A method and device for extrusion integrated molding of high-strength alloy template

By performing edge detection and dark channel processing on the rolling greyscale image of the rolling piece, the enhancement coefficient is determined, and the low bright pixel points are enhanced, which solves the edge blur problem caused by lubricating coolant and improves the rolling piece thickness recognition and control accuracy .

CN118710675BActive Publication Date: 2024-11-05SHAANXI SHOULV ALU FORMWORK TECH CO LTD CO
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411178352.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-27
Publication Date
2024-11-05
Estimated Expiration
2044-08-27

AI Technical Summary

Technical Problem

During the extrusion integral molding of high-strength alloy templates, the lubricating coolant causes blurred edges of the rolled parts, affecting monitoring accuracy and control accuracy.

Method used

By performing edge detection, dark channel processing and fitting edge analysis on the rolling rolling grayscale image, the image virtualization factor and random factor are determined, the enhancement coefficient is calculated, and the low bright pixel points are enhanced to eliminate the lubricating coolant. Influence.

Benefits of technology

It improves the accuracy of rolled piece thickness identification and the feedback control accuracy of the intelligent production system, and improves the process monitoring and control level.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118710675B_ABST
    Figure CN118710675B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of image enhancement technology, and specifically to an extrusion integrated molding method and device for high-strength alloy templates, by obtaining a grayscale image of rolled product, and determining an image virtualization factor and an image random factor; determining an enhancement coefficient according to the image virtualization factor and the image random factor, and performing enhancement processing on low-brightness pixels in the grayscale image of rolled product according to the enhancement coefficient, thereby obtaining an enhanced grayscale image of rolled product; and performing extrusion integrated molding control according to the enhanced grayscale image of rolled product. The present invention eliminates the influence of lubricating coolant on the identification of rolled product edge information, and improves the monitoring accuracy and control accuracy of the rolled product rolling process.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of image enhancement, and in particular to an extrusion integrated molding method and device for a high-strength alloy template. Background Art

[0002] The high-strength alloy formwork system consists of a panel system, a support system, a fastening system and an accessory system. The panel system is made of extruded one-piece alloy steel profiles, replacing traditional wooden formwork and metal formwork. Traditional building formwork is made of metal processing technology to achieve the required shape and size, while high-strength alloy formwork is shaped by extrusion one-piece molding technology. It has the characteristics of light weight, high strength, high yield strength, overall stability, easy assembly and disassembly, and multiple reuse.

[0003] Extrusion one-piece molding technology has extremely high precision requirements, so it is generally equipped with a machine vision system to intelligently monitor, analyze and feedback the entire production process, and then adjust the extrusion process settings and equipment parameters in real time. However, during the monitoring process, in order to reduce the friction between the roller and the rolled piece and prevent damage to the surface of the rolled piece during rolling, it is necessary to continuously spray lubricating coolant (usually a mixture of water and lubricating oil) during the rolling process. These liquids will cause the edges of the rolled piece to blur and deform. Therefore, the intelligent production system is always unable to obtain accurate information on the edges of the rolled piece, and thus cannot accurately monitor the real-time changes in the thickness of the rolled piece, which is not conducive to the intelligent control and feedback loop of the extrusion one-piece molding system. Summary of the invention

[0004] The purpose of the present invention is to provide an extrusion one-piece forming method and device for high-strength alloy templates, which is used to solve the problem that the lubricating coolant in the existing one-piece forming rolling process will blur the edge of the rolled piece, resulting in inaccurate monitoring results, thereby affecting the reliable control of the extrusion one-piece forming.

[0005] In order to solve the above technical problems, in a first aspect, the present invention provides an extrusion integrated molding method for a high-strength alloy template, comprising the following steps:

[0006] Acquire a grayscale image of a rolled piece, perform edge detection on the grayscale image of the rolled piece, and obtain an upper edge line and a lower edge line of the rolled piece;

[0007] Segmenting a workpiece entrance region image and a workpiece exit region image from the workpiece rolling grayscale image, taking the workpiece entrance region image and the workpiece exit region image as a target analysis image, performing dark channel processing on the target analysis image, and obtaining a dark channel workpiece upper edge line and a dark channel workpiece lower edge line;

[0008] Determine the upper fitting edge of the rolled piece and the lower fitting edge of the rolled piece by combining the grayscale change of the pixels between the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece in the target analysis image, and the grayscale change of the pixels between the lower edge line of the rolled piece and the lower edge line of the dark channel rolled piece;

[0009] Determine the image virtualization factor by combining the grayscale difference between the upper edge line of the rolled piece, the upper edge line of the dark channel rolled piece and the upper fitted edge of the rolled piece in the rolled piece entrance area image and the rolled piece exit area image, and determine the image random factor by combining the grayscale change of the pixel points and the pixel grayscale of the lower fitted edge of the rolled piece in the rolled piece entrance area image and the rolled piece exit area image;

[0010] Determine an enhancement coefficient according to the image virtualization factor and the image random factor, and perform enhancement processing on low-brightness pixel points in the rolled product rolling grayscale image according to the enhancement coefficient, so as to obtain the rolled product rolling grayscale image after enhancement processing;

[0011] The extrusion integrated molding control is performed according to the rolled piece rolling grayscale image after the enhancement processing.

[0012] In combination with the first aspect, in some possible implementations, obtaining the upper edge line of the dark channel rolled piece and the lower edge line of the dark channel rolled piece includes:

[0013] Performing minimum filtering on the target analysis image to obtain a dark channel image;

[0014] The rolled piece area in the dark channel image is acquired, and edge segmentation is performed on the rolled piece area to obtain an upper edge line of the dark channel rolled piece and a lower edge line of the dark channel rolled piece.

[0015] In combination with the first aspect, in some possible implementations, determining the upper fitting edge of the rolled piece and the lower fitting edge of the rolled piece includes:

[0016] Determine the grayscale minimum values ​​of all pixels on the line connecting two corresponding pixel points between the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece in the target analysis image, perform curve fitting on the pixel points corresponding to these grayscale minimum values, and obtain the fitted edge on the rolled piece;

[0017] Determine the grayscale minimum values ​​of all pixels on the line connecting two corresponding pixel points between the lower edge line of the rolled piece and the lower edge line of the dark channel rolled piece in the target analysis image, perform curve fitting on the pixel points corresponding to these grayscale minimum values, and obtain the lower fitted edge of the rolled piece.

[0018] In combination with the first aspect, in some possible implementations, determining the image virtualization factor includes:

[0019] Determine the average grayscale value of the corresponding pixel points between the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece in the target analysis image, and determine the grayscale difference between the average grayscale value and the grayscale value of the corresponding pixel points on the fitted edge of the rolled piece;

[0020] Determine the blur factor corresponding to the target analysis image by combining the overall distribution level of all the grayscale differences;

[0021] The image virtualization factor is determined by comprehensively considering the virtualization factors corresponding to the image of the rolling stock entrance area and the image of the rolling stock exit area, and the image virtualization factor is a normalized value.

[0022] In combination with the first aspect, in some possible implementations, segmenting the rolled piece entrance area image and the rolled piece exit area image from the rolled piece rolling grayscale image includes:

[0023] Segmenting the rolled product rolling grayscale image into a rolled product inlet side image and a rolled product outlet side image;

[0024] The image on the inlet side of the rolled piece and the image on the outlet side of the rolled piece are both regarded as an object image, and according to the grayscale change of the pixel points in the object image, the liquid sliding deviation of each edge pixel point on the upper edge line of the rolled piece in the object image is determined;

[0025] Determining a contact point on the upper edge line of the rolled piece in the object image according to the liquid sliding deviation;

[0026] According to the contact points on the upper edge line of the rolled piece in the image of the rolled piece entrance side, the rolled piece entrance area image is segmented from the image of the rolled piece entrance side, and according to the contact points on the upper edge line of the rolled piece in the image of the rolled piece exit side, the rolled piece exit area image is segmented from the image of the rolled piece exit side.

[0027] In combination with the first aspect, in some possible implementations, determining the liquid sliding deviation of each edge pixel point on the upper edge line of the rolled piece in the object image includes:

[0028] Determine a pixel point sequence corresponding to each edge pixel point on the upper edge line of the rolled piece in the object image, and determine a grayscale mean and a grayscale standard deviation of all the pixel points in the pixel point sequence;

[0029] The liquid slippage deviation of each edge pixel point on the upper edge line of the rolled piece in the object image is determined based on the grayscale mean difference between each edge pixel point and its previous edge pixel point, the grayscale mean difference between each edge pixel point and its subsequent edge pixel point, and the grayscale mean and grayscale standard deviation of the subsequent edge pixel point.

[0030] In combination with the first aspect, in some possible implementations, determining the contact point on the upper edge line of the rolled piece in the object image includes:

[0031] Determine the liquid slippage deviation slope of each edge pixel point on the upper edge line of the rolled piece in the object image according to the fluctuation change of the liquid slippage deviation of each edge pixel point on the upper edge line of the rolled piece in the object image;

[0032] On the upper edge line of the rolled piece in the object image, the edge pixel point corresponding to the maximum value of the liquid sliding deviation slope is determined as the contact point on the upper edge line of the rolled piece in the object image.

