Method for identifying and tracking materials in forging process

Abstract

The invention relates to a method for identifying and tracking materials in a forging process, which comprises the following steps of: S1, after the materials enter a factory, putting the materials into a storage area through a travelling crane; s2, furnace number information corresponding to each material is obtained; s3, real-time position information and process information of each material are determined through the target detection model and the motion trail of the travelling crane, and furnace number information of the materials is bound with the real-time position information and the process information; s4, detecting whether the object hoisted by the crane is a material or not through the target detection model; if the target detection model detects that the object hoisted by the crane is a material; if yes, the materials are positioned in real time according to the motion trail of the crane in the running process of the crane, and the real-time position information of the materials is updated in real time; and meanwhile, the process experienced by the material is identified through the target detection model, and the process information of the material is updated in real time. The materials are tracked through the target detection model and the motion trail of the travelling crane, and the materials on the production site are identified and tracked in real time.

Description

Technical Field

[0001] The invention relates to the technical field of forging material identification and tracking, and in particular to a material identification and tracking method in a forging process. Background Art

[0002] In the forging production process, the material is first loaded into the heating furnace by the overhead crane for heating. After heating, the material is transferred to the forging machine for forging. When the temperature of the forging drops below the process requirements, it needs to be put back into the heating furnace for secondary heating, and the above operations are repeated until the forging reaches the shape and size required by the process. After forging, heat treatment is performed according to the process requirements.

[0003] The traditional rolling process has many online and continuous processes, and the material identification and tracking of the continuous process can be completed by fixed sensors. The forging process is mainly discrete and non-continuous processes, and the discrete process has the characteristics of complexity, diversity and uncertainty of the process flow (quality factors will also affect the processing flow), that is, the material does not flow according to a fixed process, and there is a situation of repeated flow between the processes, which greatly increases the difficulty of material tracking and makes it difficult to track and identify the material. At the same time, in order to increase the efficiency of the heating furnace and the heat treatment furnace, there are multiple layers of materials stacked in the furnace, making the materials more difficult to identify and track.

[0004] Therefore, there is an urgent need for a technology that can identify and track materials in the forging process in real time. Summary of the invention

[0005] In order to identify and track materials in a forging process, the present application provides a method for identifying and tracking materials in a forging process.

[0006] The material identification and tracking method in the forging process provided in this application adopts the following technical solution:

[0007] A method for identifying and tracking materials during a forging process, the method comprising:

[0008] S1. After the materials enter the factory, the materials are placed in the storage area by driving;

[0009] S2. Obtain the furnace number information corresponding to each of the materials and enter it into a database; the furnace number information corresponds to the materials one by one;

[0010] S3. Determine the real-time location information and process information of each of the materials through the target detection model and the motion trajectory of the vehicle, and bind the furnace number information of the material with the real-time location information and the process information;

[0011] S4, when the material is hoisted from the storage area to any process by the crane, when the material is transferred between two processes by the crane, or when the material is hoisted back to the storage area by the crane, the object detection model is used to detect whether the object hoisted by the crane is the material;

[0012] If the target detection model detects that the object hoisted by the crane is the material, the material is located in real time according to the movement trajectory of the crane during the operation of the crane, and the real-time position information of the material is updated in real time; at the same time, the process that the material has gone through is identified through the target detection model, and the process information of the material is updated in real time;

[0013] If the target detection model detects that the object being hoisted by the crane is not the material, the real-time location information and the process information remain unchanged.

[0014] Optionally, in step S2, each of the materials has a unique identification code; the identification code includes the furnace number information corresponding to the material; and the furnace number information corresponding to the material can be obtained by scanning the identification code.

[0015] Optionally, in steps S3 and S4, the target detection model includes an intelligent analysis host and multiple visual cameras; the intelligent analysis host is communicatively connected to each of the visual cameras; the visual cameras are arranged at the production site according to different shooting angles, so as to capture images of the production site; the intelligent host processes the images of the production site and identifies the materials hoisted on the crane.

[0016] Optionally, the intelligent host processes the image of the production site and identifies the material hoisted on the crane, including: performing a bird's-eye view transformation on the image of the production site; the bird's-eye view transformation uniformly transforms the image of the production site captured by the visual camera at different angles into a bird's-eye view.

[0017] Optionally, the intelligent host processes the image of the production site and identifies the material hoisted on the crane, and further includes: performing target detection on the image of the production site after the bird's-eye view transformation; the target detection includes training a target detection model through deep learning; identifying the material in the bird's-eye view through the target detection model; the target detection model is used to detect and locate the overall contour of the material.

