safety assessments or scrapping decisions. Furthermore, the introduction of visual quantitative measurement can reduce the interference of human subjective factors, making the inspection results more consistent, standardized, and objective.

[0111] In some embodiments, the thread inspection method for high-temperature alloy fasteners described above performs visual measurement of micro-damage to obtain geometrical dimension parameters of the micro-damage, including:

[0112] Based on the position coordinates, the corresponding visual inspection area is located in the second image set; the first boundary outline of the micro-damage is extracted in the visual inspection area; and the geometric dimension parameters of the micro-damage are calculated based on the first boundary outline, the image acquisition resolution, and the calibration model. The calibration model is used to characterize the mapping relationship between the pixel coordinates in the image acquisition module and the actual space coordinates.

[0113] The second image set, understood as a collection of precisely aligned and multi-angled images without perturbation enhancement, exhibits high geometric and pixel accuracy. Using the position coordinates output from the perturbation-enhanced images and pixel position mapping, the micro-damage is located within the second image set and the visual inspection area is extracted. Boundary contours of the visual inspection area are extracted, and the image resolution and calibration model are combined to achieve high-precision measurement of the micro-damage's geometric dimensions.

[0114] The calibration model is used to describe the mapping relationship between the image pixel coordinate system and the actual physical space coordinate system. By photographing a calibration plate with known size features, a conversion relationship between pixels and actual space coordinates can be established, so that the measurement data in the image can accurately reflect the true physical size of the target.

[0115] In one embodiment, the position coordinates are x=410 and y=285, and these position coordinates are used as the positioning basis and mapped to the corresponding image in the second image set. The area near the position coordinates, for example, 100*100 pixels, is extracted as the visual detection area. In the visual detection area, the first boundary outline of the microcrack is extracted using edge detection or image segmentation. If the image acquisition resolution is 8μm / pixel and the crack length is 92 pixels, the corresponding actual length is: .in, px represents pixels, Indicates micrometers.

[0116] In the above-mentioned thread detection method for high-temperature alloy fasteners, visual measurement is performed on a second image set with higher geometric realism based on the coordinate information of micro-damage output by the disturbance-enhanced image, which has the advantages of balancing recognition accuracy and measurement precision. The disturbance-enhanced image improves the recognizability of micro-damage by introducing texture, contrast or structural enhancement, making it easier for the image recognition model to accurately locate the micro-damaged area. The second image set does not introduce disturbance processing, maintaining the original geometric relationship and high-precision pixel information of the image, making it suitable as a basic image for fine measurement. Through spatial mapping across image sets, while ensuring recognition sensitivity, measurement errors caused by deformation caused by image enhancement are avoided, thereby achieving a coordinated unity of high recognition rate and high measurement accuracy.

[0117] In some embodiments, the thread inspection method for high-temperature alloy fasteners described above performs visual measurement of micro-damage to obtain geometrical dimension parameters of the micro-damage, including:

[0118] According to the position coordinates, the second boundary contour of the micro-damage is extracted from the disturbance-enhanced image.

[0119] It can be understood that the image recognition model detects and locates the possible micro-damage based on the input disturbance-enhanced image and then outputs the position coordinates. Figure 1 A local area is detected, and an edge detection algorithm (eg, a Canny edge detection algorithm) is used in the local area to obtain a clear boundary contour of the micro-lesion (ie, the second boundary contour).

[0120] An image transformation model is constructed based on the second image set, the phase difference map and the disturbance-enhanced image to obtain a pixel position transformation correction factor during the image enhancement process; the image transformation model is used to characterize the pixel position change between the disturbance-enhanced image and the second image set.

[0121] It can be understood that since the pixel position of the disturbance-enhanced image is transformed relative to the second image set during the generation process, there will be errors in measuring the geometric dimensions directly based on the disturbance-enhanced image. Therefore, it is necessary to establish an image transformation model to correct the geometric dimension parameters of the second boundary contour obtained directly based on the disturbance-enhanced image.

[0122] Specifically, first, based on the second image set and the disturbance-enhanced image, structural features such as image key points, edge information and / or texture gradients are extracted. Furthermore, features representing the phase intensity and direction distribution in the phase difference map are introduced to enhance the characterization capability of the slightly perturbed area. The above-mentioned various features can be combined to form joint features through feature map splicing, weighted fusion or multi-channel parallel connection. Afterwards, based on the joint features, an affine transformation or non-rigid transformation can be used to establish a pixel position mapping relationship between the disturbance-enhanced image and the second image set, that is, to construct an image transformation model.

