Titanium alloy bar quality detection method and system

Abstract

The present invention relates to the field of data processing technology, and specifically discloses a titanium alloy bar quality detection method and system, which realizes the improvement of the efficiency of defect detection of the internal structure of the titanium alloy bar, and improves the accuracy of the quality detection of the titanium alloy bar. The specific method includes: first, using an X-ray detection device to collect the transverse X-ray image and the longitudinal X-ray image of the bar according to the projection strategy, and extracting the first grayscale image of the transverse X-ray image and the second grayscale image of the longitudinal X-ray image respectively; then, the first grayscale image and the second grayscale image are matched with the target image position to obtain the degree of deviation of the target image position containing abnormal information; then, the projection strategy is adjusted according to the degree of deviation of the target image position to obtain the adjusted first grayscale image and the second grayscale image; finally, the adjusted first grayscale image and the second grayscale image are analyzed for abnormal information recognition and the bar quality detection result is obtained.

Description

Technical Field

[0001] The invention relates to the technical field of data processing, and in particular to a titanium alloy bar quality detection method and system. Background Art

[0002] As an important metal material, titanium alloy bars are widely used in aerospace, marine engineering, chemical industry and other fields. With the development of these fields, traditional quality inspection methods that rely solely on manual visual inspection and sampling inspection are not only inefficient but also prone to misjudgment.

[0003] Although existing non-destructive testing technologies, such as ultrasonic testing, X-ray testing, eddy current testing, etc., can detect internal defects, their ability to detect tiny defects is limited. For titanium alloy bars with higher production precision requirements, tiny defects can easily pose a major safety hazard to the subsequent normal use of titanium alloy bars.

[0004] Common X-ray inspection forms an image by observing the change in X-ray intensity after passing through an object, thereby showing the internal structure and defects of titanium alloy bars. X-ray inspection that only uses a single direction is slow, resulting in increased time costs; and direct X-ray inspection lacks the sensitivity to detect tiny defects inside titanium alloy bars, resulting in unqualified quality of the obtained titanium alloy bars. Summary of the invention

[0005] The purpose of the present invention is to provide a titanium alloy bar quality detection method and system to solve the following technical problems:

[0006] How to improve the efficiency of defect detection of the internal structure of titanium alloy bars and improve the accuracy of quality inspection of titanium alloy bars.

[0007] The purpose of the present invention can be achieved through the following technical solutions:

[0008] A titanium alloy bar quality detection method, the method comprising:

[0009] S1, collecting a transverse X-ray image and a longitudinal X-ray image of the bar material according to a projection strategy by an X-ray detection device, and extracting a first grayscale image of the transverse X-ray image and a second grayscale image of the longitudinal X-ray image respectively;

[0010] S2, performing target image position matching on the first grayscale image and the second grayscale image to obtain a target image position deviation degree containing abnormal information;

[0011] S3, adjusting the projection strategy according to the degree of position deviation of the target image to obtain an adjusted first grayscale image and a second grayscale image;

[0012] S4. Perform abnormal information recognition and analysis on the adjusted first grayscale image and the second grayscale image and obtain the bar quality inspection result.

[0013] Preferably, the projection strategy includes:

[0014] Set the initial projection focus radius, initial exposure frequency and initial exposure interval duration;

[0015] And adjust the initial projection focus radius, initial exposure frequency and initial exposure interval duration according to the degree of deviation of the target image position.

[0016] Preferably, S2 includes:

[0017] Obtaining a coincidence line between the first grayscale image and the second grayscale image;

[0018] Calculate the pixel average of the coincident lines and compare the pixel average of the coincident lines with the standard pixel threshold of the first grayscale image and the second grayscale image:

[0019] If the pixel average value of the coincident line belongs to the standard pixel threshold, it is judged that there is no abnormal information;

[0020] If the pixel average value of the coincident line does not belong to the standard pixel threshold, it is determined that there is abnormal information and the first grayscale image and the second grayscale image where the coincident line is located are identified as the target image, and further judgment is made:

[0021] If the pixel average value of the coincident lines is greater than the standard pixel threshold, it is judged that the position deviation of the target image is too large.

