Casting digital radiographic detection full-automatic system and detection method

Abstract

The invention relates to a casting digital ray detection full-automatic system and a detection method in the technical field of casting detection, and the casting digital ray detection full-automatic system and the detection method are used for realizing multi-angle, rapid and high-sensitivity titanium alloy casting digital ray detection and intelligent image evaluation. The invention relates to a titanium alloy casting film radiographic inspection system and method, and aims to solve the problems of single-direction transillumination, low detection speed and wrong film evaluation / missing detection of an orienting machine in traditional titanium alloy casting film radiographic inspection. Firstly, the detection efficiency in the digital radiographic detection process of the titanium alloy casting is improved, and the detection bottleneck of traditional film radiographic detection is greatly reduced; secondly, process compiling and automatic image acquisition of the titanium alloy casting are realized, and assistance and interference of human factors are reduced; and finally, intelligent evaluation of the image is realized, the defect detection rate is greatly improved, and the quality risk caused by mistake and missing detection due to human factors is reduced.

Description

Technical Field

[0001] This application relates to the field of casting inspection technology, and in particular to a fully automated digital radiographic inspection system and method for castings. Background Technology

[0002] Currently, the internal quality inspection of titanium alloy castings mainly relies on traditional film radiographic inspection technology. This technology uses film as the recording medium and requires a cumbersome process to obtain the film. Inspectors judge whether there are defects in the casting by observing the images on the film. This leads to limitations and bottlenecks in the application of the technology, such as limited inspection accuracy, low inspection efficiency, and large interference from subjective factors in the inspection results. Summary of the Invention

[0003] In view of the problems existing in the background technology, this application provides a fully automatic digital radiographic inspection system and method for castings to solve the production bottlenecks and technical problems such as defects, omissions, and subjective experience of human factors in the existing traditional film radiographic inspection. At the same time, it aims to deepen the application mode of digital radiography in the existing industrial field and accelerate the application process of digital radiography in the inspection of titanium alloy castings.

[0004] According to one aspect of the present invention, a fully automated digital radiographic inspection system for castings is provided, comprising a casting support platform, a radiographic machine, a flat panel detector, a first attitude adjustment component, a second attitude adjustment component, an image acquisition system, and an intelligent evaluation system; the radiographic machine and the flat panel detector are arranged opposite to each other, the casting support platform is located between the radiographic machine and the flat panel detector, the first attitude adjustment component is used to adjust the spatial position of the casting support platform between the radiographic machine and the flat panel detector, the second attitude adjustment component is used to adjust the position and angle of the connecting axis of the radiographic machine and the flat panel detector relative to the casting support platform, the image acquisition system is used to acquire radiographic images of the casting using the radiographic machine and the flat panel detector, and the intelligent evaluation system is used to evaluate and rate the radiographic images of the casting.

[0005] In some embodiments of the present invention, the first attitude adjustment component includes an X-axis drive unit extending in the horizontal direction, a Z-axis drive unit extending in the vertical direction, a first rotation unit, and two clamping jaws; the X-axis drive unit is used to mount the Z-axis drive unit and drive the Z-axis drive unit to perform linear reciprocating motion along the X-axis, the Z-axis drive unit is used to mount the two clamping jaws and drive the two clamping jaws to perform linear reciprocating motion along the Z-axis, and the first rotation unit is disposed between the casting support platform and the two clamping jaws and drives the casting support platform to perform circumferential motion around its own axis.

[0006] In some embodiments of the present invention, the second attitude adjustment component includes a C-shaped structure, a Y-axis drive unit extending horizontally and perpendicular to the X-axis drive unit, a second rotation unit, and a third rotation unit; the X-ray machine is mounted above the C-shaped structure, and the flat panel detector is mounted below the C-shaped structure; the Y-axis drive unit is used to mount the axis of the C-shaped structure and drive the C-shaped structure to perform linear reciprocating motion along the Y-axis; the second rotation unit is located between the Y-axis drive unit and the C-shaped structure and drives the C-shaped structure to rotate along the C-shaped direction; the third rotation unit is located between the second rotation unit and the C-shaped structure and drives the C-shaped structure to rotate along a direction perpendicular to the vector direction of the C-shaped structure.

[0007] In some embodiments of the present invention, the image acquisition system includes an image acquisition module and an image filtering module; the image acquisition module includes a process programming and manufacturing unit and an automatic image acquisition unit, the process programming and manufacturing unit is used to initially program the casting process, and the automatic image acquisition unit is used to automatically acquire radiographic images of the same type of casting according to the programmed process; the image filtering module is configured with various filtering functions and window width and window level.

[0008] In some embodiments of the present invention, the intelligent evaluation system includes a real-time image reading module, an automatic image evaluation module, an image calibration learning module, and an image inspection report generation module; the real-time image reading module is responsible for image reading and uploading, the automatic image evaluation module is responsible for image evaluation, the image calibration learning module is responsible for defect calibration learning, and the image inspection report generation module automatically evaluates and generates an inspection report based on the inspection standard and digital reference standard film and rates the defects.