[0123] Furthermore, pixel position transformation correction factors, such as pixel displacement vector field, affine transformation matrix or nonlinear deformation function, are extracted from the image transformation model as the correction basis for geometric backtracing of micro-damage contours.

[0124] The geometric size parameters of the micro-damage are calculated according to the second boundary contour, the image acquisition resolution and the correction factor.

[0125] It can be understood that the pixel position transformation correction factor is used to perform geometric correction on the second boundary contour of the micro-lesion in the disturbance-enhanced image. Based on the known image acquisition resolution, the pixel units of the corrected boundary coordinates are converted into physical units to obtain the geometric size parameters of the micro-lesion.

[0126] In some embodiments, the thread inspection method for high-temperature alloy fasteners described above directly obtains a second boundary contour based on the perturbation-enhanced image, significantly improving the recognizability and boundary clarity of microdamages, thereby enhancing the accuracy of boundary extraction. The introduction of a correction factor effectively corrects for the spatial positional deviations introduced by the enhancement process in the second image set, balancing detection sensitivity and measurement accuracy, ultimately improving the reliability of thread microdamage identification and quantitative characterization.

[0127] It should be understood that although Figure 1-Figure 2 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. In addition, Figure 1-Figure 2 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0128] In one embodiment, Figure 3 As shown, a thread detection system 200 for high-temperature alloy fasteners is provided, comprising:

[0129] An image acquisition module 21 is configured to obtain a first image set of the thread to be inspected; the thread to be inspected is a thread on a high-temperature alloy fastener, and the first image set is a collection of images obtained by imaging the thread to be inspected from different observation angles;

[0130] A registration module 22 is used to register the image area of ​​the thread to be detected in the first image set to obtain a second image set;

[0131] A construction module 23 is configured to construct a phase difference map based on the second image set; the phase difference map is used to characterize the relative phase difference of the same thread region under different observation angles;

[0132] A generating module 24 is configured to extract disturbance features of the thread to be detected using a direction-sensitive filter bank based on the second image set and the phase difference map, and obtain a disturbance-enhanced image based on the disturbance features;

[0133] Image recognition model 25 is used to identify whether there is micro-damage on the surface of the thread to be inspected based on the disturbance-enhanced image and output the recognition result; the image recognition model is constructed based on a convolutional neural network. Micro-damage refers to surface defects that are invisible under visual observation conditions. The types of micro-damage include thermal fatigue microcracks, oxidation pits, surface corrosion and stress corrosion cracks.

[0134] In the above-mentioned thread detection system 200 for high-temperature alloy fasteners, by acquiring thread images from multiple observation angles and performing registration, a phase difference map is constructed, which effectively enhances the characterization capability of micro-damages on complex curved surfaces and improves the stability of the detection results. The disturbance features are extracted by combining a direction-sensitive filter group and a disturbance-enhanced image is generated, making small disturbances easier to identify and significantly improving the accuracy of defect recognition. Finally, the disturbance-enhanced image is automatically recognized using an image recognition model based on a convolutional neural network, which has good generalization performance and efficient processing capabilities. In summary, the key technical problems of low detection efficiency, poor stability of detection results, and insufficient micro-damage recognition accuracy in the existing technology are solved, and it is suitable for the high-reliability detection needs of micro-damages on the surface of high-temperature alloy fasteners.

[0135] In some embodiments, the image recognition model 25 of the thread inspection system 200 for high-temperature alloy fasteners described above includes an input module, a feature adaptation module, a candidate box generation module, and a position regression module. The input module is used to input the disturbance-enhanced image into the image recognition model. The feature adaptation module is used to perform feature mapping processing on the disturbance-enhanced image and output a feature map. The candidate box generation module is used to generate candidate inspection areas based on the feature map. The position regression module is used to perform regression processing on the candidate inspection areas and output the location coordinates of the micro-damage; the recognition result includes the location coordinates.

[0136] In some embodiments, the above-mentioned thread detection system 200 for high-temperature alloy fasteners, the image recognition model 25 also includes a type discrimination module, which is used to identify the damage type of micro-damage based on the position coordinates and the corresponding candidate detection area, and output the position coordinates and the corresponding damage type; the recognition result includes the position coordinates and the corresponding damage type.

