An online X-ray nondestructive testing method and system

Abstract

The present invention discloses an online X-ray nondestructive flaw detection method and system belonging to the field of flaw detection technology. The method includes: when the part to be detected triggers an infrared beam emitted downward by a diffuse reflection light barrier, obtaining an image to be detected containing structural features and inputting it into a pre-trained image model for matching. If the image matching value corresponding to the image to be detected obtained by matching is greater than the image matching threshold, the match is passed and the interference of the structural features is removed. The local dynamic threshold comparison and regional grayscale value comparison of the area to be identified are generated to determine whether the part to be detected is qualified. The present invention realizes segmented asynchronous high-definition online detection of defects, strains, fractures, etc. of the part to be detected by performing model matching, removing interference, and calculating the brightness and grayscale of the identification area on the image to be detected, thereby reducing manpower input and shortening the overall detection time, thereby improving production quality and production efficiency.

Description

Technical Field

[0001] The present invention belongs to the technical field of flaw detection and relates to an online X-ray non-destructive flaw detection method and a system thereof. Background Art

[0002] In today's industrial production, products with appearance requirements need to be inspected during the production process to check whether there are defects such as blemishes, scratches, cracks, etc. on the surface of the products to prevent them from entering the market. However, manual inspections on traditional assembly lines are often prone to omissions, missed inspections, wrong inspections, and other inadequate results, causing certain production impacts.

[0003] Most existing X-ray scanners on the market use linear array scanning, which has low spatial resolution and clarity. This makes it difficult to detect and identify minor defects on the inspection parts, leading to frequent missed or incorrect detections. Flatbed X-ray scanners, however, require stitching of large objects due to the size of the flatbeds before detecting defects. This stitching process can reduce overall image clarity, and delaying defect detection until stitching is complete can significantly impact the overall inspection process, reducing efficiency and failing to meet the speed of an assembly line.

[0004] Therefore, engineers and technicians in this field need to develop a high-definition online X-ray non-destructive testing method and system that can quickly and automatically learn and achieve segmented asynchronous detection without splicing the entire test piece image. It can detect defects, strains, fractures and other positions on the test piece to improve production efficiency. Summary of the Invention

[0005] The present invention provides an online X-ray nondestructive testing method and system thereof, aiming to solve the problems raised in the background technology and realize segmented asynchronous high-definition online detection of defects, strains, fractures and other positions of the test parts, so as to reduce manpower input and shorten the overall detection time, thereby improving production quality and production efficiency.

[0006] To achieve the above object, the present invention provides an online X-ray non-destructive testing method, which includes the following steps:

[0007] Acquisition step: When the object to be inspected triggers the infrared beam emitted downward by the diffuse reflection light barrier, the structural features of the object to be inspected are collected by an X-ray light source and input into the image detection system to obtain an image to be inspected corresponding to the object to be inspected containing the structural features. The structural features include the outer contour, inner hole contour, size, sink, fold, height, and surface texture;

[0008] Matching step: inputting the image to be detected containing the structural features into a pre-trained image model for matching. If the image matching value corresponding to the image to be detected obtained by matching is greater than the image matching threshold, the matching is successful;

[0009] Processing step: removing interference from structural features in the image to be detected that has passed the matching, and generating a region to be identified contained in the image to be detected;

[0010] Calculation step: Calculate the local dynamic threshold contrast and regional grayscale value contrast of the area to be identified through image processing. When the local dynamic threshold contrast is greater than a first preset ratio, and / or the regional grayscale value contrast is greater than a second preset ratio, it indicates that the area to be identified is a defect, a strain or a fracture area, and the part to be inspected is judged to be unqualified.

[0011] Preferably, before the collecting step, the method further comprises:

[0012] Conveying step: The parts to be inspected are placed on the production line, and the conveyor belt drives the parts to be inspected into the inspection channel, above which a diffuse reflection light barrier, an X-ray light source and an image processing system are installed.

[0013] Preferably, after the calculation step, the method further includes:

[0014] Marking step: marking the unqualified position of the identification area where the unqualified part is located and generating unqualified information, and removing it from the production line when entering the next station.

[0015] Preferably, the calculation of the to-be-identified area by image processing includes regional grayscale enhancement, grayscale value comparison, and local dynamic threshold comparison, wherein the formula is as follows:

[0016] Regional grayscale enhancement: dst = round((src0–src1)*coef)+src0

[0017] Where src0 represents input image 0, src1 represents input image 1, dst represents the processing result, coef represents the enhancement coefficient, and round represents rounding.

