Mobile small workpiece DR detection device and self-adaptive exposure method thereof

Abstract

The invention discloses a movable small workpiece DR detection device and a self-adaptive exposure method thereof, relates to the technical field of nondestructive testing, and is particularly suitable for welding seam detection of small workpieces. The device comprises a movable bearing mechanism, a detection execution mechanism and a self-adaptive exposure control system, and high-precision and high-efficiency DR detection of small workpieces is realized through software and hardware collaborative design. Wherein the adaptive exposure control system dynamically adjusts exposure parameters based on a deep learning algorithm, and solves the problems of difficult focusing and unstable image quality in traditional detection; a movable structure and an automatic adjusting mechanism improve the detection flexibility and the operation convenience, and compared with a traditional film detection device and a common DR device, the detection efficiency and the defect recognition precision are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of non-destructive testing, in particular to a mobile digital radiography (DR) detection device suitable for small workpieces and an adaptive exposure method. BACKGROUND

[0002] In the field of radiographic testing, traditional radiographic film photography detection technology has been the mainstream for a long time, and its operation process highly depends on the experience accumulation and subjective judgment of the detection personnel. The detection personnel needs to arrange the transmission according to the relevant standards and process documents, but due to the small workpieces involved in industrial production, the shape is complex, the size span is large, from a few millimeters to a few tens of centimeters, and the surface morphology is diverse, such as curved surface and special-shaped structure, so it is easy to cause focusing deviation and distance measurement error in actual operation. For example, when detecting the weld of a tubular workpiece with a diameter difference of only a few millimeters, adjusting the relative position of the X-ray source and the workpiece according to experience often leads to deviation of the transmission angle from the optimal value, resulting in distortion of the defect image or missed detection. At the same time, the film photography technology needs to go through multiple processes such as darkroom processing, developing and fixing, which not only takes a long time, but also usually takes more than 1 hour for single workpiece detection, and the film is easy to be damaged during storage and transportation due to environmental factors, which seriously restricts the detection efficiency and the reliability of the results.

[0003] Although the conventional DR digital imaging detection technology saves the film processing link through digital means and realizes instant image acquisition, it still has significant limitations in the detection of small workpieces. On the one hand, the exposure parameters, such as tube voltage, tube current and exposure time, are mostly fixed and preset modes, and cannot be dynamically adapted to the material characteristics of the workpiece, such as metal density difference and thickness distribution, such as locally thickened weld joints. For example, when detecting an aluminum alloy test plate with a thickness of 2mm and a steel test plate with a thickness of 5mm, if the same exposure parameters are used, the former is easy to cause image overexposure and loss of details due to excessive X-ray energy, and the latter may cause image overexposure and defect submersion due to insufficient energy. On the other hand, the fixation and positioning of small workpieces are difficult: the workpiece clamping mechanism of existing devices has poor universality, and the fixation stability of special-shaped workpieces is insufficient, which may cause image blur due to slight shaking during detection; and the relative position of the X-ray source and the detector is fixed, which makes it difficult to flexibly adjust the focal length according to the size of the workpiece, further aggravating the instability of the imaging quality.

[0004] These technical bottlenecks have seriously restricted the application of DR detection in the field of small workpieces, not only the detection efficiency is low and the labor intensity is large, but also the defect recognition accuracy is difficult to meet the quality control requirements of high-end manufacturing fields. Therefore, it is a key problem to be solved in the field of non-destructive testing to develop a small workpiece DR detection device with flexible mobility, parameter adaptive adjustment capability, and the ability to significantly improve the detection efficiency and accuracy. SUMMARY

[0005] The present application aims to solve the following problems existing in the detection of small workpieces by traditional ray detection and conventional DR detection: the detection device is fixed, it is difficult to adapt to the mobile detection needs in different scenarios; the exposure parameters are fixed, it is impossible to dynamically adjust according to the material and thickness of the workpiece, resulting in unstable image quality; manual operation depends on experience, the focusing accuracy is low, and the detection efficiency is low.

[0006] To solve the above problems, the present application provides a mobile small workpiece DR detection device, which comprises a movable bearing mechanism, a detection execution mechanism and an adaptive exposure control system.

