Titanium alloy bar production quality visual detection method and system

Abstract

The application relates to the technical field of visual detection, in particular to a titanium alloy bar production quality visual detection method and system. The method comprises the following steps: periodically acquiring a gray image of a to-be-detected bar in a cooling process, dividing the to-be-detected bar into a high-temperature sensitive stage, a high-temperature cooling stage and a low-temperature stage according to surface gray values; determining a standard degree according to the gray distribution, the bending degree and the length of the high-temperature sensitive stage, and screening abnormal bars with cracks; in the high-temperature cooling stage, analyzing the oxidation effect according to the gray change of the abnormal bars, and determining an initial crack area; in the low-temperature stage, detecting the cracks of the abnormal bars in the gray image, determining a real crack area according to the similarity between the detected crack area and the initial crack area; and determining the production quality of the bar according to the real crack area. The method can realize more accurate and higher-precision crack positioning, and effectively improve the accuracy of titanium alloy bar quality detection.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of visual detection, and in particular to a titanium alloy bar production quality visual detection method and system. BACKGROUND

[0002] Titanium alloy bars become core raw materials in the fields of aerospace and high-end equipment manufacturing due to excellent strength-weight ratio and corrosion resistance, and the quality of the titanium alloy bars directly relates to the reliability of terminal products. In the whole-process production of the titanium alloy bars, quality inspection is a key link for controlling product quality, and visual-based detection technology is a core support.

[0003] In the production process of the titanium alloy bars, most quality detection stages in the industry are concentrated in the latter part of the production process, such as after finish rolling or after product annealing. This "end detection" mode may cause surface micro-cracks generated in the rough rolling process to be partially covered due to high-temperature plastic deformation, oxidation layer covering or work hardening when entering the finish rolling process, so that crack characteristics are weakened or even disappear, thereby increasing detection difficulty and improving the risk of missed detection, reducing crack positioning accuracy, and reducing the accuracy of titanium alloy bar quality detection. SUMMARY

[0004] To solve the technical problem that the "end detection" mode increases detection difficulty and improves the risk of missed detection, reduces crack positioning accuracy, and reduces the accuracy of titanium alloy bar quality detection in the related art, the application provides a titanium alloy bar production quality visual detection method and system, and the technical solution is as follows:

[0005] The application provides a titanium alloy bar production quality visual detection method, and the method comprises the following steps:

[0006] Periodically acquire a gray image of a to-be-detected bar in a cooling process, and divide the to-be-detected bar into a high-temperature sensitive stage, a high-temperature cooling stage and a low-temperature stage according to a surface gray value;

[0007] According to the gray distribution, the bending degree and the length of the to-be-detected bar in the high-temperature sensitive stage, the standard degree of the to-be-detected bar in the gray image is determined; and according to the standard degree of the to-be-detected bar in each gray image, an abnormal bar with a crack is screened;

[0008] In the high-temperature cooling stage, the oxidation effect is analyzed according to the gray change of the abnormal bar at different sampling moments, and the initial crack area of the abnormal bar in the high-temperature cooling stage is determined; and in the low-temperature stage, gap detection is performed on the abnormal bar in the gray image, and the real crack area is determined according to the similarity between the detected gap area and the initial crack area;

[0009] The production quality of the bar is determined according to the real crack area.

[0010] Further, the dividing the to-be-detected rod into a high-temperature sensitive stage, a high-temperature cooling stage and a low-temperature stage according to the surface gray value comprises:

[0011] The dividing temperature of the high-temperature sensitive stage and the high-temperature cooling stage is 800 degrees Celsius, and the dividing temperature of the high-temperature cooling stage and the low-temperature stage is 500 degrees Celsius;

[0012] The gray value of the standard rod in the gray image at 800 degrees Celsius and 500 degrees Celsius is determined as a first gray threshold and a second gray threshold, respectively;

[0013] When the gray mean value of the to-be-detected rod is greater than the first gray threshold, it is determined that the to-be-detected rod is in the high-temperature sensitive stage;

[0014] When the gray mean value of the to-be-detected rod is less than or equal to the first gray threshold and greater than the second gray threshold, it is determined that the to-be-detected rod is in the high-temperature cooling stage;

[0015] When the gray mean value of the to-be-detected rod is less than or equal to the second gray threshold, it is determined that the to-be-detected rod is in the low-temperature stage.