[0018] Local dynamic threshold comparison: for bright areas, io>=it+shift; for dark areas, io<=it-shift; for areas with stable grayscale values, it-shift<=io<=it+shift; for areas with large grayscale changes, io<=it-shift or io>=it+shift;

[0019] Among them, io represents the original image, it represents the threshold image, and shift represents the threshold;

[0020] Gray value comparison: dst = src > th for bright areas and dst = src for dark areas <th;

[0021] Among them, src represents the original image, th represents the threshold, and dst represents the processing result.

[0022] Preferably, the matching step includes:

[0023] According to the image model, matching detection can be performed on the detection image within a set angle range;

[0024] According to the position, angle, distortion and deformation information after the image model matching detection, the image model is transformed so that the image model matches the corresponding position of the image to be tested;

[0025] Performing correction operations such as removing, filling, shifting, scaling, or filtering image features of the image to be detected according to the image model;

[0026] Performing right angle, arc and inner hole correction on the image to be detected according to the image model; and

[0027] The image to be detected is corrected by region according to the image model.

[0028] Preferably, the image model training step includes:

[0029] Acquire several parts to be inspected and take photos of them, perform grayscale processing on the pictures of the parts to be inspected, and generate images to be inspected corresponding to the parts to be inspected;

[0030] After feature marking the image to be detected, an image detection area and an image feature marking area corresponding to the marked image to be detected are obtained; and

[0031] The labeled image features are trained to generate single-layer or multi-layer feature information to obtain the image model.

[0032] Preferably, when the acquisition area of ​​the part to be inspected exceeds the area irradiated downward by the X-ray light source, the segmented image acquisition method is used to automatically calculate the moving distance according to the length of the inspection part. After the inspection part is moved to the appropriate position, the X-ray light source is used to acquire the various areas contained in the part to be inspected and input them into the image detection system for detection.

[0033] Preferably, the image detection system performs detection including:

[0034] After the X-ray light source has collected the first region reg1 included in the object to be inspected;

[0035] The following judgment operations are performed while performing image model matching on the first image to be detected corresponding to the first region reg1, removing interference from structural features in the image to be detected, calculating the regional brightness value and regional grayscale value of the identification area, and judging whether the current processing area of ​​the part to be detected is qualified:

[0036] Operation 1: If the inspection piece is too long, before completing the acquisition of the entire inspection piece, the moving distance is automatically calculated based on the length of the inspection piece, and after the inspection piece is moved to a suitable position, the second region reg2 included in the inspection piece is acquired by the X-ray light source;

[0037] Operation 2: If the acquisition area of ​​the test piece is within the area irradiated downward by the X-ray light source, the conveyor belt is started to wait for the next test piece until the acquisition of the next test piece is completed, and the area reg2 of the second test piece is obtained;

[0038] While calculating the model matching of the second region to be identified reg2, removing interference, and judging whether the position is qualified, the collection process of the next region or the next detection part is started according to the judgment of operation 1 and operation 2.

[0039] In addition, the present invention also provides an online X-ray non-destructive flaw detection system, which uses the online X-ray non-destructive flaw detection method described in any one of the above, and the system includes:

[0040] An acquisition module is configured to acquire structural features of the part to be inspected through an X-ray light source when the part to be inspected triggers an infrared beam emitted downward by a diffuse reflection light barrier, and input the captured features into an image detection system to obtain an image to be inspected corresponding to the part to be inspected, including structural features such as outer contour, inner hole contour, size, depression, fold, height, and surface texture.

[0041] A matching module is used to input the obtained image to be detected containing structural features into a pre-trained image model for matching. If the image matching value corresponding to the image to be detected obtained by matching is greater than the image matching threshold, the matching is successful;

[0042] a processing module, configured to remove interference from structural features in the image to be detected that has passed the matching, and generate an area to be identified contained in the image to be detected;

[0043] a calculation module, configured to calculate, through image processing, a local dynamic threshold contrast and a regional grayscale value contrast of the area to be identified; when the local dynamic threshold contrast is greater than a first preset ratio, and / or the regional grayscale value contrast is greater than a second preset ratio, indicating that the area to be identified is a defective, strained, or broken area, the part to be inspected is judged to be unqualified;

[0044] The marking module is used to mark the unqualified position of the identification area where the unqualified part to be tested is located and generate unqualified information, and remove it from the production line when entering the next workstation.

