Method and system for intelligently detecting cross section characteristics of round hole type material of composite rod wire

Abstract

The invention discloses an intelligent detection method and system for cross section characteristics of a round hole type material of a composite rod wire. The detection method comprises the following steps: calculating a first image pixel scale according to an image of a calibration plate; obtaining an edge image according to the material type cross section image; calculating the material type circumference length of each pixel point at the edge of the outer layer; according to the first image pixel scale and the horizontal leftmost pixel point and the horizontal rightmost pixel point of the outer-layer edge, calculating the circumferential broadening and the endpoint coordinates of the material type; calculating the height of the material type according to the endpoint coordinates of the circumferential broadening of the material type; calculating the coating thickness of each pixel point at the edge of the core part according to the edge image and the material type circumferential width; and calculating a core detection area and a total detection area, and calculating a coating detection area according to the core detection area and the total detection area. According to the invention, intelligent and high-precision detection of multiple characteristic data of the round hole type cross section of the composite rod wire is realized, and accurate data support is provided for optimization of the process and the hole type of the composite rod wire.

Description

Technical Field

[0001] The present invention belongs to the technical field of workpiece feature detection, and particularly relates to an intelligent detection method and system for the cross-sectional features of a round-hole-shaped profile of composite bar and wire materials. Background Art

[0002] Stainless steel-carbon steel composite bar and wire materials, that is, a composite steel material with stainless steel on the outer layer and carbon steel in the core, have good corrosion resistance of the outer stainless steel and a price about one-third of that of stainless steel, with significant advantages. The preparation process of stainless steel-carbon steel composite bar and wire materials includes non-oxidizing blanking of stainless steel pipes and carbon steel round bars, rolling forming of billets, pickling, etc. This process has initially realized industrial test production of stainless steel composite steel bars, stainless steel composite sucker rods, stainless steel clad grounding rods, etc.

[0003] During the rolling process of stainless steel-carbon steel composite bar and wire materials, compared with the rolling of traditional single-metal bar and wire materials, there are differences in the mechanical properties of the two metals, resulting in problems such as differences in the elongation rate of the profile, skewness, and uneven distribution of the stainless steel cladding. Generally, the head and tail samples cut off by the flying shear during the rolling process of composite bar and wire materials are all round-hole-shaped profiles. Taking the existing threaded steel rolling in a certain steel plant as an example, the rolling passes of threaded steel with different sizes are different, and the rolling mills are distributed horizontally and vertically in sequence. The rolling is usually divided into rough, medium, (pre) finish rolling. For example, rolling straight threaded steel with a specification of Φ28mm requires a total of 14 passes, with 6-4-4 passes for rough-medium-finish rolling respectively, and a total of 2 flying shears; rolling coiled threaded steel with a specification of Φ12mm requires a total of 22 passes, with 6-6-6-4 passes for rough-medium-pre finish-finish rolling respectively, and a total of 3 flying shears. The profiles obtained by the flying shears are all round-hole-shaped profiles.

[0004] For the profiles obtained due to faults such as flying shears, emergency stops, or mill steel piling during the rolling process of composite bar and wire materials, the on-site technical personnel's description of the characteristics of the profile (such as the cladding area and thickness distribution of the stainless steel cladding) by the naked eye is qualitative. Although subtle differences can be distinguished, it is difficult to clarify the specific magnitude, and it is easily affected by factors such as light, experience, subjective psychology, and vision, making it difficult to achieve accurate, standard, and objective evaluation of its characteristics. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent detection method and system for the cross-sectional features of a round-hole-shaped profile of composite bar and wire materials to solve the problem that it is difficult to achieve accurate, standard, and objective evaluation of the profile characteristics by qualitatively describing the profile characteristics with the naked eye in the traditional technology.

[0006] The present invention solves the above technical problems through the following technical solutions: An intelligent detection method for the cross-sectional features of a round-hole-shaped profile of composite bar and wire materials includes:

[0007] Obtaining a calibration plate image and a cross-sectional image of the profile;

[0008] Calculate the first image pixel scale based on the calibration plate image;

[0009] Obtain an edge image from the cross-sectional image of the material type, where the edge image includes an outer edge and a core edge;

[0010] Calculate the circumferential length of the material type for each pixel point of the outer edge based on the edge image;

[0011] Calculate the circumferential width expansion of the material type and the coordinates of its endpoints based on the first image pixel scale and the leftmost and rightmost horizontal pixel points of the outer edge;

[0012] Calculate the height of the material type based on the coordinates of the endpoints of the circumferential width expansion of the material type;

[0013] Calculate the cladding thickness for each pixel point of the core edge based on the edge image and the circumferential width expansion of the material type;

[0014] Calculate the core detection area and the total detection area based on the edge image, and calculate the cladding detection area based on the core detection area and the total detection area.

[0015] Furthermore, the process of obtaining the calibration plate image and the cross-sectional image of the material type is as follows:

[0016] Obtain a cross-sectional specimen of the material type;

[0017] Use a polishing machine to polish the cross-sectional specimen of the material type, then use a 4% nitric acid alcohol corrosion solution to corrode the polished cross-sectional specimen of the material type, and then clean and dry it to obtain a sample;

[0018] Place the calibration plate on one side of the sample, keep the width expansion direction of the sample horizontal, and collect the calibration plate and sample images;

[0019] Crop and segment the calibration plate and sample images to obtain the calibration plate image and the cross-sectional image of the material type.

[0020] Furthermore, calculating the first image pixel scale based on the calibration plate image includes:

[0021] Preprocess the calibration plate image; where the calibration plate image includes multiple calibration circles;

[0022] Extract multi-connected regions from the preprocessed calibration plate image, and measure the area of each multi-connected region and the major axis length and minor axis length of the ellipse having the same normalized second-order central moment as each multi-connected region;

[0023] Extract the multi-connected region i that simultaneously satisfies condition one and condition two: Condition one: ; Condition two: ; Among them, and respectively represent the areas of the multiply connected regions i and j; represents the area precision threshold; represents the number of multiply connected regions extracted from the preprocessed calibration plate image; and respectively represent the major axis length and minor axis length of the ellipse having the same normalized second-order central moment as the multiply connected region i; Calculate the first image pixel scale according to the multiply connected regions that simultaneously meet Condition 1 and Condition 2, and the specific calculation formula is: ; Among them, represents the first image pixel scale; represents the diameter of the calibration circle on the calibration plate; n represents the number of multiply connected regions that simultaneously meet Condition 1 and Condition 2.

[0024] Furthermore, calculate the material type circumferential length of each pixel point on the outer edge according to the edge image, including:

[0025] Take the leftmost horizontal pixel point on the outer edge as the first point, sort all the pixel points on the outer edge, and record the serial number of the highest vertical pixel point as Nb', the serial number of the rightmost horizontal pixel point as Nc', and the serial number of the lowest vertical pixel point as Nd';

[0026] Calculate the distance between each pixel point on the outer edge and other pixel points, and the specific calculation formula is:

[0027] ;

[0028] Among them, represents the distance between the i-th pixel point and the j-th pixel point on the outer edge, , , represents the number of pixel points on the outer edge; represents the coordinates of the i-th pixel point on the outer edge; represents the coordinates of the j-th pixel point on the outer edge;

[0029] Find the maximum distance between each pixel point on the outer edge and other pixel points, and determine whether the maximum distance is greater than the first reference distance. If not, use the first reference distance as the maximum distance; among them, the determination method of the first reference distance is:

[0030] If all the pixel points on the outer edge are sorted in the clockwise direction and i < Nc', the first reference distance is the distance between the i-th pixel point and the lowest vertical pixel point; if all the pixel points on the outer edge are sorted in the clockwise direction and i > Nc', the first reference distance is the distance between the i-th pixel point and the highest vertical pixel point;

[0031] If all the pixel points on the outer edge are sorted in the counterclockwise direction and i < Nc', the first reference distance is the distance between the i-th pixel point and the highest vertical pixel point; if all the pixel points on the outer edge are sorted in the counterclockwise direction and i > Nc', the first reference distance is the distance between the i-th pixel point and the lowest vertical pixel point;

[0032] The maximum distance between each pixel point on the outer edge and other pixel points is the circumferential length of the material type of each pixel point on the outer edge.

[0033] Furthermore, calculating the circumferential width expansion of the material type and its endpoint coordinates according to the first image pixel scale and the leftmost horizontal pixel point and the rightmost horizontal pixel point of the outer edge includes:

[0034] Taking the leftmost horizontal pixel point of the outer edge as the first point, sorting all the pixel points on the outer edge, and recording the serial number of the rightmost horizontal pixel point as Nc';

[0035] Determining the connection line between the leftmost horizontal pixel point and the rightmost horizontal pixel point of the outer edge;

[0036] According to the connection line, determining the serial number deviation between the endpoint serial numbers of the circumferential width expansion of the material type and the serial numbers of the leftmost horizontal pixel point and the rightmost horizontal pixel point of the outer edge, and further determining the endpoint serial numbers of the circumferential width expansion of the material type;

[0037] Calculating the circumferential width expansion of the material type according to the endpoint coordinates of the circumferential width expansion of the material type and the first image pixel scale, and the specific calculation formula is:

[0038] ;

[0039] Among them, represents the circumferential width expansion of the material type; 、 respectively represent the coordinates of the two endpoints of the circumferential width expansion of the material type; represents the first image pixel scale.