[0007] Further, the detector assembly comprises a detector telescopic arm and a flat panel detector; the detector telescopic arm is telescopic along the X axis, and the effective stroke is 500mm; the flat panel detector is made of amorphous silicon material, the pixel size is ≤100μm, and the image resolution is ≥300lp / cm; Further, the workpiece positioning mechanism comprises a longitudinal rail, a transverse rail, a liftable platform and a workpiece clamping device; the longitudinal rail and the transverse rail are perpendicular and cross each other, and are driven to move in two dimensions by an electric control system; the liftable platform is liftable along the Z axis, and the maximum stroke is 162mm; the workpiece clamping device has a maximum opening of 200mm, and clamps the workpiece by a manual rocker.

[0008] Further, the X-ray emitting mechanism comprises an X-ray machine fixing bracket and an X-ray machine; the fixing bracket locks the X-ray machine through two semicircular clamps with a diameter of 250mm and four butterfly nuts; the tube voltage of the X-ray machine is adjustable in the range of 50-225KV, the tube current is adjustable in the range of 0.1-5mA, and the exposure time is adjustable in the range of 0.1-10s.

[0009] Further, the hardware of the adaptive exposure control system comprises an intelligent control box and an electric control system operation table; the intelligent control box is integrated with an FPGA real-time processing chip, and the response time is milliseconds; the operation table is provided with 10 control buttons for respectively controlling the power switch, the detector telescoping, the platform lifting and moving; the adaptive exposure control system comprises an adaptive algorithm module, the algorithm module is based on PID control and reinforcement learning, adjusts the exposure amount in a closed loop based on the target gray scale, and forms an optimal strategy database through long-term iteration.

[0010] In another aspect, the present application also provides an adaptive exposure control method for DR detection, applied to the above-mentioned device, comprising the steps of: S1, receiving inputted work information, including workpiece material attribute, thickness parameter and detector sensitivity; S2, predicting initial exposure parameters based on the work information through a machine learning model; S3, performing a first low-dose scan using the initial exposure parameters to obtain a pre-scan image; S4, performing quality analysis on the pre-scan image, calculating gray scale distribution, key area contrast and image signal-to-noise ratio related parameter indicators, and judging whether they meet the standards; S5, if any indicator does not meet the standard, dynamically adjusting the exposure parameters according to the type of non-standard; S7, performing formal scanning using the standard parameters.

[0011] Further, in the step S4, calculating the gray scale distribution, key area contrast and image signal-to-noise ratio related parameter indicators, and judging whether they meet the standards specifically includes: whether the gray scale distribution is in the middle interval of the dynamic range of the detector; whether the key area contrast is greater than 0.1; and whether the image signal-to-noise ratio is greater than 30dB, the key area being defined as the weld area or the workpiece defect sensitive area. The dynamic range of the detector is the range of light signal intensity that can be effectively captured. If the gray scale distribution deviates to the low end of the dynamic range, it will be too dark to cause the details of the workpiece, such as small defects, to be overwhelmed by noise; if it deviates to the high end, it will be too bright to cause the details of the bright area to be lost due to signal saturation, such as the edge information of thin-walled workpieces. Controlling the gray scale distribution in the middle interval can ensure that the light and dark areas of the workpiece are within the linear response range of the detector, maximizing the retention of image details and providing complete gray scale gradient information for defect identification, such as pores and cracks. At the same time, the gray scale distribution in the middle interval leaves sufficient adjustment space for subsequent image processing. Even if there is slight parameter fluctuation, it can be further optimized through software algorithms to avoid the "tuning" problem caused by extreme gray scale distribution, enhancing the robustness of the image. The contrast of the key area directly reflects the gray scale difference between the defect and the workpiece matrix. When the contrast is >0.1, the boundary between the defect and the surrounding normal tissue is clear and distinguishable, which can effectively avoid the missed detection of defects due to insufficient contrast, such as small cracks being "blurred" and covered. The signal-to-noise ratio (SNR) reflects the ratio of image signal to noise, and 30dB means that the signal strength is more than 1000 times the noise. At this threshold, the effective signal of the workpiece, such as the defect outline and material difference, can significantly suppress random noise, such as detector electronic noise and radiation quantum noise, to avoid noise being misjudged as a defect or covering the real defect.