[0016] Further, the determining the standard degree of the to-be-detected rod in the gray image according to the gray distribution, the bending degree and the length of the to-be-detected rod in the high-temperature sensitive stage comprises:

[0017] The variance of the gray value of all pixel points in the to-be-detected rod region is calculated as a first standard index;

[0018] The second standard index of the bending influence is determined according to the bending degree of the to-be-detected rod shape;

[0019] The absolute value of the difference between the length of the to-be-detected rod and the standard length is taken as a third standard index;

[0020] The first standard index, the second standard index and the third standard index are combined to determine the standard degree of the to-be-detected rod, wherein the first standard index, the second standard index and the third standard index are negatively correlated with the standard degree, and the value of the standard degree is a normalized value.

[0021] Further, the determining the second standard index according to the bending degree of the to-be-detected rod shape comprises:

[0022] The to-be-detected rod region is subjected to morphological erosion processing to determine a skeleton line;

[0023] The skeleton line is subjected to straight line fitting to obtain a fitting straight line;

[0024] The distance between each pixel point on the skeleton line and the fitting straight line is determined, and the sum of the distances between all pixel points and the fitting straight line is taken as the second standard index.

[0025] Further, the method for screening abnormal rods with cracks according to the standard degree of each gray image of the rods to be detected comprises:

[0026] The rods to be detected with the standard degree less than the preset standard threshold are regarded as abnormal rods.

[0027] Further, the method for analyzing the oxidation effect according to the gray changes of the abnormal rods at different sampling moments to determine the initial crack area of the abnormal rods in the high-temperature cooling stage comprises:

[0028] The gray values of the same pixel point at different sampling moments are determined, and the gray sequence is obtained according to the time sequence; the first-order difference processing is performed on the gray sequence to obtain the difference sequence;

[0029] The elements with the values greater than the preset difference threshold in the difference sequence are regarded as oxidation influence elements;

[0030] The duration corresponding to the oxidation influence elements in the gray sequence is determined as the oxidation time;

[0031] The ratio of the sum value of all the oxidation influence elements to the oxidation time is normalized and processed as an oxidation effect index;

[0032] The pixel points with the oxidation effect index greater than the preset effect threshold are regarded as crack pixel points, and the area composed of the crack pixel points is regarded as the initial crack area.

[0033] Further, the method for detecting the crack of the abnormal rods in the gray image according to the similarity between the detected crack area and the initial crack area to determine the real crack area comprises:

[0034] The edge detection is performed on the abnormal rods in the gray image to determine the edge texture of the non-rod contour as the crack area;

[0035] The initial crack area matched with each crack area is determined according to the morphological center distance between the crack area and the initial crack area;

[0036] The real crack area is determined according to the extension direction difference and the area overlap degree between the crack area and the matched initial crack area.

[0037] Further, the method for determining the initial crack area matched with each crack area according to the morphological center distance between the crack area and the initial crack area comprises:

[0038] The morphological center points of each crack area and the initial crack area are determined, the Euclidean distance between the morphological center points of the crack area and the initial crack area is regarded as the analysis distance, and the initial crack area matched with the corresponding crack area is regarded as the initial crack area with the minimum analysis distance of the crack area.

[0039] Further, according to a difference between an extension direction of the gap region and the matched initial crack region, the real crack region is determined, comprising:

[0040] Linear fitting is respectively performed on the gap region and the matched initial crack region to obtain fitting straight lines, an included angle between the fitting straight lines of the gap region and the initial crack region is calculated, and a ratio of the included angle to 180 degrees is taken as a direction analysis index;

[0041] In the gap region and the matched initial crack region, a ratio of a same pixel point quantity to a union of all pixel points in the two regions is determined as a region overlap degree;

[0042] A difference between the region overlap degree and the direction analysis index is calculated and normalized as a crack judgment index;

[0043] The crack region with the crack judgment index greater than a preset judgment threshold is taken as the real crack region.

[0044] On the other hand, a titanium alloy bar production quality visual detection system is also provided, the system comprising a memory, a processor and a computer program stored in the memory and running on the processor, and the processor implements the steps of the method according to any one of the preceding embodiments when executing the computer program.