[0018] Optionally, the bird's-eye view generated by the images of the production site captured by each of the visual cameras is aggregated to form a global bird's-eye view; the global bird's-eye view is used to display the position of each of the materials in the production site.

[0019] Optionally, after the target detection model detects the contours of each of the materials, the contours of the materials detected in each of the visual cameras are projected into the global bird's-eye view in combination with the angle of the visual camera.

[0020] Optionally, after the target detection model detects the outline of the material, it identifies the key features of the outline; the key features do not change with the changes in the appearance of the material, so that the target detection model can still identify and track the material after the appearance of the material changes after undergoing a certain process.

[0021] Optionally, in steps S3 and S4, the motion trajectory of the crane is obtained by a first sensor, a second sensor, a first encoder, a second encoder and a calculation module; the first sensor, the second sensor, the first encoder and the second encoder are all communicatively connected to the calculation module; the first sensor is installed on the crane trolley, and is used to collect displacement data of the crane trolley in the X-axis and Y-axis directions; the second sensor is installed on the trolley of the crane, and is used to collect displacement data of the trolley of the crane in the X-axis and Y-axis directions; the first encoder is installed on the crane trolley, and is used to collect displacement data of the crane trolley in the Z-axis direction; the second encoder is installed on the trolley of the crane, and is used to collect displacement data of the trolley of the crane in the Z-axis direction; the calculation module is used to calculate the motion trajectory of the crane according to the displacement data of the crane trolley in the X-axis, Y-axis and Z-axis directions, and the displacement data of the trolley of the crane in the X-axis, Y-axis and Z-axis directions; after the target detection model detects the material hoisted by the crane, the real-time position information of the material can be obtained according to the motion trajectory of the crane.

[0022] Optionally, in step S1, the storage area is divided into multiple partitions; each partition is used to store one type of material; multiple stacks are set in each partition; a limiting portion for placing the material is set in the stack; the shape of the limiting portion is adapted to the shape of the material to be placed therein, so as to limit the material placed therein.

[0023] As described above, the present invention has at least the following beneficial effects:

[0024] The material identification and tracking method in the forging process of the present application identifies the material through the target detection model, and uses the target detection model as an aid to track the material in combination with the motion trajectory of the vehicle, and can automatically identify and track each material on the production site in real time. Compared with the existing manual methods of identifying and tracking materials, it is convenient to trace the quality of materials in the forging process, improve production efficiency and product quality, and reduce production costs, thereby meeting the needs of modern manufacturing for efficient and high-quality production. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a flow chart of an embodiment of the present invention for reflecting a method for identifying and tracking materials during a forging process. DETAILED DESCRIPTION

[0026] The following is an explanation of the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the features in the following embodiments and the embodiments can be combined with each other without conflict. It should also be understood that the terms used in the embodiments of the present invention are intended to describe specific embodiments, rather than to limit the scope of protection of the present invention. The test methods for which specific conditions are not specified in the following examples are usually carried out under conventional conditions or according to the conditions recommended by the manufacturers.

[0027] It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not used to limit the limiting conditions for the implementation of the present invention, so they have no substantial technical significance. Any modification of the structure, change in the proportional relationship, or adjustment of the size should still fall within the scope of the technical content disclosed by the present invention without affecting the effects and purposes that can be achieved by the present invention. At the same time, the terms such as "upper", "lower", "left", "right", "middle" and "one" quoted in this specification are only for the convenience of description, and are not used to limit the scope of the implementation of the present invention. The change or adjustment of their relative relationship should also be regarded as the scope of the implementation of the present invention without substantially changing the technical content.

[0028] Please refer to Figure 1 The present application first discloses a method for identifying and tracking materials during a forging process. The method comprises the following steps:

[0029] S1. After the materials enter the factory, they are placed in the storage area by driving;

[0030] S2. Obtain the furnace number information corresponding to each material and enter it into the database; the furnace number information corresponds to the material one by one;

[0031] S3. Determine the real-time location information and process information of each material through the target detection model and the motion trajectory of the vehicle, and bind the furnace number information of the material with the real-time location information and process information;

[0032] S4: When materials are hoisted from the storage area to any process by a crane, when materials are transferred between two processes by a crane, or when materials are hoisted back to the storage area by a crane, the object detection model is used to detect whether the objects hoisted by the crane are materials;

[0033] If the target detection model detects that the object being hoisted by the crane is a material, the material is located in real time according to the movement trajectory of the crane during the operation of the crane, and the real-time location information of the material is updated in real time. At the same time, the target detection model is used to identify the process that the material has gone through, and the process information of the material is updated in real time.