[0022] Preferably, the method for adjusting the projection strategy in S3 includes:

[0023] If the position deviation of the target image is too large, reduce the initial projection focus radius, increase the initial exposure frequency, and delay the initial exposure interval:

[0024] Reduced projection focus radius ;

[0025] Increased exposure frequency ;

[0026] The duration of the exposure interval after the delay ;

[0027] in, is the initial projection focus radius, is the initial exposure frequency, is the duration of the initial exposure interval; is the first correction coefficient, is the second correction coefficient, is the third correction factor; is the pixel average of the coincident lines, is the standard pixel threshold.

[0028] Preferably, the process of performing abnormal information recognition and analysis on the adjusted first grayscale image and the second grayscale image in S4 is:

[0029] The boundary lines of the adjusted first grayscale image and the second grayscale image and the contours of their abnormal regions are respectively extracted by using the Canny algorithm;

[0030] Collect the boundary lines of the first grayscale image and the second grayscale image and the contours of their abnormal regions to obtain the area of ​​the abnormal region of the first grayscale image , the abnormal area of ​​the second grayscale image and the perimeter of the abnormal area , , through the formula , respectively calculate and obtain the abnormal area structure values ​​of the first grayscale image and the second grayscale image , ;

[0031] By calculating the abnormal region structure value of the first grayscale image with abnormal region contour in the continuous time period The data set, the abnormal area structure value of the second grayscale image The data set is used to reconstruct a three-dimensional data model, obtain the three-dimensional shape of the point cloud showing the abnormal area, and output the structural features of the abnormal area;

[0032] The structural characteristics of the abnormal area include the noise parameters of the abnormal area , abnormal area volume parameters .

[0033] Preferably, the method of obtaining the bar quality test result in S4 is:

[0034] By formula Calculate the detection coefficient ;

[0035] in, is the first preset weight coefficient, is a second preset weight coefficient; is the standard noise parameter in normal area; Preset deviation value for regional noise; It is the standard volume parameter of normal area; is a preset function.

[0036] Preferably, it also includes:

[0037] The detection coefficient The preset detection coefficient threshold interval To compare:

[0038] like ∈ , it is judged that there is a defect, the defect area is small, and the bar quality is qualified;

[0039] like > , it is judged that there are defects, the defect area is large, and the bar quality is unqualified;

[0040] like < , it is judged that there are no defects and the bar quality is excellent.

[0041] A titanium alloy bar quality detection system, used to implement a titanium alloy bar quality detection method, comprising:

[0042] An image acquisition module, used to acquire a transverse X-ray image and a longitudinal X-ray image of the bar material according to a projection strategy through an X-ray detection device, and extract a first grayscale image of the transverse X-ray image and a second grayscale image of the longitudinal X-ray image respectively;

[0043] An image processing module is used to perform target image position matching on the first grayscale image and the second grayscale image to obtain a target image position deviation degree containing abnormal information;

[0044] and for adjusting the projection strategy according to the degree of position deviation of the target image to obtain an adjusted first grayscale image and a second grayscale image;

[0045] The data analysis module is used to identify and analyze the abnormal information of the adjusted first grayscale image and the second grayscale image and obtain the bar quality detection result.

[0046] Beneficial effects of the present invention:

[0047] (1) The present invention follows a pre-set projection strategy, which reflects the process of projecting the interior of the rod at different angles, by collecting internal images from the horizontal and vertical directions of X-ray irradiation, and extracting a first grayscale image from the collected horizontal X-ray image and a second grayscale image from the vertical X-ray image; and the horizontal image and the vertical image will respectively display the internal structure of the rod at different angles; which is conducive to discovering the manifestation of minor problems inside the rod according to actual detection standards.

[0048] (2) The present invention adjusts the projection strategy of the X-ray detection equipment according to the degree of position deviation of the target image, reduces and eliminates the position deviation problem by designing the adjustment process of the projection strategy, and enables the subsequently collected images to more accurately reflect the internal structure of the rod; by performing abnormal information recognition and analysis on the adjusted first grayscale image and the second grayscale image and calculating the abnormal information recognition result, the quality of the rod is evaluated by obtaining the detection coefficient result, and the information such as the position, size, shape, etc. of the abnormal area may be identified according to the detection result, and the overall quality of the rod is accurately evaluated based on qualitative and quantitative analysis.