[0009] In some embodiments of the present invention, the fully automated digital radiographic inspection system for castings further includes a real-time simulation module. The real-time simulation module is mounted on the first attitude adjustment assembly along with the X-ray machine and the flat panel detector. The real-time simulation module is used to simulate the size and shape of the casting and generate multiple inspection surfaces of the casting, while simultaneously forming a real-time running simulation image. The image acquisition system can sequentially adjust the position of the casting support platform relative to the X-ray machine and the flat panel detector according to the order of the inspection surfaces to take pictures of the casting in real time.

[0010] In some embodiments of the present invention, the fully automated digital radiographic inspection system for castings further includes an information input module, which is capable of automatically identifying casting information and automatically matching the information input image acquisition system.

[0011] According to another aspect of the present invention, a digital radiographic inspection method for castings is provided. The digital radiographic inspection method for castings is based on the above-mentioned fully automatic digital radiographic inspection system for castings. The digital radiographic inspection method for castings includes the following steps: placing the casting on a casting support platform; based on a first attitude adjustment component and a second attitude adjustment component, an image acquisition system uses a X-ray machine and a flat panel detector to take pictures of the casting from all directions to obtain radiographic images of the casting; and an intelligent evaluation system evaluates and rates the radiographic images of the casting.

[0012] In some embodiments of the present invention, before photographing the casting, the casting information is identified and automatically matched to the input image acquisition system.

[0013] In some embodiments of the present invention, before photographing the casting, a simulated casting image is acquired, and the simulated casting image is used to generate multiple inspection surfaces. The image acquisition system adjusts the position of the casting support platform relative to the X-ray machine and the flat panel detector in real time according to the order of the inspection surfaces to photograph the casting.

[0014] Compared with the prior art, the present invention achieves the following technical effects: This invention uses digital X-rays to replace traditional film-based inspection methods, significantly improving inspection efficiency and reducing film costs. Simultaneously, it employs a five-axis linkage automated inspection system, overcoming the limitations of previous single-axis operation in digital X-ray equipment, greatly saving inspection time, reducing the limitations of casting inspection, and expanding the product's application range. This invention also incorporates a data acquisition and filtering system in conjunction with the automated inspection system, enabling batch product inspection through automatic identification of inspection surfaces and the creation of process plans. This significantly improves image acquisition and storage time, overcoming the previous method of single-piece, single-storage inspection of digital X-rays, thus increasing inspection efficiency. Furthermore, this invention utilizes an intelligent image evaluation system to assess and generate reports from the acquired images, reducing interference from subjective human experience and increasing the defect detection rate. This invention is the first to integrate a mechanical system, a data acquisition and filtering system, and an intelligent image evaluation system, making it highly valuable for casting manufacturing enterprises and significantly solving production bottlenecks and defect identification problems in X-ray inspection. Attached Figure Description

[0015] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a schematic diagram of the overall structure of the fully automated digital X-ray inspection system for castings of the present invention; Figure 2This is a schematic diagram of the mechanical system of the present invention; Figure 3 This is a top view of the mechanical system of the present invention.

[0016] Figure 4 This is a flowchart of the digital X-ray inspection method for castings according to the present invention.

[0017] The labels in the attached diagram are as follows: 100. Workpiece platform module; 101. Clamping jaws; 102. First running track; 103. Casting support platform; 200. C-arm module; 201. C-structure; 202. X-ray machine; 203. Flat panel detector; 204. Second running track; 205. Track; 300. Real-time simulation module; 301. Area array infrared sensor; 400. Information input module; 401. Laser barcode scanner; 402. Touchscreen; 500. Image acquisition system; 600. Intelligent evaluation system. Detailed Implementation

[0018] It should be understood that the described embodiments are merely some, not all, of the embodiments in this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.

[0019] In the following description, when referring to the accompanying drawings, the same numbers in different drawings denote the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0020] In the description of this application, it should be understood that the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances. Furthermore, in the description of this application, unless otherwise stated, "multiple" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship.

[0021] This application discloses a fully automated digital radiographic inspection system for castings. For example... Figures 1-4As shown, the fully automated digital radiographic inspection system for castings includes a mechanical system, an image acquisition system 500, and an intelligent evaluation system 600. The image acquisition system 500 communicates with the mechanical system via a high-speed data cable, while the intelligent evaluation system 600 connects to the image acquisition system 500 via a local area network. The image acquisition system 500 is used to acquire radiographic images of the castings using the mechanical system, and the intelligent evaluation system 600 is used to evaluate and grade the radiographic images of the castings.