[0034] If the target detection model detects that the object being hoisted by the crane is not a material, the real-time location information and process information will not change.

[0035] In the forging process, the main material used is steel ingot, so the implementation principle of this application is explained below using steel ingot as an example:

[0036] The furnace number information is the information given to the steel ingot after it is produced, and each steel ingot has a different furnace number information. The real-time location information represents the real-time location of the steel ingot in the production site. The process information represents the processes that the steel ingot has gone through, such as upsetting, drawing, punching, cutting, etc. After the furnace number information, the real-time location information corresponding to the furnace number information and the process information are determined, the identification and tracking of materials in the forging process are realized.

[0037] For step S2, each material has a unique identification code. The identification code contains the furnace number information corresponding to the material. By scanning the identification code, the furnace number information corresponding to the material can be obtained.

[0038] Specifically, the unique identification code can be a barcode, a QR code, etc. By scanning the identification code, the furnace number information corresponding to each steel ingot can be identified, so that the furnace number information can be uploaded to the database. The method of identifying the identification code can be that after the steel ingot is put into storage, the staff can hold a barcode scanner connected to the database to scan the identification code of the steel ingot one by one, and upload the corresponding furnace number information to the database.

[0039] For steps S3 and S4, at the production site, the objects hoisted by the crane are not only steel ingots, but sometimes other objects such as materials and tools need to be hoisted. It is not very meaningful to identify and track other objects such as materials and tools. Therefore, when hoisting objects by the crane, it is first necessary to identify whether the object hoisted by the crane is a steel ingot through the target detection model. If it is a steel ingot, the steel ingot is tracked. If it is not a steel ingot, it is not necessary to track the object through step S4.

[0040] The target detection model includes an intelligent analysis host and multiple visual cameras. The intelligent analysis host is connected to each visual camera. The visual cameras are arranged at the production site according to different shooting angles to capture images of the production site. The intelligent host processes the images of the production site and identifies the materials hoisted on the crane.

[0041] Specifically, before identifying and tracking the steel ingots, it is necessary to first arrange the visual cameras. The arrangement of the visual cameras requires that all the visual cameras can be combined to capture the images of the production site without blind spots. Therefore, the arrangement of the visual cameras not only needs to have different shooting angles in the horizontal direction, but also in the vertical direction. After the visual camera captures the image of the production site, the intelligent host processes the image of the production site and identifies the steel ingots.

[0042] The intelligent host processes the image of the production site and identifies the materials hoisted on the crane, including: performing a bird's-eye view transformation on the image of the production site and performing target detection on the image of the production site after the bird's-eye view transformation.

[0043] Bird's-eye view transformation transforms the images of the production site captured by visual cameras at different angles into a unified bird's-eye view. Object detection includes training the object detection model through deep learning. The object detection model is used to identify the materials in the bird's-eye view. The object detection model is used to detect and locate the overall contour of the material.

[0044] Specifically, since the shooting angles of the images of the production site captured by each visual camera are inconsistent, it is difficult to train and recognize these images directly for the target detection model, resulting in low recognition efficiency. Therefore, in order to improve the efficiency of target detection model training and recognition, it is necessary to pre-process the images of the production site captured by each visual camera. In this embodiment, after each visual camera captures the images of the production site, these images are first transformed into a bird's-eye view through an algorithm.

[0045] After the bird's-eye view transformation, the images of the production site captured by each visual camera are converted to the same perspective, which facilitates the target detection model to extract, segment and match features from the image, thereby improving the efficiency of target detection model training and recognition.

[0046] In some other embodiments of the present application, the image can be denoised and smoothed by median filtering, mean filtering, Gaussian filtering, etc., or the image can be sharpened by Laplace filtering, Sobel operator, Canny operator, etc., and the image can be further pre-processed to further improve the efficiency of target detection model training and recognition.

[0047] In this embodiment, the target detection model uses the YOLO model (other models such as Vision Transformer can also be used) for target recognition, and the SORT algorithm is used for target tracking. When training the target detection model, a certain number of pre-processed images of the production site are selected as samples and input into the target detection model. The images of the production site can also be subjected to operations such as mirror transformation, random cropping, and random rotation, so as to expand the scale of the sample, increase the diversity of the training data, and thus improve the generalization ability of the target detection model.

[0048] Through the target detection model, the steel ingots in the images of the production site captured by the visual camera can be identified and tracked continuously.