[0045] The present application has the following beneficial effects:

[0046] The present application periodically acquires a gray scale image of the bar to be detected in the cooling process, and divides the bar to be detected into a high-temperature sensitive stage, a high-temperature cooling stage and a low-temperature stage according to the gray scale value. Firstly, by performing quality detection on the cooling process, compared with the quality detection of the finished product in the related art, the quality detection link is moved to the transition stage between rough rolling and finish rolling, which can effectively capture the difference of the surface change of the bar, accurately identify the surface texture, and complete accurate positioning when the crack has not been complicated by the deformation of the finish rolling. In the high-temperature sensitive stage, the gray scale distribution, the bending degree and the length are combined to determine the abnormal bar with cracks. In the high-temperature cooling stage, the oxidation effect is analyzed according to the gray scale change to determine the initial crack region. In the low-temperature stage, gap detection is performed, and the real crack region is determined in combination with the similarity. Therefore, by combining the characteristic performance of the bar at different stages in the cooling process, crack analysis in different dimensions is performed to realize more accurate and higher-precision crack positioning, and the accuracy of the quality detection of the titanium alloy bar is effectively improved. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, and the advantages thereof, a brief introduction will be given to the drawings required in the description of the embodiments or the prior art. Obviously, the drawings in the following description only show some embodiments of the present application, and for those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0048] Figure 1 A flow chart of a titanium alloy bar production quality visual detection method provided by an embodiment of the present application;

[0049] Figure 2 A walking beam transport field scene schematic diagram provided by an embodiment of the present application;

[0050] Figure 3 An image acquisition schematic diagram provided by an embodiment of the present application. DETAILED DESCRIPTION

[0051] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined purposes, the following describes the specific embodiments, structure, features and effects of the titanium alloy bar production quality visual detection method and system according to the present application in combination with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0052] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0053] The following specifically describes the specific scheme of the titanium alloy bar production quality visual detection method provided by the present application in combination with the drawings.

[0054] Please refer to Figure 1 which shows a flow chart of a titanium alloy bar production quality visual detection method provided by an embodiment of the present application. The method comprises:

[0055] S101: periodically acquire a gray scale image of the bar to be detected in the cooling process, and divide the bar to be detected into a high-temperature sensitive stage, a high-temperature cooling stage and a low-temperature stage according to the surface gray scale value.

[0056] In the titanium alloy bar production process, the quality detection stage in the industry is mostly concentrated in the later stage of the production process, such as after finish rolling or after product annealing. This 'end detection' mode may cause surface micro-cracks generated in the rough rolling process to be partially covered due to high temperature plastic deformation, oxidation layer covering or work hardening when entering the finish rolling process, resulting in weakening or even disappearance of crack characteristics, increasing detection difficulty and increasing the risk of missed detection. Therefore, the quality detection link is moved to the transition stage between rough rolling and finish rolling to realize crack detection of the bar.

[0057] In the embodiment of the present application, the titanium alloy bar after rough rolling is placed in the V-shaped groove of the walking beam conveyor, and is sequentially transported backward with the stepping action of the conveyor. During the stationary interval of the conveyor, the bar is in a stationary state and is naturally cooled. The natural cooling process is the cooling process. In the embodiment of the present application, referring to Figure 2 and Figure 3 , Figure 2 The scene schematic diagram of the walking beam conveyor provided by one embodiment of the present application, Figure 3 The image acquisition schematic diagram provided by one embodiment of the present application; a high frame rate camera can be configured directly above the walking beam conveyor for acquiring the original image of the bar to be detected. Then, the original image is subjected to mean gray value processing to obtain a gray value image, which is well known to those skilled in the art and will not be described here.

[0058] When the bar just comes out of the furnace, the bar body will cause high brightness due to high temperature, and with natural cooling and surface oxidation, the gray value gradually decreases. Therefore, in order to accurately distinguish, the cooling process of the bar to be detected is divided into three stages: high-temperature sensitive stage, high-temperature cooling stage and low-temperature stage in the embodiment of the present application.