[0022] like Figure 2 and Figure 3 As shown, the mechanical system includes a workpiece platform module 100 and a C-arm module 200; the workpiece platform module 100 includes a casting support platform 103 and a first posture adjustment component; the C-arm module 200 includes a X-ray machine 202, a flat panel detector 203, and a second posture adjustment component; wherein, the X-ray machine 202 and the flat panel detector 203 are arranged opposite to each other, and the casting support platform 103 is located between the X-ray machine 202 and the flat panel detector 203; the first posture adjustment component is used to adjust the spatial position of the casting support platform 103 between the X-ray machine 202 and the flat panel detector 203; the second posture adjustment component is used to adjust the position and angle of the connecting axis of the X-ray machine 202 and the flat panel detector 203 relative to the casting support platform 103; the image acquisition system 500 is used to acquire X-ray images of the casting using the X-ray machine 202 and the flat panel detector 203.

[0023] In this invention, the casting is placed on the casting support platform 103. The relative synchronous movement of the casting support platform 103 and the C-arm module 200 enables automated digital X-ray inspection of the casting. The image acquisition system 500 then acquires and processes the images. Finally, the intelligent evaluation system 600 identifies defects in the images, rates them, and generates inspection reports, improving the defect detection rate. This results in high reliability and accuracy of the inspection results, more stable inspection consistency, and allows for multi-angle, high-sensitivity, and rapid automated digital X-ray inspection. It effectively solves the bottleneck problems of traditional film X-ray inspection, as well as issues such as missed or incorrect defects and subjective experience due to human factors. At the same time, it deepens the application of digital X-ray in existing industrial fields and accelerates the application process of digital X-ray in the inspection of titanium alloy castings.

[0024] In some embodiments of the present invention, such as Figure 2 and Figure 3As shown, the first attitude adjustment assembly includes an X-axis drive unit extending horizontally, a Z-axis drive unit extending vertically, a first rotation unit, and two gripping jaws 101. The X-axis drive unit is used to mount the Z-axis drive unit and drive the Z-axis drive unit to perform linear reciprocating motion along the X-axis. The Z-axis drive unit is used to mount the two gripping jaws 101 and drive the two gripping jaws 101 to perform linear reciprocating motion along the Z-axis. The first rotation unit is located between the casting support platform 103 and the two gripping jaws 101 and drives the casting support platform 103 to perform circumferential motion around its own axis.

[0025] In this embodiment, the Z-axis drive unit and the X-axis drive unit can carry the gripper 101 to grasp the casting support platform 103 and move it up and down and back and forth; the gripper 101 grasps the casting support platform 103 through the first rotation unit and drives the casting support platform 103 to move circumferentially around its own axis.

[0026] In some embodiments of the present invention, the X-axis drive unit includes, but is not limited to, a tracked linear transmission structure and a slider rail type 205 linear transmission structure.

[0027] Preferably, the X-axis drive unit adopts a tracked linear transmission structure, such as... Figure 2 and Figure 3 As shown, the X-axis drive unit is the first running track 102 arranged in front and behind.

[0028] In some embodiments of the present invention, the Z-axis drive unit includes, but is not limited to, a tracked linear transmission structure and a slider rail type 205 linear transmission structure.

[0029] Preferably, the Z-axis drive unit adopts a track-type linear transmission structure, such as... Figure 2 As shown, the Z-axis drive unit is the second running track 204 arranged vertically.

[0030] In some embodiments of the present invention, such as Figure 2 As shown, the first rotating unit can grip the casting support platform 103 by means of directional wheels set on the upper and lower sides of the gripping claw 101 along the circumferential direction, and drive the casting support platform 103 to rotate in the forward and reverse directions by friction / pulling between the motor output shaft and the circumferential surface of the casting support platform 103.

[0031] In some embodiments of the present invention, the casting support platform 103 may be designed to meet the size requirements of the casting to be inspected.

[0032] In some embodiments of the present invention, the casting support platform 103, the clamping claw 101, and the first posture adjustment component of the workpiece platform module 100 may be made of carbon fiber material or composite material that does not affect X-ray detection, so as to ensure that the effective image is not disturbed during the detection process.

[0033] In some embodiments of the present invention, such as Figure 2 and Figure 3 As shown, the second attitude adjustment assembly includes a C-shaped structure 201, a Y-axis drive unit extending horizontally and perpendicular to the X-axis drive unit, a second rotation unit, and a third rotation unit. The X-ray machine 202 is mounted above the C-shaped structure 201, and the flat panel detector 203 is mounted below the C-shaped structure 201. The X-ray machine 202 and the flat panel detector 203 are in a relatively perpendicular position, making them coaxial. The Y-axis drive unit is used to mount the axis of the C-shaped structure 201 and drive the C-shaped structure 201 to perform linear reciprocating motion along the Y-axis. The second rotation unit is located between the Y-axis drive unit and the C-shaped structure 201 and drives the C-shaped structure 201 to rotate along the C-shaped direction. The third rotation unit is located between the second rotation unit and the C-shaped structure 201 and drives the C-shaped structure 201 to rotate along a direction perpendicular to the vector direction of the C-shaped structure 201.

