A method for billet face recognition

Abstract

The present application relates to a kind of billet face detection method, belong to the detection field of billet state.The six faces of the billet are divided into four turnover faces and two end faces, including opposite two rolling faces in four turnover faces, four long edges of four turnover faces are respectively processed or shaped into different edge shape of folding edge, at least one of four top corners in one end face of billet is photographed in set photographing orientation, the shape of top corner in photo is analyzed to match the long edge corresponding to the top corner of photographing in the photographing orientation, the position of four turnover faces in current photographing orientation is determined according to this, and the face of billet is recognized.The face recognition efficiency is improved, and the face recognition accuracy of billet is guaranteed, and the combination of face recognition and turnover is applied to pre-rolling heating, significantly improves the product quality of rolled material.

Description

Technical Field

[0001] This invention relates to an automatic surface recognition detection method for billets. Background Technology

[0002] Due to the hydrostatic pressure of molten steel, the original aspect ratio of the continuously cast billet frame can be as high as 1.5, leading to problems such as frame segregation, which cannot be eliminated by ordinary rolling. During slow cooling, transportation, and hoisting, the billet may be flipped, causing the rolling surface of the billet to be on the side instead of the top surface when it enters the rolling mill. The face-checking rolling process requires that the billet be face-checked before entering the heating furnace, and the billet be flipped according to the face-checking result, with the rolling surface facing upwards, before being pushed into the heating furnace for heating, and then entering the rolling process. In the process of identifying and flipping the billets, most of the flipping was originally done in the billet warehouse using the most primitive manual identification method. Due to the hot working environment and strong heat radiation, the labor intensity of personnel working on-site for a long time was high and there were corresponding safety hazards. Secondly, the manual identification and flipping of billets was subject to some human factors, such as human absence, human error, and human misjudgment. In particular, the identification and flipping effect was not good during the night shift, resulting in unstable quality of the rolled billets from continuous casting.

[0003] Therefore, it is necessary to invent an automatic billet face recognition and detection method in front of the furnace to solve this problem, replace manual face recognition, realize the automation of the production line, improve the operating rate, and stabilize the quality of rolled products.

[0004] Currently, there are few control methods for billet turning. Patent publication number CN115673272A discloses a control method and system for turning billets separately in a continuous casting machine. By designing a control unit to determine the casting stream to be turned according to preset rules, it can realize the function of turning billets from each casting stream separately. Only after the casting stream to be turned is determined will the turning unit at that casting stream perform the turning operation; turning units at other casting streams can only continue to wait, preventing simultaneous turning. The method for determining the casting stream to be turned: each casting stream is marked with a corresponding ID value. Determining the casting stream to be turned according to preset rules includes: acquiring the ID value corresponding to each casting stream; timing according to a preset interval and outputting an intermediate variable signal; accumulating the intermediate variable signal and recording the accumulated value; comparing the accumulated value with the most recently acquired ID value corresponding to the casting stream; if the accumulated value is the same as the ID value corresponding to the casting stream, then the casting stream is determined as the casting stream to be turned. Summary of the Invention

[0005] The purpose of this invention is to provide an automatic billet face detection method, which can perform face detection of billets before they enter the heating furnace to determine whether the rolling surface is the upper surface. When the rolling surface is detected as the side, the billet is flipped using an existing billet turning machine. After flipping 90°, the rolling surface is turned upwards. This application solves the problems of high labor intensity and corresponding safety hazards associated with manual billet face detection, while improving work efficiency and ensuring the accuracy of billet face detection and turning. It ensures that the billet is heated and rolled with the rolling surface facing upwards, thereby improving the product rolling quality of the rolled material.

[0006] The technical solution of this application:

[0007] A method for identifying the surface of a billet involves dividing the six sides of the billet into four flipped surfaces and two end surfaces. The four flipped surfaces include two opposite rolled surfaces. The four long sides of the four flipped surfaces are respectively processed or formed into different folded edge shapes. At least one of the four apex corners of one end surface of the billet is photographed at a set photographing position. The shape of the apex corner in the photograph is analyzed to match the long side corresponding to the photographed apex corner at that photographing position. Based on this, the position of the four flipped surfaces at the current photographing position is determined, thus completing the billet surface identification.

[0008] As one of the preferred applications of this application, the billet is a continuously cast billet. The four inner corners of the copper tube in the crystallizer of the continuous casting equipment have the same shape as the folded edge shape corresponding to the four long sides, thus directly obtaining a billet with a specific folded edge shape. No additional processing of the billet is required.

[0009] As one of the preferred embodiments of this application, the four long sides of the four flipped surfaces are respectively processed or formed into arc surfaces, and the arc surfaces of adjacent sides have different central angles. Compared with other folded edge shapes, arc surfaces are more conducive to forming during the continuous casting process of square billets.