[0059] Further, in some embodiments of the present application, the bar to be detected is divided into high-temperature sensitive stage, high-temperature cooling stage and low-temperature stage according to the surface gray value, comprising: determining that the division temperature of the high-temperature sensitive stage and the high-temperature cooling stage is 800 degrees Celsius, and the division temperature of the high-temperature cooling stage and the low-temperature stage is 500 degrees Celsius; determining the gray value of the standard bar in the gray value image under the surface temperature of 800 degrees Celsius and 500 degrees Celsius as the first gray threshold value and the second gray threshold value, respectively; when the gray value mean of the bar to be detected is greater than the first gray threshold value, it is determined to be in the high-temperature sensitive stage; when it is less than or equal to the first gray threshold value and greater than the second gray threshold value, it is determined to be in the high-temperature cooling stage; when it is less than or equal to the second gray threshold value, it is determined to be in the low-temperature stage.

[0060] That is to say, the actual stage is divided by temperature, and the gray value corresponding to the temperature is used as the threshold value to realize the gray stage division on the image, and all the gray images of the bar to be detected are divided into high-temperature sensitive stage, high-temperature cooling stage and low-temperature stage.

[0061] It should be noted that different stages have different performances, when the bar is in the high temperature sensitive stage, the bar just completes rough rolling, the surface temperature is extremely high, the crack generated in the rolling process has not been completely closed due to high temperature plastic deformation, and therefore it is easy to find; and in the middle of the cooling stage of the bar in the high temperature cooling stage, the gradually formed oxide film enhances the characteristic performance of the crack defect area; in the low temperature stage, the surface state of the bar tends to be stable, the defect morphology no longer changes with the temperature, and the final confirmation and analysis of the defect are facilitated.

[0062] Based on the above logic, a multi-stage crack quality detection process is realized.

[0063] S102: According to the gray distribution, bending degree and length of the bar to be detected in the high temperature sensitive stage, the standard degree of the bar to be detected in the gray image is determined; according to the standard degree of the bar to be detected in each gray image, the abnormal bar with cracks is screened.

[0064] Because in the high temperature sensitive stage, the normal bar body will have a higher gray value due to high temperature, and the crack area will have a more obvious gray difference with the normal high temperature bar body due to the non-uniform heat dissipation corresponding to itself, and the bar body will have a length inconsistency and bending phenomenon in this stage. At present, the bar body is just formed and still has great plasticity, and the crack will cause uneven stress, resulting in more obvious bending phenomenon.

[0065] In the embodiment of the application, the gray distribution, bending degree and length are combined to screen the abnormal bar with cracks.

[0066] Further, in some embodiments of the application, according to the gray distribution, bending degree and length of the bar to be detected in the high temperature sensitive stage, the standard degree of the bar to be detected in the gray image is determined, including: calculating the gray value variance of all pixel points in the bar to be detected area as a first standard index; according to the bending degree of the bar to be detected, a second standard index of bending influence is determined; the absolute value of the difference between the length of the bar to be detected and the standard length is taken as a third standard index; the standard degree of the bar to be detected is determined by combining the first standard index, the second standard index and the third standard index, wherein the first standard index, the second standard index and the third standard index are all negatively correlated with the standard degree, and the value of the standard degree is a normalized value.

[0067] Among them, the first standard index, the second standard index and the third standard index are all index information under the corresponding characteristics.

[0068] The first standard index represents the discrete characteristics of the gray value, and because under the influence of the crack, the crack area will have a more obvious gray difference, therefore, the greater the variance of the gray value, the more obvious the crack influence.

[0069] The second standard index represents the bending feature, and the bending degree can be analyzed by performing linear fitting and according to the difference of the fitted straight line, or directly determining the curvature of the bar through image recognition.

[0070] Further, in some embodiments of the present application, the second standard index is determined according to the bending degree of the shape of the bar to be detected, including: performing morphological erosion processing on the region of the bar to be detected to determine a skeleton line; performing linear fitting on the skeleton line to obtain a fitted straight line; determining the distance between each pixel point on the skeleton line and the fitted straight line, and taking the sum of the distances between all pixel points and the fitted straight line as the second standard index.

[0071] The morphological erosion processing is a technology known in the art, and the second standard index is determined by the difference between the skeleton line and the fitted straight line, that is, the greater the value of the second standard index, the higher the bending degree of the corresponding bar to be detected, and the more obvious the crack influence.

[0072] The third standard index represents the length difference feature, and the length consistency is also an important influencing factor for bar quality detection. The absolute value of the difference between the length of the bar to be detected and the standard length is taken as the third standard index, and the greater the value of the third standard index, the greater the difference between the length and the standard, and the worse the quality of the bar.