[0040] Furthermore, the detection method further includes correcting the circumferential width expansion of the material type, specifically including:

[0041] Obtaining the actual width expansion of the cross-section of the material type, and calculating the width expansion error rate according to the actual width expansion of the cross-section of the material type and the circumferential width expansion of the material type;

[0042] Determine whether the spreading error rate is less than the error rate threshold. If not, calculate the second image pixel scale according to the endpoint coordinates of the circumferential spreading of the stock shape and the actual spreading of the cross-section of the stock shape. The specific calculation formula is:

[0043] ;

[0044] wherein, represents the second image pixel scale; represents the actual spreading of the cross-section of the stock shape; and respectively represent the coordinates of the two endpoints of the circumferential spreading of the stock shape;

[0045] Calculate the corrected circumferential spreading of the stock shape according to the second image pixel scale and the endpoint coordinates of the circumferential spreading of the stock shape. The specific formula is:

[0046] ;

[0047] wherein, represents the corrected circumferential spreading of the stock shape.

[0048] Furthermore, calculate the height of the stock shape according to the endpoint coordinates of the circumferential spreading of the stock shape, specifically including:

[0049] Determine the abscissa of the midpoint in the spreading direction according to the endpoint coordinates of the circumferential spreading of the stock shape;

[0050] Determine the abscissa value range according to the set abscissa deviation and the abscissa of the midpoint;

[0051] Determine the upper intersection point number and the lower intersection point number of the vertical line passing through each abscissa within the abscissa value range with the outer edge, and then determine the ordinate corresponding to each abscissa within the abscissa value range and the midpoint in the spreading direction;

[0052] Determine the coordinates of the upper groove bottom arc center point corresponding to each midpoint according to the vertical distance difference between the upper groove bottom arc center point of the roll pass and the actual midpoint in the spreading direction, and determine the coordinates of the lower groove bottom arc center point corresponding to each midpoint according to the vertical distance difference between the lower groove bottom arc center point of the roll pass and the actual midpoint in the spreading direction;

[0053] Set the value range of the upper intersection point number and the lower intersection point number;

[0054] Calculate the distance between each upper intersection point within the value range of the upper intersection point number and the corresponding upper groove bottom arc center point, and the distance between each lower intersection point within the value range of the lower intersection point number and the corresponding lower groove bottom arc center point. All the distances form a distance set;

[0055] Calculate the variance of each distance in the distance set;

[0056] Select the middle point corresponding to the distance with the smallest variance as the center point of the stock shape;

[0057] Determine the upper intersection point and the lower intersection point of the vertical line passing through the center point of the stock shape and the outer edge, and calculate the height of the stock shape based on the upper intersection point and the lower intersection point.

[0058] Furthermore, calculate the cladding thickness of each pixel point on the core edge according to the edge image and the circumferential spread of the stock shape, including:

[0059] Take the left intersection points of the line where the circumferential spread of the stock shape is located with the core edge and the outer edge as the first points of the core edge and the outer edge, and sort all the pixel points of the core edge and the outer edge respectively;

[0060] Take the distance between the first point of the core edge and the first point of the outer edge as the second reference distance;

[0061] Calculate the distance between each pixel point on the core edge and other pixel points on the outer edge; wherein, the other pixel points on the outer edge refer to the pixel points except the first pixel point on the outer edge;

[0062] Find the minimum distance between each pixel point on the core edge and other pixel points on the outer edge;

[0063] If the minimum distance is less than the second reference distance, the minimum distance is the cladding thickness of the corresponding pixel point on the core edge; if the minimum distance is greater than or equal to the second reference distance, the second reference distance is the cladding thickness of the corresponding pixel point on the core edge.

[0064] Furthermore, the detection method further includes calculating the circumferential evaluation index of the stock shape, and then evaluating the circumferential characteristics of the stock shape; wherein, the circumferential evaluation index of the stock shape includes the relative reduction of the current pass, the relative spread of the current pass, the circumferential non-uniformity coefficient of the stock shape circumference, the circumferential fluctuation coefficient of the stock shape circumference, the local weakness index of the stock shape circumference, the ear size, the proportion of the cladding detection area in the current pass, the total area error rate, the core area error rate, the cladding area error rate, and the elongation coefficient of the current pass;

[0065] Calculate the relative reduction of the current pass according to the height of the stock shape and the height of the billet before rolling;

[0066] Calculate the relative spread of the current pass according to the circumferential spread of the stock shape and the width spread of the billet before rolling;

[0067] Calculate the maximum material type circumferential length, minimum material type circumferential length, average value of material type circumferential length, and standard deviation of material type circumferential length based on the circumferential length of each pixel point on the outer edge. Use the ratio of the maximum material type circumferential length to the minimum material type circumferential length as the circumferential non-uniformity coefficient of the material type circumference, use the ratio of the standard deviation of the material type circumferential length to the average value of the material type circumferential length as the circumferential fluctuation coefficient of the material type circumference, and use the ratio of the minimum material type circumferential length to the average value of the material type circumferential length as the local weakness index of the material type circumference;

[0068] Calculate the ear size of the material type cross-section based on the material type circumference spread and the theoretically designed spread;

[0069] Calculate the proportion of the clad layer detection area in the current pass based on the clad layer detection area and the total detection area;

[0070] Calculate the total area error rate based on the theoretically designed total area and the total detection area, calculate the core area error rate based on the theoretically designed core area and the core detection area, and calculate the clad layer area error rate based on the theoretically designed clad layer area and the clad layer detection area;

[0071] Calculate the elongation coefficient of the current pass based on the cross-sectional area of the billet and the total detection area.

[0072] Further, the detection method further includes calculating a material type clad layer evaluation index to evaluate the characteristics of the material type clad layer; wherein, the material type clad layer evaluation index includes the circumferential non-uniformity coefficient of the clad layer, the circumferential fluctuation coefficient of the clad layer, the local weakness index of the clad layer, the number of wave peaks, and the number of wave valleys;

[0073] Calculate the maximum clad layer thickness, minimum clad layer thickness, average value of clad layer thickness, and standard deviation of clad layer thickness based on the clad layer thickness of each pixel point on the core edge. Use the ratio of the maximum clad layer thickness to the minimum clad layer thickness as the circumferential non-uniformity coefficient of the clad layer, use the ratio of the standard deviation of the clad layer thickness to the average value of the clad layer thickness as the circumferential fluctuation coefficient of the clad layer, and use the ratio of the minimum clad layer thickness to the average value of the clad layer thickness as the local weakness index of the clad layer;

[0074] Draw a clad layer thickness curve based on the clad layer thickness of each pixel point on the core edge, and draw a horizontal line in the clad layer thickness curve. Determine the number of wave peaks and the number of wave valleys based on the horizontal line.

[0075] Based on the same concept, the present invention further provides an intelligent detection system for the cross-sectional characteristics of a composite bar and wire round pass material type, and the detection system includes:

[0076] An acquisition unit for acquiring a calibration plate image and a material type cross-sectional image; obtaining an edge image according to the material type cross-sectional image, wherein the edge image includes an outer edge and a core edge;

[0077] The first calculation unit is configured to calculate a first image pixel scale based on the calibration plate image;

[0078] The second calculation unit is configured to calculate the circumferential length of each pixel point of the outer edge based on the edge image;

[0079] The third calculation unit is configured to calculate the width expansion of the profile circumference and the coordinates of its end points based on the first image pixel scale and the leftmost and rightmost horizontal pixel points of the outer edge;

[0080] The fourth calculation unit is configured to calculate the profile height based on the coordinates of the end points of the width expansion of the profile circumference;

[0081] The fifth calculation unit is configured to calculate the cladding thickness of each pixel point of the core edge based on the edge image and the width expansion of the profile circumference;

[0082] The sixth calculation unit is configured to calculate the core detection area and the total detection area based on the edge image, and calculate the cladding detection area based on the core detection area and the total detection area.

[0083] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0084] The present invention can quantitatively describe the characteristics of the profile cross-section such as the circumferential length, width expansion, height, cladding thickness, and cladding area, solves the problem of inaccurate detection of the profile cross-section characteristics of the round-hole profile by manual on-site, realizes the intelligent and high-precision detection of multiple characteristic data of the round-hole cross-section of the composite bar wire, and provides accurate data support for the process and pass optimization of the composite bar wire.