[0137] In some embodiments, in the thread inspection system 200 for high-temperature alloy fasteners, the first image set captured by the image acquisition module 21 includes a corresponding visible light image and infrared image for each observation angle. The registration module 22 includes a fusion submodule for fusing the corresponding visible light image and infrared image for each observation angle to obtain a fused image at each observation angle; and a registration submodule for registering the image regions of the threads to be inspected in each fused image to obtain a second image set.

[0138] In some embodiments, as Figure 4 As shown, the thread detection system 200 for high-temperature alloy fasteners further includes a measurement module 26 for visually measuring micro-damages to obtain geometric size parameters of the micro-damages.

[0139] In some embodiments, the measurement module 26 of the thread inspection system 200 for high-temperature alloy fasteners includes: a positioning submodule for locating a corresponding visual inspection area in the second image set based on position coordinates; a first extraction submodule for extracting a first boundary contour of microdamages within the visual inspection area; and a first calculation submodule for calculating geometric dimensional parameters of the microdamages based on the first boundary contour, image acquisition resolution, and a calibration model; the calibration model is used to characterize the mapping relationship between pixel coordinates in the image acquisition module and actual spatial coordinates.

[0140] The measurement module 26 may be composed of the submodules in the previous embodiment; in other embodiments, the measurement module 26 may include: a second extraction submodule, used to extract the second boundary contour of the micro-damage in the disturbance-enhanced image according to the position coordinates; a correction submodule, used to construct an image transformation model based on the second image set, the phase difference map and the disturbance-enhanced image, and obtain the pixel position transformation correction factor in the image enhancement process; the image transformation model is used to characterize the pixel position change between the disturbance-enhanced image and the second image set; a second calculation submodule, used to calculate the geometric size parameters of the micro-damage based on the second boundary contour, the image acquisition resolution and the correction factor.

[0141] The specific limitations of the thread detection system for high-temperature alloy fasteners can be found in the limitations of the thread detection method for high-temperature alloy fasteners described above and will not be further elaborated here. Each module in the aforementioned thread detection system for high-temperature alloy fasteners can be implemented in whole or in part through software, hardware, or a combination thereof. Each of these modules can be embedded in or independent of a processor in a computer device in hardware form, or stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0142] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0143] Obtaining a first image set of the thread to be inspected; the thread to be inspected is a thread on a high-temperature alloy fastener, and the first image set is a collection of images obtained by imaging the thread to be inspected from different observation angles;

[0144] Registering the image area of ​​the thread to be detected in the first image set to obtain a second image set;

[0145] Based on the second image set, a phase difference map is constructed; the phase difference map is used to characterize the relative phase difference of the same thread area under different observation angles;

[0146] Based on the second image set and the phase difference image, a direction-sensitive filter bank is used to extract disturbance features of the thread to be detected, and a disturbance-enhanced image is obtained based on the disturbance features;

[0147] Based on the disturbance-enhanced image, the image recognition model identifies whether there is micro-damage on the surface of the thread to be inspected and outputs the recognition result; the image recognition model is constructed based on a convolutional neural network. Micro-damage refers to surface defects that are invisible under visual observation conditions. The types of micro-damage include thermal fatigue microcracks, oxidation pits, surface corrosion and stress corrosion cracks.

[0148] It can be understood that in addition to the memory and processor involved above, the above-mentioned computer device also includes other software and hardware components not listed in this specification. The specific components can be determined according to the model of the specific computer device in different application scenarios. This specification will not list them one by one in detail.

[0149] In one embodiment, when executing the computer program, the processor may further implement the additional steps or sub-steps in each embodiment of the above-mentioned thread detection method for high-temperature alloy fasteners.

[0150] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same and similar parts between the various embodiments can be referenced to each other.

[0151] The foregoing description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be readily conceived by a person skilled in the art within the technical scope disclosed herein are intended to be encompassed within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A thread detection method for high temperature alloy fasteners, characterized in that: include: Obtaining a first image set of a thread to be inspected; the thread to be inspected is a thread on a high-temperature alloy fastener, and the first image set is a collection of images obtained by imaging the thread to be inspected from different observation angles; registering the image region of the to-be-detected thread in the first image set to obtain a second image set; constructing a phase difference map based on the second image set; the phase difference map is used to characterize the relative phase difference of the same thread area under different observation angles; extracting disturbance features of the to-be-detected thread using a direction-sensitive filter bank based on the second image set and the phase difference map, and obtaining a disturbance-enhanced image based on the disturbance features; Based on the disturbance-enhanced image, the image recognition model identifies whether there is micro-damage on the surface of the thread to be inspected and outputs a recognition result; the image recognition model is constructed based on a convolutional neural network, and the micro-damage refers to surface defects that are invisible under visual observation conditions. The types of micro-damage include thermal fatigue microcracks, oxidation pits, surface corrosion and stress corrosion cracks.