[0034] In this embodiment, using the second and third rotation units, the C-shaped structure 201 can rotate along the C-shaped direction or perpendicular to the vector direction of the C-shaped structure 201, with a maximum rotation angle of 180°, which facilitates the imaging of castings with various complex structures. During operation, the C-shaped structure 201 is controlled by the Y-axis drive unit to move left and right.

[0035] In some embodiments of the present invention, the Y-axis drive unit includes a track 205, and the bottom of the C-shaped structure 201 is slidably connected to the track 205. Specifically, the C-shaped structure 201 can be driven by a motor to move left and right on the track 205, so that the X-ray center beam of the X-ray machine 202 can 100% cover the casting support platform 103, so as to be able to detect castings of different sizes.

[0036] In some embodiments of the present invention, the second rotating unit may adopt a rotating shaft-bearing structure. For example, the outer ring of the bearing is slidably connected to the track 205 through a mounting seat, the rotating shaft is interference-fitted with the inner ring of the bearing and fixed with the C-shaped structure 201, and the rotating shaft is driven to rotate by a motor. When the motor rotates in the forward or reverse direction, its output shaft drives the rotating shaft and the C-shaped structure 201 to rotate in a direction perpendicular to the vector direction of the C-shaped structure 201.

[0037] In some embodiments of the present invention, the third rotating unit may adopt a structure of arc-shaped slide rail-slider. For example, the C-shaped structure 201 is mounted on the above-mentioned mounting base through the arc-shaped slide rail structure, so that the C-shaped structure 201 is slidably connected to the arc-shaped slide rail through a number of sliders, and the C-shaped structure 201 is driven by a motor to rotate along the C-shaped direction on the arc-shaped slide rail.

[0038] In some embodiments of the present invention, such as Figure 2As shown, the C-shaped structure 201 can be designed as a square C (i.e., a "U" shape) or a letter C shape, and its design requirements need to meet certain bearing capacity, high stability, high structural stiffness, etc.

[0039] In some embodiments of the present invention, in the relative position of the C-shaped structure 201 in the vertical state, the second running track 204 can carry the clamping jaw 101 to drive the casting bearing platform 103 to move up and down to adjust the focal length; the first running track 102 is used to move back and forth to adjust the relative position between the casting and the C-shaped structure 201, so that the ray center beam covers the casting bearing platform 103; at the same time, the C-shaped structure 201 can also ensure that the ray center beam covers the casting bearing platform 103 by moving left and right on the track 205; in the relative position of the C-shaped structure 201 in other angular states, the relative position between the casting bearing platform 103 and the flat panel detector 203 can be adjusted by the front and back, up and down of the casting bearing platform 103 and the left and right and rotation of the C-shaped structure 201, so as to reduce the distance from the object to the flat panel detector 203 and realize the full-automatic detection of the digital ray system.

[0040] In some embodiments of the present invention, the image acquisition system 500 includes an image acquisition module and an image filtering module; wherein, the image acquisition module includes a process programming and production unit and an image automatic acquisition unit. The process programming and production unit is used to perform process preparation on the casting initially, and the image automatic acquisition unit is used to automatically obtain the radiographic images of the same kind of castings according to the prepared process; the image filtering module is configured with various filtering functions and window widths and window levels.

[0041] In an embodiment of the present invention, the process programming and production unit first selects a kind of casting for detection and sequentially stores and accesses images according to the corresponding partitions; secondly, makes a detection process according to the acquired images and edits the images in sequence; in this production process, the position information of the mechanical system attached to the stored and accessed images is retrieved to realize process preparation; finally, when detecting this kind of casting again, the prepared process is directly retrieved, the image can be automatically acquired, and the detected images are automatically saved according to the path.

[0042] In some embodiments of the present invention, the technical solution can meet the batch detection of castings according to the prepared process, that is, the batch automatic detection of the same kind of products is carried out by using the pre-prepared process to be applicable to 100% detection of a large number of products.

[0043] In an embodiment of the present invention, the image filtering module is configured with various filtering functions and window widths and window levels. During the process of making the process, the stored and accessed images are pre-adjusted with appropriate filtering and window widths and window levels, and the adjusted image filtering information is brought into the detected images during image acquisition without secondary adjustment of filtering for image evaluation.

[0044] In some embodiments of the present invention, the technical solution can realize the adaptive image filtering algorithm proposed in this solution, which is an adaptive bilateral filtering algorithm based on gray-level gradient. To facilitate defect identification, the core principle of the adaptive bilateral filtering algorithm adopted in this solution is to calculate the gray-level gradient of each pixel in the image. When the gray-level gradient is greater than a preset threshold, it is determined to be a defect edge region, and the smoothing intensity of the filtering kernel is reduced to preserve the detail edges. When the gray-level gradient is less than the preset threshold, it is determined to be a background noise region, and the smoothing intensity of the filtering kernel is increased to remove noise.