[0085] The present invention uses the actual width expansion of the profile cross-section to perform accuracy inspection and data correction on the detected profile circumference width expansion, improves the detection accuracy of the profile width expansion, solves the problem that it is difficult to accurately detect the width expansion and other characteristic data of the round-hole profile cross-section during the rolling process of the composite bar wire, realizes the intelligent detection of the characteristic data of the round-hole profile cross-section of the composite bar wire, with good stability, high accuracy, and fast speed. Description of the Drawings

[0086] In order to more clearly illustrate the technical solutions of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only one embodiment of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0087] Figure 1 It is a flowchart of the intelligent detection method for the characteristics of the round-hole profile cross-section of the composite bar wire in the first embodiment of the present invention;

[0088] Figure 2It is a schematic diagram of the hot rolling production line and profile acquisition of Φ12 mm composite spiral ribbed steel bars in the second embodiment of the present invention;

[0089] Figure 3 It is an image of the calibration plate and the sample F1 of the No. 1 flying shear after corrosion in the second embodiment of the present invention;

[0090] Figure 4 It is an image of the calibration plate after pretreatment in the second embodiment of the present invention;

[0091] Figure 5 It is the calibration circle selected according to Condition 1 and Condition 2 in the second embodiment of the present invention;

[0092] Figure 6 It is an image of the core carbon steel filling in the second embodiment of the present invention;

[0093] Figure 7 It is an image of the edge of the core carbon steel in the second embodiment of the present invention;

[0094] Figure 8 It is an image of the outer edge in the second embodiment of the present invention;

[0095] Figure 9 It is an edge image in the second embodiment of the present invention;

[0096] Figure 10 It is a characteristic data diagram in the second embodiment of the present invention;

[0097] Figure 11 It is a profile circumferential length curve in the second embodiment of the present invention;

[0098] Figure 12 It is a distance curve from the outer edge to the center point of the profile in the second embodiment of the present invention;

[0099] Figure 13 It is a clad layer thickness curve in the second embodiment of the present invention;

[0100] Figure 14 It is an image of the calibration plate and the round hole profile S1 of the No. 1 flying shear of Φ16 mm composite steel bars after corrosion in the third embodiment of the present invention;

[0101] Figure 15 It is a clad layer thickness change curve in the third embodiment of the present invention. Detailed implementation manners

[0102] Next, in combination with the accompanying drawings in the embodiments of the present invention, the technical solutions in the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without making creative efforts shall fall within the protection scope of the present invention.

[0103] The technical solution of the present invention will be described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0104] Embodiment 1

[0105] As Figure 1 shown, the intelligent detection method for the cross-sectional characteristics of the round-hole pass of the composite bar and wire rod provided by the embodiment of the present invention includes the following steps:

[0106] Step 1: Obtain a calibration plate image and a cross-sectional image of the pass.

[0107] During the rolling process in the industrial site of composite bar and wire rod, an intermediate-pass round-hole pass is obtained by flying shear or emergency stop, and then a cross-sectional specimen of the pass is obtained by wire cutting along the normal plane of the rolling direction of the pass. The thickness of the specimen is 5 - 10 mm, which is polished using a polishing machine, corroded with a 4% nitric acid alcohol corrosion solution for 1 - 2 minutes, slowly washed in water and dried to obtain a sample.

[0108] Place the calibration plate on one side of the sample, and keep the width expansion direction of the sample as horizontal as possible. Use a high-definition camera to collect the calibration plate and sample images. Use image processing software to extract the calibration plate image and the cross-sectional image of the pass from the calibration plate and sample images.

[0109] The present invention uses the difference in corrosion resistance of bimetals and a uniform light source to eliminate the influence of ambient light and has strong anti-interference ability.

[0110] Step 2: Calculate the first image pixel scale according to the calibration plate image.

[0111] In the specific embodiment of the present invention, calculating the first image pixel scale according to the calibration plate image specifically includes:

[0112] Step 2.1: Preprocess the calibration plate image; wherein, the calibration plate image contains multiple calibration circles.

[0113] In this embodiment, the preprocessing of the calibration plate image sequentially includes grayscale conversion, median filtering, and binary inversion processing.

[0114] Step 2.2: Extract multi-connected regions from the preprocessed calibration plate image, and measure the area of each multi-connected region and the major axis length and minor axis length of the ellipse having the same normalized second-order central moment as each multi-connected region.

[0115] The multi-connected region in this embodiment is an 8-connected region. The 8-connected regions in the calibration plate image include not only the calibration circles, but also the bounding boxes, etc. The number of 8-connected regions is greater than the number of calibration circles. In the MATLAB software, the function regionprops is used to measure the attributes of each 8-connected region, such as area, centroid, minimum bounding rectangle, lengths of the major and minor axes of the region, etc.

[0116] Step 2.3: Extract the multi-connected region i that simultaneously satisfies Condition 1 and Condition 2:

[0117] Condition 1: ;

[0118] Condition 2: ;

[0119] where , represent the areas of the multi-connected regions i and j, respectively; represents the area accuracy threshold; represents the number of multi-connected regions extracted from the preprocessed calibration plate image; , represent the major axis length and minor axis length of the ellipse having the same normalized second-order central moment as the multi-connected region i, respectively.

[0120] In this embodiment, the area accuracy threshold is set to 0.1. Some calibration circles are screened out through the area limitation of Condition 1 and the bounding box limitation of Condition 2.

[0121] Step 2.4: Calculate the first image pixel scale according to the multi-connected regions that simultaneously satisfy Condition 1 and Condition 2. The specific calculation formula is:

[0122] (1)

[0123] where represents the first image pixel scale; represents the diameter of the calibration circle in the calibration plate; n represents the number of multi-connected regions that simultaneously satisfy Condition 1 and Condition 2.

[0124] Step 3: Obtain the edge image according to the cross-sectional image of the material type.

[0125] The core edge and the outer edge are extracted from the cross-sectional image of the material type through an edge detection algorithm, and then the core edge and the outer edge are combined to obtain the edge image.

[0126] Step 4: Calculate the material type circumferential length of each pixel point of the outer edge according to the edge image.

[0127] In a specific embodiment of the present invention, calculating the circumferential length of the material type for each pixel point of the outer edge based on the edge image specifically includes:

[0128] Step 4.1: Taking the leftmost horizontal pixel point of the outer edge as the first point and denoting it as Na'(1), sorting all the pixel points of the outer edge, and denoting the serial number of the highest vertical pixel point as Nb', the serial number of the rightmost horizontal pixel point as Nc', and the serial number of the lowest vertical pixel point as Nd'.

[0129] Step 4.2: Calculating the distance between each pixel point of the outer edge and other pixel points, and the specific calculation formula is:

[0130] (2)

[0131] Wherein, represents the distance between the i-th pixel point and the j-th pixel point of the outer edge, , , represents the number of pixel points of the outer edge; represents the coordinates of the i-th pixel point of the outer edge; represents the coordinates of the j-th pixel point of the outer edge.

[0132] Step 4.3: Finding the maximum distance between each pixel point of the outer edge and other pixel points, and determining whether the maximum distance is greater than the first reference distance. If not, taking the first reference distance as the maximum distance.

[0133] For the i-th pixel point of the outer edge, finding the maximum value in it and denoting it as , that is, it represents the maximum distance between the i-th pixel point of the outer edge and other pixel points. If is less than the first reference distance, then taking the first reference distance as .

[0134] In this embodiment, the determination method of the first reference distance is:

[0135] If all the pixel points of the outer edge are sorted in the clockwise direction and i < Nc', then the first reference distance is the distance between the i-th pixel point and the lowest vertical pixel point, that is, the distance between the i-th pixel point and Nd'; if all the pixel points of the outer edge are sorted in the clockwise direction and i > Nc', then the first reference distance is the distance between the i-th pixel point and the highest vertical pixel point, that is, the distance between the i-th pixel point and Nb'.

[0136] If all the pixel points on the outer edge are sorted in the counterclockwise direction and i < Nc', the first reference distance is the distance between the i-th pixel point and the highest vertical pixel point, that is, the distance between the i-th pixel point and Nb'; if all the pixel points on the outer edge are sorted in the counterclockwise direction and i > Nc', the first reference distance is the distance between the i-th pixel point and the lowest vertical pixel point, that is, the distance between the i-th pixel point and Nd'.

[0137] Step 4.4: The maximum distance between each pixel point on the outer edge and other pixel points is the circumferential length of the material type for each pixel point on the outer edge.

[0138] Find the maximum value from the circumferential lengths of the material types of all pixel points on the outer edge and denote it as , which represents the maximum circumferential length of the material type, and determine the serial number of the pixel point corresponding to the maximum circumferential length of the material type . Similarly, find the minimum circumferential length of the material type and its corresponding pixel point serial number.

[0139]

[0140] Step 5: Calculate the circumferential width expansion of the material type and the coordinates of its endpoints according to the first image pixel scale and the leftmost horizontal pixel point and the rightmost horizontal pixel point on the outer edge.