[0045] Preferably, the method further includes:

[0046] The conveying module is used to place the parts to be inspected on the production line. The conveyor belt drives the parts to be inspected into the inspection channel. A diffuse reflection light barrier, an X-ray light source and an image processing system are installed above the inspection channel.

[0047] The beneficial effects of the present invention compared to the prior art are as follows:

[0048] The present invention provides an online X-ray nondestructive flaw detection method and system. When an object to be detected triggers an infrared beam emitted downward by a diffuse reflection barrier, the X-ray light source captures the structural features of the object to be detected and inputs them into an image detection system. An image to be detected corresponding to the object to be detected, containing the structural features, is obtained and input into a pre-trained image model for matching. If the image matching value corresponding to the image to be detected obtained by matching is greater than an image matching threshold, the match is successful, interference from the structural features is removed, and a region to be identified contained in the image to be detected is generated. Image processing is then used to calculate a local dynamic threshold contrast and a regional grayscale value contrast of the region to be identified. When the local dynamic threshold contrast is greater than a first preset ratio and / or the regional grayscale value contrast is greater than a second preset ratio, indicating that the region to be identified is a defect, strain, or fracture region, the object to be detected is judged as unqualified. By performing model matching, removing interference, and calculating the brightness and grayscale of the identified region on the image to be detected, the present invention achieves segmented, asynchronous, high-definition online detection of defects, strains, fractures, and other locations on the object to be detected, thereby reducing manpower input and shortening overall detection time, thereby improving production quality and efficiency.

[0049] In order to more clearly illustrate the structural features and effects of the present invention, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 1 is a schematic diagram of a first process flow of a preferred embodiment of the online X-ray non-destructive testing method of the present invention;

[0051] Figure 2 Schematic diagram of a second process flow of a preferred embodiment of the online X-ray non-destructive testing method of the present invention;

[0052] Figure 3 Schematic diagram of the program modules of the online X-ray non-destructive testing system of the present invention;

[0053] Figure 4Schematic diagram of the collection and detection process in the online X-ray non-destructive testing method of the present invention; DETAILED DESCRIPTION

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meanings as commonly understood by those skilled in the art to which this application belongs. The terms used in the specification of the application are only for the purpose of describing specific embodiments and are not intended to limit this application. The terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of this application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0055] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0056] To achieve the above object, the present invention provides an online X-ray non-destructive testing method. Figure 1 FIG. 1 is a schematic diagram of a first process flow of a preferred embodiment of an online X-ray nondestructive testing method of the present invention. The testing method includes the following steps:

[0057] S1: When the part to be inspected triggers the infrared beam emitted downward by the diffuse reflection light barrier, the structural features of the part to be inspected are collected by an X-ray light source and input into the image detection system to obtain the image to be inspected corresponding to the part to be inspected containing the structural features. The structural features include the outer contour, inner hole contour, size, sink, fold, height and surface texture.

[0058] In this embodiment, the part to be inspected can be a circular, rectangular, or irregular metal casting, sheet metal, or plastic plate. The X-ray light source utilizes the ability of X-rays (which can also be gamma rays or other high-energy rays) to penetrate materials. Due to the different absorption and scattering effects of materials on X-rays, the sensitivity of the film varies, thus forming images of varying blackness on the film, which can be used to determine material defects. Of course, X-rays can also be replaced by gamma rays or other high-energy rays. The diffuse reflection light barrier is an infrared beam installed on both sides of the X-ray light source and emits downward. When the part to be inspected contacts the infrared beam, the diffuse reflection light barrier sends a signal to the X-ray light source indicating that the part to be inspected has entered. The X-ray light source then begins to collect structural features of the part to be inspected and inputs them into the image detection system to obtain an image of the part to be inspected. The structural features include the outer contour, inner hole contour, size, depression, folded edge, height, and surface texture.

[0059] Furthermore, when the collection area of ​​the part to be inspected exceeds the area irradiated downward by the X-ray light source, the segmented image collection method is used to automatically calculate the moving distance according to the length of the inspection part. After the inspection part is moved to the appropriate position, the X-ray light source is used to collect the various areas included in the part to be inspected and input them into the image detection system for detection.