2. The thread detection method for high temperature alloy fasteners according to claim 1, characterized in that: Based on the disturbance enhanced image, the image recognition model identifies whether there is micro-damage on the thread surface to be inspected, and outputs a recognition result, including: Inputting the disturbance-enhanced image into the image recognition model; the image recognition model includes a feature adaptation module, a candidate box generation module and a position regression module; The feature adaptation module performs feature mapping processing on the disturbance enhanced image and outputs a feature map; The candidate frame generation module generates a candidate detection area according to the feature map; The position regression module performs regression processing on the candidate detection area and outputs the position coordinates of the micro-lesion; the recognition result includes the position coordinates.

3. The thread detection method for high temperature alloy fasteners according to claim 2, characterized in that: The image recognition model also includes a type discrimination module, which is used to identify the damage type of the micro-damage based on the position coordinates and the corresponding candidate detection area, and output the position coordinates and the corresponding damage type; the recognition result includes the position coordinates and the corresponding damage type.

4. The thread detection method for high temperature alloy fasteners according to claim 2, characterized in that: In the first image set, each observation angle includes a corresponding visible light image and infrared image; The registering the image region of the to-be-detected thread in the first image set to obtain the second image set includes: For each observation angle, the corresponding visible light image and the infrared image are fused to obtain a fused image at each observation angle; The image regions of the threads to be detected in the fused images are registered to obtain the second image set.

5. The thread detection method for high temperature alloy fasteners according to any one of claims 2 to 4, characterized in that: After outputting the recognition result, the method further includes: Visual measurement is performed on the micro-damage to obtain geometric size parameters of the micro-damage.

6. The thread detection method for high temperature alloy fasteners according to claim 5, characterized in that: The visual measurement of the micro-damage to obtain geometric size parameters of the micro-damage includes: Locating a corresponding visual detection area in the second image set according to the position coordinates; Extracting a first boundary contour of the micro-lesion in the visual inspection area; The geometric size parameters of the micro-lesion are calculated according to the first boundary contour, image acquisition resolution and a calibration model; the calibration model is used to characterize the mapping relationship between pixel coordinates in an image acquisition module and actual space coordinates.

7. The thread detection method for high temperature alloy fasteners according to claim 5, characterized in that: The visual measurement of the micro-damage to obtain geometric size parameters of the micro-damage includes: Extracting a second boundary contour of the micro-lesion from the disturbance-enhanced image according to the position coordinates; constructing an image transformation model based on the second image set, the phase difference map, and the disturbance-enhanced image to obtain a pixel position transformation correction factor during the image enhancement process; the image transformation model is used to characterize the pixel position change between the disturbance-enhanced image and the second image set; The geometric size parameters of the micro-lesion are calculated according to the second boundary contour, the image acquisition resolution and the correction factor.

8. A thread detection system for high temperature alloy fasteners, characterized in that: include: An image acquisition module is configured to obtain a first image set of the thread to be inspected; the thread to be inspected is a thread on a high-temperature alloy fastener, and the first image set is a collection of images obtained by imaging the thread to be inspected from different observation angles; A registration module, configured to register the image region of the to-be-detected thread in the first image set to obtain a second image set; A construction module is used to construct a phase difference map based on the second image set; the phase difference map is used to characterize the relative phase difference of the same thread area under different observation angles; a generating module, configured to extract disturbance features of the to-be-detected thread using a direction-sensitive filter bank based on the second image set and the phase difference map, and obtain a disturbance-enhanced image based on the disturbance features; An image recognition model is used to identify whether there is micro-damage on the surface of the thread to be inspected based on the disturbance-enhanced image, and output a recognition result; the image recognition model is constructed based on a convolutional neural network, the micro-damage refers to surface defects that are invisible under visual observation conditions, and the types of micro-damage include thermal fatigue microcracks, oxidation pits, surface corrosion, and stress corrosion cracks.

9. The thread detection system for high temperature alloy fasteners according to claim 8, characterized in that: Also includes: The measuring module is used to perform visual measurement on the micro-damage to obtain geometric size parameters of the micro-damage.

10. A computer device comprising a memory and a processor, characterized in that: The memory stores a computer program, and when the processor executes the computer program, the steps of the thread detection method for a high-temperature alloy fastener according to any one of claims 1 to 7 are implemented.

Images