[0140] In the specific implementation manner of the present invention, calculating the circumferential width expansion of the material type and the coordinates of its endpoints according to the first image pixel scale and the leftmost horizontal pixel point and the rightmost horizontal pixel point on the outer edge includes:

[0141] Step 5.1: Take the leftmost horizontal pixel point on the outer edge as the first point and denote it as Na'(1), sort all the pixel points on the outer edge, and denote the serial number of the highest vertical pixel point as Nb', the serial number of the rightmost horizontal pixel point as Nc', and the serial number of the lowest vertical pixel point as Nd'.

[0142] Step 5.2: Determine the connection line between the leftmost horizontal pixel point and the rightmost horizontal pixel point on the outer edge, that is, connect the point labeled Na'(1) and the point labeled Nc'.

[0143] Step 5.3: Determine the sequence number deviation between the sequence numbers of the endpoints of the circumferential width expansion of the material type and the sequence numbers of the leftmost horizontal pixel point Na'(1) and the rightmost horizontal pixel point Nc' on the outer edge according to the connection line in Step 5.2, and then determine the sequence numbers of the endpoints of the circumferential width expansion of the material type.

[0144] During the rolling process of composite bar wire, the shape of the round pass is a flat ellipse or a vertical ellipse. The line connecting the leftmost horizontal pixel point and the rightmost horizontal pixel point on the outer edge basically coincides with the circumferential spread or the height of the pass shape, but there is a certain error. Taking the No. 1 flying shear as an example, the pass shape is a flat ellipse, and the line connecting the leftmost horizontal pixel point and the rightmost horizontal pixel point on the outer edge is basically consistent with the circumferential spread of the pass shape. Plot the serial number Na'(1) of the leftmost horizontal pixel point, the serial number Nc' of the rightmost horizontal pixel point on the outer edge and the connecting line in the cross-sectional image of the pass shape. By observing and analyzing, it is determined that the serial number deviation between the end point serial number of the circumferential spread of the pass shape and the serial number Na'(1) of the leftmost horizontal pixel point and the serial number Nc' of the rightmost horizontal pixel point on the outer edge is , , then the end point serial numbers of the circumferential spread of the pass shape are:

[0145] , (3)

[0146] Among them, , respectively represent the left end point serial number and the right end point serial number of the circumferential spread of the pass shape. , When it is a positive value, it means clockwise upward adjustment; , When it is a negative value, it means counterclockwise downward adjustment.

[0147] Step 5.4: Calculate the circumferential spread of the pass shape according to the end point coordinates of the circumferential spread of the pass shape and the first image pixel scale. The specific calculation formula is:

[0148] (4)

[0149] Among them, represents the circumferential spread of the pass shape; , respectively represent the coordinates of the two end points of the circumferential spread of the pass shape; represents the first image pixel scale.

[0150] In order to improve the detection accuracy of the circumferential spread of the pass shape and other characteristic data, the present invention also corrects the circumferential spread of the pass shape, specifically including:

[0151] Step 5.5: Obtain the actual spread of the cross section of the pass shape, and calculate the spread error rate according to the actual spread of the cross section of the pass shape and the circumferential spread of the pass shape.

[0152] Use a vernier caliper to measure the actual spread of the cross section of the sample in step 1. The calculation formula of the spread error rate is:

[0153] (5)

[0154] Among them, represents the spreading error rate, and B represents the actual spreading of the cross-section of the material type.

[0155] Step 5.6: Determine the spreading error rate Whether it is less than the error rate threshold. If so, the calculated circumferential spreading of the material type has high reliability and accuracy, and there is no need to correct the circumferential spreading of the material type ; if not, calculate the second image pixel scale according to the endpoint coordinates of the circumferential spreading of the material type and the actual spreading of the cross-section of the material type. The specific calculation formula is:

[0156] (6)

[0157] Among them, represents the second image pixel scale.

[0158] Step 5.7: Calculate the corrected circumferential spreading of the material type according to the second image pixel scale and the endpoint coordinates of the circumferential spreading of the material type. The specific formula is:

[0159] (7)

[0160] Among them, represents the corrected circumferential spreading of the material type. The corrected circumferential spreading of the material type is basically equal to the actual spreading B of the cross-section of the material type.

[0161] Step 6: Calculate the height of the material type according to the endpoint coordinates of the circumferential spreading of the material type.

[0162] In the specific embodiment of the present invention, calculating the height of the material type according to the endpoint coordinates of the circumferential spreading of the material type specifically includes:

[0163] Step 6.1: Determine the abscissa of the midpoint in the spreading direction according to the endpoint coordinates of the circumferential spreading of the material type.

[0164] Taking the left endpoint of the circumferential spreading of the material type as the first point, sort all the pixel points on the outer edge. The serial number of the right endpoint of the circumferential spreading of the material type is denoted as , then the abscissa of the midpoint in the spreading direction is:

[0165] (8)

[0166] Among them, represents the abscissa of the midpoint in the spreading direction; represents the abscissa of the left endpoint of the circumferential spreading of the material type; represents the abscissa of the right endpoint of the circumferential spreading of the material type; represents the rounding function.

[0167] Step 6.2: Determine the abscissa value range according to the set abscissa deviation and the abscissa of the midpoint.

[0168] The set abscissa deviation is denoted as , then the abscissa value range is ( , ).

[0169] Step 6.3: Determine the upper intersection point number and the lower intersection point number of the vertical line passing through each abscissa within the abscissa value range with the outer edge, and then determine the ordinate corresponding to each abscissa within the abscissa value range and the midpoint in the spreading direction.

[0170] Each abscissa within the abscissa value range is denoted as , , , the vertical line passing through each abscissa intersects with the outer edge, and the upper intersection point number and the lower intersection point number are denoted as and respectively. The ordinate corresponding to each abscissa is , where and represent the ordinates of the upper and lower intersection points of the vertical line passing through the abscissa with the outer edge respectively. The coordinates of each midpoint in the spreading direction are ( , ).

[0171] Step 6.4: Determine the coordinates of the upper groove bottom arc center point corresponding to each midpoint according to the vertical distance difference between the upper groove bottom arc center point of the roll pass and the actual midpoint in the spreading direction, and determine the coordinates of the lower groove bottom arc center point corresponding to each midpoint according to the vertical distance difference between the lower groove bottom arc center point of the roll pass and the actual midpoint in the spreading direction.

[0172] The vertical distance difference between the upper groove bottom arc center point of the roll pass and the actual midpoint in the spreading direction is equal to the vertical distance difference between the lower groove bottom arc center point of the roll pass and the actual midpoint in the spreading direction. Let this vertical distance difference be , and the vertical distance difference can be obtained according to the roll pass drawing.

[0173] The coordinates of the upper groove bottom arc center point O1(k) corresponding to each midpoint k are , and the coordinates of the lower groove bottom arc center point O2(k) corresponding to each midpoint k are , where PX represents the image pixel scale. If the circumferential spreading of the material type meets the accuracy requirements and no correction is required, then PX is ; if the circumferential spreading of the material type does not meet the accuracy requirements and correction is required, then PX is 。

[0174] Step 6.5: Set the value ranges of the upper intersection point number and the lower intersection point number.

[0175] In this embodiment, let the value ranges of the upper intersection point number and the lower intersection point number of the vertical line passing through each abscissa within the abscissa value range and the outer edge be K. Then the upper intersection point number is , and the lower intersection point number is , 。

[0176] Step 6.6: Calculate the distances between each upper intersection point within the value range of the upper intersection point number and the corresponding upper slot bottom arc center point, and the distances between each lower intersection point within the value range of the lower intersection point number and the corresponding lower slot bottom arc center point. All the distances form a distance set.

[0177] Step 6.7: Calculate the variance of each distance in the distance set.

[0178] Step 6.8: Select the middle point corresponding to the distance with the minimum variance as the center point of the material shape, that is, select the k corresponding to the distance with the minimum variance. At this time, k is denoted as , The corresponding abscissa and ordinate are used as the center point of the material shape. That is, the coordinates of the center point of the material shape are ( , ).

[0179] Step 6.9: Determine the upper and lower intersection points of the vertical line passing through the center point of the material shape and the outer edge. Calculate the height of the material shape according to the upper and lower intersection points. The specific calculation formula is:

[0180] (9)

[0181] Wherein, represents the height of the material shape; and respectively represent the upper and lower intersection points of the vertical line passing through the center point of the material shape and the outer edge. The pixel scale selected when calculating the circumferential spread of the material shape in Step 5 is used as the pixel scale when calculating the height of the material shape. The height error rate can be calculated according to the height of the material shape and the actual measured height.

[0182] In this embodiment, the distances from the center point of the material shape to each pixel point of the outer edge are also calculated. The specific calculation formula is:

[0183] (10)

[0184] Wherein, represents the distance from the center point of the material shape to each pixel point of the outer edge; Pixels representing the outer edge; Represents the center point of the material type. The PX in formula (10) is determined by the PX in formula (9), that is, if the height of the material type is calculated according to Calculated, then the PX in formula (10) is ; if the height of the material type is calculated according to Calculated, then the PX in formula (10) is .