[0033] In combination with the first aspect, in some possible implementations, determining the image random factor includes:

[0034] Normalizing the liquid slippage deviation of each edge pixel point on the upper edge line of the rolled piece in the target analysis image to obtain a first normalized value;

[0035] Normalizing the grayscale value of each edge pixel point on the lower fitting edge of the rolled piece in the target analysis image to obtain a second normalized value;

[0036] Determine a sub-random factor according to a difference between the first normalized value of each edge pixel point on the upper edge line of the rolled piece in the target analysis image and the second normalized value of a corresponding edge pixel point on the lower fitting edge of the rolled piece in the target analysis image;

[0037] The image random factor is determined by comprehensively considering the sub-random factors corresponding to the image of the rolling piece entrance area and the image of the rolling piece exit area, and the image random factor is a normalized value.

[0038] In combination with the first aspect above, in some possible implementations, the step of determining low-brightness pixel points in the grayscale image of the rolled product includes:

[0039] Determine the target analysis area corresponding to the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece, and the target analysis area corresponding to the lower edge line of the rolled piece and the lower edge line of the dark channel rolled piece in the target analysis image;

[0040] Determine the grayscale histogram of all pixels in each target analysis area, and determine the average grayscale level in the grayscale histogram; determine the low-brightness pixels in the rolled product rolling grayscale image based on the pixels corresponding to all grayscale levels less than the average grayscale level in the grayscale histogram.

[0041] In order to solve the above technical problems, in a second aspect, the present invention also provides an extrusion integrated molding device for a high-strength alloy template, comprising:

[0042] The rolling piece edge line acquisition module is used to: acquire the rolling piece grayscale image, perform edge detection on the rolling piece grayscale image, and obtain the upper edge line and the lower edge line of the rolling piece;

[0043] A dark channel edge line acquisition module is used to: segment the roll entry area image and the roll exit area image from the roll rolling grayscale image, use the roll entry area image and the roll exit area image as a target analysis image, perform dark channel processing on the target analysis image, and obtain a dark channel roll upper edge line and a dark channel roll lower edge line;

[0044] A fitting edge line acquisition module is used to determine the upper fitting edge of the rolled piece and the lower fitting edge of the rolled piece by combining the pixel grayscale changes between the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece in the target analysis image, and the pixel grayscale changes between the lower edge line of the rolled piece and the lower edge line of the dark channel rolled piece;

[0045] An image factor acquisition module is used to: determine the image virtualization factor by combining the grayscale difference between the upper edge line of the rolled piece, the upper edge line of the dark channel rolled piece and the upper fitted edge of the rolled piece in the rolled piece entrance area image and the rolled piece exit area image, and determine the image random factor by combining the grayscale change of the pixel points and the pixel grayscale of the lower fitted edge of the rolled piece in the rolled piece entrance area image and the rolled piece exit area image;

[0046] An enhancement module, used to: determine an enhancement coefficient according to the image virtualization factor and the image random factor, and enhance low-brightness pixel points in the rolled product rolling grayscale image according to the enhancement coefficient, so as to obtain an enhanced rolled product rolling grayscale image;

[0047] The control module is used to: perform extrusion integrated molding control according to the rolled piece rolling grayscale image after the enhancement processing.

[0048] To solve the above technical problems, in a third aspect, the present invention further provides an extrusion integrated molding system for a high-strength alloy template, comprising a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, so that the device executes the method in the above first aspect or any possible implementation of the first aspect.

[0049] To solve the above technical problems, in a fourth aspect, the present invention further provides a computer program product, which includes: a computer program code, when the computer program code runs on a computer, enables the computer to execute the method in the above first aspect or any possible implementation of the first aspect.

[0050] To solve the above technical problems, in a fifth aspect, the present invention further provides a computer-readable storage medium, which stores a computer program code. When the computer program code runs on a computer, the computer executes the method in the above first aspect or any possible implementation of the first aspect.

[0051] The present invention has the following beneficial effects: firstly, edge detection is performed on the grayscale image of rolled workpiece to obtain the edge of the rolled workpiece, i.e., the upper edge line of the rolled workpiece and the lower edge line of the rolled workpiece; at the same time, the rolled workpiece entrance area image and the rolled workpiece exit area image are segmented from the grayscale image of rolled workpiece, and the two images are subjected to dark channel processing to obtain the dark channel rolled workpiece edge, i.e., the upper edge line of the dark channel rolled workpiece and the lower edge line of the dark channel rolled workpiece. Since the edge of the rolled workpiece obtained by the former is the outward expansion edge after the cooling lubricating liquid covers the rolled workpiece and becomes blurred, and the edge of the dark channel rolled workpiece obtained by the latter is the edge eroded by the dark effect, the real edge of the rolled workpiece should be located between the dark channel rolled workpiece edge and the rolled workpiece edge. At this time, the pixel grayscale changes between the upper edge line of the rolled workpiece and the upper edge line of the dark channel rolled workpiece, and the pixel grayscale changes between the lower edge line of the rolled workpiece and the lower edge line of the dark channel rolled workpiece are combined to determine the rolled workpiece fitting edge, i.e., the upper fitting edge of the rolled workpiece and the lower fitting edge of the rolled workpiece. Since the fitted edge of the rolled piece is obtained by fitting the pixel points that are least affected by the liquid coverage, the grayscale of the upper edge line of the rolled piece, the upper edge line of the dark channel rolled piece, and the pixel points on the upper fitted edge of the rolled piece are compared to evaluate the static blur of the rolled piece edge when affected by the lubricating coolant coverage and determine the image virtualization factor; at the same time, the grayscale change of the pixel points in the rolled piece entrance area image and the rolled piece exit area image and the pixel grayscale of the lower fitted edge of the rolled piece are combined to evaluate the dynamic blur of the lower edge of the rolled piece when affected by the lubricating coolant coverage and determine the image random factor. Combined with the image virtualization factor and the image random factor, the enhancement coefficient is determined, that is, the edge loss of the rolled piece in the grayscale image of the rolled piece rolling, and all low-brightness pixels in the edge area of ​​the rolled piece are enhanced according to the edge loss of the rolled piece to eliminate the influence of the lubricating coolant on the recognition of the edge information of the rolled piece, thereby improving the monitoring accuracy of the rolled piece rolling process, such as the recognition accuracy of the rolled piece thickness, and ultimately improving the feedback control accuracy of the intelligent production system and the process level. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0053] Figure 1 It is a schematic diagram of rolling an alloy steel material by a roller according to an embodiment of the present invention;

[0054] Figure 2 A basic flow diagram of an extrusion integrated molding method for a high-strength alloy template according to an embodiment of the present invention;

[0055] Figure 3 is a schematic diagram of a final grayscale image of a rolled product according to an embodiment of the present invention;

[0056] Figure 4 It is a schematic diagram of the elastic deformation zone, plastic deformation zone and elastic recovery zone when the rolled piece passes through the rollers according to an embodiment of the present invention;

[0057] Figure 5 For the embodiment of the present invention Figure 3 A schematic diagram of an edge image of a final rolled product rolling grayscale image;

[0058] Figure 6 It is a schematic diagram of a liquid sliding deviation polyline corresponding to some edge pixel points on the upper edge line of the rolled piece in the object image in an embodiment of the present invention;

[0059] Figure 7 For the embodiment of the present invention Figure 3 Schematic diagram of edge detection results of a rolled product rolling grayscale image after enhanced processing corresponding to a final rolled product rolling grayscale image;

[0060] Figure 8 The present invention is a schematic structural diagram of an extrusion integrated molding system for a high-strength alloy template according to an embodiment of the present invention. DETAILED DESCRIPTION

[0061] In order to clearly illustrate the technical features of the present invention, the present invention is described in detail below through specific implementation methods in conjunction with the accompanying drawings.

[0062] Embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present invention. It should be understood that the drawings and embodiments of the present invention are only for exemplary purposes and are not intended to limit the scope of protection of the present invention.

[0063] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0064] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0065] It should be noted that the concepts such as "first" and "second" mentioned in the present invention are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0066] It should be noted that the modifications of "one" and "multiple" mentioned in the present invention are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly indicated in the context, they should be understood as "one or more". In the description of the present invention, unless otherwise specified, "multiple" refers to two or more than two, and other quantifiers are similar; "at least one item", "one or more items" or similar expressions refer to any combination of these items, including any combination of single items or plural items. For example, at least one item a can represent any number of a; for another example, one or more items of a, b and c can represent: a, b, c, ab, ac, bc, or abc, where a, b, c can be single or multiple; "and / or" is a kind of association relationship that describes the associated objects, indicating that there can be three relationships, for example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural.

[0067] Although operations or steps are described in a specific order in the drawings in the embodiments of the present invention, it should not be understood that it is required to perform these operations or steps in the specific order shown or in a serial order, or that all the operations or steps shown must be performed to obtain the desired results. In the embodiments of the present invention, these operations or steps may be performed in series; these operations or steps may also be performed in parallel; or some of these operations or steps may be performed.

[0068] At the same time, it is understood that the data involved in the technical solution of the present invention (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of the relevant laws, regulations and relevant provisions. Unless otherwise defined, all technical and scientific terms used in the present invention have the same meanings as those generally understood by technicians in the technical field of the present invention, and all parameters or indicators in the formulas involved in the present invention are normalized values ​​that eliminate the influence of dimensions.