[0049] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0051] Figure 1 This is a step diagram of a titanium alloy bar quality inspection method of the present invention;

[0052] Figure 2 This is a module diagram of a titanium alloy bar quality inspection system of the present invention. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] See also Figure 1 As shown, the present invention is a method for detecting the quality of titanium alloy bars, the method comprising:

[0055] S1, collecting a transverse X-ray image and a longitudinal X-ray image of the bar material according to a projection strategy by an X-ray detection device, and extracting a first grayscale image of the transverse X-ray image and a second grayscale image of the longitudinal X-ray image respectively;

[0056] S2, performing target image position matching on the first grayscale image and the second grayscale image to obtain a target image position deviation degree containing abnormal information;

[0057] S3, adjusting the projection strategy according to the degree of position deviation of the target image to obtain an adjusted first grayscale image and a second grayscale image;

[0058] S4. Perform abnormal information recognition and analysis on the adjusted first grayscale image and the second grayscale image and obtain the bar quality inspection result.

[0059] In the above technical scheme, the present invention designs a method for quality inspection of titanium alloy bars, specifically by first using an X-ray inspection device, and according to a pre-set projection strategy, the projection strategy reflects the process of projecting the inside of the bar at different angles, and this projection angle and direction are set in advance according to demand, by irradiating the bar with X-rays, internal images are collected from the transverse and longitudinal directions of the X-ray irradiation, and a first grayscale image is extracted from the collected transverse X-ray image, and a second grayscale image is extracted from the longitudinal X-ray image; wherein, the grayscale image is a black and white image, in which the brightness value of each pixel represents the X-ray absorption rate of the point, thereby reflecting the density distribution inside the bar; and the transverse image and the longitudinal image will respectively display the internal structure of the bar at different angles; it is conducive to discovering the manifestation of minor problems inside the bar according to actual detection standards.

[0060] Then, by matching the position of the target image, the first grayscale image and the second grayscale image are matched and analyzed for their positional relationship, which usually involves a convolutional neural network automatically learning the representation information of the image to determine the position of the target image in two directions. For the needs of subsequent abnormality detection, the positions of the abnormal areas that appear simultaneously inside the actual rod are usually consistent in two directions; and the degree of deviation of the position of the target image is obtained by calculation and judgment based on this image containing abnormal information. Specifically, through the matching algorithm, for the target image containing abnormal information, the severity of the abnormality can be evaluated by calculating the degree of position deviation in the first grayscale image and the second grayscale image. This can usually be achieved by calculating the structure of the abnormal area in the target image. For example, the position difference of the wireframe of the target image in the two images can be calculated; by calculating the degree of position deviation of the target image in two directions, the analysis result of the degree of position deviation will be used as a reference for subsequent adjustment of the projection strategy.

[0061] Next, the projection strategy of the X-ray detection equipment is adjusted according to the degree of position deviation of the target image. The position deviation problem is reduced and eliminated by designing the adjustment process of the projection strategy, and the subsequently collected images can more accurately reflect the internal structure of the rod. The specific way in which the adjusted projection strategy obtains information is by re-collecting the transverse X-ray image and the longitudinal X-ray image of the rod, and extracting a new first grayscale image and a second grayscale image.

[0062] Finally, by performing abnormal information recognition and analysis on the adjusted first grayscale image and the second grayscale image; by using image processing algorithms such as convolutional neural networks to detect and locate abnormal areas, such as the degree of small cracks, inclusions, and pores; and based on the results of calculated abnormal information recognition, it is possible to evaluate the quality of the bar by obtaining the detection coefficient results, identify information that may include the location, size, shape, etc. of the abnormal area based on the detection results, and accurately evaluate the overall quality of the bar based on qualitative and quantitative analysis results.

[0063] As an implementation mode of the present invention, the projection strategy includes:

[0064] Set the initial projection focus radius, initial exposure frequency and initial exposure interval duration;

[0065] And adjust the initial projection focus radius, initial exposure frequency and initial exposure interval duration according to the degree of deviation of the target image position.

[0066] In the above technical scheme, the projection strategy in this embodiment ensures the specific projection method and setting related parameters for projecting the inside of the rod, mainly: the initial projection focus radius is set according to the imaging requirements of the rod and the equipment performance, and the appropriate focus radius size can be selected; the exposure time includes the exposure frequency of the rod and the exposure interval duration of the rod, which usually needs to be adjusted according to factors such as the material, thickness, required image quality and power of the X-ray source of the detected rod to ensure that the internal situation can be clearly detected and displayed according to the conventional projection efficiency.

