A method for detecting forming defects of titanium alloy

Abstract

The present invention discloses a titanium alloy forming defect detection method, which relates to the field of metal detection technology, and includes the following steps: obtaining image data and sensor data, performing denoising and preprocessing, constructing a defect recognition model, extracting and classifying defect features, confirming and analyzing defects, performing multi-scale analysis, accurately marking defects based on the results of defect classification, and presenting the detection results through a visual display. The titanium alloy forming defect detection method can comprehensively obtain various defect information in the forming process by fusing data and image features from different sensors. Multimodal data fusion makes defect recognition not only limited to image information, but also takes into account physical parameters directly related to the forming process, effectively improving the recognition precision and accuracy of complex defects.

Description

Technical Field

[0001] The invention relates to the technical field of metal detection, in particular to a titanium alloy forming defect detection method. Background Art

[0002] Titanium alloys are widely used in aerospace, automotive, medical and other fields due to their excellent mechanical properties, corrosion resistance and high temperature performance. In the forming process of titanium alloys, defects such as cracks, pores and inclusions often appear. These defects not only affect the mechanical properties of titanium alloys, but may also cause product failure, and even endanger safety in severe cases. Therefore, timely and accurate detection and identification of defects in the titanium alloy forming process is crucial to ensure the quality of the finished product.

[0003] Existing titanium alloy forming defect detection methods mainly rely on traditional non-destructive detection technologies, such as ultrasonic testing, X-ray testing, infrared thermal imaging, etc. Although these methods have high detection accuracy under certain conditions, their application scope is usually limited. Traditional detection methods often have the following shortcomings:

[0004] Low detection accuracy: Many traditional technologies can only detect relatively significant defects and have difficulty identifying subtle defects such as tiny cracks and surface flaws.

[0005] The testing process is cumbersome and manual: Many methods require manual operation or analysis and are easily affected by the operator's experience and technical level, resulting in poor accuracy and consistency of test results. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a titanium alloy forming defect detection method to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] In a first aspect, an embodiment of the present invention provides a titanium alloy forming defect detection method, comprising the following steps:

[0009] S1, synchronously acquire image data and sensor data of the titanium alloy forming process to obtain image data and sensor data;

[0010] S2, denoising and preprocessing the obtained image data;

[0011] S3. Construct and train a defect recognition model based on a convolutional neural network (CNN). By inputting a large number of annotated titanium alloy forming defect images, the image features of different defects in the titanium alloy forming process are learned.

[0012] S4, inputting the denoised and preprocessed image data into the trained defect recognition model to extract and classify defect features;

[0013] S5. Confirm and analyze the defects based on the classified defect types and sensor data;

[0014] S6. Use the fused sensor data to perform multi-scale analysis on the classified defect types and further classify the defects at different scales;

[0015] S7. Based on the defect classification results in step S6, the defects are accurately marked and the detection results are displayed in a visual manner.

[0016] To further optimize the technical solution, in step S1, during the titanium alloy forming process, a high-precision industrial camera is used to capture real-time images of the forming area of ​​the titanium alloy, and multiple sensor data are collected simultaneously, and the sensor data include temperature, pressure, and stress data; the real-time collected image data is used to capture subtle changes on the surface and inside of the titanium alloy, including surface defects such as cracks, pores, and poor welding that occur during the forming process.

[0017] To further optimize the technical solution, in step S2, the image data is subjected to denoising and preprocessing, including:

[0018] Gaussian blur and median filtering are used to remove irrelevant background noise and enhance the edge features of the image to improve the clarity of the image;

[0019] Perform geometric transformations on images, including rotation, cropping, and scaling, to ensure that the obtained image data meets standardized input requirements.

[0020] To further optimize the technical solution, in step S4, defect feature extraction is performed based on the defect recognition model, including:

[0021] ;

[0022] in,

[0023] is the extracted image features;

[0024] is the input image;

[0025] Training parameters for CNN model;

[0026] Conduct defect classification, including:

[0027] ;

[0028] in,

[0029] For the Classification results of class defects;

[0030] is the classifier weight;

[0031] is the bias term;

[0032] The extracted image features are matched with the features in the existing standard defect library to determine whether there are defects in the image, classify the defects and determine the defect types, and identify the defect types of cracks and pores that appear in the forming process.

[0033] To further optimize the technical solution, in step S5, the sensor data provides environmental information of the titanium alloy forming process, including temperature, pressure, and stress; excessive temperature causes pore defects, and excessive pressure or stress causes cracks to form;

[0034] Further defect confirmation and analysis is performed by combining sensor data with the classified defect type.