[0060] It should be noted that the collection area of ​​the part to be detected exceeds the area irradiated downward by the X-ray light source means that the length or width of the part to be detected exceeds the diameter of the area irradiated downward by the X-ray light source. The part to be detected can be a short part to be detected or a long part to be detected. For a long part to be detected, it can be divided into multiple detection areas.

[0061] The image detection system performs detection including:

[0062] After the X-ray light source has collected the first region reg1 included in the object to be inspected;

[0063] The following judgment operations are performed while performing image model matching on the first image to be detected corresponding to the first region reg1, removing interference from structural features in the image to be detected, calculating the regional brightness value and regional grayscale value of the identification area, and judging whether the current processing area of ​​the part to be detected is qualified:

[0064] Operation 1: If the inspection piece is too long, before completing the acquisition of the entire inspection piece, the moving distance is automatically calculated based on the length of the inspection piece, and after the inspection piece is moved to a suitable position, the second region reg2 included in the inspection piece is acquired by the X-ray light source;

[0065] Operation 2: If the acquisition area of ​​the test piece is within the area irradiated downward by the X-ray light source, the conveyor belt is started to wait for the next test piece until the acquisition of the next test piece is completed, and the area reg2 of the second test piece is obtained;

[0066] While calculating the model matching, interference removal and determining whether the position is qualified for the second region to be recognized reg2, start the acquisition process of the next region or the next test piece according to the judgments of operation 1 and operation 2.

[0067] In this embodiment, as Figure 4 shown, it is a schematic diagram of the acquisition and detection processing in the online X-ray non-destructive testing method of the present invention. The asynchronous processing mode is used for simultaneous detection of multiple regions. Specifically, during the movement of the test piece to be detected and the acquisition of structural features, the previously acquired test image is synchronously processed, which can reduce the overall time of the test piece to be detected. For example, assuming that t1 is the scanning time and t2 is the detection time, the total time required for each test piece image in the synchronous processing mode is T = t1 + t2, while in the asynchronous processing mode, the acquisition and processing are separated. After the image of the test piece to be detected (for example, the first region to be detected) is acquired, the image acquisition operation of the next test piece to be detected (for example, the second region to be detected of the same test) can be performed without waiting for the result. Thus, it can be seen that the total time required for the image acquisition and detection of the test piece to be detected in the asynchronous processing mode is T < t1 + t2. In the production of pipeline operations, it can shorten the overall detection time, which is beneficial to pipeline operations and improve production quality and production efficiency.

[0068] S2: Input the obtained test image containing structural features into a pre-trained image model for matching. If the image matching value corresponding to the obtained test image is greater than the image matching threshold, the matching is passed.

[0069] Further, the S2 includes:

[0070] The test image can be matched and detected within a set angle range according to the image model;

[0071] According to the information of the position, angle, distortion and deformation after the matching detection of the image model, the image model is transformed so that the image model matches the corresponding position of the待测图像;

[0072] Perform operations such as removal, filling, offset, scaling or screening correction on the image features of the test image according to the image model;

[0073] Perform corrections on the right angle, arc and inner hole of the test image according to the image model; and

[0074] Perform regional correction on the test image according to the image model.

[0075] Further, the training steps of the image model include:

[0076] A1: Acquire several parts to be inspected and take photos of them, perform grayscale processing on the photos of the parts to be inspected, and generate images to be inspected corresponding to the parts to be inspected;

[0077] A2: After feature marking the image to be detected, an image detection area and an image feature marking area corresponding to the marked image to be detected are obtained; and

[0078] A3: Training the labeled image features to generate single-layer or multi-layer feature information to obtain the image model.

[0079] S3: removing interference from the structural features in the image to be detected that have passed the matching, and generating a region to be identified contained in the image to be detected.

[0080] For example, in an optional embodiment, the image to be inspected obtained from a rectangular sheet metal part to be inspected includes the image's contour, size, inner hole, folded edge, surface texture, convex hull and sunken platform, etc., and may include an image with defects, scratches, fractures and other positions, wherein the image features include the contour, surface texture, size, holes, folded edges, convex hull and sunken platform on the part to be inspected. After removing these image features, what remains is the area to be identified.

[0081] S4: Calculate the local dynamic threshold contrast and regional grayscale value contrast of the area to be identified through image processing. When the local dynamic threshold contrast is greater than a first preset ratio, and / or the regional grayscale value contrast is greater than a second preset ratio, it indicates that the area to be identified is a defect, a strain or a fracture area, and the part to be inspected is judged to be unqualified.