[0067] As an implementation mode of the present invention, S2 includes:

[0068] Obtaining a coincidence line between the first grayscale image and the second grayscale image;

[0069] Calculate the pixel average of the coincident lines and compare the pixel average of the coincident lines with the standard pixel threshold of the first grayscale image and the second grayscale image:

[0070] If the pixel average value of the coincident line belongs to the standard pixel threshold, it is judged that there is no abnormal information;

[0071] If the pixel average value of the coincident line does not belong to the standard pixel threshold, it is determined that there is abnormal information and the first grayscale image and the second grayscale image where the coincident line is located are identified as the target image, and further judgment is made:

[0072] If the pixel average value of the coincident lines is greater than the standard pixel threshold, it is judged that the position deviation of the target image is too large.

[0073] In the above technical solution, the acquisition method of the degree of position deviation in step S2 of this embodiment is to obtain the feature points extracted from the two grayscale images according to the feature point detection algorithm such as SIFT, and these feature points have local uniqueness and can represent the key structure in the image. Through the matched feature points, the geometric transformation required to transform the first image into the second image, such as affine transformation or perspective transformation, can be estimated. The estimated transformation is applied to align the first image to the coordinate system of the second image, and in the aligned image, the coincidence line can be extracted by using the edge detection algorithm. In this way, according to the feature change acquisition method of the first grayscale image and the second grayscale image, it can be known that there must be an intersecting "coincidence line" between the real plane images, and then the pixel average value of the coincidence line is calculated and the pixel average value of the coincidence line is compared with the standard pixel threshold of the first grayscale image and the second grayscale image to determine whether the coincidence line meets the pixel standard requirements set by the detection, and then determine the image setting of the target image, that is, the first grayscale image and the second grayscale image, and then determine the degree of position deviation of the target image.

[0074] As an implementation mode of the present invention, the method for adjusting the projection strategy in S3 includes:

[0075] If the position deviation of the target image is too large, reduce the initial projection focus radius, increase the initial exposure frequency, and delay the initial exposure interval:

[0076] Reduced projection focus radius ;

[0077] Increased exposure frequency ;

[0078] The duration of the exposure interval after the delay ;

[0079] in, is the initial projection focus radius, is the initial exposure frequency, is the duration of the initial exposure interval; is the first correction coefficient, is the second correction coefficient, is the third correction factor; is the pixel average of the coincident lines, is the standard pixel threshold.

[0080] In the above technical solution, in this embodiment, the position deviation of the target image is adjusted to optimize the clarity of the currently projected rod by reducing the focal radius of the initial projection and increasing the exposure frequency and exposure interval duration.

[0081] Among them, the correction factor , , They are all obtained after rationality analysis and adjustment based on the standard pixel threshold interval difference. Therefore, the projection strategy is executed through the adjusted projection focus radius, exposure frequency and exposure interval duration, thereby achieving the effect of improving the accuracy of analysis and judgment while adaptively reducing the adjustment computing power requirements of the early X-ray projection detection.

[0082] As an implementation mode of the present invention, the process of performing abnormal information recognition and analysis on the adjusted first grayscale image and the second grayscale image in S4 is:

[0083] The boundary lines of the adjusted first grayscale image and the second grayscale image and the contours of their abnormal regions are respectively extracted by using the Canny algorithm;

[0084] Collect the boundary lines of the first grayscale image and the second grayscale image and the contours of their abnormal regions to obtain the area of ​​the abnormal region of the first grayscale image , the abnormal area of ​​the second grayscale image and the perimeter of the abnormal area , , through the formula , respectively calculate and obtain the abnormal area structure values ​​of the first grayscale image and the second grayscale image , ;

[0085] By calculating the abnormal region structure value of the first grayscale image with abnormal region contour in the continuous time period The data set, the abnormal area structure value of the second grayscale image The data set is used to reconstruct a three-dimensional data model, obtain the three-dimensional shape of the point cloud showing the abnormal area, and output the structural features of the abnormal area; the structural features of the abnormal area include the noise parameters of the abnormal area , abnormal area volume parameters .

[0086] In the above technical scheme, the boundary lines and abnormal area contours of the first grayscale image and the second grayscale image containing abnormal information are respectively extracted by the Canny algorithm. Here, the threshold segmentation is to compare the pixels in the image with one or more thresholds according to their grayscale values, divide the image into different areas, and extract the region of interest according to the result of the threshold segmentation to obtain the abnormal area contour and boundary lines; then substitute them into the convolutional neural network model for analysis, by collecting the boundary lines and abnormal area contours of the first grayscale image and the second grayscale image as input, output the abnormal area area of ​​the first grayscale image, the abnormal area area of ​​the second grayscale image and the perimeter of the abnormal area; and realize the acquisition of structural features of the abnormal area through the structural calculation formula.