[0012] Further, in step S5, the exposure parameters are dynamically adjusted according to the type of non-compliance, specifically, when the overall overexposure, the tube voltage is reduced by 10% or the tube current is reduced by 15%; when the local underexposure, the partition increases the tube current value; when the signal-to-noise ratio is insufficient, the exposure time is extended by 20%. For different image quality problems such as overexposure, local underexposure, and insufficient signal-to-noise ratio, the specific parameters of tube voltage, tube current, and exposure time are accurately matched and adjusted to avoid blind adjustment. Using a quantitative adjustment ratio can reduce the number of parameter iterations, quickly make the image compliant, and shorten the detection cycle. When overexposed, the radiation energy or intensity is reduced, when underexposed, the local enhancement is increased, and when the noise is high, the exposure is extended, which cooperatively ensures that the image gray scale, contrast, and signal-to-noise ratio meet the standards and improves the defect recognition accuracy.

[0013] Further, the machine learning model in step S2 is generated by training historical detection data to associate the mapping relationship between the workpiece attributes and the optimal exposure parameters.

[0014] Further, in step S5, the dynamic adjustment adopts a PID control algorithm to perform closed-loop adjustment based on the target gray value.

[0015] Further, after step S6, it further includes: storing the compliant parameters and corresponding workpiece attributes into a database; when detecting similar workpieces, directly calling the parameters in the database to perform formal scanning.

[0016] The beneficial effects of the present application are: Strong mobility and scene adaptability: all components are integrated on a flat trolley with a universal fixable wheel, which can be flexibly moved to different scenes such as workshops and laboratories, solving the limitations of traditional detection devices, and is especially suitable for small workpiece detection in multiple stations.

[0017] Stable and controllable image quality: through the adaptive exposure control system, the tube voltage, tube current, and exposure time are dynamically adjusted based on PID control and reinforcement learning to ensure that the image gray scale distribution is in the middle of the dynamic range, the weld contrast is >0.1, and the signal-to-noise ratio SNR is >30dB, avoiding overexposure / underexposure problems caused by differences in workpiece material and thickness, and greatly improving the defect recognition accuracy.

[0018] Significant improvement in detection efficiency: the darkroom processing link of the traditional film method is omitted, the single workpiece detection time is greatly reduced, and the database stores historical optimal parameters, so that similar workpieces can directly call the parameters, and the detection efficiency is greatly improved compared with the conventional DR device; the workpiece two-dimensional movement and lifting positioning are realized through the electric translation table, the focal length of the probe telescopic arm can be accurately adjusted, and the visual operation table and manual clamping mechanism are matched to reduce the dependence on manual experience and labor intensity. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1It is the front view of the mobile small workpiece DR detection device in the application.

[0020] Figure 2 It is the side view of the mobile small workpiece DR detection device in the application.

[0021] Figure 3 It is the top view of the mobile small workpiece DR detection device in the application.

[0022] Figure 4 It is the flow chart of the adaptive exposure control method for DR detection in the application.

[0023] 1-probe telescopic arm, 2-flat panel detector, 3-workpiece clamping device, 4-workpiece clamping device rocker, 5-liftable platform, 6-transverse rail, 7-ray machine fixed support longitudinal rail, 8-probe telescopic arm support, 9-flat panel trolley, 10-universal fixable wheel, 11-electric control system operation table, 12-ray machine fixed support, 13-ray machine. DETAILED DESCRIPTION

[0024] The specific embodiments of the present application will be further described in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present application, but not to limit the scope of the present application.

[0025] It should be understood that the term "comprising" as used in the specification and the appended claims indicates the presence of the recited features, integers, steps, operations, elements, and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0026] For the sake of simplicity, only the parts related to the present application are shown in the drawings, which do not represent the actual structure of the product. In addition, in order to make the drawing simple and easy to understand, in some drawings, only one of the components with the same structure or function is shown schematically, or only one of them is marked. In this text, "one" not only means "only one", but also means "more than one" situation.

[0027] It should be further understood that the term "and / or" used in the present application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes these combinations.

[0028] In the embodiments shown in the drawings, the indications of direction, such as up, down, left, right, front and back, are not absolute but relative to explain the structure and movement of various components of the present application. When these components are in the position shown in the drawings, these descriptions are appropriate. If the position of these components changes, the indications of direction also change accordingly.

[0029] Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0030] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the specific implementation methods of the present invention will be described below with reference to the accompanying drawings. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings and other implementation methods can be obtained based on these drawings without any creative effort.