[0073] The above specific analysis of the first standard index, the second standard index and the third standard index can realize the calculation of the standard degree. It should be noted that the positive correlation relationship represents that there is a same direction change relationship between the independent variable and the dependent variable, that is, the greater the independent variable, the greater the dependent variable; the negative correlation relationship represents that there is a reverse change relationship between the independent variable and the dependent variable, that is, the smaller the independent variable, the greater the dependent variable; the specific form of the positive correlation relationship and the negative correlation relationship is determined by actual application, and the present application does not make special limitation.

[0074] At the same time, in order to facilitate operation, all index data involved in operation in the embodiments of the present application are subjected to data preprocessing, and then the dimension influence is cancelled. The specific means for removing the dimension influence is a technology known to those skilled in the art, which is not limited here.

[0075] Therefore, in the embodiments of the present application, the product value of the first standard index, the second standard index and the third standard index can be directly calculated, the reciprocal of the product value is subjected to maximum and minimum value normalization processing, and the standard degree is taken as the standard degree.

[0076] In the embodiment of the present application, the standard degree represents the standard coefficient of the rod to be detected in the three dimensions of gray scale distribution, bending degree and length. The greater the value of the standard degree is, the more normal the corresponding rod to be detected performs in the high-temperature sensitive stage. Therefore, in the embodiment of the present application, the rod to be detected with a standard degree less than a preset standard threshold is regarded as an abnormal rod.

[0077] The preset standard threshold is a threshold value of the standard degree. Specifically, it may be, for example, 0.6. That is, the rod to be detected with a standard degree less than 0.6 is regarded as an abnormal rod.

[0078] S103: In the high-temperature cooling stage, the oxidation effect is analyzed according to the gray scale change of the abnormal rod at different sampling time points, and the initial crack area of the abnormal rod in the high-temperature cooling stage is determined; in the low-temperature stage, the abnormal rod in the gray scale image is subjected to crack detection, and the real crack area is determined according to the similarity between the detected crack area and the initial crack area.

[0079] When the rod enters the high-temperature cooling stage, the surface continuity of the rod caused by the crack is destroyed, and the exposed internal metal matrix forms a large number of high-activity oxidation reaction sites. These sites react with oxygen first in the temperature range of the high-temperature cooling stage. At the same time, the crack area exposes more surface, and the high-temperature air retained in the interior further accelerates the formation of the oxidation film. The formation of the oxidation film can cause the gray scale change. Therefore, the oxidation analysis can be directly performed according to the gray scale change, so as to determine the initial crack area.

[0080] Further, in some embodiments of the present application, the oxidation effect is analyzed according to the gray scale change of the abnormal rod at different sampling time points, and the initial crack area of the abnormal rod in the high-temperature cooling stage is determined, which includes: determining the gray scale values of the same pixel point at different sampling time points, and arranging the gray scale values in time sequence to obtain a gray scale sequence; performing first-order difference processing on the gray scale sequence to obtain a difference sequence; regarding the elements with values greater than a preset difference threshold in the difference sequence as oxidation influence elements; determining the corresponding duration of the oxidation influence elements in the gray scale sequence as oxidation time; normalizing the ratio of the sum value of all oxidation influence elements to the oxidation time as an oxidation effect index; regarding the pixel points with oxidation effect indexes greater than a preset effect threshold as crack pixel points, and regarding the area composed of the crack pixel points as the initial crack area.

[0081] The oxidation reaction can make the gray scale of the rod surface decrease rapidly, therefore, in the embodiment of the present application, for the same pixel point, the gray scale sequence is determined, and the differential sequence is analyzed, the gray scale change in the differential sequence is partly due to the gray scale change caused by the normal temperature drop, and the other part is due to the gray scale change caused by the surface oxidation, and the gray scale change caused by the surface oxidation is more intense, so that the gray scale decrease effect is more obvious, therefore, in the embodiment of the present application, the preset differential threshold is set to realize the acquisition of the oxidation influence element.

[0082] The preset differential threshold is a threshold value of the element in the differential sequence, specifically, for example, 2, that is, at adjacent two sampling moments (which can be specifically 10-30 frames per second high frame shooting, and each frame is taken as a sampling moment), the gray scale value changes by 2 values, that is, the gray scale change caused by the oxidation effect is taken as the oxidation influence element.