[0069] First, the application scenario of the present invention is explained. More than 90% of the plastic processing of metal materials, especially alloy materials, is completed through rolling. The rolling process is a process in which the friction between the workpiece and the roller pulls the workpiece between rollers with different rotation directions to cause plastic deformation. The alloy material of the high-strength alloy template workpiece is high-strength low-alloy steel, which adopts the latest generation of cold rolling roller pressing integrated molding technology, and has the advantages of high dimensional accuracy, good surface quality, superior mechanical properties, and high material utilization.

[0070] In the cold rolling integrated forming technology, the alloy steel material of the high-strength alloy template rolled piece is fed into the rolling device. The rolling device has multiple sets of rollers, each set of rollers consists of two upper and lower rollers. The rolled piece is pulled into the gap between the rollers by the friction between the rollers. After continuous rolling by multiple sets of rollers, the rolled piece is elongated and compacted, as shown below. Figure 1 After continuous rolling, a panel of high-strength alloy template can be obtained.

[0071] In order to reduce the friction between the rollers and the workpiece and prevent damage to the workpiece surface, it is necessary to continuously spray lubricating coolant during the rolling process. The lubricating coolant is sprayed in the contact area between the rollers and the workpiece material, usually at the entrance and exit of each set of rollers. It must be sprayed evenly at an appropriate angle, and no splashing of lubricating coolant should occur, to ensure that the rollers and the workpiece material can be fully cooled and lubricated on the entire contact surface. Among them, this process achieves the spraying requirements by reasonably designing the nozzle, adjusting the installation height of the nozzle, and debugging multiple times, which will not be elaborated here.

[0072] In order to monitor and adjust the rolling process of the rolled piece throughout the whole process, the intelligent production system sets a high-definition camera at each roll to take high-definition images of the rolled piece passing through each set of rolls from a side view, and processes and analyzes the images to adjust the extrusion process settings and equipment parameters in real time. However, since the lubricating coolant will drip from the rolled plate surface, the lubricating coolant will affect the accurate extraction of the edge information of the rolled piece, thereby interfering with the accurate identification of the rolled piece thickness, reducing the monitoring accuracy of the rolled piece rolling process, and thus affecting the extrusion process control and feedback.

[0073] In order to solve the problem that the lubricating coolant in the above-mentioned existing one-piece rolling process of rolled pieces will blur the edge of the rolled piece, resulting in inaccurate monitoring results, and further affecting the reliable control of extrusion one-piece forming, an embodiment of the present invention provides an extrusion one-piece forming method and device for a high-strength alloy template, by performing edge detection on the rolled piece grayscale image, a double edge of the rolled piece, namely, an upper edge line of the rolled piece and a lower edge line of the rolled piece is obtained; at the same time, the rolled piece grayscale image is segmented into a rolled piece entrance area image and a rolled piece exit area image, and dark channel corrosion is performed on the rolled piece entrance area image and the rolled piece exit area image to obtain another double edge of the rolled piece, namely, an upper edge line of the dark channel rolled piece and a lower edge line of the dark channel rolled piece; the fitted upper and lower edges are obtained according to the area between the two double edges, namely, the upper fitting edge of the rolled piece and the lower fitting edge of the rolled piece edge; then determine the image virtualization factor according to the grayscale difference between the upper edge line of the rolled piece in the image of the rolled piece entrance area and the image of the rolled piece exit area, the upper edge line of the rolled piece in the dark channel and the upper fitted edge of the rolled piece, and determine the image random factor according to the grayscale change of the pixel points in the image of the rolled piece entrance area and the image of the rolled piece exit area and the lower fitted edge of the rolled piece; combine the image virtualization factor and the image random factor to determine the enhancement coefficient, that is, the edge loss of the rolled piece in the grayscale image of the rolled piece rolling, and enhance all low-brightness pixels in the edge area of ​​the rolled piece according to the edge loss of the rolled piece to eliminate the influence of the lubricating coolant on the recognition of the edge information of the rolled piece, improve the recognition accuracy of the thickness of the rolled piece, thereby improving the monitoring accuracy of the rolled piece rolling process and the feedback control accuracy of the intelligent production system, and improving the process level.

[0074] The following will introduce in detail a method and device for extrusion integrated molding of a high-strength alloy template provided by an embodiment of the present invention in conjunction with the accompanying drawings.

[0075] Figure 2 A basic flow chart of a method for extrusion integral molding of a high-strength alloy template provided by an embodiment of the present invention is shown, such as Figure 2 As shown, the method specifically comprises the following steps:

[0076] Step S100: Acquire a grayscale image of a rolled piece, perform edge detection on the grayscale image of the rolled piece, and obtain an upper edge line and a lower edge line of the rolled piece.

[0077] In the extrusion integrated molding process of high-strength alloy templates, the high-definition camera set at each roller is used to collect the rolling images of the rolled pieces, and the collected rolling images of the rolled pieces are sent to the processing system of the intelligent production system. The processing system analyzes and processes the rolling images of the rolled pieces, and performs extrusion integrated molding control based on the analysis and processing results. The specific process of analyzing and processing the rolling images of the rolled pieces will be introduced in detail below in combination with specific steps.

[0078] First, the processing system performs grayscale processing on the rolled product rolling image to obtain a rolled product rolling grayscale image. Since the specific implementation steps of the image grayscale processing belong to the conventional technology in this field, they are not repeated here. The semantic segmentation algorithm is used to perform mask processing on the background in the rolled product rolling grayscale image, so as to obtain the final rolled product rolling grayscale image. Figure 3 A schematic diagram of a final grayscale image of a rolled product in an embodiment of the present invention is given.

[0079] Theoretically, when the workpiece passes through the roller, the position where the workpiece just touches the roller in the feeding direction on the left side is called the elastic deformation zone, the position where the workpiece touches the roller in the extrusion direction on the right side is called the elastic recovery zone, and the area between the elastic deformation zone and the elastic recovery zone is called the plastic deformation zone. Figure 4 As shown, 1 represents the elastic deformation zone, 2 represents the plastic deformation zone, and 3 represents the elastic recovery zone.

[0080] Since the plastic deformation zone is completely compacted by the rollers, the deformation process of the rolled piece occurs outside the plastic deformation zone, and the intelligent production system needs to monitor the deformation process of the rolled piece. However, during the monitoring process, due to the coverage of the lubricating coolant, the deformation area cannot be determined based on the clear edge, so the real rolled piece edge needs to be enhanced.

[0081] Considering that the lubricating coolant evenly sprayed on the workpiece surface almost slides vertically from the side in the area far away from the roller (if the side of the workpiece is smooth, the sliding path is close to vertical), but when the contact roller deformation begins to occur, the workpiece is squeezed and thinned, and the liquid will flow to the low potential position, and the sliding path of the lubricating coolant between the elastic deformation zone and the plastic deformation zone will be inclined. In addition, due to the close contact between the roller and the workpiece, the liquid will be blocked by the contact point between the roller and the workpiece when flowing to the low potential position, causing the liquid to accumulate behind the contact point; therefore, the liquid sliding path between the elastic deformation zone and the plastic deformation zone is inclined and has the largest liquid flow. Therefore, according to the liquid sliding performance, the boundary line of the plastic deformation zone can be obtained, and then according to the obtained boundary line of the plastic deformation zone, the blur of the workpiece boundary can be accurately evaluated, and then the contrast of the edge area can be enhanced, and finally a grayscale image of the workpiece rolling with clear edges can be obtained.

[0082] To this end, canny edge detection is performed on the final rolled product rolling grayscale image obtained above, and then all discontinuous edges are connected completely according to the principle of nearest connection, so as to obtain an edge image. Figure 5 Given Figure 3 The edge image of the final grayscale image of the rolled product. Figure 5 It can be seen that there is obvious liquid sag on the lower edge line of the rolled piece, which will affect the recognition of the actual rolled piece edge.

[0083] The trained neural network is used to identify the upper and lower edges of the rolled piece in the edge image, and finally the upper and lower edge lines of the rolled piece can be obtained. Since the upper edge of the rolled piece is less disturbed by the liquid, its edge segmentation result is relatively good.

[0084] Step S200: Segment the workpiece entrance area image and the workpiece exit area image from the workpiece rolling grayscale image, use both the workpiece entrance area image and the workpiece exit area image as a target analysis image, perform dark channel processing on the target analysis image, and obtain a dark channel workpiece upper edge line and a dark channel workpiece lower edge line.

[0085] Because the lubricating coolant is evenly sprayed on the workpiece surface, the lubricating coolant almost slides vertically from the side in the area far away from the roller (if the side of the workpiece is smooth, the sliding path is close to vertical), but when the workpiece contacts the roller and begins to deform, the workpiece is squeezed and thinned, and the liquid will flow to the low potential position, and the sliding path of the lubricating coolant between the elastic deformation zone and the plastic deformation zone will be inclined; in addition, due to the close contact between the roller and the workpiece, the liquid will be blocked by the contact point between the roller and the workpiece when flowing to the low potential position, causing the liquid to accumulate after the contact point, so the liquid sliding path between the elastic deformation zone and the plastic deformation zone is inclined and has the largest liquid flow. Therefore, according to the liquid sliding performance, the boundary of the plastic deformation zone can be obtained, and based on the boundary, the image of the workpiece entrance area and the image of the workpiece exit area can be segmented from the grayscale image of the workpiece rolling, so as to facilitate the subsequent acquisition of the blurring of the edge of the workpiece, and then feedback to enhance the contrast of the edge area.