[0031] like Figures 1-3 As shown, the mobile small workpiece DR inspection device in this embodiment includes: a flatbed trolley 9 with universal, fixable wheels 10 at the bottom; a detector telescopic arm 1 (fixed to a detector telescopic arm bracket 8), with a flatbed detector 2 connected to its end; a longitudinal track 7 spanning the flatbed trolley 9, with a transverse track 6 on it, and a lifting platform 5 mounted on the transverse track 6, with a workpiece clamping device 3 and a manual crank 4 on the platform; an X-ray machine mounting bracket 12 mounted on the flatbed trolley 9, with an X-ray machine 13 fixed on it, the focal point being at the same height as the center of the flatbed detector 2; and an electrical control system control panel 11 integrated into the flatbed trolley 9, connecting all electric mechanisms. The electrical control system control panel 11 can electrically control the device. The control panel has 10 buttons, which control the power switch, detector extension / retraction, lifting platform elevation / retraction, vertical movement of the lifting platform, and horizontal movement of the lifting platform.

[0032] In a preferred embodiment, the above-mentioned device further includes an adaptive exposure technology system, comprising hardware components (X-ray source module, detector array, motion control unit, and intelligent control box) and software system (adaptive algorithm module, human-machine interface, and database management). The hardware component includes a micro-focus X-ray source with adjustable tube voltage (50-225KV), tube current (0.1-5mA), and exposure time (0.1-10s), supporting rapid parameter switching. The detector array employs a high-resolution amorphous silicon flat panel detector (pixel size ≤100μm) with real-time signal acquisition capabilities. The motion control unit utilizes a high-precision electric translation stage (positioning accuracy ±0.5mm), enabling automatic or manual workpiece positioning and scanning. The intelligent control box integrates an FPGA real-time processing chip, achieving millisecond-level parameter response. The software system's adaptive algorithm module is based on a deep learning-based exposure parameter prediction model. The human-machine interface uses a visual operation interface, supporting parameter setting and detection result analysis. The database management system can store detection parameters and historical data for different workpieces.

[0033] On the basis of the above scheme, preferably, in the hardware composition, the adjustable tube voltage of the ray source module can be adjusted in the range of 50-225KV with a step of 1KV; the adjustable tube current can be adjusted in the range of 0.1-5mA with a step of 0.1mA; and the exposure time can be adjusted in the range of 0.1-10s with a step of 0.01s.

[0034] On the basis of the above scheme, preferably, in the hardware composition, the detector array can guarantee that the image resolution is greater than or equal to 300 lp / cm, and the detection accuracy of defects is greater than or equal to 0.1 mm.

[0035] On the basis of the above scheme, preferably, in the hardware composition, the accurate control of the motion control unit can guarantee that the detection accuracy of defects is greater than or equal to 0.1 mm.

[0036] On the basis of the above scheme, preferably, in the hardware composition, the intelligent control box integrates an FPGA real-time processing chip, and can realize millisecond-level parameter response to guarantee that the detection time of each product is controlled within 30s, thereby greatly improving the detection efficiency.

[0037] On the basis of the above scheme, preferably, in the software system, the adaptive algorithm module adopts PID control, takes the target gray scale as a reference, and adjusts the exposure amount in a closed loop; and after long-term iteration, an optimal strategy database can be formed through reinforcement learning.

[0038] On the basis of the above scheme, preferably, in the software system, the man-machine interaction interface adopts a visual operation interface, and information such as the material and thickness of the workpiece and the sensitivity of the detector is input; based on the ML model of historical data, the system outputs the initial KV, mA and exposure time. Through the pre-scanning image quality evaluation index: whether the gray scale distribution is in the middle of the dynamic range; the contrast of the key area (weld) is greater than 0.1; and the signal-to-noise ratio SNR is greater than 30Db (to avoid noise from drowning defects). When the overall overexposure occurs, the adjustment strategy is to reduce the KV value by 10% or the mA value by 15% to reduce the photon flux; when the local (weld) underexposure occurs, the adjustment strategy is to increase the mA value in the partition to increase the local photon density; and when the noise is too high, the adjustment strategy is to prolong the exposure time by 20% to improve the signal accumulation. Such a cycle adjustment is continued until the image index requirements are met.

[0039] On the basis of the above scheme, preferably, in the software system, the database management system automatically stores the optimal exposure parameters of products of different materials and specifications, which can be called at any time during the detection of subsequent similar products.