[0045] In some embodiments of the present invention, the technical solution described herein can satisfy the image acquisition system 500 to also have a function of periodically and automatically calibrating the detector. When the geometric sharpness exceeds the error under the same identification conditions, self-calibration is initiated, and the calibration result is stored in the current file directory for execution.

[0046] In some embodiments of the present invention, the intelligent evaluation system 600 includes an image real-time reading module, an image automatic evaluation and calibration learning module, and an image detection report generation module; the image automatic evaluation and calibration learning module includes an image automatic evaluation module and an image calibration learning module.

[0047] In this invention, the real-time image reading module is responsible for image reading and uploading, the automatic image evaluation module is responsible for image evaluation, the image calibration learning module is responsible for defect calibration learning, and the image inspection report generation module automatically evaluates and generates inspection reports based on inspection standards and digital reference standard film, and rates the defects.

[0048] In this invention, the real-time image reading module automatically reads images along the corresponding paths sequentially based on the image's detection state and uploads them to the automatic image evaluation and calibration learning module. The automatic image evaluation module automatically identifies typical defects based on the uploaded images and performs qualitative and quantitative analysis. The automatic image evaluation module only needs to identify and mark defects in areas where the grayscale value of the casting projection meets the requirements; other areas do not require identification. The image calibration learning module adds the identified defects to a deep learning library, continuously learning through continuous detection and expanding the database based on this learning, thus training the module and reducing the false detection / false detection rate. The image detection report generation module outputs a report based on the automatic image evaluation results and the detector pixel coordinates.

[0049] In some embodiments of the present invention, the report should not be editable, modified, or deleted; the report should contain image partitions, corresponding defect types, sizes, and ratings based on the output of standard radiographic films.

[0050] In some embodiments of the present invention, the intelligent evaluation system 600 includes an intelligent image evaluation system, a defect intelligent rating system, an automatic inspection report generation module, a defect data statistical analysis module, and an image evaluation verification module. The technical solution satisfies the following: the intelligent image evaluation system reads image data from an image acquisition folder and performs automatic defect identification; the defect intelligent rating system has a built-in titanium alloy defect feature library and a multi-level standard mapping table to automatically determine defect levels; the automatic inspection report generation module associates the acceptance standard database and standard film library to generate reports with clause references and image comparisons; the defect data statistical analysis module statistically analyzes defects by batch / part number, outputs defect trend charts based on casting information (casting time, process execution time, etc.), and analyzes abnormal information; the image evaluation verification module enables multi-mode push of evaluation results, manual verification, and feeds the verified defect data back into the intelligent image evaluation system for training.

[0051] In some embodiments of the present invention, the technical solution described herein can satisfy the requirement that the intelligent image evaluation system adopts a dual-channel lightweight YOLOv8 network. Channel 1 enhances the edge features of low-contrast areas through an attention mechanism, while Channel 2 integrates a titanium alloy-specific noise reduction operator to process high-noise backgrounds, thereby identifying various types of minute defects on titanium alloy castings.

[0052] In some embodiments of the present invention, the technical solution described herein can satisfy the requirements of the intelligent defect rating system to construct a three-dimensional mapping model of "defect features - acceptance criteria - reference film". The first-level mapping matches the acceptance criteria clauses, the second-level mapping calls the standard film library for visual comparison and outputs the rating confidence level, and the third-level mapping marks the defects on the simulated casting image based on the identification of the morphology of the scanned casting and the inspection surface.

[0053] In some embodiments of the present invention, the technical solution can satisfy the requirement that the automatic generation module for the test report includes a standard clause reference unit and an image comparison and annotation unit; the standard clause reference unit can automatically extract the clauses corresponding to the acceptance standards and defects; the image comparison and annotation unit displays the defect image and the defect images of the same type in the standard film library on the same screen, and adds grayscale profile lines and dimension annotations.

[0054] In some embodiments of the present invention, the technical solution can satisfy the defect data statistical analysis module including a structured database and an LSTM prediction unit: the structured database uses MongoDB to store defect data according to a multi-dimensional index of "batch-part number-defect type-defect size", wherein the process parameters include casting time and process execution time; the LSTM prediction unit can output the defect occurrence rate change curve of future N (adjustable) batches and generate the correlation between defect type and process parameters.

[0055] In some embodiments of the present invention, the technical solution can satisfy the following: the image assessment and review module includes a confidence level push unit, an incremental training unit, and a federated learning fine-tuning unit. The confidence level push unit pushes the intelligent assessment results to the corresponding inspector terminals according to the confidence level intervals. The incremental training unit automatically adds the reviewed defect images (including manually corrected type and grade labels) to the training dataset. The federated learning fine-tuning unit fine-tunes the model parameters monthly based on the incremental dataset, and does not disclose the original enterprise data during the training process.