[0185] Step 7: Calculate the cladding thickness of each pixel point on the core edge according to the edge image and the circumferential spread of the material type.

[0186] In a specific embodiment of the present invention, calculating the cladding thickness of each pixel point on the core edge according to the edge image and the circumferential spread of the material type includes:

[0187] Step 7.1: Take the intersection points of the line where the circumferential spread of the material type is located with the left side of the core edge and the outer edge as the first points of the core edge and the outer edge, and sort all the pixel points of the core edge and the outer edge respectively.

[0188] Exemplarily, take the intersection point of the line where the circumferential spread of the material type is located with the left side of the core edge as the first point of the core edge, and sort all the pixel points of the core edge in the clockwise direction; take the intersection point of the line where the circumferential spread of the material type is located with the left side of the outer edge as the first point of the outer edge, and sort all the pixel points of the outer edge in the clockwise direction.

[0189] Step 7.2: Take the distance between the first point of the core edge and the first point of the outer edge as the second reference distance.

[0190] Step 7.3: Calculate the distance between each pixel point on the core edge and other pixel points on the outer edge; where the other pixel points on the outer edge refer to the pixel points other than the first pixel point on the outer edge.

[0191] Step 7.4: Find the minimum distance between each pixel point on the core edge and other pixel points on the outer edge.

[0192] For each pixel point on the core edge, find the minimum value from the distances between this pixel point and each other pixel point on the outer edge, then this minimum value is the minimum distance between this pixel point and other pixel points on the outer edge.

[0193] Step 7.5: Determine whether the minimum distance between each pixel point on the core edge and other pixel points on the outer edge is less than the second reference distance.

[0194] If the minimum distance is less than the second reference distance, the minimum distance is the cladding thickness of the corresponding pixel point on the core edge; if the minimum distance is greater than or equal to the second reference distance, the second reference distance is the cladding thickness of the corresponding pixel point on the core edge. Thus, the cladding thickness of each pixel point on the core edge can be calculated.

[0195] Step 8: Calculate the core detection area and the total detection area based on the edge image, and calculate the cladding detection area based on the core detection area and the total detection area.

[0196] Fill the core edge in the edge image, and then calculate the filled area to obtain the core detection area; fill the outer edge in the edge image, and then calculate the filled area to obtain the total detection area; the cladding detection area is equal to the difference between the total detection area and the core detection area.

[0197] Step 9: Calculate the evaluation index of the billet shape circumference, and then evaluate the characteristics of the billet shape circumference.

[0198] In this embodiment, the evaluation index of the billet shape circumference includes the relative reduction of the current pass, the relative spread of the current pass, the circumferential non-uniformity coefficient of the billet shape circumference, the circumferential fluctuation coefficient of the billet shape circumference, the local weakness index of the billet shape circumference, the ear size, the ratio of the cladding detection area of the current pass, the total area error rate, the core area error rate, the cladding area error rate, and the elongation coefficient of the current pass.

[0199] Calculate the relative reduction of the current pass according to the billet shape height and the height of the blank before rolling. The specific calculation formula is:

[0200] (11)

[0201] Where, represents the relative reduction of the current pass; represents the height of the blank before rolling, which is obtained by measurement; h represents the billet shape height, which is calculated in Step 6.

[0202] Calculate the relative spread of the current pass according to the circumferential spread of the billet shape and the circumferential spread of the blank before rolling. The specific calculation formula is:

[0203] (12)

[0204] Where, represents the relative spread of the current pass; represents the circumferential spread of the billet shape, which is calculated in Step 5. When the circumferential spread of the billet shape needs to be corrected, the circumferential spread of the billet shape here is the corrected circumferential spread of the billet shape; represents the circumferential spread of the blank before rolling, which is obtained by measurement.

[0205] Calculate the maximum material type circumferential length, minimum material type circumferential length, average value of material type circumferential length, and standard deviation of material type circumferential length based on the circumferential length of each pixel point on the outer edge. Use the ratio of the maximum material type circumferential length to the minimum material type circumferential length as the circumferential non-uniformity coefficient of the material type circumference, the ratio of the standard deviation of the material type circumferential length to the average value of the material type circumferential length as the circumferential fluctuation coefficient of the material type circumference, and the ratio of the minimum material type circumferential length to the average value of the material type circumferential length as the local weakness index of the material type circumference.

[0206] In this embodiment, when the circumferential non-uniformity coefficient of the material type circumference is less than 1.3, it indicates that the circumferential direction of the material type circumference is uniform; when the circumferential non-uniformity coefficient of the material type circumference is greater than or equal to 1.3, it indicates that the circumferential direction of the material type circumference is non-uniform. When the circumferential fluctuation coefficient of the material type circumference is less than 0.1, it indicates that the circumferential fluctuation of the material type circumference is small; when the circumferential non-uniformity coefficient of the material type circumference is greater than or equal to 0.1, it indicates that the circumferential fluctuation of the material type circumference is large. When the local weakness index of the material type circumference is greater than 0.8, it indicates that the local part of the material type circumference is not weak; when the local weakness index of the material type circumference is less than or equal to 0.8, it indicates that the local part of the material type circumference is weak.

[0207] Calculate the size of the fin on the cross-section of the material type according to the spread of the material type circumference and the theoretically designed spread. The specific calculation formula is:

[0208] (13)

[0209] Where, represents the size of the fin on the cross-section of the material type; represents the spread of the material type circumference, which is obtained by step 5. When the spread of the material type circumference needs to be corrected, the spread of the material type circumference here is the corrected spread of the material type circumference; represents the theoretically designed spread.

[0210] Calculate the proportion of the cladding detection area in the current pass according to the cladding detection area and the total detection area. The specific calculation formula is:

[0211] (14)

[0212] Where, represents the proportion of the cladding detection area in the current pass, represents the cladding detection area, represents the total detection area.

[0213] Let the outer diameter of the outer layer of the composite bar and wire be 、the thickness be , the cross-sectional area of the outer layer of the initial blank be , and the cross-sectional area of the initial blank be . Then the calculation formula for the proportion of the outer layer cross-sectional area is:

[0214] (15)

[0215] Wherein, represents the proportion of the outer cross-sectional area.

[0216] Calculate the total area error rate Serr according to the theoretically designed total area and the total detected area, calculate the core area error rate Sterr according to the theoretically designed core area and the core detected area, and calculate the cladding area error rate Sberr according to the theoretically designed cladding area and the cladding detected area.

[0217] Calculate the current pass elongation coefficient according to the blank cross-sectional area and the total detected area. The specific calculation formula is:

[0218] (16)

[0219] Wherein, represents the current pass elongation coefficient.

[0220] Step 10: Calculate the profile cladding evaluation index, and then evaluate the profile cladding characteristics.

[0221] In this embodiment, the profile cladding evaluation index includes the circumferential non-uniformity coefficient of the cladding, the circumferential fluctuation coefficient of the cladding, the local weakness index of the cladding, the number of wave peaks, and the number of wave valleys.

[0222] Calculate the maximum cladding thickness, the minimum cladding thickness, the average cladding thickness, and the standard deviation of the cladding thickness according to the cladding thickness of each pixel point at the core edge. Take the ratio of the maximum cladding thickness to the minimum cladding thickness as the circumferential non-uniformity coefficient of the cladding, take the ratio of the standard deviation of the cladding thickness to the average cladding thickness as the circumferential fluctuation coefficient of the cladding, and take the ratio of the minimum cladding thickness to the average cladding thickness as the local weakness index of the cladding.

[0223] In this embodiment, in this embodiment, when the circumferential non-uniformity coefficient of the cladding is less than 1.3, it indicates that the cladding is circumferentially uniform; when the circumferential non-uniformity coefficient of the cladding is greater than or equal to 1.3, it indicates that the cladding is circumferentially non-uniform. When the circumferential fluctuation coefficient of the cladding is less than 0.1, it indicates that the circumferential fluctuation of the cladding is small; when the circumferential non-uniformity coefficient of the cladding is greater than or equal to 0.1, it indicates that the circumferential fluctuation of the cladding is large. When the local weakness index of the cladding is greater than 0.8, it indicates that the cladding is not locally weak; when the local weakness index of the cladding is less than or equal to 0.8, it indicates that the cladding is locally weak.

[0224] In order to better characterize the thickness uniformity of the profile cladding, draw a cladding thickness curve according to the cladding thickness of each pixel point at the core edge, and draw a horizontal line (average value of the cladding thickness ± deviation) in the cladding thickness curve, and determine the number of wave peaks and the number of wave valleys based on the horizontal line. Evaluate the cladding uniformity according to the number of wave peaks and the number of wave valleys.

