A method for removing the transversely bundled steel strips of a steel coil

Abstract

A method for removing transversely bundled steel straps from steel coils includes: S1, placing the steel coil at the unbundling station using a lifting device, establishing a three-dimensional model using a vision scanning system, and placing the three-dimensional model within a pre-established three-dimensional spatial coordinate system to determine the precise location of the steel coil; S2, performing image recognition using an image processing module, classifying and defining the edge-enclosed area using pre-input region classification information, and fitting the recognized image with the three-dimensional model to determine the specific spatial coordinates of the steel coil, the bundled straps, and the bundled clips; this method intelligently determines the optimal unbundling position and automatically completes the unbundling operation, while simultaneously using a pneumatic pull-out clamp to remove the bundled straps, preventing the transverse bundled straps from remaining stuck inside the steel coil after unbundling and affecting subsequent operations, thereby improving unbundling efficiency and reducing working time.

Description

Technical Field

[0001] This invention relates to the field of steel coil leveling line technology, and more specifically to a method for removing the transversely bundled steel strips of a steel coil. Background Technology

[0002] As a traditional manufacturing industry, the steel industry is a pillar industry of the national economy. Steel coil unbundling, a crucial part of this industry, still suffers from severe reliance on manual labor, low automation, and low production efficiency. Currently, there are patents in China for intelligent unbundling in this field, but all suffer from immature design schemes, limited identification methods, and low actual identification accuracy. Furthermore, they only provide limited intelligence for a single process and have not truly solved the problem of low production efficiency, thus failing to achieve large-scale adoption and application.

[0003] After steel coils are finished being rolled, steel mills usually bundle them with steel straps for easy transportation. In addition to bundling the coils radially, they are also bundled laterally. This improves safety during transportation, as the coils will not unravel even if they fall from the transport vehicle. However, this makes it difficult to unwind the coils. The bundling steel straps need to be removed before unwinding. Traditional methods can remove the radial bundling steel straps, but the lateral bundling steel straps still need to be removed manually. Summary of the Invention

[0004] To address the aforementioned problems and overcome the shortcomings of existing technologies, this invention provides a method for removing the transversely bundled steel strips from steel coils.

[0005] To achieve the above objectives, the present invention provides a method for removing the transversely bundled steel strips of steel coils. The device used includes a visual scanning system, an image processing module, a laser displacement sensor, a debundling assembly, and a debundling assembly, which are respectively installed on the foundation on both sides of the debundling station.

[0006] The unbundling assembly includes a robotic arm and pneumatic scissors. The pneumatic scissors are mounted at the end of the robotic arm and can move freely in three-dimensional space.

[0007] The strapping assembly includes a robotic arm and a pneumatic strapping clamp, which is mounted at the end of the robotic arm and can move freely in three-dimensional space.

[0008] The visual scanning system includes at least two camera mechanisms, a supplementary lighting mechanism, and a control component for controlling the camera mechanisms to capture images. The control component controls the supplementary lighting mechanism to automatically turn on during image capture.

[0009] The visual scanning system also includes at least two laser scanning mechanisms. The laser scanning mechanisms acquire the three-dimensional coordinate data of the surface of the object being measured at the unpacking station, and then build and store a three-dimensional model in the form of point cloud data after denoising, registration and stitching.

[0010] The image processing module includes an image preprocessing unit, an image feature extraction unit, and an image analysis unit;

[0011] The image preprocessing unit receives multiple images acquired by the visual scanning system, compares the multiple images after image preprocessing, and selects the image with the most obvious image features as the local image.

[0012] The image feature extraction unit receives the local image, reduces image noise to achieve image smoothing, and then performs convolution operations on the image in the horizontal and vertical directions to obtain the gradient intensity and gradient direction of each pixel. For each pixel, based on its gradient direction, it checks whether it is a local maximum value of the gradient in that direction. If not, the pixel is suppressed. An edge threshold is set, and pixels with gradient intensity greater than the edge threshold are identified as edge pixels. The edge pixels are then connected to form a complete edge.