[0056] In some embodiments of the present invention, the technical solution can satisfy the requirement that channel one of the dual-channel lightweight YOLOv8 network enhances the edge features of low-contrast regions (grayscale difference ≤20) through an attention mechanism, while channel two uses a titanium alloy-specific noise reduction operator (the size of the filter kernel is dynamically adjusted to 3×3-7×7 according to the intensity of the ray noise) to process the high-noise background.

[0057] In some embodiments of the present invention, the technical solution can satisfy the following: the first-level mapping of the three-dimensional mapping model matches the defect level clauses of the acceptance standard by the defect geometric parameters (pore diameter ≥ 0.2 mm, crack length ≥ 0.5 mm); the second-level mapping calculates the similarity between the defect image and the standard film by SSIM (structural similarity index); when the similarity is ≥ 85%, the rating confidence is increased by 10%-20%.

[0058] In some embodiments of the present invention, the technical solution described herein can satisfy the requirement that the federated learning fine-tuning unit adopts a model parameter encryption transmission mechanism, only aggregating the model gradient update values ​​of each detection node, without acquiring the original defect image data, thus ensuring the privacy and security of enterprise data.

[0059] In some embodiments of the present invention, such as Figure 1 As shown, the mechanical system also includes a real-time simulation module 300. The real-time simulation module 300 is mounted on the first attitude adjustment assembly along with the X-ray machine 202 and the flat panel detector 203. The real-time simulation module 300 is used to simulate the size and shape of the casting and generate multiple inspection surfaces of the casting. At the same time, it forms a real-time running simulation image. The image acquisition system 500 can adjust the position of the casting support platform 103 relative to the X-ray machine 202 and the flat panel detector 203 in sequence according to the inspection surface order to take pictures of the casting in real time. This facilitates the monitoring of the movement position during the inspection process and provides a safe operating range for the mechanical system.

[0060] In some embodiments of the present invention, such as Figure 2 As shown, the real-time simulation module 300 includes an array infrared sensor 301, which is installed inside the C-shaped structure 201. The array infrared sensor 301 is used to simulate the size and shape of the casting.

[0061] In some embodiments of the present invention, the technical solution can satisfy the following: after the casting moves to the initial running position, a signal is sent, the area array infrared sensor 301 is activated to identify the shape and size of the casting, the casting is 100% identified by the rotation of the casting support platform 103, a simulated casting image is generated in the mechanical system, and N detection surfaces are generated on the casting.

[0062] In some embodiments of the present invention, the technical solution can satisfy the following: after the simulated casting identification is completed, the image acquisition system 500 is turned on, and the mechanical system is adjusted in real time to take pictures according to the basic operating principles of X-ray inspection, such as 100% coverage of the inspection surface, perpendicularity to the inspection surface, and single-wall radiography, so as to simplify the operation procedure and avoid repeated adjustments.

[0063] In some embodiments of the present invention, the technical solution can satisfy the requirement that, in subsequent batch inspection of castings, when the simulated casting generated after identification by the area array infrared sensor 301 has a similarity of more than 99% to the result of the first simulated casting, and the infrared identification of the casting's unique identifier code is consistent, the process for compiling the unique identifier code will be automatically retrieved and executed, and the inspection result will be saved to the unique identifier code's inspection file. During the execution process, it is necessary to identify whether the image grayscale standard deviation meets the requirements.

[0064] The aforementioned technical solution can satisfy the following conditions: when the casting simulation result is less than 99% and the unique identifier is consistent, then a second scan is performed; when the casting simulation result is less than 99% and the unique identifier is inconsistent, then the feeding procedure and area array scan are re-executed.

[0065] In some embodiments of the present invention, the technical solution can meet the requirements of the image processing function based on the window width and window level and image geometric sharpness initially set by the image acquisition system 500, according to the acquired image of the corresponding detection surface, to perform adaptive filtering processing on the image so that the image grayscale value is in a position conducive to observing defect values; if the image grayscale value does not meet the requirements, it is made to meet them by increasing the voltage or current; it should be noted that the voltage and current are not necessarily better the higher they are, and should be automatically adjusted according to the grayscale standard deviation.

[0066] In some embodiments of the present invention, the technical solution can satisfy the requirement that after the test is completed, the process is compiled sequentially according to the order of the test surfaces, and a process execution risk report is generated and stored in the process file with the corresponding identification code.

[0067] In some embodiments of the present invention, such as Figure 1 As shown, the mechanical system also includes an information input module 400, which can automatically identify casting information and automatically match the information input image acquisition system 500.

[0068] In some embodiments of the present invention, such asFigure 1 As shown, the information input module 400 includes a laser barcode scanner 401 and a touch screen 402.

[0069] In some embodiments of the present invention, the technical solution can enable operators to identify the information attached to the casting through the laser barcode scanner 401 and transmit the identified corresponding information to the information window of the image acquisition system 500 to generate process documents and inspection documents with unique identifiers. The lower-level storage logic of the inspection documents can be selected according to an easy-to-execute scheme, such as date-batch number-...-part number, etc., to trace the inspection results.