[0225] The intelligent detection system for the cross-sectional characteristics of the round-hole pass of the composite bar and wire rod provided by the embodiment of the present invention includes:

[0226] An acquisition unit, configured to acquire a calibration plate image and a cross-sectional image of the pass; obtain an edge image according to the cross-sectional image of the pass, where the edge image includes an outer edge and a core edge;

[0227] A first calculation unit, configured to calculate a first image pixel scale according to the calibration plate image;

[0228] A second calculation unit, configured to calculate the circumferential length of the pass for each pixel point of the outer edge according to the edge image;

[0229] A third calculation unit, configured to calculate the circumferential spread of the pass and the coordinates of its end points according to the first image pixel scale and the leftmost horizontal pixel point and the rightmost horizontal pixel point of the outer edge;

[0230] A fourth calculation unit, configured to calculate the height of the pass according to the coordinates of the end points of the circumferential spread of the pass;

[0231] A fifth calculation unit, configured to calculate the cladding thickness for each pixel point of the core edge according to the edge image and the circumferential spread of the pass;

[0232] A sixth calculation unit, configured to calculate the core detection area and the total detection area according to the edge image, and calculate the cladding detection area according to the core detection area and the total detection area.

[0233] In some specific embodiments of the present invention, the intelligent detection system for the cross-sectional characteristics of the round-hole pass of the composite bar and wire rod may incorporate the features of the intelligent detection method for the cross-sectional characteristics of the round-hole pass of the composite bar and wire rod in Embodiment 1 of the present invention, and vice versa.

[0234] Embodiment 2

[0235] The spiral ribbed wire rod refers to the coiled ribbed steel bars with a diameter of 6-12 mm rolled on the high-speed wire rod production line. Among the steel bars used in concrete structures, about 20-25% are thin-diameter steel bars with a nominal diameter less than 12 mm. They are mainly used as the stressed steel bars for components such as slabs and walls, and as stirrups, erection bars, and structural bars in beam and column components.

[0236] The stainless steel composite spiral ribbed steel bar, hereinafter referred to as the composite spiral ribbed wire rod, is a composite bar and wire rod with a stainless steel outer layer and a carbon steel core. It has properties such as high strength, high toughness, and excellent corrosion resistance. At the same time, it can also save a large amount of precious resources such as nickel and chromium in stainless steel, thereby improving economic benefits and having broad development prospects in engineering applications.

[0237] First, purchase 316L stainless steel seamless pipes produced by standardization (Φ159×9000, wall thickness 6mm, in line with GB / T20878 - 2007) and HRB400E carbon steel mandrels (Φ155×9000mm, processed to Φ146.90mm) to prepare composite round billets.

[0238] Carry out Φ12mm composite spiral rib hot rolling industrial tests in the high - speed wire rod plant of a steel mill. This production line has a total of 22 rolling mills (1–22#), which are distributed horizontally and vertically in turn. The rough - middle - pre - finishing - finishing rolling has 6 - 6 - 6 - 4 passes respectively. The flying shears for rough - middle - pre - finishing are between the 6–7#, 12–13#, and 18–19# rolling mills respectively, and are denoted as the 1–3# flying shears in turn, as Figure 2 shown. Figure 2 Among them, 1 represents the calibration circle, 2 represents the calibration circle with a diameter of 4mm, 3 represents the core edge, 4 represents the outer layer edge, and 5 represents the ear. The 1–2# rolling mills use box - shaped passes, the 3–22# rolling mills use oval - round passes, and the finished pass of the 22# rolling mill meets the requirements of GB / T1499.2 - 2018. When rolling Φ12mm composite spiral rib products, the on - site hot rolling process is as follows:

[0239] Heating temperature and time: Soaking section 1060±40℃, heating for 2.5h; Starting rolling temperature: greater than 1025±30℃; Sample acquisition: Obtain 1–3# flying shear samples (F1, F2, and F3) through emergency stop on - site, and obtain Φ12mm composite spiral rib finished products Fp on the cooling bed; Sample cooling method: The 1–3# flying shear samples are air - cooled to room temperature, and the finished products are water - cooled to room temperature.

[0240] Taking the 1# flying shear sample F1 during the hot rolling process of Φ12mm composite spiral rib steel bars as an example for explanation:

[0241] Corrode the 1# flying shear sample F1 with a 4% nitric acid alcohol corrosion solution for 1 - 2 minutes, and slowly wash and dry it in water. Select a calibration circle diameter D0 of 4mm, and select a yellow PVC board as the background board for the sample cell.

[0242] As Figure 3 shown is the image of the calibration board and the corroded 1# flying shear sample F1. In addition, the actual width spread B measured by a vernier caliper is 83.64mm, and the actual height H is 75.54mm. Cut and divide the Figure 3 image into a calibration board image and a cross - section image of the material shape. Figure 4 Shows the pre - processed calibration board image, which has 49 calibration circles, and the area accuracy threshold is set to 0.1. Select 18 calibration circles through the area limitation of condition 1 and the border limitation of condition 2 in Example 1, as Figure 5 shown. Calculate the first image pixel scale according to step 2 in Example 1 is 0.0414 mm / pixel.

[0243] The cross-sectional image of the corroded material shape is first binarized, subjected to closing operation, and filling treatment to obtain the filled image of the core carbon steel, as Figure 6 shown, and then edge detection is performed to obtain the edge image of the core carbon steel, as Figure 7 shown.

[0244] For the cross-sectional image of the corroded material shape, the contrast is enhanced, and an opening operation is performed on it to obtain the background image. The difference between the two is used to obtain the foreground image and binarize it, removing the Figure 6 filled area of the core carbon steel shown, and then the connected domain with the largest area is found. This connected domain is the stainless steel cladding. After filling the cladding, its outer edge is obtained, as Figure 8 shown. The edge image of the core carbon steel and the outer edge image of the stainless steel cladding are combined, and the result is as Figure 9 shown, that is, the edge image is obtained.

[0245] According to the material shape circumferential spread b calculated in step 5 of Embodiment 1, it is 86.13 mm, and the spread error rate is 2.98%. The material shape circumferential spread b is corrected. At this time, the second image pixel scale is 0.0402 mm / pixel. Taking the left end point of the corrected material shape circumferential spread as the first point, all the pixel points of the outer edge are re-ordered. Through the second image pixel scale , according to step 6 in Embodiment 1 of the present invention, calculate the material shape height, the material shape center point, and the distance from each pixel point of the outer edge to the material shape center point. Then, according to step 4 in Embodiment 1, recalculate the material shape circumferential length of each pixel point of the outer edge. According to step 6 in Embodiment 1, calculate the material shape height, the material shape center point, and the distance from each pixel point of the outer edge to the material shape center point. According to step 7 in Embodiment 1, calculate the cladding thickness of each pixel point of the core edge, etc. Draw a graph based on these calculated characteristic data, as Figure 10 shown. Figure 10 In it, 6 represents the clockwise direction, 7 represents the connection line between the leftmost horizontal pixel point and the rightmost horizontal pixel point of the outer edge, 8 represents the material shape circumferential spread line, 9 represents the minimum material shape circumferential length, 10 represents the material shape height, 11 represents the maximum cladding thickness, 12 represents the minimum cladding thickness, 13 represents the upper groove bottom arc, 14 represents the lower groove bottom arc, and 15 represents the horizontal connection line at the end (referring to the end close to the spread direction) of the tangent lines on both sides of the upper groove bottom (the circular hole type uses the tangent line of the groove bottom arc radius to connect). Table 1 shows the obtained characteristic data.

[0246] Table 1 Cross-sectional characteristic data of the Φ12 composite spiral bar 1# flying shear circular hole type material shape

[0247] The circumferential spread b' of the stock shape after correction in this embodiment is 83.64 mm, which is equal to the actual spread B in the cross-section; through the second image pixel scale The calculated height h of the stock shape is 75.32 mm, and the error rate from the actual height H is 0.29%, meeting the accuracy requirements.

[0248] The core detection area and the total detection area are 3872.69 mm 2 and 4570.70 mm 2 respectively. Therefore, the clad detection area is equal to 698.01 mm 2 . The height of the billet before rolling is 159 mm, and the relative reduction ratio in the current pass is 52.63%; the spread of the billet before rolling is 159 mm, and the relative spread ratio in the current pass is -47.38%; the theoretically designed spread is 75.3, and the ear size is 4.17 mm.

[0249] The proportion of the outer cross-sectional area is 14.52%, and the proportion of the clad detection area in the current pass is 15.27%, which can be used for the analysis of clad rheological differences.

[0250] According to the circular pass stock shape area calculation method in rolling theory, the theoretically designed total area of the F1 sample of the current No. 1 flying shear is 4547.00 mm 2 , and according to the ratio, the theoretically designed core area is 3881.28 mm 2 , and the theoretically designed clad area is 660.44 mm 2 . According to the theoretically designed clad area and the clad detection area, the clad detection area of the F1 sample of the No. 1 flying shear increases by 37.57 mm 2 . Calculate that the error rates of the total area, core area, and clad area in the current pass are 0.52%, 0.22%, and 5.7% respectively. The total detection area is 4570.70 mm 2 , and the elongation coefficient in the current pass of the rolling process is 4.3441.