[0049] More specifically, at the same time, the images of the production site captured by each visual camera are transformed into a bird's-eye view, and the images after the bird's-eye view transformation are summarized to obtain a global bird's-eye view. After the target detection model detects the contours of each ingot, it projects the contours of the ingots detected by each visual camera into the global bird's-eye view in combination with the calibrated camera parameters (camera angle), and uses a video tracking algorithm to associate each frame of the video taken by the visual camera with the contour information and image features of the ingots detected by the target detection model in the previous and next frames in real time, so that the real-time position information and process information corresponding to each ingot in the database can be updated in real time.

[0050] When the steel ingot undergoes certain processes at the production site, its shape will change. In order to prevent the target detection model from being affected by the change in the shape of the steel ingot, the target detection model detects the outline of the steel ingot and identifies the key features of the outline. The key features do not change with the change in the shape of the steel ingot, so that the target detection model can still identify and track the steel ingot after the shape of the steel ingot changes after a certain process. On this basis, in order to more accurately identify the steel ingots at the production site, a visual camera is installed at the entrance and exit of the equipment corresponding to each process.

[0051] In steps S3 and S4, the motion trajectory of the vehicle is obtained by the first sensor, the second sensor, the first encoder, the second encoder and the calculation module. The first sensor, the second sensor, the first encoder and the second encoder are all connected to the calculation module block for communication. The first sensor is installed on the vehicle's trolley for collecting displacement data in the X-axis and Y-axis directions of the vehicle's trolley. The second sensor is installed on the trolley of the vehicle for collecting displacement data in the X-axis and Y-axis directions of the trolley of the vehicle. The first encoder is installed on the vehicle's trolley for collecting displacement data in the Z-axis direction of the vehicle's trolley. The second encoder is installed on the trolley of the vehicle for collecting displacement data in the Z-axis direction of the trolley of the vehicle. The calculation module is used to calculate the motion trajectory of the vehicle based on the displacement data in the X-axis, Y-axis and Z-axis directions of the vehicle's trolley and the displacement data in the X-axis, Y-axis and Z-axis directions of the trolley of the vehicle. After the target detection model detects the material hoisted by the vehicle, the real-time position information of the material can be obtained according to the motion trajectory of the vehicle.

[0052] The first sensor and the second sensor can be Graham wires, which have the advantages of long measuring range and strong environmental adaptability. The first sensor and the second sensor can also be laser displacement sensors and absolute encoders, which can be selected according to the actual needs of the production site.

[0053] Specifically, the crane needs to move according to the transfer operation instruction, and the transfer operation instruction is issued by the control system of the production site. For example, when the steel ingot with furnace number information a completes process A, it needs to be lifted from process A to process B to carry out process B. At this time, the control system will issue a transfer operation instruction to the crane: lift the steel ingot with furnace number information a from process A to process B. Therefore, the starting point and end point of the crane's motion trajectory can be obtained according to the transfer operation instruction, and then combined with the displacement data of the X-axis, Y-axis, and Z-axis directions of the large car, and the displacement data of the X-axis, Y-axis, and Z-axis directions of the small car, and a certain point on the production site is used as the origin, the vertical direction is used as the Z-axis, and the two mutually perpendicular straight lines in the horizontal plane are used as the X-axis and Y-axis to establish a spatial rectangular coordinate system, the motion trajectory of the crane can be accurately calculated, and then supplemented by the recognition and tracking of the target detection model, each steel ingot on the production site can be recognized and tracked.

[0054] In step S1, the storage area is divided into a plurality of partitions. Each partition is used to store a material. A plurality of stacks are set in each partition. A limiting portion for placing materials is set in the stack. The shape of the limiting portion is adapted to the shape of the material to be placed therein, so as to limit the material placed therein.

[0055] After the storage area is divided into multiple partitions, the steel ingots can be classified and managed according to their types and sizes. And because each partition is provided with a stack and a limiter is provided in the stack, the steel ingot will not be displaced after being placed in the stack. After the steel ingot is hoisted to the stack by the crane, the real-time position information of the steel ingot has been determined based on the motion trajectory of the crane, the recognition and tracking of the target detection model. If the steel ingot rolls during storage, its real-time position information has not changed, but the actual position has shifted. When the overhead crane hoists the steel ingot again, accidents are prone to occur due to inaccurate real-time position information. The limiter in the stack can effectively prevent the position of the steel ingot from shifting during storage.

[0056] The material identification and tracking method in the forging process of the present application identifies the material through the target detection model, and uses the target detection model as an aid to track the material in combination with the motion trajectory of the vehicle, and can automatically identify and track each material on the production site in real time. Compared with the existing manual methods of identifying and tracking materials, it is convenient to trace the quality of materials in the forging process, improve production efficiency and product quality, and reduce production costs, thereby meeting the needs of modern manufacturing for efficient and high-quality production.