[0070] In some embodiments of the present invention, the technical solution can satisfy the requirement that, during the operation of the mechanical system, an interlocked spherical operating range with minimum safety distance is formed based on the relative positions of the simulated casting image, the casting support platform 103, the X-ray machine 202, and the flat panel detector 203. When adjusting the relative positions of the casting support platform 103 with the X-ray machine 202 and the flat panel detector 203, the five-axis linkage of the mechanical system can be realized, avoiding the mode of repeated single-axis operation and identification of safety distance.

[0071] In some embodiments of the present invention, the technical solution can satisfy the following: after the casting has been inspected, the casting support platform 103 completes initialization and operation, sends out the casting, and sends in the next piece of the product.

[0072] In some embodiments of the present invention, the technical solution described herein can satisfy the need to maintain the consistency of test results and the effectiveness of process execution, improve the timeliness of batch testing, and ensure the uniformity of adjusting the spatial position of castings using a robotic arm.

[0073] This embodiment also proposes a digital radiographic inspection method for castings. This method utilizes the aforementioned fully automated digital radiographic inspection system for castings to perform radiographic inspection of the castings. Figure 4 As shown, the digital radiographic inspection method for castings includes the following steps: 1) The casting is placed on the casting support platform 103. Based on the first attitude adjustment component and the second attitude adjustment component, the image acquisition system 500 uses the X-ray machine 202 and the flat panel detector 203 to take pictures of the casting from all directions to obtain X-ray images of the casting. 2) The intelligent evaluation system 600 evaluates and rates the radiographic images of castings.

[0074] In some embodiments of the present invention, before photographing the casting, the casting information is identified and automatically matched into the image acquisition system 500.

[0075] In some embodiments of the present invention, before photographing the casting, a simulated casting image is acquired, and multiple inspection surfaces are generated from the simulated casting image. The image acquisition system 500 adjusts the position of the casting support platform 103 relative to the X-ray machine 202 and the flat panel detector 203 in real time according to the order of the inspection surfaces to photograph the casting.

[0076] In some embodiments of the present invention, during the first inspection of the casting, the casting is placed on a workpiece platform. During the process planning, after radiography is completed on one inspection surface, the casting support platform 103 automatically rotates to the next inspection surface position. At this time, the vertical and horizontal positions of the casting support platform 103 and the left-right and rotational positions of the C-shaped structure 201 are automatically adjusted according to the relative positions of the inspection surface, the X-ray machine 202, and the detector, and the final position of the movement is recorded. During batch inspection, automated inspection can be performed by executing the process.

[0077] In one embodiment of the present invention, such as Figure 4 As shown, the digital radiographic inspection method for castings includes the following steps: First, the casting is placed on the casting support platform 103. By rotating the casting support platform 103 and the C-shaped structure 201, the casting is scanned using an area array infrared sensor to generate a simulated image of the casting, identify the detection surface, and simulate the real-time state of the mechanical system.

[0078] Secondly, based on the number of inspection surfaces, the process is pre-programmed, and image acquisition begins. During the process, in order to adapt to imaging of different casting thicknesses, adaptive image filtering is required. Process images are generated one by one and stored in the process folder, and the inspection folder is generated simultaneously.

[0079] Third, batch inspection of similar products begins. Before inspection, the consistency between the casting markings and casting morphology and the simulated casting image must be compared. If discrepancies are found, the feeding procedure and area scan are re-executed, serving as a reminder to the operator to ensure the casting is being inspected correctly. If consistency is found, the lead door is closed, single-piece automated inspection begins, and the inspection results are saved to the inspection folder.

[0080] Fourth, the intelligent evaluation system reads the image, begins automatic image evaluation and rating, and pushes the results to manual review. Finally, the review results are fed back to the defect training library for calibration and deep learning; simultaneously, an inspection report is generated based on the manual review results, and defect statistical analysis is completed based on the inspection report results.

[0081] The digital X-ray inspection technology of this invention features fast imaging speed, high resolution, and good contrast, greatly improving the defect detection rate and inspection efficiency. Furthermore, the digital X-ray technology stores images digitally, making image storage, retrieval, transmission, and sharing extremely convenient. This invention further enhances the efficiency and accuracy of titanium alloy casting inspection through a fully automated system. In this process, no human interference is required, and the system can automatically complete each inspection step. Compared with traditional film inspection methods, the fully automated inspection system of this invention can improve inspection efficiency by several times or even tens of times, greatly meeting the requirements of large-scale production for inspection accuracy. This system employs image processing technology and artificial intelligence algorithms such as deep learning, possessing powerful intelligent analysis and judgment capabilities. Through learning and analyzing a large amount of inspection data, the intelligent algorithm continuously optimizes its judgment model, greatly improving the accuracy and reliability of inspection and reducing the interference of subjective factors in the inspection results. Simultaneously, this invention can automatically generate detailed inspection reports based on the inspection results, providing a strong basis for subsequent production improvements and quality control.