[0251] Draw the curves of the circumferential length of the stock shape, the distance from the outer edge to the center point of the stock shape and the clad thickness change, as Figures 11 to 13 shown, where Nb', Nc', Nd' respectively represent the vertically highest pixel point, horizontally rightmost pixel point, and vertically lowest pixel point of the outer edge, and Nb' 芯 , Nc' 芯 , Nd' 芯 respectively represent the vertically highest pixel point, horizontally rightmost pixel point, and vertically lowest pixel point of the core edge, and D FCIt represents the cladding thickness. The quantitative evaluation indexes of the circumferential dimension of the material shape and the cladding non-uniformity are shown in Table 2, which can provide data support for the subsequent rolling of composite bar wire rods.

[0252] Table 2 Quantitative evaluation indexes of the circumferential dimension of the material shape and the cladding non-uniformity

[0253] Example 3

[0254] Standardized 304 stainless steel welded pipes were purchased from the market, with an outer diameter of Φ168×9000mm and a wall thickness of 8mm. The welds were welded with 304 stainless steel electrodes; the HRB400E carbon steel mandrel had a size of Φ155×9000mm and was processed to Φ151.90mm to prepare a composite round billet.

[0255] An industrial hot rolling small batch production test of Φ16mm composite steel bars was carried out in a certain steel plant. Among them, there were 18 rolling mills in total (denoted as 1–18# respectively), distributed horizontally and vertically in turn, with 6 passes for rough, medium and finish rolling respectively. The flying shears for rough and medium rolling were between the 6–7# and 12–13# rolling mills respectively, denoted as 1 and 2# flying shears. During the rolling process, the heads and tails of the composite round billets were cut off respectively. Samples S1 and S2 were obtained from the 1 and 2# flying shears on the rolling site, and a finished product sample Sp was obtained on the cooling bed. According to the cross-section and longitudinal section diagrams of the finished product sample Sp of the composite steel bars, it can be seen that there is obvious non-uniformity in the cladding.

[0256] The 1# flying shear sample S1 was corroded with a 4% nitric acid alcohol corrosion solution for 1–2 minutes, slowly washed in water and dried. The calibrated circle diameter D0 can be selected as 4mm, and the sample cell selects a yellow PVC board as the background board, as Figure 14 shown are the images of the calibrated board and the 1# flying shear round hole-shaped material S1 after corrosion. In addition, the actual width spread B of the cross-section was measured with a vernier caliper to be 82.60mm, and the actual height H was 76.48mm.

[0257] The corrected width spread b' of the material shape was 82.60mm, the corrected height h' of the material shape was 75.82mm, and the height error rate was 0.86%. It can be seen that the height measurement error after correction is extremely small. The maximum value, minimum value, mean value and standard deviation of the cladding thickness are 5.24mm, 2.10mm, 3.61mm and 0.5930mm respectively. The cladding thickness change curve is as Figure 15 shown, where D FC represents the cladding thickness. The cladding evaluation indexes of the S1 material shape are: the circumferential non-uniformity coefficient of the cladding is 2.495, the circumferential fluctuation coefficient of the cladding is 0.164, the local weakness index of the cladding is 0.582, and the number of wave peaks and wave valleys is 15.

[0258] The core detection area and the total detection area are 3784.78 mm 2 and 4635.66 mm 2 respectively. Therefore, the cladding detection area is equal to 850.88 mm 2 . The height of the billet before rolling is 168 mm, and the relative reduction of the current pass is 54.87%; the spread of the billet before rolling is 168 mm, and the relative spread of the current pass is -50.83%.

[0259] The proportion of the outer layer cross-sectional area is 18.14%, and the proportion of the cladding detection area of the current pass is 18.36%, which can be used for the analysis of the rheological differences of the cladding. The total detection area is 4635.66 mm 2 , and the elongation coefficient of the current pass in the rolling process is 4.7818. The elongation coefficients of the 1-6# rolling mills at the industrial hot rolling site are: 1.23×1.220×1.365×1.292×1.288×1.339 = 4.5641. The actual elongation is greater than the on-site experimental value, which provides data support for the optimization of the elongation coefficient in subsequent industrial production.

[0260] The specific embodiments disclosed above are only for the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or variations, which should be covered within the protection scope of the present invention.

Claims

1. An intelligent detection method for the cross-sectional characteristics of the round-hole pass shape of composite bar and wire materials, characterized in that, The detection method includes: Obtaining a calibration plate image and a cross-sectional image of the material type; Calculating a first image pixel scale based on the calibration plate image; Obtaining an edge image from the cross-sectional image of the material type, where the edge image includes an outer edge and a core edge; Calculating the circumferential length of the material type for each pixel point of the outer edge based on the edge image; Calculating the circumferential width expansion of the material type and the coordinates of its end points based on the first image pixel scale and the leftmost and rightmost horizontal pixel points of the outer edge; Calculating the height of the material type based on the coordinates of the end points of the circumferential width expansion; Calculating the cladding thickness for each pixel point of the core edge based on the edge image and the circumferential width expansion; Calculating the core detection area and the total detection area based on the edge image, and calculating the cladding detection area based on the core detection area and the total detection area.

2. The intelligent detection method for the cross-sectional characteristics of the round-hole pass of the composite bar wire according to claim 1, characterized in that, The process of obtaining the calibration plate image and the cross-sectional image of the material type is as follows: Obtaining a cross-sectional specimen of the material type; Using a polishing machine to polish the cross-sectional specimen of the material type, then using a 4% nitric acid alcohol corrosion solution to corrode the polished cross-sectional specimen of the material type, and then cleaning and drying to obtain a sample; Placing the calibration plate on one side of the sample, keeping the width expansion direction of the sample horizontal, and collecting the calibration plate and sample images; Cropping and segmenting the calibration plate and sample images to obtain the calibration plate image and the cross-sectional image of the material type.

3. The intelligent detection method for the cross-sectional characteristics of the round-hole pass of the composite bar wire according to claim 1, wherein, Calculating the first image pixel scale based on the calibration plate image includes: Preprocessing the calibration plate image; where the calibration plate image includes multiple calibration circles; Extracting multi-connected regions from the preprocessed calibration plate image, and measuring the area of each multi-connected region and the major axis length and minor axis length of the ellipse having the same normalized second-order central moment as each multi-connected region; Extracting the multi-connected region i that simultaneously satisfies Condition 1 and Condition 2: Condition 1: ; Condition 2: ; Among them, and represent the areas of the multiply connected regions i and j, respectively; represents the area precision threshold; represents the number of multiply connected regions extracted from the preprocessed calibration plate image; and represent the major axis length and minor axis length of the ellipse having the same normalized second-order central moment as the multiply connected region i, respectively; Calculating the first image pixel scale based on the extracted multi-connected region that simultaneously satisfies Condition 1 and Condition 2, and the specific calculation formula is: ； Among them, represents the first image pixel scale; represents the diameter of the calibration circle in the calibration plate; n represents the number of multi-connected regions that simultaneously satisfy Condition 1 and Condition 2 and are extracted.

4. The intelligent detection method for the cross-sectional characteristics of the round-hole pass of the composite bar wire according to claim 1, characterized in that, Calculating the circumferential length of the material type for each pixel point of the outer edge based on the edge image includes: Taking the leftmost horizontal pixel point of the outer edge as the first point, sorting all pixel points of the outer edge, and recording the serial number of the highest vertical pixel point as Nb', the serial number of the rightmost horizontal pixel point as Nc', and the serial number of the lowest vertical pixel point as Nd'; Calculating the distance between each pixel point of the outer edge and other pixel points, and the specific calculation formula is: ； Among them, represents the distance between the i-th pixel point and the j-th pixel point on the outer edge, , , represents the number of pixel points on the outer edge; represents the coordinates of the i-th pixel point on the outer edge; represents the coordinates of the j-th pixel point on the outer edge; Finding the maximum distance between each pixel point of the outer edge and other pixel points, and determining whether the maximum distance is greater than the first reference distance. If not, taking the first reference distance as the maximum distance; where the determination method of the first reference distance is: If all pixel points of the outer edge are sorted in the clockwise direction and i < Nc', the first reference distance is the distance between the i-th pixel point and the lowest vertical pixel point; if all pixel points of the outer edge are sorted in the clockwise direction and i > Nc', the first reference distance is the distance between the i-th pixel point and the highest vertical pixel point; If all the pixel points on the outer edge are sorted in the counterclockwise direction and i < Nc', the first reference distance is the distance between the i-th pixel point and the highest vertical pixel point; if all the pixel points on the outer edge are sorted in the counterclockwise direction and i > Nc', the first reference distance is the distance between the i-th pixel point and the lowest vertical pixel point; The maximum distance between each pixel point on the outer edge and other pixel points is the material type circumferential length of each pixel point on the outer edge.