[0010] As one of the preferred embodiments of this application, two cameras are used to take pictures of the two apex corners of one end face. The shape of the two photographed apex corners is analyzed to match the two long sides corresponding to the two photographed apex corners at that photographing position. Based on this, the positions of the four flipped faces at the current photographing position are determined, thus completing the billet face identification. Face identification using the two apex corners of the end face can significantly improve the accuracy of the face identification results compared to relying on a single apex corner.

[0011] As one of the preferred embodiments of this application, the folded edge shapes of the four long sides of the four flipped surfaces satisfy the condition that the opposite angles are the same and the adjacent angles are different. That is, the four vertices of the square billet end face also satisfy the condition that the opposite angles are the same and the adjacent angles are different. This is beneficial for designing adjacent edge folded edge shapes with significant differences, thereby facilitating identification and matching.

[0012] As one of the preferred embodiments of this application, two cameras are used to take pictures of two adjacent apex corners of one end face, and the shape of the two photographed apex corners is analyzed to match the two long sides corresponding to the two photographed apex corners under the photographing position. Based on this, the positions of the four flipped faces under the current photographing position are determined, and the billet face recognition is completed.

[0013] The above method is applied to the control unit to realize the automatic control of the billet face detection method. The control unit includes a camera shooting module, an image processing module, a data communication module and a monitoring and alarm module. The camera shooting module is used to take pictures of the top corner of the billet end face at a set position, and multiple images are obtained for one top corner.

[0014] The image processing module is used to analyze multiple images to obtain the average value of the feature parameters of the apex shape, and use the average value of the feature parameters of the apex shape to match the actual value of the feature parameters of the four apex shapes of the end face. If the difference between the two is within the error allowable range, the matching is completed, the identification of the photographed apex and the actual apex of the end face is completed, and the position of the four flipped surfaces is determined.

[0015] The data communication module is used to receive the billet arrival trigger signal to trigger the face recognition operation, and to send the face recognition result signal.

[0016] In one preferred embodiment of this application, the data communication module is signal-connected to the billet turning device. Based on the face recognition result signal, the billet turning device receives a trigger signal to turn the billet over to adjust the positions of the four turning faces. The aforementioned billet turning equipment is an existing device, connected to the data communication module via its control box. The control box is triggered to turn the current billet, and the data communication module transmits the face recognition result signal to the control box. The control box then generates a billet turning trigger signal based on the face recognition result signal. Alternatively, the data communication module can directly generate a billet turning trigger signal based on the face recognition result and send it to the control box of the billet turning equipment.

[0017] As one of the preferred embodiments of this application, the four sides are designed as arc surfaces, and the apex shape of the corresponding end face is an arc. When identifying the apex, the central angle is used as a feature parameter.

[0018] Compared with the prior art, the advantages of the present invention are: it replaces manual identification of the billet surface, improves identification efficiency and ensures the accuracy of billet identification, and the combination of identification and flipping is applied to pre-rolling heating, which can ensure that the billet is heated and rolled in the desired rolling surface position, thereby significantly improving the product quality of the rolled material. Attached Figure Description

[0019] Figure 1 This is a schematic diagram illustrating the principle of the automatic surface recognition detection method for billets in an embodiment of the present invention;

[0020] Figure 2This is a flowchart illustrating the automatic surface recognition detection method for billets in an embodiment of the present invention;

[0021] Figure 3 This is a schematic diagram of the end face of a billet and a schematic diagram of the detection position in the automatic billet recognition detection method of this invention.

[0022] Figure 4 This is a top view schematic diagram of the copper tube of the continuous casting tooling crystallizer in an embodiment of the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] This embodiment provides an automatic surface recognition and detection device and method for billets in front of the furnace, and the specific embodiment is as follows:

[0025] Based on the process requirements of billet face recognition rolling and the layout of equipment on the steel rolling production line, an automatic billet face recognition detection device and method in front of the furnace was designed.

[0026] See Figure 1 As shown in the schematic diagram, the automatic billet recognition detection device in front of the furnace in this embodiment consists of four parts: a camera shooting module 100, an image processing module 101, a data communication module 102, and a monitoring and alarm module 103.

[0027] See Figure 2 As shown in the flowchart, the automatic surface recognition detection method for billets in front of the furnace in this embodiment is completed in seven steps: receiving the billet incoming signal S200, acquiring the image to be detected S201, performing image enhancement and optimization S202, identifying the edge of the billet image S203, acquiring the average actual round angle S204, performing surface recognition detection judgment S205, and outputting the surface recognition detection result S206.

[0028] See Figure 3 As shown in the billet end face diagram and detection position diagram, in this embodiment, the four apex corners of the billet end face are rounded corners with different central angles. The opposite angles of the rounded corners are equal, while the adjacent angles are not equal. The camera module 100 is respectively set at the upper left and upper right of the billet's material feeding direction. Camera 1 is used to detect the upper right rounded angle R1 of the billet end face, and camera 2 is used to detect the upper left rounded angle R2 of the billet end face.

[0029] See Figure 4The top view of the copper tube in the continuous casting fixture is shown in the schematic diagram. The inner angles of the selected copper tube in the continuous casting fixture are also R1 and R2, and the opposite angles are equal while the adjacent angles are not equal.