[0086] In an embodiment of the present invention, the steps of segmenting the rolling piece entrance area image and the rolling piece exit area image from the rolling piece rolling grayscale image include:

[0087] Segmenting the rolled product rolling grayscale image into a rolled product inlet side image and a rolled product outlet side image;

[0088] The image on the inlet side of the rolled piece and the image on the outlet side of the rolled piece are both regarded as an object image, and according to the grayscale change of the pixel points in the object image, the liquid sliding deviation of each edge pixel point on the upper edge line of the rolled piece in the object image is determined;

[0089] Determining a contact point on the upper edge line of the rolled piece in the object image according to the liquid sliding deviation;

[0090] According to the contact points on the upper edge line of the rolled piece in the image of the rolled piece entrance side, the rolled piece entrance area image is segmented from the image of the rolled piece entrance side, and according to the contact points on the upper edge line of the rolled piece in the image of the rolled piece exit side, the rolled piece exit area image is segmented from the image of the rolled piece exit side.

[0091] For the above steps, since the directions of the roll entrance and exit are opposite, and the rolling direction in the grayscale image of the rolled product rolling is vertically downward in the embodiment of the present invention, the grayscale image of the rolled product rolling is divided into left and right side images along the center line, the left image is the image of the rolled product entrance side, and the right image is the image of the rolled product exit side. In the left image, i.e., the image of the rolled product entrance side, the number of columns is divided along the direction from left to right, i.e., the first column, the second column, the third column, etc. from left to right; in the right image, i.e., the image of the rolled product exit side, the number of columns is divided along the direction from right to left, i.e., the first column, the second column, the third column, etc. from right to left.

[0092] Since the lubricating coolant is a translucent mixed liquid with a slightly turbid characteristic, the higher the grayscale value and the smaller the grayscale gradient, the thicker and more turbid the "water film" representing the liquid sliding down, and vice versa, the thinner and more transparent the "water film". In order to express the liquid sliding down situation in each column of the rolled piece inlet side image and the rolled piece outlet side image, the rolled piece inlet side image and the rolled piece outlet side image are both taken as an object image. The embodiment of the present invention takes the left image, i.e., the rolled piece inlet side image, as an example, and takes each edge pixel point on the upper edge line of the rolled piece in the rolled piece inlet side image (one edge pixel point corresponds to one column) as the starting point of liquid sliding down, and traverses downward along the column direction with the liquid sliding starting point as the starting point to obtain the pixel point sequence corresponding to each edge pixel point, and obtains the grayscale mean and grayscale standard deviation of all pixels in the pixel point sequence.

[0093] When the left image, i.e., the image on the entrance side of the rolled piece, is taken as the object image, the liquid slippage deviation of each edge pixel point on the upper edge line of the rolled piece in the object image can be determined based on the grayscale mean difference between each edge pixel point on the upper edge line of the rolled piece and its previous edge pixel point, the grayscale mean difference between each edge pixel point and its subsequent edge pixel point, and the grayscale mean and grayscale standard deviation of the subsequent edge pixel point.

[0094] In the embodiment of the present invention, the liquid sliding deviation of each edge pixel point on the edge line of the rolled piece in the object image is determined, and the corresponding calculation formula is:

[0095] ;

[0096] Indicates the first Liquid sliding deviation of edge pixels; Indicates the first The grayscale mean of all pixels in the pixel sequence corresponding to the edge pixel; Indicates the first The grayscale mean of all pixels in the pixel sequence corresponding to the edge pixel; Indicates the first The grayscale mean of all pixels in the pixel sequence corresponding to the edge pixel; Indicates the first The grayscale standard deviation of all pixels in the pixel sequence corresponding to the edge pixels.

[0097] In the above calculation formula, the first term The first The ratio of the grayscale mean to the grayscale standard deviation of all pixels in the pixel sequence corresponding to the edge pixel. The larger the ratio, the higher the grayscale standard deviation of all pixels in the pixel sequence corresponding to the edge pixel. The larger the total amount of liquid sliding down in the corresponding column of edge pixels, the larger the total amount of liquid sliding down in the corresponding column of edge pixels. The more likely the edge pixel is to be the liquid accumulation position behind the real contact point, the The edge pixel point is located at the previous adjacent The more edge pixels there are, the more likely they are to be the contact point columns, which is the boundary of the plastic deformation zone.

[0098] In the above calculation formula, the second term The first The edge pixel and the The difference between the grayscale mean values ​​of the edge pixels divided by the The edge pixel and the -1 is the absolute value of the difference between the grayscale mean values ​​of the edge pixels. The larger the numerator, the The total amount of sliding liquid corresponding to the edge pixel is greater than that of the The corresponding column of edge pixels represents the direction of liquid sliding down to the The corresponding column of edge pixels is tilted; the larger the denominator, the The edge pixel and the The greater the difference in the total amount of liquid sliding down the corresponding columns of edge pixels, the greater the difference in the absolute value. Greater than When The corresponding column of edge pixels is tilted (tilted toward the feeding direction); Less than When -1 edge pixel point corresponds to a tilted column (tilted in the opposite direction of the feed direction, which may be caused by the rough side of the rolled piece). In either case, as long as the numerator is greater than the denominator, it means that the liquid is in the first The corresponding column of edge pixels has an obvious tendency to tilt toward the feeding direction of the rolled piece. Therefore, the second term is combined with the first term to calculate the product of the two. The larger the product, the more serious the liquid sliding direction tilts to the right. The larger the liquid sliding deviation is, the larger the edge pixel is.

[0099] According to the above method, when the left image, i.e., the image on the entrance side of the rolled piece, is taken as the object image, the liquid sliding deviation of each edge pixel point on the upper edge line of the rolled piece in the object image can be determined, that is, the liquid sliding deviation of each edge pixel point on the upper edge line of the rolled piece in the left image can be determined.

[0100] For the right image, i.e., the image at the exit of the rolled piece, the rolled piece will have deformation rebound in the elastic recovery zone, so the thickness of the rolled piece outside the exit is also greater than the thickness of the rolled piece in the plastic deformation zone, i.e., the low potential point at the exit is on the left side of the contact point, and the liquid sliding direction at the exit contact point should be tilted to the left. Since the column division directions of the left and right images are set oppositely, the liquid sliding deviation of each edge pixel on the upper edge line of the rolled piece in the object image can be determined in the same manner as when the left image, i.e., the image at the entrance of the rolled piece, is used as the object image. When the right image, i.e., the image at the exit of the rolled piece, is used as the object image, the liquid sliding deviation of each edge pixel on the upper edge line of the rolled piece in the object image can be determined. That is, the liquid sliding deviation of each edge pixel on the upper edge line of the rolled piece in the right image can be determined.

[0101] When the left image, i.e., the image on the inlet side of the rolled piece, is used as the object image, according to the above method of determining the liquid sliding deviation of each edge pixel point on the upper edge line of the rolled piece in the object image, it can be known that the liquid sliding deviation reflects the degree to which the liquid sliding direction tilts to the right. When the liquid sliding deviation changes the most, it means that the degree to which the liquid sliding direction tilts to the right is greater, and the point with the greatest degree of change is the contact point position on the upper edge line of the rolled piece in the object image. Therefore, according to the change of the liquid sliding deviation of each edge pixel point on the upper edge line of the rolled piece in the object image, the contact point on the upper edge line of the rolled piece in the object image can be determined.

[0102] In an embodiment of the present invention, the liquid slippage deviation slope of each edge pixel point on the upper edge line of the rolled piece in the object image is determined based on the fluctuation change of the liquid slippage deviation of each edge pixel point on the upper edge line of the rolled piece in the object image. Then, on the upper edge line of the rolled piece in the object image, the edge pixel point corresponding to the maximum value of the liquid slippage deviation slope is determined as the contact point on the upper edge line of the rolled piece in the object image. In the process of determining the liquid slippage deviation slope of each edge pixel point on the upper edge line of the rolled piece in the object image, the column number of the column where each edge pixel point on the upper edge line of the rolled piece in the object image is located is used as the horizontal coordinate, and the liquid slippage deviation of each edge pixel point is used as the vertical coordinate. The horizontal coordinate and the vertical coordinate corresponding to each edge pixel point can determine a coordinate point, which is called the mapping point of the corresponding edge pixel point in the two-dimensional coordinate system. For the mapping points of each edge pixel point on the upper edge line of the rolled piece in the object image in the two-dimensional coordinate system, two adjacent mapping points are connected to obtain a mapping point polyline, also called the liquid slippage deviation polyline. The difference between the vertical coordinates of each mapping point on the liquid slippage deviation polyline and its previous mapping point is determined to obtain the slope value of each mapping point on the liquid slippage deviation polyline. The slope value is the liquid slippage deviation slope of the corresponding edge pixel point on the upper edge line of the rolled piece in the object image. Figure 6 A schematic diagram of the liquid slippage deflection line corresponding to some edge pixels on the upper edge line of the rolled piece in the object image is given.

[0103] By the above method, when the left image, i.e., the image at the entry side of the rolled piece, is used as the object image, the contact point on the upper edge line of the rolled piece in the object image can be determined. Since the contact point is located in the entry area of ​​the rolled piece, the contact point can be recorded as the entry end contact point of the rolled piece. The left side of the column where the entry end contact point of the rolled piece is located is the elastic deformation area where the original rolled piece thickness undergoes elastic deformation, and the right side is the plastic deformation area.