5. The intelligent detection method for the cross-sectional characteristics of the round-hole pass shape of the composite bar wire according to claim 1, characterized in that Calculate the material type circumferential width expansion and its endpoint coordinates according to the first image pixel scale and the leftmost horizontal pixel point and the rightmost horizontal pixel point of the outer edge, including: taking the leftmost horizontal pixel point of the outer edge as the first point, sorting all the pixel points on the outer edge, and recording the serial number of the rightmost horizontal pixel point as Nc'; Determine the connection line between the leftmost horizontal pixel point and the rightmost horizontal pixel point of the outer edge; determine the serial number deviation between the endpoint serial numbers of the material type circumferential width expansion and the serial numbers of the leftmost horizontal pixel point and the rightmost horizontal pixel point of the outer edge according to the connection line, and then determine the endpoint serial numbers of the material type circumferential width expansion; Calculate the material type circumferential width expansion according to the endpoint coordinates of the material type circumferential width expansion and the first image pixel scale. The specific calculation formula is: ； Among them, represents the circumferential spread of the stock shape; , respectively represent the coordinates of the two endpoints of the circumferential spread of the stock shape; represents the first image pixel scale.

6. The intelligent detection method for the cross-sectional characteristics of the round-hole pass of the composite bar wire according to claim 5, characterized in that, The detection method further includes correcting the material type circumferential width expansion, specifically including: Obtain the actual width expansion of the material type cross-section, and calculate the width expansion error rate according to the actual width expansion of the material type cross-section and the material type circumferential width expansion; Judge whether the width expansion error rate is less than the error rate threshold. If not, calculate the second image pixel scale according to the endpoint coordinates of the material type circumferential width expansion and the actual width expansion of the material type cross-section. The specific calculation formula is: ； Among them, represents the second image pixel scale; represents the actual spread of the cross-section of the material type; , respectively represent the coordinates of the two endpoints of the circumferential spread of the material type; Calculate the corrected material type circumferential width expansion according to the second image pixel scale and the endpoint coordinates of the material type circumferential width expansion. The specific formula is: ； Among them, represents the circumferential spread of the stock shape after correction.

7. The intelligent detection method for the cross-sectional characteristics of the round-hole pass shape of the composite bar wire according to claim 1, wherein Calculate the material type height according to the endpoint coordinates of the material type circumferential width expansion, specifically including: Determine the abscissa of the midpoint in the width expansion direction according to the endpoint coordinates of the material type circumferential width expansion; Determine the abscissa value range according to the set abscissa deviation and the abscissa of the midpoint; Determine the upper intersection point serial number and the lower intersection point serial number of the vertical line passing through each abscissa within the abscissa value range and the outer edge, and then determine the ordinate corresponding to each abscissa within the abscissa value range and the midpoint in the width expansion direction; Determine the coordinates of the upper groove bottom arc center point corresponding to each midpoint according to the vertical distance difference between the upper groove bottom arc center point of the roll pass and the actual midpoint in the width expansion direction, and determine the coordinates of the lower groove bottom arc center point corresponding to each midpoint according to the vertical distance difference between the lower groove bottom arc center point of the roll pass and the actual midpoint in the width expansion direction; Set the value range of the upper intersection point serial number and the lower intersection point serial number; Calculate the distance between each upper intersection point within the value range of the upper intersection point serial number and the corresponding upper groove bottom arc center point, and the distance between each lower intersection point within the value range of the lower intersection point serial number and the corresponding lower groove bottom arc center point. All the distances form a distance set; Calculate the variance of each distance in the distance set; Select the midpoint corresponding to the distance with the smallest variance as the material type center point; Determine the upper and lower intersection points of the vertical line passing through the center point of the stock shape with the outer edge, and calculate the height of the stock shape based on the upper and lower intersection points.

8. The intelligent detection method for the cross-sectional characteristics of the round-hole pass profile of the composite bar wire according to claim 1, wherein Calculate the cladding thickness of each pixel point on the core edge according to the edge image and the circumferential spread of the stock shape, including: Take the intersection points of the line where the circumferential spread of the stock shape is located with the left side of the core edge and the outer edge as the first points of the core edge and the outer edge, and sort all the pixel points of the core edge and the outer edge respectively; Take the distance between the first point of the core edge and the first point of the outer edge as the second reference distance; Calculate the distance between each pixel point of the core edge and other pixel points of the outer edge; among them, the other pixel points of the outer edge refer to the pixel points except the first pixel point of the outer edge; Find the minimum distance between each pixel point of the core edge and other pixel points of the outer edge; If the minimum distance is less than the second reference distance, then the minimum distance is the cladding thickness of the corresponding pixel point of the core edge; if the minimum distance is greater than or equal to the second reference distance, then the second reference distance is the cladding thickness of the corresponding pixel point of the core edge.

9. The intelligent detection method for the cross-sectional characteristics of the round-hole pass of the composite bar wire according to any one of claims 1 to 8, characterized in that, The detection method further includes calculating the circumferential evaluation index of the stock shape, and then evaluating the circumferential characteristics of the stock shape; among them, the circumferential evaluation index of the stock shape includes the relative reduction of the current pass, the relative spread of the current pass, the circumferential non-uniformity coefficient of the stock shape circumference, the circumferential fluctuation coefficient of the stock shape circumference, the local weakness index of the stock shape circumference, the size of the ear, the proportion of the cladding detection area in the current pass, the total area error rate, the core area error rate, the cladding area error rate, and the elongation coefficient of the current pass; Calculate the relative reduction of the current pass according to the height of the stock shape and the height of the billet before rolling; Calculate the relative spread of the current pass according to the circumferential spread of the stock shape and the width spread of the billet before rolling; Calculate the maximum stock shape circumference length, the minimum stock shape circumference length, the average value of the stock shape circumference length, and the standard deviation of the stock shape circumference length according to the stock shape circumference length of each pixel point on the outer edge. Take the ratio of the maximum stock shape circumference length to the minimum stock shape circumference length as the circumferential non-uniformity coefficient of the stock shape circumference, take the ratio of the standard deviation of the stock shape circumference length to the average value of the stock shape circumference length as the circumferential fluctuation coefficient of the stock shape circumference, and take the ratio of the minimum stock shape circumference length to the average value of the stock shape circumference length as the local weakness index of the stock shape circumference; Calculate the size of the ear on the cross-section of the stock shape according to the circumferential spread of the stock shape and the theoretically designed spread; Calculate the proportion of the cladding detection area in the current pass according to the cladding detection area and the total detection area; Calculate the total area error rate according to the theoretically designed total area and the total detection area, calculate the core area error rate according to the theoretically designed core area and the core detection area, and calculate the cladding area error rate according to the theoretically designed cladding area and the cladding detection area; Calculate the elongation coefficient of the current pass according to the cross-sectional area of the billet and the total detection area.

10. The intelligent detection method for the cross-sectional characteristics of the round-hole pass shape of the composite bar wire according to any one of claims 1 to 8, characterized in that The detection method further includes calculating the cladding evaluation index of the stock shape, and then evaluating the cladding characteristics of the stock shape; among them, the cladding evaluation index of the stock shape includes the circumferential non-uniformity coefficient of the cladding, the circumferential fluctuation coefficient of the cladding, the local weakness index of the cladding, the number of wave peaks, and the number of wave valleys; Calculate the maximum cladding thickness, minimum cladding thickness, average cladding thickness, and standard deviation of cladding thickness based on the cladding thickness of each pixel point at the core edge. Use the ratio of the maximum cladding thickness to the minimum cladding thickness as the circumferential non-uniformity coefficient of the cladding, the ratio of the standard deviation of cladding thickness to the average cladding thickness as the circumferential fluctuation coefficient of the cladding, and the ratio of the minimum cladding thickness to the average cladding thickness as the local weakness index of the cladding; Draw a cladding thickness curve based on the cladding thickness of each pixel point at the core edge, and draw a horizontal line in the cladding thickness curve. Determine the number of wave peaks and wave valleys based on the horizontal line.

11. An intelligent detection system for the cross-sectional characteristics of the round pass shape of composite bar and wire materials, characterized in that, The detection system includes: An acquisition unit, configured to acquire a calibration plate image and a cross-sectional image of the material type; obtain an edge image from the cross-sectional image of the material type, where the edge image includes an outer edge and a core edge; A first calculation unit, configured to calculate a first image pixel scale based on the calibration plate image; A second calculation unit, configured to calculate the circumferential length of the material type of each pixel point of the outer edge based on the edge image; A third calculation unit, configured to calculate the circumferential width expansion of the material type and the coordinates of its endpoints based on the first image pixel scale and the leftmost horizontal pixel point and the rightmost horizontal pixel point of the outer edge; A fourth calculation unit, configured to calculate the height of the material type based on the coordinates of the endpoints of the circumferential width expansion of the material type; A fifth calculation unit, configured to calculate the cladding thickness of each pixel point of the core edge based on the edge image and the circumferential width expansion of the material type; A sixth calculation unit, configured to calculate the core detection area and the total detection area based on the edge image, and calculate the cladding detection area based on the core detection area and the total detection area.

Images