[0083] The oxidation influence element represents the oxidation behavior at adjacent two sampling moments, therefore, the corresponding time interval can be taken as the time interval corresponding to the oxidation phenomenon, the differential sequence corresponding to a pixel point is traversed to obtain the corresponding duration of all oxidation influence elements in the gray scale sequence as the overall oxidation time.

[0084] The sum value of all oxidation influence elements is calculated, the greater the value, the more intense the oxidation reaction, and the ratio of the sum value to the oxidation time is normalized as an oxidation effect index, the oxidation effect index represents the obvious degree of the oxidation effect, the greater the value of the oxidation effect index, the more intense the oxidation reaction, and the shorter the time, and it is more likely to represent the oxidation effect caused by the crack, the pixel point with the oxidation effect index greater than a preset effect threshold is taken as a crack pixel point, and the region composed of the crack pixel points is taken as an initial crack region.

[0085] The preset effect threshold is a threshold value of the oxidation effect index, which can be specifically, for example, 0.8 in the embodiment of the present application, and is adjusted according to the actual rod oxidation characteristics, which is not limited.

[0086] If the oxide film formed in the high-temperature cooling stage is locally damaged due to the step beam type transport machine, new metal surface may be exposed to contact with oxygen in a short time, temporarily accelerating oxidation and forming a "pseudo-crack signal". Therefore, after completing the dynamic oxidation behavior screening in the high-temperature cooling stage, further determination is needed in the low-temperature stage with stable effect.

[0087] Further, in some embodiments of the present application, the crack detection is performed on the abnormal bar in the gray image, and the real crack area is determined according to the similarity between the detected crack area and the initial crack area, including: performing edge detection on the abnormal bar in the gray image, and determining that the edge texture of the non-bar contour is a crack area; determining the initial crack area matched with each crack area according to the morphological center distance between the crack area and the initial crack area; and determining the real crack area according to the extension direction difference and the area overlap degree between the crack area and the matched initial crack area.

[0088] As the titanium alloy bar temperature decreases into the low temperature stage, the oxide film growth tends to be stable, and the detection signal depending on the dynamic oxidation rate in the early stage may tend to be flat. At this time, the nature of the defect needs to be defined through the static geometric characteristics of the cracking of the oxide film. Because the oxide film at the real crack is structurally weak due to the defect, it is easy to further expand along the original crack when moving on the transport machine, and the formed crack has clear position inheritance and direction consistency, that is, the crack position is highly overlapped with the crack position detected in the low temperature stage, and the trend is in a continuous straight line; and the oxide film fracture caused by mechanical stress during transportation is mostly in the form of irregular network or point cracking, and the distribution direction is random and has no relevance to the initial crack area detected in the early stage.

[0089] Therefore, first, edge detection is performed on the abnormal bar in the gray image, the edge texture of the non-bar contour is determined to be a crack area, and matching analysis is performed based on morphological characteristics to determine the initial crack area matched with the crack area.

[0090] Further, in some embodiments of the present application, the initial crack area matched with each crack area is determined according to the morphological center distance between the crack area and the initial crack area, including: determining the morphological center point of each crack area and the initial crack area, taking the Euclidean distance between the morphological center points of the crack area and the crack area as the analysis distance; and taking the crack area with the minimum analysis distance from the crack area as the initial crack area matched with the corresponding crack area.

[0091] The morphological center point is well known to those skilled in the art, and will not be described here. The analysis distance represents the objective position distance between two areas. The smaller the value, the higher the matching degree. Therefore, the minimum analysis distance is used as a condition to determine the initial crack area matched with the corresponding crack area.

[0092] After determining the matched initial crack area, it is necessary to exclude the influence of the pseudo-crack phenomenon caused by local damage. Because the pseudo-crack is usually irregularly diffused, it can be screened according to the extension direction.

[0093] Further, in some embodiments of the present application, the real crack region is determined according to the difference between the extension directions of the gap region and the matched initial crack region, including: performing linear fitting on the gap region and the matched initial crack region respectively to obtain fitting straight lines, calculating the included angle between the fitting straight line of the gap region and the fitting straight line of the crack region, and taking the ratio of the included angle to 180 degrees as a direction analysis index; determining the ratio of the number of the same pixel points in the gap region and the matched initial crack region to the union of all pixel points in the two regions as a region overlap degree; calculating the difference between the region overlap degree and the direction analysis index, and normalizing the difference as a crack judgment index; and taking the crack region with the crack judgment index greater than a preset judgment threshold as the real crack region.