[0104] Similarly, when the right image, i.e., the image at the exit of the rolled piece, is used as the object image, the contact point on the upper edge line of the rolled piece in the object image can be determined. Since the contact point is located in the exit area of ​​the rolled piece, the contact point can be recorded as the exit contact point of the rolled piece. The left side of the column where the exit contact point of the rolled piece is located is the plastic deformation zone, and the right side is the elastic recovery zone.

[0105] The left image, i.e., the image of the column where the contact point of the rolling piece inlet end is located and all the columns to the left of the image on the left side, is obtained, and this image is called the rolling piece inlet region image; at the same time, the right image, i.e., the image of the column where the contact point of the rolling piece outlet end is located and all the columns to the right of the image on the right side, is obtained, and this image is called the rolling piece outlet region image. Thus, the segmentation of the rolling piece inlet region image and the rolling piece outlet region image from the rolling piece rolling grayscale image is realized.

[0106] The image of the rolled piece entrance area and the image of the rolled piece exit area are both regarded as a target analysis image. Taking the image of the rolled piece entrance area as the target analysis image as an example, dark channel processing is performed on the target analysis image to obtain the edge of the rolled piece in the dark channel, that is, the upper edge line of the rolled piece in the dark channel and the lower edge line of the rolled piece in the dark channel. Subsequently, by combining the upper edge line and the lower edge line of the rolled piece in the target analysis image, the upper and lower fitting edges of the rolled piece can be obtained based on the double edges.

[0107] In an embodiment of the present invention, the upper edge line and the lower edge line of the dark channel rolled piece are obtained, and the implementation steps include:

[0108] Performing minimum filtering on the target analysis image to obtain a dark channel image;

[0109] The rolled piece area in the dark channel image is acquired, and edge segmentation is performed on the rolled piece area to obtain an upper edge line of the dark channel rolled piece and a lower edge line of the dark channel rolled piece.

[0110] For the above steps, when the image of the rolling stock entrance area is used as the target analysis image, the minimum filtering algorithm is used to filter the target analysis image. The filter size of the minimum filtering algorithm is 11*11, and the image after filtering is called the dark channel image. Due to the influence of the minimum filtering, the edge of the rolling stock in the dark channel image will be eroded by the expansion of the dark effect. The dark channel image is superpixel segmented, and the rolling stock area is determined based on each superpixel block obtained by superpixel segmentation. In an embodiment of the present invention, in the camera's shooting angle of view of the rolling stock, a coordinate position in the middle of the side of the rolling stock is arbitrarily selected, and then the superpixel area containing the coordinate position is recorded as the rolling stock area. Then, the upper and lower edges of the rolling stock area are obtained by using the canny edge segmentation algorithm, and the upper and lower edges of the rolling stock area obtained by segmentation are called the upper edge line of the dark channel rolling stock and the lower edge line of the dark channel rolling stock. It should be understood that in other achievable methods, a neural network can also be used to directly segment the rolling stock area in the dark channel image, which is not limited here.

[0111] Similarly, when the image of the rolling stock exit area is used as the target analysis image, the dark channel rolling stock upper edge line and the dark channel rolling stock lower edge line of the target analysis image can also be determined in the same manner.

[0112] Step S300: Determine the upper fitting edge of the rolled piece and the lower fitting edge of the rolled piece by combining the pixel grayscale changes between the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece in the target analysis image, and the pixel grayscale changes between the lower edge line of the rolled piece and the lower edge line of the dark channel rolled piece.

[0113] Since the edges of the dark channel workpiece in the above dark channel image, namely the upper edge line and the lower edge line of the dark channel workpiece, are the edges eroded by the dark effect, the thickness of the workpiece determined by the edges of the dark channel workpiece in the dark channel image will be smaller than the actual thickness of the workpiece, and the edges of the workpiece obtained by canny edge detection on the grayscale image of workpiece rolling, namely the upper edge line and the lower edge line of the workpiece, are the outward expanded edges blurred by the cooling lubricating liquid covering the workpiece. At this time, the thickness of the workpiece determined by the upper edge line and the lower edge line of the workpiece will be greater than the actual thickness of the workpiece. Therefore, in the grayscale image of workpiece rolling, the upper edge line and the lower edge line of the dark channel workpiece will be wrapped by the upper edge line and the lower edge line of the workpiece. At this time, the actual edge of the workpiece should be located between the edge of the dark channel workpiece and the edge of the workpiece obtained by canny edge detection.

[0114] For the dark channel workpiece edge and the workpiece edge obtained by canny edge detection, since any low-brightness pixel between these two edges may be the real workpiece edge, the upper fitting edge and the lower fitting edge of the workpiece are determined based on the change in pixel grayscale between the two edges in the target analysis image.

[0115] In an embodiment of the present invention, the upper fitting edge and the lower fitting edge of the rolled piece are determined, and the implementation steps include:

[0116] Determine the grayscale minimum values ​​of all pixels on the line connecting two corresponding pixel points between the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece in the target analysis image, perform curve fitting on the pixel points corresponding to these grayscale minimum values, and obtain the fitted edge on the rolled piece;

[0117] Determine the grayscale minimum values ​​of all pixels on the line connecting two corresponding pixel points between the lower edge line of the rolled piece and the lower edge line of the dark channel rolled piece in the target analysis image, perform curve fitting on the pixel points corresponding to these grayscale minimum values, and obtain the lower fitted edge of the rolled piece.

[0118] For the above steps, when the image of the rolling stock entrance area is used as the target analysis image, two edge pixel points located in the same column on the upper edge line of the rolling stock and the upper edge line of the dark channel rolling stock in the target analysis image are determined, and the two edge pixel points are called corresponding pixel points. All the pixel points located in the same column and between the two corresponding pixel points are formed into a pixel point column according to the distribution order of the pixel points, and the grayscale minimum pixel point of the pixel point column is determined, that is, the pixel point corresponding to the grayscale minimum value of all the pixels in the pixel point column, so that the grayscale minimum point between the upper edge line of the rolling stock and the upper edge line of the dark channel rolling stock in each column of the target analysis image can be obtained, and then all the grayscale minimum points are curve-fitted to obtain the upper fitting edge of the rolling stock. Similarly, for the lower edge line of the rolling stock and the lower edge line of the dark channel rolling stock in the target analysis image, the grayscale minimum point between the lower edge line of the rolling stock and the lower edge line of the dark channel rolling stock in each column of the target analysis image can also be obtained, and then all the grayscale minimum points are curve-fitted to obtain the lower fitting edge of the rolling stock. The upper fitting edge of the rolled piece and the lower fitting edge of the rolled piece are collectively referred to as the fitting edge of the rolled piece.

[0119] Similarly, when the image of the exit area of ​​the rolled piece is used as the target analysis image, the upper fitting edge of the rolled piece and the lower fitting edge of the rolled piece of the target analysis image can also be determined in the same manner.

[0120] Step S400: Determine the image virtualization factor by combining the grayscale difference between the upper edge line of the workpiece, the upper edge line of the dark channel workpiece and the upper fitted edge of the workpiece in the workpiece entrance area image and the workpiece exit area image, and determine the image random factor by combining the grayscale change of the pixel points in the workpiece entrance area image and the workpiece exit area image and the pixel grayscale of the fitted edge below the workpiece.

[0121] For the images of the entry area and the exit area of ​​the rolled piece, the upper edge of the rolled piece only has the problem of edge blurring caused by the "water film" coverage. The blurring can be regarded as the static blurring of the edge of the rolled piece. Therefore, the image blurring factor can be calculated to evaluate the degree of static blurring of the edge of the rolled piece.

[0122] Considering that the real edge of the rolled piece is between the edge of the dark channel rolled piece and the edge of the rolled piece obtained by canny edge detection on the grayscale image of the rolled piece, and the fitted edge of the rolled piece is obtained by fitting the pixels that are least affected by the liquid coverage, the image virtualization factor can be determined by comparing the grayscale conditions of the pixels on the upper edge line of the rolled piece, the upper edge line of the dark channel rolled piece, and the upper fitted edge of the rolled piece in the same column.

[0123] In an embodiment of the present invention, the image virtualization factor is determined, and the implementation steps include:

[0124] Determine the average grayscale value of the corresponding pixel points between the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece in the target analysis image, and determine the grayscale difference between the average grayscale value and the grayscale value of the corresponding pixel points on the fitted edge of the rolled piece;

[0125] Determine the blur factor corresponding to the target analysis image by combining the overall distribution level of all the grayscale differences;

[0126] The image virtualization factor is determined by comprehensively considering the virtualization factors corresponding to the image of the rolling stock entrance area and the image of the rolling stock exit area, and the image virtualization factor is a normalized value.

[0127] For the above steps, in the embodiment of the present invention, when the image of the rolling stock entrance area is used as the target analysis image, the deblurring factor corresponding to the target analysis image is determined, and the corresponding calculation formula is:

[0128] ;

[0129] in, It indicates the blur factor corresponding to the target analysis image when the image of the rolling stock entrance area is used as the target analysis image; Represents the gray value of the pixel point located on the upper edge line of the rolled piece in the sth column of the target analysis image; Represents the gray value of the pixel point located on the upper edge line of the dark channel rolled piece in the sth column of the target analysis image; Represents the gray value of the pixel point located on the fitting edge of the rolled piece in the sth column of the target analysis image; Indicates the number of columns of pixels in the target analysis image.