[0040] The device in this embodiment is fixed to the support before detection, so that the focal point center is aligned with the center position of the flat panel detector. After the surface oil and impurities are removed through artificial cleaning treatment, the detected product is manually placed on the clamping device, the rocker is operated to clamp the workpiece, the detector telescopic button is operated to set the focal length, the platform longitudinal movement button and the transverse movement button are operated to make the workpiece close to the flat panel detector, and the basic information such as the workpiece material (such as metal density), thickness range, and detector sensitivity (such as CMOS / flat panel type) is input. The algorithm module calculates the optimal exposure parameters (KV, mA, exposure time) according to the input parameters. The X-ray machine is operated, and high-precision DR imaging exposure is performed on the detection part according to the calculated parameters. Then the image is automatically processed for noise reduction and contrast enhancement. The parameters are adjusted to form a database, which facilitates rapid detection of subsequent similar products. Since the positions of the workpiece, the detector, and the X-ray machine are relatively stable, the operation is simple, and therefore the imaging quality and detection efficiency can be greatly improved.

[0041] Figure 4 The adaptive exposure control method for DR detection in this embodiment includes the following steps: (a) receiving input work information, including workpiece material properties, thickness parameters, and detector sensitivity; (b) predicting initial exposure parameters based on the work information through a machine learning model; (c) performing a first low-dose scan using the initial exposure parameters to obtain a pre-scan image; (d) performing quality analysis on the pre-scan image and calculating the following indicators: - whether the gray scale distribution is in the middle interval of the detector dynamic range; - whether the contrast of the key area is greater than 0.1; - whether the image signal-to-noise ratio is greater than 30 dB; (e) if any indicator does not meet the standard, dynamically adjust the exposure parameters according to the type of non-compliance: - when the overall overexposure, reduce the tube voltage by 10% or the tube current by 15%; - when the local underexposure, increase the tube current value in the partition; - when the signal-to-noise ratio is insufficient, extend the exposure time by 20%; (f) repeat steps (c)-(e) until all quality indicators meet the standard; (g) perform a formal scan using the standard parameters.

[0042] As a preferred embodiment, the machine learning model in step (b) is generated by training historical detection data, and the mapping relationship between workpiece attributes and optimal exposure parameters is associated.

[0043] As a preferred embodiment, the dynamic adjustment in step (e) adopts a PID control algorithm to perform closed-loop adjustment based on the target gray value.

[0044] As a preferred embodiment, step (g) further comprises: - storing the qualified parameters and corresponding workpiece attributes into a database; - when detecting the same type of workpiece, directly calling the parameters in the database to perform formal scanning.

[0045] As a preferred embodiment, the critical area in step (d) is defined as the weld area or the defect-sensitive area of the workpiece. Specific embodiments

[0046] Aluminum alloy welding test plate detection (thickness 3mm) Workpiece information: aluminum alloy welding test plate (size 100mm x 50mm x 3mm), weld width 2mm, need to detect internal porosity and slag defects in the weld.

[0047] Detection steps: Clamping and positioning: place the test plate on the workpiece clamping device 3, clamp it by the manual rocker 4 (opening width 50mm); operate the electric control system console 11 to control the longitudinal rail 7 and the transverse rail 6 to move, so that the weld area is aligned with the center of the flat panel detector 2; adjust the lift platform 5 (Z axis) to keep the distance between the workpiece and the detector at 100mm.

[0048] Parameter input and initial prediction: input "material = aluminum alloy, thickness = 3mm, detector sensitivity = high" through the human-computer interaction interface, and the machine learning model outputs the initial parameters based on historical data: tube voltage 80KV, tube current 1.2mA, exposure time 0.5s.

[0049] Adaptive exposure adjustment: after the first low-dose scan, the image analysis shows that "weld contrast = 0.08, not qualified", the system starts the local underexposure adjustment strategy, and the partition increases the tube current to 1.5mA, after the second scan, the contrast reaches 0.12, meeting the requirements.

[0050] Formal scanning results: using the qualified parameters (80KV, 1.5mA, 0.5s) for scanning, the image signal-to-noise ratio is 35dB, and 0.15mm porosity defects are clearly identified; the whole process takes 22s, and the parameters are stored in the database.

[0051] It should be pointed out finally that the above embodiments are only used to illustrate the technical solutions of the present application but not to limit it. Although the present application has been described in detail with reference to the above embodiments, it should be understood by those skilled in the art that the specific embodiments of the present application can be modified or replaced equivalently without departing from the spirit and scope of the present application, and any modification or equivalent replacement should be covered in the protection scope of the claims of the present application.