[0094] The angle value obtained by performing linear fitting determines the consistency of the extension direction, and the smaller the value of the direction analysis index, the more consistent the extension direction. The region overlap degree further analyzes the overlap of the matched gap region and the initial crack region, and the larger the value of the region overlap degree, the higher the overlap degree of the gap region and the matched initial crack region, that is, the more likely it is a real crack that has existed all the time. Therefore, in the embodiments of the present application, the difference between the region overlap degree and the direction analysis index is calculated and normalized as a crack judgment index, and the larger the value of the crack judgment index, the more consistent it is with the characteristics of the real crack.

[0095] The preset judgment threshold is a threshold value of the crack judgment index, which may be specifically 0.5 in the embodiments of the present application, so as to screen the real crack region.

[0096] S104: determining the production quality of the bar according to the real crack region.

[0097] The real crack region represents the region where the crack is generated, and therefore, the larger the area of the real crack region, the worse the production quality of the bar. The area of the real crack region can be directly combined with other characteristics such as its morphology and position to serve as a production quality detection result.

[0098] The application divides the detected rod into a high-temperature sensitive stage, a high-temperature cooling stage and a low-temperature stage according to the gray value by periodically acquiring the gray image of the rod in the cooling process. Firstly, by performing quality detection on the cooling process, compared with the quality detection of the finished product in the related art, the quality detection link is moved to the transition stage between rough rolling and finish rolling, which can effectively capture the difference of the surface change of the rod, accurately identify the surface texture, and complete accurate positioning when the crack has not been complicated by finish rolling deformation. In the high-temperature sensitive stage, the abnormal rod with cracks is determined by combining the gray distribution, the bending degree and the length. In the high-temperature cooling stage, the oxidation effect is analyzed according to the gray change to determine the initial crack area. In the low-temperature stage, the gap detection is performed to determine the real crack area by combining the similarity. Thus, by combining the characteristic performance of the rod in different stages of the cooling process, the crack analysis in different dimensions is performed to realize more accurate and higher-precision crack positioning, and the accuracy of the quality detection of the titanium alloy rod is effectively improved.

[0099] In another aspect, a titanium alloy rod production quality visual detection system is also provided, which comprises a memory, a processor and a computer program stored in the memory and running on the processor, and the processor implements the steps of the method according to any one of the preceding aspects when executing the computer program.

[0100] It should be noted that the above-mentioned sequence of the embodiments of the application is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0101] Each of the embodiments in the specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment mainly describes the difference from other embodiments.

Claims

1. A visual inspection method for the production quality of titanium alloy bars, characterized in that, The method includes: The grayscale images of the bar under test are periodically acquired during the cooling process, and the bar under test is divided into high-temperature sensitive stage, high-temperature cooling stage and low-temperature stage according to the surface grayscale value. Based on the grayscale distribution, curvature, and length of the bar under test during the high-temperature sensitive stage, the standardization of the bar under test in the grayscale image is determined; based on the standardization of the bar under test in each grayscale image, abnormal bars with cracks are screened. During the high-temperature cooling stage, the oxidation effect is analyzed based on the grayscale changes of the abnormal bars at different sampling times to determine the initial crack area of ​​the abnormal bars during the high-temperature cooling stage. During the low-temperature stage, gap detection is performed on the abnormal bars in the grayscale image, and the actual crack area is determined based on the similarity between the detected gap area and the initial crack area. The production quality of the bar stock is determined based on the actual crack area. The methods for dividing the high-temperature sensitive stage, the high-temperature cooling stage, and the low-temperature stage include: The dividing temperature between the high-temperature sensitive stage and the high-temperature cooling stage is set at 800 degrees Celsius, and the dividing temperature between the high-temperature cooling stage and the low-temperature stage is set at 500 degrees Celsius. The grayscale values ​​of standard bars with surface temperatures of 800 degrees Celsius and 500 degrees Celsius in the grayscale image were determined as the first grayscale threshold and the second grayscale threshold, respectively. When the average gray value of the bar to be tested is greater than the first gray value threshold, it is determined to be in the high temperature sensitive stage; When the grayscale value is less than or equal to the first grayscale threshold and greater than the second grayscale threshold, it is determined that the system is in the high-temperature cooling stage. When the grayscale value is less than or equal to the second grayscale threshold, it is determined to be in the low-temperature stage; Methods for determining the standard degree of the bar material to be detected in a grayscale image include: Calculate the variance of grayscale values ​​of all pixels in the region of the bar to be tested, and use it as the first standard indicator; The second standard index of the bending effect is determined based on the degree of bending in the shape of the bar to be tested; The absolute value of the difference between the length of the bar to be tested and the standard length is used as the third standard indicator. The standard degree of the bar to be tested is determined by combining the first standard index, the second standard index, and the third standard index, wherein the first standard index, the second standard index, and the third standard index are all negatively correlated with the standard degree, and the standard degree is a normalized value.