[0087] As an implementation mode of the present invention, the method of obtaining the bar quality test result in S4 is:

[0088] By formula Calculate the detection coefficient ;

[0089] in, is the first preset weight coefficient, is a second preset weight coefficient; is the standard noise parameter in normal area; Preset deviation value for regional noise; It is the standard volume parameter of normal area; is a preset function.

[0090] In the above technical solution, the results of the bar quality inspection were calculated and analyzed in this embodiment, and according to the formula Calculate the detection coefficient ; By obtaining the detection coefficient, it is judged whether the current bar quality detection meets the requirements, and the main factors constituting the bar quality are the size of the abnormal structural features inside the bar and the noise fuzziness of the abnormal area. In this embodiment, the abnormal area noise parameters obtained by abnormal information identification , abnormal area volume parameters To conduct further analysis, and then reflect the impact of abnormal conditions on the quality of carbon rods.

[0091] in, , , , are obtained after fitting based on empirical test data, and , All are greater than 0, and will not be described in detail here; It is preset in advance based on the historical database and is obtained through real-time analysis based on the actual training network. It will not be described in detail here; preset function It is a detection function set according to historical data to ensure that the detection coefficient is within a specific reasonable range and the preset function It is set based on the influence of the historical abnormal characteristics of the material of the current bar and the standard parameters of the abnormal characteristics list of actual problems; for example, the data quality analysis function used in common data quality assessment models.

[0092] As an embodiment of the present invention, it also includes:

[0093] The detection coefficient The preset detection coefficient threshold interval To compare:

[0094] like ∈ , it is judged that there is a defect, the defect area is small, and the bar quality is qualified;

[0095] like > , it is judged that there are defects, the defect area is large, and the bar quality is unqualified;

[0096] like < , it is judged that there are no defects and the bar quality is excellent.

[0097] In the above technical solution, the detection coefficient is compared in this embodiment. If the preset threshold is exceeded, it is judged that the defect area has a serious problem and the bar quality inspection does not meet the standard; the preset detection coefficient threshold range It is obtained by fitting empirical data.

[0098] See also Figure 2 As shown, a titanium alloy bar quality detection system is used to implement a titanium alloy bar quality detection method, comprising:

[0099] An image acquisition module, used to acquire a transverse X-ray image and a longitudinal X-ray image of the bar material according to a projection strategy through an X-ray detection device, and extract a first grayscale image of the transverse X-ray image and a second grayscale image of the longitudinal X-ray image respectively;

[0100] An image processing module is used to perform target image position matching on the first grayscale image and the second grayscale image to obtain a target image position deviation degree containing abnormal information;

[0101] and for adjusting the projection strategy according to the degree of position deviation of the target image to obtain an adjusted first grayscale image and a second grayscale image;

[0102] The data analysis module is used to identify and analyze the abnormal information of the adjusted first grayscale image and the second grayscale image and obtain the bar quality detection result.

[0103] In the above technical scheme, the image acquisition module is used to obtain grayscale images of different angles of the rod inspected by X-ray, the target image is matched according to the grayscale images at different angles, and the degree of deviation is judged according to the position of the target image containing abnormal information, and the projection strategy is adjusted according to the degree of deviation to accurately determine the clarity of the internal structure of the rod. Finally, the data analysis module is used to identify and analyze the abnormal information data of the adjusted grayscale image, and finally the result of the rod quality inspection is obtained, and the qualified status of the rod is screened according to the inspection result.

[0104] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device, equipment, and non-volatile computer storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0105] The above specific embodiments of this specification are described. Other embodiments are within the scope of the attached documents. In some cases, the actions or steps recorded in this application can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0106] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined in this application, they shall all fall within the protection scope of the present invention.