Claims

1. A mobile small workpiece DR inspection device, characterized in that, It includes a movable support mechanism, a detection execution mechanism, and an adaptive exposure control system; the movable support mechanism is a flatbed trolley with omnidirectional fixed wheels; the detection execution mechanism includes a detector assembly, a workpiece positioning mechanism, and an X-ray emission mechanism, all integrated on the flatbed trolley; the adaptive exposure control system is used to dynamically adjust exposure parameters and control the coordinated operation of each mechanism.

2. The apparatus according to claim 1, characterized in that, The detector assembly includes a telescopic arm and a flat panel detector; the telescopic arm extends and retracts along the X-axis with an effective stroke of 500 mm; the flat panel detector is made of amorphous silicon with a pixel size ≤100 μm and an image resolution ≥300 lp / cm; the workpiece positioning mechanism includes a longitudinal track, a transverse track, a lifting platform, and a workpiece clamping device; the longitudinal track and the transverse track intersect perpendicularly and are driven by an electronic control system to achieve two-dimensional movement; the lifting platform moves and retracts along the Z-axis with a maximum stroke of 162 mm; the workpiece clamping device has a maximum opening of 200 mm and clamps the workpiece using a manual crank.

3. The apparatus according to claim 1, characterized in that, The X-ray emitting mechanism includes an X-ray machine mounting bracket and an X-ray machine; the mounting bracket locks the X-ray machine in place using two 250mm diameter semi-circular clamps and four butterfly nuts; the tube voltage of the X-ray machine is adjustable from 50-225KV, the tube current is adjustable from 0.1-5mA, and the exposure time is adjustable from 0.1-10s.

4. The apparatus according to claim 1, characterized in that, The hardware of the adaptive exposure control system includes an intelligent control box and an electronic control system operating console. The intelligent control box integrates an FPGA real-time processing chip with a response time in milliseconds. The operating console has 10 control buttons to control the power switch, detector extension and retraction, platform lifting and moving. The adaptive exposure control system includes an adaptive algorithm module, which is based on PID control and reinforcement learning. It uses the target grayscale as a reference to adjust the exposure in a closed loop and iterates over a long period to form an optimal strategy database.

5. An adaptive exposure control method for DR detection, applied to the apparatus of any one of claims 1-4, comprising the steps of: S1 receives input working information, including workpiece material properties, thickness parameters, and detector sensitivity; S2, Based on the working information, the initial exposure parameters are predicted using a machine learning model; S3, Perform the first low-dose scan using the initial exposure parameters to obtain a pre-scan image; S4, perform quality analysis on the pre-scanned image, calculate grayscale distribution, key area contrast and image signal-to-noise ratio related parameters, and determine whether they meet the standards; S5. If any indicator fails to meet the standard, the exposure parameters will be dynamically adjusted according to the type of failure. S6, perform the formal scan using the compliant parameters.

6. The method according to claim 5, characterized in that, In step S4, calculating the grayscale distribution, key area contrast, and image signal-to-noise ratio related parameters and determining whether they meet the standards specifically includes: whether the grayscale distribution is in the middle range of the detector's dynamic range; and whether the key area contrast is greater than 0.

1. Whether the image signal-to-noise ratio is greater than 30dB, the key area is defined as the weld area or the workpiece defect sensitive area.

7. The method according to claim 5, characterized in that, In step S5, dynamically adjusting the exposure parameters according to the type of non-compliance is specifically as follows: when the overall exposure is overexposed, reduce the tube voltage by 10% or the tube current by 15%; when the local exposure is underexposed, increase the tube current value in the corresponding area; when the signal-to-noise ratio is insufficient, extend the exposure time by 20%.

8. The method according to claim 5, characterized in that, The machine learning model described in step S2 is generated by training on historical detection data and associates the mapping relationship between workpiece attributes and optimal exposure parameters.

9. The method according to claim 5, characterized in that, The dynamic adjustment in step (e) uses a PID control algorithm to perform closed-loop adjustment based on the target gray value.

10. The method as described in claim 5, characterized in that, Step S7 is followed by: storing the compliance parameters and corresponding workpiece attributes in the database; when detecting similar workpieces, directly calling the parameters in the database to perform a formal scan.

Images