2. The visual inspection method for the production quality of titanium alloy bars as described in claim 1, characterized in that, The determination of the second standard index based on the degree of curvature in the shape of the bar to be tested includes: Morphological etching is performed on the area of ​​the bar to be tested to determine the skeleton line; The skeleton lines are fitted with straight lines to obtain the fitted straight lines; Determine the distance between each pixel on the skeleton line and the fitted line, and use the sum of the distances between all pixels and the fitted line as the second standard indicator.

3. The visual inspection method for the production quality of titanium alloy bars as described in claim 1, characterized in that, The step of screening for abnormal bars with cracks based on the standard degree of the bar material to be tested in each grayscale image includes: Bars whose standard level is less than a preset standard threshold are considered abnormal bars.

4. The visual inspection method for the production quality of titanium alloy bars as described in claim 1, characterized in that, The analysis of the oxidation effect based on the grayscale changes of the abnormal rod at different sampling times, and the determination of the initial crack region of the abnormal rod during the high-temperature cooling stage, includes: Determine the grayscale values ​​of the same pixel at different sampling times, and arrange them in time sequence to obtain a grayscale sequence; perform first-order difference processing on the grayscale sequence to obtain a difference sequence; Elements in the difference sequence whose values ​​are greater than a preset difference threshold are considered as oxidation-affecting elements; Determine the duration of the oxidation effect on the element in the grayscale sequence, and use it as the oxidation time; The sum of all oxidation-affecting elements and the ratio of oxidation time were normalized to serve as an indicator of oxidation effectiveness. Pixels whose oxidation effect index is greater than a preset effect threshold are designated as crack pixels, and the area formed by the crack pixels is designated as the initial crack area.

5. The visual inspection method for the production quality of titanium alloy bars as described in claim 1, characterized in that, Crack detection is performed on abnormal bars in grayscale images. Based on the similarity between the detected crack areas and the initial crack areas, the true crack areas are determined, including: Edge detection is performed on abnormal bars in grayscale images to identify gap regions as the edge textures of non-bar contours. Based on the morphological center distance between the gap region and the initial crack region, an initial crack region matching each gap region is determined. The actual crack region is determined based on the difference in the extension direction and the degree of overlap between the gap region and the matching initial crack region.

6. The visual inspection method for the production quality of titanium alloy bars as described in claim 5, characterized in that, Based on the morphological center distance between the gap region and the initial crack region, an initial crack region matching each gap region is determined, including: Determine the morphological center point of each fissure region and the initial crack region, and use the Euclidean distance between the morphological center points of the fissure region and the crack region as the analysis distance; the crack region with the smallest analysis distance to the fissure region is taken as the initial crack region that matches the corresponding fissure region.

7. The visual inspection method for the production quality of titanium alloy bars as described in claim 5, characterized in that, The true crack region is determined based on the difference in the extension direction between the gap region and the matching initial crack region, including: Linear fitting is performed on the gap region and the matching initial crack region to obtain the fitted line. The angle between the fitted line of the gap region and the fitted line of the crack region is used as the ratio of the angle to 180 degrees as the direction analysis index. In the gap region and the matching initial crack region, the ratio of the same number of pixels to the union of all pixels in both regions is determined as the region overlap. The difference between the regional overlap and the direction analysis index is calculated, normalized, and used as an indicator for crack judgment. Crack areas where the crack judgment index is greater than the preset judgment threshold are considered as real crack areas.

8. A visual inspection system for the production quality of titanium alloy bars, the system comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 7.

Images