[0057] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A method for identifying and tracking materials during a forging process, characterized in that: The method comprises the following steps: S1. After the materials enter the factory, the materials are placed in the storage area by driving; S2. Obtain the furnace number information corresponding to each of the materials and enter it into a database; the furnace number information corresponds to the materials one by one; S3. Determine the real-time location information and process information of each of the materials through the target detection model and the motion trajectory of the vehicle, and bind the furnace number information of the material with the real-time location information and the process information; S4, when the material is hoisted from the storage area to any process by the crane, when the material is transferred between two processes by the crane, or when the material is hoisted back to the storage area by the crane, the object detection model is used to detect whether the object hoisted by the crane is the material; If the target detection model detects that the object hoisted by the crane is the material, the material is located in real time according to the movement trajectory of the crane during the operation of the crane, and the real-time position information of the material is updated in real time; at the same time, the process that the material has gone through is identified through the target detection model, and the process information of the material is updated in real time; If the target detection model detects that the object being hoisted by the crane is not the material, the real-time location information and the process information remain unchanged.

2. The method according to claim 1, characterized in that In the step S2, Each of the materials has a unique identification code; the identification code contains the furnace number information corresponding to the material; By scanning the identification code, the furnace number information corresponding to the material can be obtained.

3. The method according to claim 1, characterized in that In the steps S3 and S4, The target detection model includes an intelligent analysis host and a plurality of visual cameras; the intelligent analysis host is communicatively connected with each of the visual cameras; The visual cameras are arranged at the production site according to different shooting angles to capture images of the production site; The intelligent host processes the images of the production site and identifies the materials hoisted on the crane.

4. The method according to claim 3, characterized in that The intelligent host processes the image of the production site and identifies the material hoisted on the crane, including: performing a bird's-eye view conversion on the image of the production site; The bird's-eye view transformation uniformly transforms the images of the production site captured by the visual camera at different angles into a bird's-eye view.

5. The method according to claim 4, characterized in that The intelligent host processes the image of the production site and identifies the material hoisted on the crane, further comprising: performing target detection on the image of the production site after the bird's-eye view transformation; The target detection includes training a target detection model through deep learning; identifying the material in the bird's-eye view through the target detection model; and the target detection model is used to detect and locate the overall contour of the material.

6. The method according to claim 4, characterized in that At the same time, the bird's-eye view generated by the images of the production site captured by each of the visual cameras is aggregated to form a global bird's-eye view; The global bird's-eye view is used to display the location of each material in the production site.

7. The method according to claim 6, characterized in that After the target detection model detects the contours of each of the materials, the contours of the materials detected by each of the visual cameras are projected into the global bird's-eye view in combination with the angle of the visual camera.

8. The method according to claim 7, characterized in that After the target detection model detects the outline of the material, it identifies the key features of the outline; the key features do not change with the changes in the shape of the material, so that the target detection model can still identify and track the material after the shape of the material changes after undergoing a certain process.

9. The method according to claim 3, characterized in that: In the steps S3 and S4, Obtaining the motion trajectory of the vehicle through a first sensor, a second sensor, a first encoder, a second encoder and a calculation module; The first sensor, the second sensor, the first encoder and the second encoder are all communicatively connected to the computing module block; The first sensor is installed on the trolley of the traveling vehicle, and is used to collect displacement data of the trolley of the traveling vehicle in the X-axis and Y-axis directions; the second sensor is installed on the trolley of the traveling vehicle, and is used to collect displacement data of the trolley of the traveling vehicle in the X-axis and Y-axis directions; The first encoder is installed on the trolley of the traveling crane, and is used to collect displacement data of the trolley of the traveling crane in the Z-axis direction; the second encoder is installed on the trolley of the traveling crane, and is used to collect displacement data of the trolley of the traveling crane in the Z-axis direction; The calculation module is used to calculate the motion trajectory of the vehicle according to the displacement data of the vehicle's trolley in the X-axis, Y-axis and Z-axis directions and the displacement data of the vehicle's trolley in the X-axis, Y-axis and Z-axis directions; After the target detection model detects the material hoisted by the crane, the real-time position information of the material can be obtained according to the motion trajectory of the crane.

10. The method according to claim 1, characterized in that In step S1, the storage area is divided into a plurality of partitions; each partition is used to store one type of material; A plurality of stacking positions are arranged in each partition; a limiting portion for placing the material is arranged in the stacking position; the shape of the limiting portion is adapted to the shape of the material to be placed therein, so as to limit the material placed therein.

Images