[0035] To further optimize the technical solution, in step S5, the defect confirmation and analysis are performed, including:

[0036] Sensor data fusion,Data fusion is as follows:

[0037] ;

[0038] in, is the fused multimodal feature vector;

[0039] is the temperature data;

[0040] is the pressure data;

[0041] is the stress data;

[0042] is the image feature;

[0043] is the weighting coefficient of each data source, reflecting its importance in defect identification;

[0044] Defect confirmation and severity analysis, the predicted results of the analysis are as follows:

[0045] ;

[0046] in,

[0047] To predict the severity of defects;

[0048] is the weight of the regression model;

[0049] is the bias term;

[0050] is an activation function that ensures that the output is non-negative.

[0051] To further optimize the technical solution, in step S6, the multi-scale analysis includes:

[0052] There are differences in the manifestations of titanium alloy forming defects at the macro and micro levels;

[0053] Perform macroscopic analysis of large-scale defects on the formed surface and detect large-area cracks and pores;

[0054] Through microscopic image analysis, we focus on detecting tiny defects, including micro cracks and micro pores.

[0055] To further optimize the technical solution, in step S6, a multi-scale analysis model is constructed, and the multi-scale analysis model uses the fused multi-modal feature vector As input, multiscale analysis and further classification are performed.

[0056] To further optimize the technical solution, the multi-scale analysis model includes:

[0057] Multiscale analysis formula:

[0058] ;

[0059] in,

[0060] Indicates The fusion results of sensor data and image features at different scales,

[0061] is the weighting coefficient for each scale, indicating the importance of each scale to defect identification;

[0062] Further defect classification is as follows:

[0063] ;

[0064] in,

[0065] For further defect classification results,

[0066] is the weight of the multi-layer perceptron MLP,

[0067] is the bias term;

[0068] By weighted aggregation of feature information at different scales, more detailed defect features, including tiny cracks or surface flaws, can be identified.

[0069] To further optimize the technical solution, in step S7, the specific location of each defect is accurately marked, the area where the defect is located is highlighted using image processing technology, and the corresponding environmental parameters during the forming process, including temperature, pressure, and stress, are superimposed on the image.

[0070] In a second aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of a titanium alloy forming defect detection method as described in the first aspect of the present invention are implemented.

[0071] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of a titanium alloy forming defect detection method as described in the first aspect of the present invention are implemented.

[0072] Compared with the prior art, the present invention provides a titanium alloy forming defect detection method, which has the following beneficial effects:

[0073] This titanium alloy forming defect detection method can fully obtain various defect information in the forming process by fusing data from different sensors (such as temperature, pressure, and stress) with image features (such as image features of surface defects such as cracks and pores). Multimodal data fusion makes defect identification not only limited to image information, but also takes into account physical parameters directly related to the forming process, effectively improving the recognition precision and accuracy of complex defects.

[0074] This method uses multi-scale analysis technology to comprehensively analyze the defects in the titanium alloy forming process from macro to micro levels. It identifies serious defects such as large-scale cracks from the macro scale, and small defects such as micro cracks and surface flaws from the micro scale. Multi-scale analysis ensures that defects of different scales can be accurately identified, avoiding the limitation that traditional methods can only detect large defects. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] 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 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.

[0076] Figure 1 A schematic diagram of a process for detecting titanium alloy forming defects proposed by the present invention;

[0077] Figure 2 A schematic flow chart of a defect recognition model in a titanium alloy forming defect detection method proposed in the present invention;

[0078] Figure 3 This is a schematic flow chart of a multi-scale analysis model in a titanium alloy forming defect detection method proposed in the present invention. DETAILED DESCRIPTION

[0079] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0080] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0081] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

[0082] Embodiment 1:

[0083] Reference Figure 1 to Figure 3 , which is the first embodiment of the present invention, provides a titanium alloy forming defect detection method, comprising the following steps:

[0084] S1. Acquire images of the titanium alloy forming process and obtain sensor data synchronously to obtain image data and sensor data.

[0085] In this embodiment, during the titanium alloy forming process, a high-precision industrial camera is used to perform real-time image acquisition of the titanium alloy forming area, and simultaneously collect a variety of sensor data, including temperature, pressure, and stress data; the real-time acquired image data is used to capture subtle changes on the surface and inside of the titanium alloy, including surface defects such as cracks, pores, and poor welding that occur during the forming process.

[0086] S2. De-noise and pre-process the obtained image data.

[0087] In this embodiment, during the titanium alloy forming process, the captured image may be affected by external light, environmental factors, equipment vibration, etc., thereby generating noise. De-noising and pre-processing of the image data include:

[0088] Gaussian blur, median filtering and other techniques are used to remove irrelevant background noise and enhance the edge features of the image to make the image clearer. After data denoising, important features in the image (such as defect edges, cracks, etc.) will be more prominent, reducing errors in subsequent analysis.