[0082] The calculation of the area to be identified by image processing includes regional grayscale enhancement, grayscale value comparison, and local dynamic threshold comparison, wherein the formula is as follows:

[0083] Regional grayscale enhancement: dst = round((src0–src1)*coef)+src0

[0084] Where src0 represents input image 0, src1 represents input image 1, dst represents the processing result, coef represents the enhancement coefficient, and round represents rounding.

[0085] Local dynamic threshold comparison: for bright areas, io>=it+shift; for dark areas, io<=it-shift; for areas with stable grayscale values, it-shift<=io<=it+shift; for areas with large grayscale changes, io<=it-shift or io>=it+shift;

[0086] Among them, io represents the original image, it represents the threshold image, and shift represents the threshold;

[0087] Gray value comparison: dst = src > th for bright areas and dst = src for dark areas <th;

[0088] Among them, src represents the original image, th represents the threshold, and dst represents the processing result.

[0089] For further reference, Figure 2 FIG. 1 is a second flow chart of a preferred embodiment of the online X-ray non-destructive testing method of the present invention, which further includes the following steps before step S1:

[0090] S0: The parts to be inspected are placed on the production line. The conveyor belt drives the parts to be inspected into the inspection channel. A diffuse reflection light barrier, an X-ray light source and an image processing system are installed above the inspection channel.

[0091] And, after the S4 step, the method further includes:

[0092] S5: Mark the unqualified position of the identification area where the unqualified part to be inspected is located and generate unqualified information, and remove it from the production line when entering the next workstation.

[0093] In addition, the present invention also provides an online X-ray non-destructive testing system, referring to Figure 3 FIG. 1 is a schematic diagram of a program module of an online X-ray nondestructive flaw detection system according to the present invention. The online X-ray nondestructive flaw detection method described in any one of the above is applied, and the system includes:

[0094] The conveying module 10 is used to place the parts to be inspected on the production line. The conveyor belt drives the parts to be inspected into the inspection channel. A diffuse reflection light barrier, an X-ray light source and an image processing system are installed above the inspection channel.

[0095] The acquisition module 20 is configured to acquire structural features of the object to be inspected by an X-ray light source when the object to be inspected triggers the infrared beam emitted downward by the diffuse reflection light barrier, and input the captured features into the image detection system to obtain an image to be inspected corresponding to the object to be inspected, which includes the structural features, such as the outer contour, inner hole contour, size, depression, fold, height, and surface texture.

[0096] The matching module 30 is used to input the image to be detected containing the structural features into a pre-trained image model for matching. If the image matching value corresponding to the image to be detected obtained by matching is greater than the image matching threshold, the matching is successful;

[0097] A processing module 40 is configured to remove interference from structural features in the image to be detected that have been matched, and generate a region to be identified contained in the image to be detected;

[0098] a calculation module 50 for calculating, through image processing, a local dynamic threshold contrast and a regional grayscale value contrast of the area to be identified; when the local dynamic threshold contrast is greater than a first preset ratio and / or the regional grayscale value contrast is greater than a second preset ratio, indicating that the area to be identified is a defect, a strain, or a fracture area, the part to be inspected is judged as unqualified;

[0099] The marking module 60 is used to mark the unqualified position of the identification area where the unqualified part to be inspected is located and generate unqualified information, so that the part is removed from the production line when entering the next workstation.

[0100] The beneficial effects of the present invention compared to the prior art are as follows:

[0101] The present invention provides an online X-ray nondestructive flaw detection method and system. When an object to be detected triggers an infrared beam emitted downward by a diffuse reflection barrier, the X-ray light source captures the structural features of the object to be detected and inputs them into an image detection system. An image to be detected corresponding to the object to be detected, containing the structural features, is obtained and input into a pre-trained image model for matching. If the image matching value corresponding to the image to be detected obtained by matching is greater than an image matching threshold, the match is successful, interference from the structural features is removed, and a region to be identified contained in the image to be detected is generated. Image processing is then used to calculate a local dynamic threshold contrast and a regional grayscale value contrast of the region to be identified. When the local dynamic threshold contrast is greater than a first preset ratio and / or the regional grayscale value contrast is greater than a second preset ratio, indicating that the region to be identified is a defect, strain, or fracture region, the object to be detected is judged as unqualified. By performing model matching, removing interference, and calculating the brightness and grayscale of the identified region on the image to be detected, the present invention achieves segmented, asynchronous, high-definition online detection of defects, strains, fractures, and other locations on the object to be detected, thereby reducing manpower input and shortening overall detection time, thereby improving production quality and efficiency.