[0082] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A fully automated digital radiographic inspection system for castings, characterized in that, The system includes a casting support platform, a X-ray machine, a flat panel detector, a first attitude adjustment component, a second attitude adjustment component, an image acquisition system, and an intelligent evaluation system. The X-ray machine and the flat panel detector are arranged opposite to each other, with the casting support platform located between them. The first attitude adjustment component is used to adjust the spatial position of the casting support platform between the X-ray machine and the flat panel detector. The second attitude adjustment component is used to adjust the position and angle of the connecting axis between the X-ray machine and the flat panel detector relative to the casting support platform. The image acquisition system is used to acquire radiographic images of the casting using the X-ray machine and the flat panel detector. The intelligent evaluation system is used to evaluate and rate the radiographic images of the casting.

2. The fully automated digital radiographic inspection system for castings according to claim 1, characterized in that, The first attitude adjustment component includes an X-axis drive unit extending horizontally, a Z-axis drive unit extending vertically, a first rotation unit, and two gripping jaws. The X-axis drive unit is used to mount the Z-axis drive unit and drive the Z-axis drive unit to perform linear reciprocating motion along the X-axis. The Z-axis drive unit is used to mount the two gripping jaws and drive the two gripping jaws to perform linear reciprocating motion along the Z-axis. The first rotation unit is located between the casting support platform and the two gripping jaws and drives the casting support platform to perform circumferential motion around its own axis.

3. The fully automated digital radiographic inspection system for castings according to claim 2, characterized in that, The second attitude adjustment component includes a C-shaped structure, a Y-axis drive unit extending horizontally and perpendicular to the X-axis drive unit, a second rotation unit, and a third rotation unit; the X-ray machine is installed above the C-shaped structure, and the flat panel detector is installed below the C-shaped structure; the Y-axis drive unit is used to mount the axis of the C-shaped structure and drive the C-shaped structure to perform linear reciprocating motion along the Y-axis; the second rotation unit is located between the Y-axis drive unit and the C-shaped structure and drives the C-shaped structure to rotate along the C-shaped direction; the third rotation unit is located between the second rotation unit and the C-shaped structure and drives the C-shaped structure to rotate along a direction perpendicular to the vector direction of the C-shaped structure.

4. The fully automated digital radiographic inspection system for castings according to claim 1, characterized in that, The image acquisition system includes an image acquisition module and an image filtering module; The image acquisition module includes a process programming and production unit and an automatic image acquisition unit. The process programming and production unit is used to initially program the casting process, and the automatic image acquisition unit is used to automatically acquire radiographic images of the same type of casting according to the programmed process. The image filtering module is equipped with various filtering functions and window width and window level.

5. The fully automated digital radiographic inspection system for castings according to claim 1, characterized in that, The intelligent evaluation system includes a real-time image reading module, an automatic image evaluation module, an image calibration learning module, and an image detection report generation module; The real-time image reading module is responsible for image reading and uploading, the automatic image evaluation module is responsible for image evaluation, the image calibration learning module is responsible for defect calibration learning, and the image inspection report generation module automatically evaluates and generates inspection reports based on the inspection standards and digital reference standard film, and rates the defects.

6. The fully automated digital radiographic inspection system for castings according to claim 1, characterized in that, It also includes a real-time simulation module, which is installed on the first attitude adjustment component along with the X-ray machine and the flat panel detector. The real-time simulation module is used to simulate the size and shape of the casting and generate multiple inspection surfaces of the casting, while simultaneously forming a real-time running simulation image. The image acquisition system can adjust the position of the casting support platform relative to the X-ray machine and the flat panel detector in real time according to the order of the inspection surfaces to take pictures of the casting.

7. The fully automated digital radiographic inspection system for castings according to claim 1, characterized in that, It also includes an information input module, which can automatically identify casting information and automatically match the information input image acquisition system.

8. A digital radiographic inspection method for castings, characterized in that, The digital radiographic inspection method for castings is based on the fully automated digital radiographic inspection system for castings according to any one of claims 1 to 7, and the digital radiographic inspection method for castings includes the following steps: The casting is placed on the casting support platform. Based on the first and second attitude adjustment components, the image acquisition system uses a X-ray machine and a flat panel detector to take pictures of the casting from all directions to obtain X-ray images of the casting. The intelligent evaluation system evaluates and rates the radiographic images of the casting.

9. The digital radiographic inspection method for castings according to claim 8, characterized in that, Before photographing the casting, the casting information is identified and automatically matched into the image acquisition system.

10. The digital radiographic inspection method for castings according to claim 8, characterized in that, Before photographing the casting, a simulated casting image is acquired, and multiple inspection surfaces are generated from the simulated casting image. The image acquisition system then adjusts the position of the casting support platform relative to the X-ray machine and the flat panel detector in real time according to the order of the inspection surfaces to photograph the casting.

Images