[0089] Perform geometric transformations on images, including rotation, cropping, and scaling, to ensure that all images meet standardized input requirements.

[0090] S3. Construct and train a defect recognition model based on convolutional neural network (CNN). By inputting a large number of annotated titanium alloy forming defect images, the image features of different defects in the titanium alloy forming process are learned.

[0091] In this embodiment, the core of this step is to train a defect recognition model based on a convolutional neural network (CNN). The model learns the performance characteristics of different defects (such as cracks, pores, poor welding, etc.) in the image by inputting a large number of annotated titanium alloy forming defect images. During the training process, the generalization ability of the model is improved through data enhancement technology to ensure that the model can accurately identify various types of defects. It combines image processing and deep learning technology to achieve high-precision and high-efficiency automated defect detection and reduce the necessity of manual intervention.

[0092] S4. Input the denoised and preprocessed image data into the trained defect recognition model to extract and classify defect features.

[0093] In this embodiment, defect feature extraction is performed based on the defect recognition model, including:

[0094] ;

[0095] in,

[0096] is the extracted image features;

[0097] is the input image;

[0098] Training parameters for CNN model;

[0099] Conduct defect classification, including:

[0100] ;

[0101] in,

[0102] For the Classification results of class defects;

[0103] is the classifier weight;

[0104] is the bias term;

[0105] After training, the defect recognition model based on convolutional neural network (CNN) begins to process the real-time image and extract key features, such as the edge, shape, texture, etc. of the defect. These features are then matched with the features in the existing standard defect library to determine whether there are defects in the image and determine the defect type. This process uses an algorithm based on similarity measurement to ensure efficient extraction and accurate matching of defects in complex backgrounds. In this way, the system can automatically identify different types of defects without manual intervention, thereby improving detection efficiency.

[0106] S5. Confirm and analyze the defects based on the classified defect types and sensor data.

[0107] In this embodiment, based on image recognition, sensor data (such as temperature, pressure, stress, etc.) can provide more information about the forming process environment. For example, too high temperature may cause pore defects, and too high pressure may cause cracks. By combining sensor data with image analysis results, the nature of the defect can be further confirmed.

[0108] Further, the defects are confirmed and analyzed, including:

[0109] Sensor data fusion,Data fusion is as follows:

[0110] ;

[0111] in, is the fused multimodal feature vector;

[0112] is the temperature data;

[0113] is the pressure data;

[0114] is the stress data;

[0115] is the image feature;

[0116] is the weighting coefficient of each data source, reflecting its importance in defect identification;

[0117] Defect confirmation and severity analysis, the predicted results of the analysis are as follows:

[0118] ;

[0119] in,

[0120] To predict the severity of defects;

[0121] is the weight of the regression model;

[0122] is the bias term;

[0123] is an activation function that ensures that the output is non-negative.

[0124] Through multi-dimensional data fusion technology, the accuracy of defect recognition can be enhanced to avoid insufficient image information or misjudgment. Especially in highly complex forming environments, sensor data is crucial for defect analysis.

[0125] S6. Use the fused sensor data to perform multi-scale analysis on the classified defect types and further classify the defects at different scales.

[0126] Titanium alloy forming defects may manifest in various forms, with differences at the macro and micro levels, so multi-scale analysis is required. First, a macro analysis is performed on the large-scale defects on the forming surface to detect large-area cracks, pores, etc.; then, through refined image analysis, the focus is on detecting small defects (such as microcracks, micropores, etc.). Multi-scale analysis can not only improve the accuracy of defect recognition, but also capture defects of different sizes more comprehensively, ensuring the high accuracy of the test results.

[0127] Construct a multi-scale analysis model, which uses the fused multi-modal feature vector As input, multiscale analysis and further classification are performed.

[0128] Furthermore, the multi-scale analysis model includes:

[0129] Multiscale analysis formula:

[0130] ;

[0131] in,

[0132] Indicates The fusion results of sensor data and image features at different scales,

[0133] is the weighting coefficient for each scale, indicating the importance of each scale to defect identification;

[0134] Further defect classification is as follows:

[0135] ;

[0136] in,

[0137] For further defect classification results,

[0138] is the weight of the multi-layer perceptron MLP,

[0139] is the bias term.

[0140] Using multi-scale analysis methods, defect features are further extracted from different scales (such as macro, micro, etc.). The goal of this process is to identify defects in the titanium alloy forming process from different scale perspectives and improve the recognition accuracy. By weighted aggregation of feature information at different scales, the model can identify more detailed defect features, such as tiny cracks or surface flaws, thereby improving the accuracy of classification.

[0141] S7. Based on the defect classification results in step S6, the defects are accurately marked and the detection results are displayed in a visual manner.