[0030] This embodiment adopts Figure 4 The four rounded corners of the square billet produced by the continuous casting tooling crystallizer copper tube are consistent with the shape of the inner corners of the copper tube of the continuous casting tooling crystallizer. A schematic diagram of the end face of the square billet to be inspected is shown below. Figure 3 As shown, the fillets on the end faces of the square billet are equal at opposite corners and unequal at adjacent corners, with the fillet angles being R1 and R2 respectively.

[0031] The camera shooting module 100 can select three different camera working modes based on different production processes through the image, including camera 1 working independently, camera 2 working independently, and camera 1 and camera 2 working simultaneously. This embodiment takes the mode of camera 1 and camera 2 working simultaneously as an example.

[0032] The camera module 100 is installed on both sides of the furnace inlet roller conveyor. It is equipped with two cameras, each mounted on a dedicated bracket, positioned on the upper left and upper right sides of the inlet roller conveyor respectively. (See attached image.) Figure 3 The detection location is shown in the schematic diagram. After the data communication module 102 receives the billet incoming signal S200, the camera imaging module 100 acquires the image to be detected S201, which is a multi-image collection including a 3D oblique view of the billet.

[0033] The image processing module 101 has the function of storing the image to be detected. It implements image enhancement and optimization S202 and identifies the edge of the billet image S203 through the algorithm, obtains the actual round angle of the end face of the billet in the image to be detected, calculates the average of the actual round angles of multiple billet end faces, and finally obtains the target data, which is the average actual round angle of the end face of the billet S204. Based on the average actual round angle of the billet end face and the target round angle of the billet end face, the billet is judged by face recognition S205. If the two data are equal or within the allowable error range, the face recognition result is correct; otherwise, it is incorrect. As shown in Figure (3), if the target round angle of the upper right of the billet end face of camera 1 is set to R2 and the target round angle of the upper left of the billet end face of camera 2 is set to R1, then when the average actual round angle of the upper right of the billet end face detected by camera 1 is R1 and the average actual round angle of the upper left of the billet end face detected by camera 2 is R2, the face recognition result signal is incorrect and the billet needs to be flipped 90°. Otherwise, it is correct and the billet does not need to be flipped. The billet can be directly put into the heating furnace.

[0034] The data communication module 102 adopts I / O communication, which is stable and reliable. It is used to receive the billet incoming signal S200 and output the surface recognition detection result S206, etc. The data communication module 102 can communicate with and organically integrate with existing steel turning equipment to realize the automatic surface recognition and turning function of billet before entering the furnace.

[0035] The monitoring and alarm module 103 features a visual human-machine interface, enabling functions such as camera working mode selection, historical record query, setting various parameters, abnormal alarms, and surface data display. Parameter categories include: target radius angle of the billet end face, allowable error range, and radius exceeding limits.

Claims

1. A method for inspecting the surface of a billet, characterized in that: The six faces of the billet are divided into four flip faces and two end faces. The four flip faces include two opposite rolling faces. The four long sides of the four flip faces are respectively processed into different folded edge shapes. At least one of the four apex corners of one end face of the billet is photographed from a set photographing position. The billet is a continuously cast billet. The shape of the four inner corners of the copper tube of the crystallizer of the continuous casting equipment is the same as the shape of the folded edge corresponding to the four long sides. The four long sides of the four flipped surfaces are respectively processed into arc surfaces, and the arc surfaces of adjacent sides have different central angles. Two cameras take pictures of two adjacent apex corners of one end face respectively. The shape of the two photographed apex corners is analyzed to match the two long sides corresponding to the two photographed apex corners under the photographing position. Based on this, the position of the four flipped surfaces under the current photographing position is determined, and the billet face recognition is completed.

2. The billet surface inspection method according to claim 1, characterized in that: The method is applied to a control unit, which includes a camera shooting module, an image processing module, a data communication module, and a monitoring and alarm module. The camera shooting module is used to take pictures of the top corner of the blank end face at a set position, and multiple images are obtained for each top corner at a given position. The image processing module is used to analyze multiple images to obtain the average value of the feature parameters of the apex shape, and use the average value of the feature parameters of the apex shape to match the actual value of the feature parameters of the four apex shapes of the end face. If the difference between the two is within the error allowable range, the matching is completed, the identification of the photographed apex and the actual apex of the end face is completed, and the position of the four flipped surfaces is determined. The data communication module is used to receive the billet arrival trigger signal to trigger the face recognition operation, and to send the face recognition result signal.

3. The billet surface inspection method according to claim 2, characterized in that: The data communication module is connected to the billet turning device. Based on the recognition result signal, the billet turning device receives a trigger signal to turn the billet over to adjust the position of the four turning surfaces.

4. The method for inspecting the surface of a billet according to claim 2, characterized in that: The apex of the end face is rounded, and the characteristic parameter is the central angle.

Images