[0130] In the above calculation formula, the average grayscale value of the corresponding pixel points (i.e., the pixel points located in the same column) between the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece in the target analysis image is calculated, and the average grayscale value is compared with the grayscale value of the pixel points located on the fitting edge of the rolled piece in the same column to obtain the grayscale difference, thereby calculating the mean square error. The larger the mean square error, the higher the degree of static blurring of the edge of the rolled piece by the water film. At this time, when the image of the rolled piece entrance area is used as the target analysis image, the larger the value of the sub-blurring factor corresponding to the target analysis image is.

[0131] According to the above method, when the image of the rolled piece entrance area is used as the target analysis image, the blur factor corresponding to the target analysis image can be determined, that is, the blur factor corresponding to the image of the rolled piece entrance area can be determined; similarly, when the image of the rolled piece exit area is used as the target analysis image, according to the same method, the blur factor corresponding to the target analysis image can also be determined, that is, the blur factor corresponding to the image of the rolled piece exit area can be determined.

[0132] The sub-blur factors corresponding to the image of the entry area of ​​the rolled piece and the image of the exit area of ​​the rolled piece are integrated, that is, the sub-blur factors corresponding to the image of the entry area of ​​the rolled piece and the image of the exit area of ​​the rolled piece are averaged, and then the average is normalized to obtain the total blur factor, which is the image virtualization factor. Among them, the specific implementation means of the normalization process can be reasonably selected according to needs, and is not specifically limited here. In the embodiment of the present invention, the norm normalization function is used for normalization.

[0133] In addition, for the images of the entrance area and the exit area of ​​the rolled piece, the clarity of the lower edge of the rolled piece is theoretically affected by the free sliding of the liquid. However, in addition to this, the shooting clarity may also be affected by factors such as astigmatism and refraction, resulting in deformation of the edge, which can be regarded as dynamic blur of the lower edge line.

[0134] Considering that the liquid slippage deviation line represents the slippage trend and path change of the liquid on the side of the rolled piece, and for the lower edge of the rolled piece, the liquid slippage trend should be negatively correlated with the quality of the fitted lower edge, because the edge of the rolled piece must be less clear at the location where the liquid slippage accumulates to a higher degree, and the "water film" is thinner at the location where the liquid slippage does not accumulate, and the edge of the rolled piece is clearer, so the two are negatively correlated. Therefore, in the image of the rolled piece entrance area and the image of the rolled piece exit area, combined with the liquid slippage deviation of each edge pixel point on the upper edge line of the rolled piece determined according to the grayscale change of the pixel points in the image, and the grayscale value of the upper edge pixel point of the lower fitted edge of the rolled piece, the image random factor can be determined, and the image random factor reflects the negative correlation between the liquid slippage trend and the quality of the fitted lower edge.

[0135] In an embodiment of the present invention, the image random factor is determined, and the implementation steps include:

[0136] Normalizing the liquid slippage deviation of each edge pixel point on the upper edge line of the rolled piece in the target analysis image to obtain a first normalized value;

[0137] Normalizing the grayscale value of each edge pixel point on the lower fitting edge of the rolled piece in the target analysis image to obtain a second normalized value;

[0138] Determine a sub-random factor according to a difference between the first normalized value of each edge pixel point on the upper edge line of the rolled piece in the target analysis image and the second normalized value of a corresponding edge pixel point on the lower fitting edge of the rolled piece in the target analysis image;

[0139] The image random factor is determined by comprehensively considering the sub-random factors corresponding to the image of the rolling piece entrance area and the image of the rolling piece exit area, and the image random factor is a normalized value.

[0140] For the above steps, in an embodiment of the present invention, when the image of the entrance area of ​​the rolled piece is used as the target analysis image, the liquid slip deviation of each edge pixel point on the upper edge line of the rolled piece in the target analysis image is normalized to obtain a first normalized value. The normalization process is: determine the maximum value of the liquid slip deviation of each edge pixel point on the upper edge line of the rolled piece, and divide the liquid slip deviation of each edge pixel point on the upper edge line of the rolled piece by the maximum value. The quotient obtained is the first normalized value.

[0141] At the same time, the grayscale values ​​of each edge pixel point on the fitted edge of the rolled piece in the target analysis image are normalized to obtain a second normalized value. The normalization process is: determine the maximum value among the grayscale values ​​of each edge pixel point on the fitted edge of the rolled piece, divide the grayscale value of each edge pixel point on the fitted edge of the rolled piece by the maximum value, and the obtained quotient is the second normalized value.

[0142] By normalizing the liquid slippage deviation of each edge pixel point on the upper edge line of the rolled piece in the target analysis image to obtain the first normalized value, and normalizing the grayscale value of each edge pixel point on the lower fitting edge of the rolled piece to obtain the second normalized value, the influence of the dimension can be eliminated, so that the mean square error can be calculated based on the difference between the first normalized value of each edge pixel point on the upper edge line of the rolled piece in the target analysis image and the second normalized value of the corresponding edge pixel point on the lower fitting edge of the rolled piece in the target analysis image, that is, the difference between the first normalized value of the edge pixel point on the upper edge line of the rolled piece located in the same column in the target analysis image and the second normalized value of the edge pixel point on the lower fitting edge of the rolled piece. The larger the value of the mean square error, the greater the negative correlation between the edge clarity of the lower fitting and the liquid slippage path, that is, the smaller the degree of dynamic blurring of the lower edge of the rolled piece. The difference obtained by subtracting the mean square error from the value 1 to correct the logical relationship is the sub-random factor of the target analysis image. The larger the value of the sub-random factor, the greater the degree of dynamic blurring of the lower edge of the rolled piece.

[0143] According to the above method, when the image of the rolled piece entrance area is used as the target analysis image, the sub-random factor corresponding to the target analysis image can be determined, that is, the sub-random factor corresponding to the image of the rolled piece entrance area can be determined; similarly, when the image of the rolled piece exit area is used as the target analysis image, the sub-random factor corresponding to the target analysis image can also be determined in the same way, that is, the sub-random factor corresponding to the image of the rolled piece exit area can be determined.

[0144] The sub-random factors corresponding to the integrated rolling piece entrance area image and the rolled piece exit area image are averaged, and the average is used as the total random factor. The total blurring factor is the image random factor.

[0145] Step S500: Determine an enhancement coefficient according to the image virtualization factor and the image random factor, and perform enhancement processing on low-brightness pixels in the rolled product rolling grayscale image according to the enhancement coefficient, so as to obtain the rolled product rolling grayscale image after enhancement processing.

[0146] By combining the image virtualization factor and the image random factor determined above, the edge loss amount of the grayscale image of the rolled product can be obtained, that is, the product value of the image virtualization factor and the image random factor is calculated, and the product value is used as the edge loss amount of the grayscale image of the rolled product. The edge loss amount is used as the enhancement coefficient of the low-brightness pixel points in the grayscale image of the rolled product, and based on the enhancement coefficient, the low-brightness pixel points in the grayscale image of the rolled product are enhanced to increase the contrast of the edge of the rolled product in the image, thereby facilitating more accurate detection of the thickness of the rolled product in the image.

[0147] In an embodiment of the present invention, the step of determining low-brightness pixel points in the grayscale image of the rolled product includes:

[0148] Determine the target analysis area corresponding to the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece, and the target analysis area corresponding to the lower edge line of the rolled piece and the lower edge line of the dark channel rolled piece in the target analysis image;

[0149] Determine the grayscale histogram of all pixels in each target analysis area, and determine the average grayscale level in the grayscale histogram; determine the low-brightness pixels in the rolled product rolling grayscale image based on the pixels corresponding to all grayscale levels less than the average grayscale level in the grayscale histogram.

[0150] For the above steps, when the image of the entrance area of ​​the rolled piece is used as the target analysis image, the minimum circumscribed rectangle of the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece in the target analysis image is determined, and the area framed by the minimum circumscribed rectangle is used as the target analysis area corresponding to the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece; at the same time, the minimum circumscribed rectangle of the lower edge line of the rolled piece and the lower edge line of the dark channel rolled piece in the target analysis image is determined, and the area framed by the minimum circumscribed rectangle is used as the target analysis area corresponding to the lower edge line of the rolled piece and the lower edge line of the dark channel rolled piece.

[0151] Obtain the grayscale values ​​of all pixels in each target analysis area, determine the grayscale histogram, determine the average of all grayscales with non-zero frequency values ​​involved in the grayscale histogram, and obtain the average grayscale. Then, determine the pixels corresponding to all grayscales with non-zero frequency values ​​involved in the grayscale histogram that are less than the corresponding average grayscale as low-brightness pixels.

[0152] According to the above method, when the image of the rolled piece entrance area is used as the target analysis image, all low-brightness pixels in the target analysis image can be determined, that is, all low-brightness pixels in the image of the rolled piece entrance area can be determined; similarly, when the image of the rolled piece exit area is used as the target analysis image, according to the same method, all low-brightness pixels in the target analysis image can also be determined, that is, all low-brightness pixels in the image of the rolled piece exit area can be determined.