[0102] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better embodiment. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0103] The technical principles of the present invention have been described above in conjunction with specific embodiments, which are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions based on the principles of the present invention fall within the scope of protection of the present invention. Those skilled in the art will be able to conceive of other specific embodiments of the present invention without inventive effort and will fall within the scope of protection of the present invention.

Claims

1. An online X-ray nondestructive testing method, characterized in that: The detection method comprises the following steps: Acquisition step: When the object to be inspected triggers the infrared beam emitted downward by the diffuse reflection light barrier, the structural features of the object to be inspected are collected by an X-ray light source and input into the image detection system to obtain an image to be inspected corresponding to the object to be inspected containing the structural features. The structural features include the outer contour, inner hole contour, size, sink, fold, height and surface texture; Matching step: inputting the image to be detected containing structural features into a pre-trained image model for matching. If the image matching value corresponding to the image to be detected obtained by matching is greater than the image matching threshold, the matching is successful. The matching step includes: According to the image model, matching detection can be performed on the detection image within a set angle range; According to the position, angle, distortion and deformation information after the image model matching detection, the image model is transformed so that the image model matches the corresponding position of the image to be tested; Performing correction operations such as removing, filling, shifting, scaling, or filtering image features of the image to be detected according to the image model; Performing right angle, arc and inner hole correction on the image to be detected according to the image model; and Performing region-by-region correction on the image to be detected according to the image model; Processing step: removing interference from structural features in the image to be detected that has passed the matching, and generating a region to be identified contained in the image to be detected; Calculation step: calculating the local dynamic threshold contrast and regional grayscale value contrast of the area to be identified through image processing; when the local dynamic threshold contrast is greater than a first preset ratio and / or the regional grayscale value contrast is greater than a second preset ratio, it indicates that the area to be identified is a defect, a strained area, or a fractured area, and the part to be inspected is determined to be unqualified; The method further includes, when the acquisition area of ​​the part to be inspected exceeds the area irradiated downward by the X-ray light source, automatically calculating the moving distance according to the length of the part to be inspected by using a segmented image acquisition method, and after the part to be inspected is moved to a suitable position, then using the X-ray light source to acquire images of various areas of the part to be inspected and inputting them into the image detection system for inspection respectively; The image detection system performs detection including: After the X-ray light source has collected the first region reg1 included in the object to be inspected; The following judgment operations are performed while performing image model matching on the first image to be detected corresponding to the first region reg1, removing interference from structural features in the image to be detected, calculating the regional brightness value and regional grayscale value of the identification area, and judging whether the current processing area of ​​the part to be detected is qualified: Operation 1: If the inspection piece is too long, before completing the acquisition of the entire inspection piece, the moving distance is automatically calculated based on the length of the inspection piece, and after the inspection piece is moved to a suitable position, the second region reg2 included in the inspection piece is acquired by the X-ray light source; Operation 2: If the detection piece acquisition area is within the area of ​​the X-ray light source irradiating downward, the conveyor is started to wait for the next detection piece until the next detection piece acquisition is completed to obtain the second detection piece area reg2; While calculating the model matching of the second region to be identified reg2, removing interference, and determining whether the position is qualified, start the acquisition process of the next region or the next detection part according to the judgments of operation 1 and operation 2.

2. The online X-ray nondestructive testing method according to claim 1, characterized in that: Before the acquisition step, it further includes: Transfer step: Place the part to be detected on the production line, and the conveyor belt drives the part to be detected into the detection channel. A diffuse reflection light barrier, an X-ray source, and an image processing system are installed above the detection channel.

3. The online X-ray nondestructive testing method according to claim 1, characterized in that: After the calculation step, it further includes: Marking step: Mark the unqualified position of the identification region where the part to be detected judged as unqualified is located and generate unqualified information, and remove it from the production line when entering the next station.