[0142] In this embodiment, after defect identification and classification, the next step is to accurately mark the specific location of each defect. Image processing technology is used to highlight the defect area, and the corresponding environmental parameters (such as temperature, pressure, etc.) during the forming process are superimposed on the image. Through real-time visual display, operators can quickly understand the distribution and type of defects in the forming process, so as to make faster adjustment decisions.

[0143] Embodiment 2:

[0144] This embodiment also provides a computer device, which is applicable to a titanium alloy forming defect detection method, including a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement a titanium alloy forming defect detection method proposed in the above embodiment.

[0145] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, a titanium alloy forming defect detection method as proposed in the above embodiment is implemented.

[0146] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0147] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0148] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0149] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk case (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0150] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit with a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit with a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0151] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A titanium alloy forming defect detection method, characterized in that: The following steps are involved: S1, synchronously acquire image data and sensor data of the titanium alloy forming process to obtain image data and sensor data; S2, denoising and preprocessing the obtained image data; S3. Construct and train a defect recognition model based on a convolutional neural network (CNN). By inputting a large number of annotated titanium alloy forming defect images, the image features of different defects in the titanium alloy forming process are learned. S4, inputting the denoised and preprocessed image data into the trained defect recognition model to extract and classify defect features; S5. Confirm and analyze the defects based on the classified defect types and sensor data; S6. Use the fused sensor data to perform multi-scale analysis on the classified defect types and further classify the defects at different scales; S7, based on the defect classification result in step S6, accurately mark the defects and display the detection results in a visual display manner; In step S5, the sensor data provides environmental information of the titanium alloy forming process, including temperature, pressure, and stress; excessive temperature causes porosity defects, and excessive pressure or stress causes cracks to form; Further defect confirmation and analysis by combining sensor data with classified defect types; In step S5, defect confirmation and analysis are performed, including: Sensor data fusion,Data fusion is as follows: ； in, is the fused multimodal feature vector; is the temperature data; is the pressure data; is the stress data; is the image feature; is the weighting coefficient of each data source, reflecting its importance in defect identification; Defect confirmation and severity analysis, the predicted results of the analysis are as follows: ； in, To predict the severity of defects; is the weight of the regression model; is the bias term; is the activation function, ensuring that the output is non-negative; In step S6, the multi-scale analysis includes: There are differences in the manifestations of titanium alloy forming defects at the macro and micro levels; Perform macroscopic analysis of large-scale defects on the formed surface and detect large-area cracks and pores; Through microscopic image analysis, focus on detecting tiny defects, including micro cracks and micro pores; In step S6, a multi-scale analysis model is constructed, and the multi-scale analysis model uses the fused multi-modal feature vector As input, multiscale analysis and further classification are performed; The multi-scale analysis model comprises: Multiscale analysis formula: ； in, Indicates The fusion results of sensor data and image features at different scales, is the weighting coefficient for each scale, indicating the importance of each scale to defect identification; Further defect classification is as follows: ； in, For further defect classification results, is the weight of the multi-layer perceptron MLP, is the bias term; By weighted aggregation of feature information at different scales, more detailed defect features, including tiny cracks or surface flaws, can be identified.

2. A titanium alloy forming defect detection method according to claim 1, characterized in that: In the step S1, during the titanium alloy forming process, a high-precision industrial camera is used to collect real-time images of the titanium alloy forming area, and multiple sensor data are collected simultaneously, and the sensor data includes temperature, pressure, and stress data; The image data collected in real time is used to capture subtle changes on the surface and inside of the titanium alloy, including surface defects such as cracks, pores, and poor welding that occur during the forming process.

3. A titanium alloy forming defect detection method according to claim 1, characterized in that: In step S2, the image data is subjected to denoising and preprocessing, including: Gaussian blur and median filtering are used to remove irrelevant background noise and enhance the edge features of the image to improve the clarity of the image; Perform geometric transformations on images, including rotation, cropping, and scaling, to ensure that the obtained image data meets standardized input requirements.

4. The titanium alloy forming defect detection method according to claim 1, characterized in that: In step S4, defect feature extraction is performed based on the defect recognition model, including: ； in, is the extracted image features; is the input image; Training parameters for CNN model; Conduct defect classification, including: ； in, For the Classification results of class defects; is the classifier weight; is the bias term; The extracted image features are matched with the features in the existing standard defect library to determine whether there are defects in the image, classify the defects and determine the defect types, and identify the defect types of cracks and pores that appear during the forming process.

5. The titanium alloy forming defect detection method according to claim 1, characterized in that: In step S7, the specific location of each defect is accurately marked, and the defect area is highlighted using image processing technology. At the same time, the corresponding environmental parameters of the forming process, including temperature, pressure, and stress, are superimposed on the image.

Images