[0153] According to the positions of all low-brightness pixels in the rolled piece entrance area image and the rolled piece exit area image, the corresponding pixels at the same positions in the rolled piece rolling grayscale image can be found. These pixels are the low-brightness pixels in the rolled piece rolling grayscale image.

[0154] The edge loss determined above is used as the enhancement coefficient of the low-brightness pixel points in the grayscale image of the rolled product, and the low-brightness pixel points in the grayscale image of the rolled product are enhanced, that is, the grayscale values ​​of the low-brightness pixel points in the grayscale image of the rolled product are reduced to times the original edge loss amount, so as to improve the contrast of the edge of the rolled product in the image area wrapped by the liquid, thereby obtaining the enhanced grayscale image of the rolled product. It should be understood that if the edge loss amount times of the grayscale value obtained when the grayscale value is reduced to times the original edge loss amount is not an integer, then the edge loss amount times of the grayscale value is taken down. Then, the enhanced grayscale image of the rolled product is laterally smoothed using Gaussian filtering to eliminate discrete dark spots, thereby obtaining the final enhanced grayscale image of the rolled product.

[0155] Step S600: performing extrusion integrated molding control according to the rolled product rolling grayscale image after the enhancement processing.

[0156] The final enhanced grayscale image of the rolled product is subjected to canny edge detection to obtain an edge image. Figure 7 Shows Figure 3 Schematic diagram of edge detection results of the rolled product rolling grayscale image after enhanced processing corresponding to the final rolled product rolling grayscale image in FIG. Figure 5 and Figure 7 It can be seen that by using the enhancement coefficient to enhance the low-brightness pixels in the grayscale image of the rolled product to improve the contrast, the edge of the rolled product can be better detected. Although there are still intermittent missing parts of the local lower edge, it is sufficient to meet the monitoring accuracy requirements of the continuous rolling process of the rolled product. The edge can be further processed by edge connection, refinement and other means to obtain the final edge image, which will not be repeated here.

[0157] Based on the edge image, the distance between the upper and lower edges of the rolled piece is measured to obtain the accurate thickness information of the rolled piece when it passes through the entrance and exit positions of each set of rollers, that is, the thickness of the rolled piece. Based on the real-time monitored thickness of the rolled piece, the parameters of the cold rolling and rolling device can be adjusted. The parameter adjustment process can be selected as needed and is not limited here. In an embodiment of the present invention, based on the real-time monitored thickness of the rolled piece, the parameters of the cold rolling and rolling device are adjusted, and the implementation process includes:

[0158] 1. Set the target thickness of the roll;

[0159] 2. Initialize the roller speed and roller pressure to the preset values;

[0160] 3. Obtain the actual thickness of the rolled plate at the entrance and exit positions, and calculate the error between the actual thickness and the set thickness;

[0161] 4. Input the thickness error into the PID controller and calculate the control output. The PID controller adjusts the output according to the thickness error size, change speed and cumulative error to achieve fast and stable control.

[0162] 5. According to the output of PID controller, adjust the roller speed and roller pressure:

[0163] If the thickness error is positive (actual thickness is less than the set thickness), increase the roll speed and / or roll pressure.

[0164] If the thickness error is negative (actual thickness is greater than the set thickness), reduce the roll speed and / or roll pressure.

[0165] 6. Feedback loop: Continuously detect the actual thickness and repeat the above steps until the error between the actual thickness and the set thickness is within an acceptable range.

[0166] 7. When the roll thickness stabilizes near the set value, the PID controller output tends to be stable, and the roll speed and roll pressure also remain stable.

[0167] The present invention analyzes the grayscale image of rolled product to determine the image virtualization factor and the image random factor, and combines the image virtualization factor and the image random factor to determine the enhancement coefficient, that is, the edge loss of the rolled product in the grayscale image of rolled product, and enhances all low-brightness pixels in the edge area of ​​the rolled product according to the edge loss of the rolled product to eliminate the influence of lubricating coolant on the identification of the edge information of the rolled product, improve the identification accuracy of the rolled product thickness, thereby improving the monitoring accuracy of the rolled product rolling process and the feedback control accuracy of the intelligent production system, and improving the process level.

[0168] Based on the same inventive concept, an embodiment of the present invention further provides an extrusion integrated molding device for a high-strength alloy template, comprising:

[0169] The rolling piece edge line acquisition module is used to: acquire the rolling piece grayscale image, perform edge detection on the rolling piece grayscale image, and obtain the upper edge line and the lower edge line of the rolling piece;

[0170] A dark channel edge line acquisition module is used to: segment the roll entry area image and the roll exit area image from the roll rolling grayscale image, use the roll entry area image and the roll exit area image as a target analysis image, perform dark channel processing on the target analysis image, and obtain a dark channel roll upper edge line and a dark channel roll lower edge line;

[0171] A fitting edge line acquisition module is used to determine the upper fitting edge of the rolled piece and the lower fitting edge of the rolled piece by combining the pixel grayscale changes between the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece in the target analysis image, and the pixel grayscale changes between the lower edge line of the rolled piece and the lower edge line of the dark channel rolled piece;

[0172] An image factor acquisition module is used to: determine the image virtualization factor by combining the grayscale difference between the upper edge line of the rolled piece, the upper edge line of the dark channel rolled piece and the upper fitted edge of the rolled piece in the rolled piece entrance area image and the rolled piece exit area image, and determine the image random factor by combining the grayscale change of the pixel points and the pixel grayscale of the lower fitted edge of the rolled piece in the rolled piece entrance area image and the rolled piece exit area image;

[0173] An enhancement module, used to: determine an enhancement coefficient according to the image virtualization factor and the image random factor, and enhance low-brightness pixel points in the rolled product rolling grayscale image according to the enhancement coefficient, so as to obtain an enhanced rolled product rolling grayscale image;

[0174] The control module is used to: perform extrusion integrated molding control according to the rolled piece rolling grayscale image after the enhancement processing.

[0175] It should be noted that the device provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device is divided into different functional modules to complete all or part of the functions described above. In addition, the device and the embodiment of the extrusion integrated molding method for high-strength alloy templates provided in the above embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0176] Based on the same inventive concept, the embodiment of the present invention also provides an extrusion integrated molding system for a high-strength alloy template, such as Figure 8As shown, the system includes: a memory 801, a processor 802, and a computer program 803 stored in the memory 801 and running on the processor 802, wherein when the processor 802 executes the computer program 803, the system can execute any one of the extrusion one-piece molding methods for high-strength alloy templates introduced above.

[0177] The embodiment of the present invention can divide the system into functional modules according to the above method example. For example, each functional module can be corresponded to, or two or more functions can be integrated into one processing module. The above integrated module can be implemented in the form of hardware. It should be noted that the division of modules in this embodiment is schematic and is only a logical function division. There may be other division methods in actual implementation.

[0178] Based on the same inventive concept, an embodiment of the present invention also provides a computer program product, which includes: computer program code, when the computer program code runs on a computer, the computer executes any one of the above-mentioned extrusion one-piece molding methods for high-strength alloy templates.

[0179] Based on the same inventive concept, an embodiment of the present invention also provides a computer-readable storage medium, which stores a computer program code. When the computer program code runs on a computer, the computer executes any one of the above-mentioned extrusion one-piece molding methods for high-strength alloy templates.

[0180] It should be noted that the above-described embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for extrusion integral molding of a high-strength alloy template, characterized in that: The following steps are involved: Acquire a grayscale image of a rolled piece, perform edge detection on the grayscale image of the rolled piece, and obtain an upper edge line and a lower edge line of the rolled piece; Segmenting a workpiece entrance region image and a workpiece exit region image from the workpiece rolling grayscale image, taking the workpiece entrance region image and the workpiece exit region image as a target analysis image, performing dark channel processing on the target analysis image, and obtaining a dark channel workpiece upper edge line and a dark channel workpiece lower edge line; Determine the upper fitting edge of the rolled piece and the lower fitting edge of the rolled piece by combining the grayscale change of the pixels between the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece in the target analysis image, and the grayscale change of the pixels between the lower edge line of the rolled piece and the lower edge line of the dark channel rolled piece; Determine the image virtualization factor by combining the grayscale difference between the upper edge line of the rolled piece, the upper edge line of the dark channel rolled piece and the upper fitted edge of the rolled piece in the rolled piece entrance area image and the rolled piece exit area image, and determine the image random factor by combining the grayscale change of the pixel points and the pixel grayscale of the lower fitted edge of the rolled piece in the rolled piece entrance area image and the rolled piece exit area image; Determine an enhancement coefficient according to the image virtualization factor and the image random factor, and perform enhancement processing on low-brightness pixel points in the rolled product rolling grayscale image according to the enhancement coefficient, so as to obtain the rolled product rolling grayscale image after enhancement processing; Performing extrusion integrated molding control according to the rolled piece rolling grayscale image after the enhancement process; Determine the image virtualization factors, including: Determine the average grayscale value of the corresponding pixel points between the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece in the target analysis image, and determine the grayscale difference between the average grayscale value and the grayscale value of the corresponding pixel points on the fitted edge of the rolled piece; Determine the blur factor corresponding to the target analysis image by combining the overall distribution level of all the grayscale differences; Determine an image virtualization factor by combining the virtualization factors corresponding to the image of the rolling piece entrance area and the image of the rolling piece exit area, wherein the image virtualization factor is a normalized value; Determine the blur factor corresponding to the target analysis image. The corresponding calculation formula is: ; in, Indicates the blur factor corresponding to the target analysis image; Represents the gray value of the pixel point located on the upper edge line of the rolled piece in the sth column of the target analysis image; Represents the gray value of the pixel point located on the upper edge line of the dark channel rolled piece in the sth column of the target analysis image; Represents the gray value of the pixel point located on the fitting edge of the rolled piece in the sth column of the target analysis image; Indicates the number of columns of pixels in the target analysis image; Determine the image random factors, including: Normalizing the liquid slippage deviation of each edge pixel point on the upper edge line of the rolled piece in the target analysis image to obtain a first normalized value; Normalizing the grayscale value of each edge pixel point on the lower fitting edge of the rolled piece in the target analysis image to obtain a second normalized value; Determine a sub-random factor according to a difference between the first normalized value of each edge pixel point on the upper edge line of the rolled piece in the target analysis image and the second normalized value of a corresponding edge pixel point on the lower fitting edge of the rolled piece in the target analysis image; Determine an image random factor by synthesizing the sub-random factors corresponding to the image of the rolling piece entrance area and the image of the rolling piece exit area, wherein the image random factor is a normalized value; Determine the sub-random factors, including: Calculating a mean square error based on a difference between a first normalized value of an edge pixel point on an upper edge line of a rolled piece located in the same column in the target analysis image and a second normalized value of an edge pixel point on a lower fitting edge of the rolled piece; The difference obtained by subtracting the mean square error from 1 is the sub-random factor of the target analysis image.