4. The online X-ray nondestructive testing method according to claim 1, characterized in that: The calculation of the region to be identified by image processing includes region gray-scale enhancement, region gray-scale value comparison, and local dynamic threshold comparison. The formulas are as follows: Region gray-scale enhancement: dst = round((src0 – src1) * coef) + src0 Where, src0 represents the input image 0, src1 represents the input image 1, dst represents the processing result, coef represents the enhancement coefficient, and round represents rounding; Local dynamic threshold comparison: bright judgment io >= it + shift, dark area judgment io <= it - shift, gray-scale value stable area judgment it - shift < io < it + shift, area with large changes judgment io <= it - shift or io >= it + shift; Where, io represents the original image, it represents the threshold image, and shift represents the threshold; Region gray-scale value comparison: bright area judgment src > th, dark area judgment src < th; Where, src represents the original image, and th represents the threshold.

5. The online X-ray nondestructive testing method according to claim 1, characterized in that: The training step of the image model includes: Take pictures of several parts to be detected, perform gray-scale processing on the pictures of the parts to be detected, and generate the detection images corresponding to the parts to be detected; After feature marking of the detection images, obtain the image detection area and the image feature marking area corresponding to the marked detection images; and Train the marked image features to generate single-layer or multi-layer feature information, and obtain the image model.

6. An online X-ray non-destructive flaw detection system, applying the online X-ray non-destructive flaw detection method according to any one of claims 1 to 5, characterized in that: The system includes: An acquisition module, which is used to collect the structural features of the part to be detected and input them into the image detection system through the X-ray source when the part to be detected triggers the infrared beam emitted downward by the diffuse reflection light barrier, and obtain the detection image containing structural features corresponding to the part to be detected. The structural features include outer contour, inner hole contour, dimensions, counterbore, hem, height, and surface texture; A matching module, which is used to input the detection image containing structural features obtained into a pre-trained image model for matching. If the image matching value corresponding to the detection image obtained by matching is greater than the image matching threshold, the matching passes. The matching step includes: The detection image can be matched and detected within a set angle range according to the image model; According to the position, angle, distortion and deformation information after the image model matching detection, the image model is transformed so that the image model matches the corresponding position of the image to be tested; Performing correction operations such as removing, filling, shifting, scaling, or filtering image features of the image to be detected according to the image model; Performing right angle, arc and inner hole correction on the image to be detected according to the image model; and Performing region-by-region correction on the image to be detected according to the image model; a processing module, configured to remove interference from structural features in the image to be detected that has passed the matching, and generate an area to be identified contained in the image to be detected; a calculation module, configured to calculate, through image processing, a local dynamic threshold contrast and a regional grayscale value contrast of the area to be identified; when the local dynamic threshold contrast is greater than a first preset ratio, and / or the regional grayscale value contrast is greater than a second preset ratio, indicating that the area to be identified is a defective, strained, or broken area, the part to be inspected is judged to be unqualified; A marking module is used to mark the unqualified position of the identification area where the unqualified part is located and generate unqualified information, and remove it from the production line when entering the next workstation; The system further includes a segmentation module for automatically calculating the moving distance according to the length of the inspection piece by using a segmented image acquisition method when the acquisition area of ​​the inspection piece exceeds the area irradiated downward by the X-ray light source. After the inspection piece is moved to a suitable position, the X-ray light source is used to acquire images of each area of ​​the inspection piece and input the images into the image detection system for detection. The asynchronous module, used for the image detection system to perform detection, includes: After the X-ray light source has collected the first region reg1 included in the object to be inspected; The following judgment operations are performed while performing image model matching on the first image to be detected corresponding to the first region reg1, removing interference from structural features in the image to be detected, calculating the regional brightness value and regional grayscale value of the identification area, and judging whether the current processing area of ​​the part to be detected is qualified: Operation 1: If the inspection piece is too long, before completing the acquisition of the entire inspection piece, the moving distance is automatically calculated based on the length of the inspection piece, and after the inspection piece is moved to a suitable position, the second region reg2 included in the inspection piece is acquired by the X-ray light source; Operation 2: If the detection piece acquisition area is within the area of ​​the X-ray light source irradiating downward, the conveyor is started to wait for the next detection piece until the next detection piece acquisition is completed to obtain the second detection piece area reg2; While calculating the model matching of the second region to be identified reg2, removing interference, and judging whether the position is qualified, the collection process of the next region or the next detection part is started according to the judgment of operation 1 and operation 2.

7. The online X-ray nondestructive testing system according to claim 6, characterized in that: The acquisition module also includes: The conveying module is used to place the parts to be inspected on the production line. The conveyor belt drives the parts to be inspected into the inspection channel. A diffuse reflection light barrier, an X-ray light source and an image processing system are installed above the inspection channel.

Images