Claims

1. A titanium alloy bar quality inspection method, characterized in that: The method comprises: S1, collecting a transverse X-ray image and a longitudinal X-ray image of the bar material according to a projection strategy by an X-ray detection device, and extracting a first grayscale image of the transverse X-ray image and a second grayscale image of the longitudinal X-ray image respectively; S2, performing target image position matching on the first grayscale image and the second grayscale image to obtain a target image position deviation degree containing abnormal information; S3, adjusting the projection strategy according to the degree of position deviation of the target image to obtain an adjusted first grayscale image and a second grayscale image; S4, performing abnormal information recognition and analysis on the adjusted first grayscale image and the second grayscale image and obtaining a bar quality inspection result; The projection strategy includes: Set the initial projection focus radius, initial exposure frequency and initial exposure interval duration; and adjusting the initial projection focus radius, initial exposure frequency, and initial exposure interval duration according to the degree of deviation of the target image position; The S2 includes: Obtaining a coincidence line between the first grayscale image and the second grayscale image; Calculate the pixel average of the coincident lines and compare the pixel average of the coincident lines with the standard pixel threshold of the first grayscale image and the second grayscale image: If the pixel average value of the coincident line belongs to the standard pixel threshold, it is judged that there is no abnormal information; If the pixel average value of the coincident line does not belong to the standard pixel threshold, it is determined that there is abnormal information and the first grayscale image and the second grayscale image where the coincident line is located are identified as the target image, and further judgment is made: If the average pixel value of the coincident line is greater than the standard pixel threshold, it is judged that the position deviation of the target image is too large; The method for adjusting the projection strategy in S3 includes: If the position deviation of the target image is too large, reduce the initial projection focus radius, increase the initial exposure frequency, and delay the initial exposure interval: Reduced projection focus radius The increased exposure frequency F = F0*∈*(P l -P ith ); The duration of the exposure interval after the delay Where R0 is the initial projection focus radius, F0 is the initial exposure frequency, T0 is the initial exposure interval duration; ε is the first correction coefficient, ∈ is the second correction coefficient, and θ is the third correction coefficient; P l is the average pixel value of the coincident line, P ith is the standard pixel threshold.

2. A titanium alloy bar quality inspection method according to claim 1, characterized in that: The process of performing abnormal information recognition and analysis on the adjusted first grayscale image and the second grayscale image in S4 is: The boundary lines of the adjusted first grayscale image and the second grayscale image and the contours of their abnormal regions are respectively extracted by using the Canny algorithm; Collect the boundary lines of the first grayscale image and the second grayscale image and the contours of their abnormal regions to obtain the area S of the abnormal region of the first grayscale image. A , the abnormal area of ​​the second grayscale image S B and the perimeter of the abnormal area C A , C B , through the formula Calculate and obtain the abnormal region structure value i of the first grayscale image and the second grayscale image respectively A 、i B ; By calculating the abnormal region structure value i of the first grayscale image with abnormal region contour in the continuous time period A The data set, the abnormal region structure value i of the second grayscale image B The data set is used to reconstruct a three-dimensional data model, obtain the three-dimensional shape of the point cloud showing the abnormal area, and output the structural features of the abnormal area; The abnormal region structural features include an abnormal region noise parameter Np and an abnormal region volume parameter Vp.

3. A titanium alloy bar quality inspection method according to claim 2, characterized in that: The method of obtaining the bar quality test result in S4 is: By formula Calculate and obtain the detection coefficient De; Among them, α is the first preset weight coefficient, β is the second preset weight coefficient; Np0 is the standard noise parameter of the normal area; ΔNp is the preset deviation value of the regional noise; Vp0 is the standard volume parameter of the normal area; and f is a preset function.

4. A titanium alloy bar quality inspection method according to claim 3, characterized in that: Also includes: The detection coefficient De is compared with the preset detection coefficient threshold interval [De A , De B ] for comparison: If De∈[De A , De B ], it is judged that there is a defect, the defect area is small, and the bar quality is qualified; If De>De B , it is judged that there are defects, the defect area is large, and the bar quality is unqualified; If De <De A , it is judged that there are no defects and the bar quality is excellent.

5. A titanium alloy bar quality inspection system, characterized in that: The method for detecting the quality of a titanium alloy bar according to any one of claims 1 to 4 comprises: An image acquisition module, used to acquire a transverse X-ray image and a longitudinal X-ray image of the bar material according to a projection strategy through an X-ray detection device, and extract a first grayscale image of the transverse X-ray image and a second grayscale image of the longitudinal X-ray image respectively; An image processing module is used to perform target image position matching on the first grayscale image and the second grayscale image to obtain a target image position deviation degree containing abnormal information; and used to adjust the projection strategy according to the degree of position deviation of the target image to obtain an adjusted first grayscale image and a second grayscale image; The data analysis module is used to identify and analyze the abnormal information of the adjusted first grayscale image and the second grayscale image and obtain the bar quality detection result.

Images