2. The method for extrusion integral molding of a high-strength alloy template according to claim 1, characterized in that: Obtaining the upper edge line and the lower edge line of the dark channel rolled piece includes: Performing minimum filtering on the target analysis image to obtain a dark channel image; The rolled piece area in the dark channel image is acquired, and edge segmentation is performed on the rolled piece area to obtain an upper edge line of the dark channel rolled piece and a lower edge line of the dark channel rolled piece.

3. The method for extrusion integral molding of a high-strength alloy template according to claim 1, characterized in that: Determine the upper fitting edge and the lower fitting edge of the rolled piece, including: Determine the grayscale minimum values ​​of all pixels on the line connecting two corresponding pixel points between the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece in the target analysis image, perform curve fitting on the pixel points corresponding to these grayscale minimum values, and obtain the fitted edge on the rolled piece; Determine the grayscale minimum values ​​of all pixels on the line connecting two corresponding pixel points between the lower edge line of the rolled piece and the lower edge line of the dark channel rolled piece in the target analysis image, perform curve fitting on the pixel points corresponding to these grayscale minimum values, and obtain the lower fitted edge of the rolled piece.

4. The method for extrusion integral molding of a high-strength alloy template according to claim 1, characterized in that: Segmenting a rolling piece entrance area image and a rolling piece exit area image from the rolling piece rolling grayscale image, including: Segmenting the rolled product rolling grayscale image into a rolled product inlet side image and a rolled product outlet side image; The image on the inlet side of the rolled piece and the image on the outlet side of the rolled piece are both regarded as an object image, and according to the grayscale change of the pixel points in the object image, the liquid sliding deviation of each edge pixel point on the upper edge line of the rolled piece in the object image is determined; Determining a contact point on the upper edge line of the rolled piece in the object image according to the liquid sliding deviation; According to the contact points on the upper edge line of the rolled piece in the image of the rolled piece entrance side, the rolled piece entrance area image is segmented from the image of the rolled piece entrance side, and according to the contact points on the upper edge line of the rolled piece in the image of the rolled piece exit side, the rolled piece exit area image is segmented from the image of the rolled piece exit side.

5. The method for one-piece extrusion molding of a high-strength alloy template according to claim 4, characterized in that: Determining the liquid sliding deviation of each edge pixel point on the upper edge line of the rolled piece in the object image includes: Determine a pixel point sequence corresponding to each edge pixel point on the upper edge line of the rolled piece in the object image, and determine a grayscale mean and a grayscale standard deviation of all the pixel points in the pixel point sequence; The liquid slippage deviation of each edge pixel point on the upper edge line of the rolled piece in the object image is determined based on the grayscale mean difference between each edge pixel point and its previous edge pixel point, the grayscale mean difference between each edge pixel point and its subsequent edge pixel point, and the grayscale mean and grayscale standard deviation of the subsequent edge pixel point.

6. The method for one-piece extrusion molding of a high-strength alloy template according to claim 4, characterized in that: Determining the contact point on the upper edge line of the rolled piece in the object image comprises: Determine the liquid slippage deviation slope of each edge pixel point on the upper edge line of the rolled piece in the object image according to the fluctuation change of the liquid slippage deviation of each edge pixel point on the upper edge line of the rolled piece in the object image; On the upper edge line of the rolled piece in the object image, the edge pixel point corresponding to the maximum value of the liquid sliding deviation slope is determined as the contact point on the upper edge line of the rolled piece in the object image.

7. The method for extrusion integral molding of a high-strength alloy template according to claim 1, characterized in that: The step of determining low-brightness pixel points in the rolled product rolling grayscale image comprises: Determine the target analysis area corresponding to the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece, and the target analysis area corresponding to the lower edge line of the rolled piece and the lower edge line of the dark channel rolled piece in the target analysis image; Determine the grayscale histogram of all pixels in each target analysis area, and determine the average grayscale level in the grayscale histogram; determine the low-brightness pixels in the rolled product rolling grayscale image based on the pixels corresponding to all grayscale levels less than the average grayscale level in the grayscale histogram.

8. An extrusion integrated molding device for high-strength alloy templates, characterized in that: include: The rolling piece edge line acquisition module is used to: acquire the rolling piece grayscale image, perform edge detection on the rolling piece grayscale image, and obtain the upper edge line and the lower edge line of the rolling piece; A dark channel edge line acquisition module is used to: segment the roll entry area image and the roll exit area image from the roll rolling grayscale image, use the roll entry area image and the roll exit area image as a target analysis image, perform dark channel processing on the target analysis image, and obtain a dark channel roll upper edge line and a dark channel roll lower edge line; A fitting edge line acquisition module is used to determine the upper fitting edge of the rolled piece and the lower fitting edge of the rolled piece by combining the pixel grayscale changes between the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece in the target analysis image, and the pixel grayscale changes between the lower edge line of the rolled piece and the lower edge line of the dark channel rolled piece; An image factor acquisition module is used to: determine the image virtualization factor by combining the grayscale difference between the upper edge line of the rolled piece, the upper edge line of the dark channel rolled piece and the upper fitted edge of the rolled piece in the rolled piece entrance area image and the rolled piece exit area image, and determine the image random factor by combining the grayscale change of the pixel points and the pixel grayscale of the lower fitted edge of the rolled piece in the rolled piece entrance area image and the rolled piece exit area image; An enhancement module, used to: determine an enhancement coefficient according to the image virtualization factor and the image random factor, and enhance low-brightness pixel points in the rolled product rolling grayscale image according to the enhancement coefficient, so as to obtain an enhanced rolled product rolling grayscale image; A control module, used to: perform extrusion integrated molding control according to the rolled piece rolling grayscale image after the enhancement process; Determine the image virtualization factors, including: Determine the average grayscale value of the corresponding pixel points between the upper edge line of the rolled piece and the upper edge line of the dark channel rolled piece in the target analysis image, and determine the grayscale difference between the average grayscale value and the grayscale value of the corresponding pixel points on the fitted edge of the rolled piece; Determine the blur factor corresponding to the target analysis image by combining the overall distribution level of all the grayscale differences; Determine an image virtualization factor by combining the virtualization factors corresponding to the image of the rolling piece entrance area and the image of the rolling piece exit area, wherein the image virtualization factor is a normalized value; Determine the blur factor corresponding to the target analysis image. The corresponding calculation formula is: ; in, Indicates the blur factor corresponding to the target analysis image; Represents the gray value of the pixel point located on the upper edge line of the rolled piece in the sth column of the target analysis image; Represents the gray value of the pixel point located on the upper edge line of the dark channel rolled piece in the sth column of the target analysis image; Represents the gray value of the pixel point located on the fitting edge of the rolled piece in the sth column of the target analysis image; Indicates the number of columns of pixels in the target analysis image; Determine the image random factors, including: Normalizing the liquid slippage deviation of each edge pixel point on the upper edge line of the rolled piece in the target analysis image to obtain a first normalized value; Normalizing the grayscale value of each edge pixel point on the lower fitting edge of the rolled piece in the target analysis image to obtain a second normalized value; Determine a sub-random factor according to a difference between the first normalized value of each edge pixel point on the upper edge line of the rolled piece in the target analysis image and the second normalized value of a corresponding edge pixel point on the lower fitting edge of the rolled piece in the target analysis image; Determine an image random factor by synthesizing the sub-random factors corresponding to the image of the rolling piece entrance area and the image of the rolling piece exit area, wherein the image random factor is a normalized value; Determine the sub-random factors, including: Calculating a mean square error based on a difference between a first normalized value of an edge pixel point on an upper edge line of a rolled piece located in the same column in the target analysis image and a second normalized value of an edge pixel point on a lower fitting edge of the rolled piece; The difference obtained by subtracting the mean square error from 1 is the sub-random factor of the target analysis image.