[0013] The image analysis unit classifies and defines the edge-enclosed regions using pre-input region classification information, and then fits the recognized image to a stereo model;

[0014] Specifically, the steps include the following:

[0015] S1. The steel coil is placed at the unbundling station using a lifting device. A three-dimensional model is created using a visual scanning system, and the three-dimensional model is placed in a pre-established three-dimensional spatial coordinate system to determine the precise location of the steel coil.

[0016] S2. Image recognition is performed through the image processing module. Based on the pre-input region classification information, the edge-enclosed region is classified and defined, and the recognized image is fitted with the 3D model to determine the specific spatial coordinates of the steel coil, strapping, and strapping buckle.

[0017] S3. After determining the spatial coordinates, the robotic arm drives the pneumatic scissors to move to the position of the strapping on the same side, and the robotic arm drives the pneumatic strapping clamp to move to the position of the strapping on the same side and clamp the strapping.

[0018] S4. The pneumatic scissors start to cut the strapping. At the same time, the laser displacement sensor captures the status of the strapping and determines whether the strapping has been cut. If it is cut, the robotic arm drives the pneumatic strapping clamp to move away from the steel coil, so that the strapping is pulled out from inside the steel coil.

[0019] S5. The camera mechanism captures the movement trajectory of the end of the cable tie after it is cut in real time, and analyzes and stores the movement trajectory in a three-dimensional coordinate set.

[0020] Furthermore, the coordinates of the cutting position of the strapping in S3 are used as marker values, and the three-dimensional spatial coordinate set in S5 is used as reference values. The marker values ​​and reference values ​​are matched one-to-one to form a database and a mapping relationship is formed. The reference values ​​of the first to Nth unbundling are compared with the size of the movement range of the end of the strapping and the length of time the strapping is completely separated from the steel coil. The optimal cutting position is selected through weight calculation, and all data is stored to form a database.

[0021] Furthermore, machine learning is performed based on the database to identify the mapping relationship between the labeled value and the reference value, as well as the optimal cutting position, to form an identification model. The identification model continuously optimizes the optimal strapping cut position through supervised learning and applies it to the N+1th cutting operation. The data after the N+1th cutting is recorded, and the calculation and optimization are continued in the above processing method.

[0022] Furthermore, the image preprocessing unit obtains the local image with the most obvious image features through noise reduction, image visualization enhancement, and lossless image compression. Image visualization enhancement includes contrast enhancement, brightness adjustment, and saturation adjustment.

[0023] Furthermore, in noise reduction, the contrast and sharpness of the image are improved by replacing each pixel value in the image with the average value of its surrounding pixels.

[0024] Furthermore, the pneumatic bundling clamp includes a body, inside which is a cylinder. The piston rod of the cylinder is connected to a crossbar, and the two ends of the crossbar are connected to a connecting rod in a hinged manner. The other end of the connecting rod is connected to a fan plate in a hinged manner, and the arc-shaped edge of the fan plate is provided with serrations.

[0025] The upper part of the body is provided with a slide, and two tooth seats are slidably arranged on the slide. The upper part of the tooth seats is fixedly connected with a clamping claw, and the two clamping claws are mirror symmetrical; the saw teeth mesh with the teeth on the tooth seats.

[0026] Furthermore, the upper part of the gripper is provided with a slope.

[0027] Furthermore, in S4, if the strapping is blocked when being pulled out due to being pressed down by the steel coil, an alarm signal is output, and the pneumatic strapping clamp stops working.

[0028] The beneficial effects of this invention are:

[0029] The method for removing the transverse binding straps of steel coils provided by this invention can intelligently determine the optimal unbundling position and then automatically complete the unbundling operation. At the same time, a pneumatic pull-out clamp is used to remove the binding straps, avoiding the transverse binding straps remaining stuck inside the steel coil after unbundling, which would affect the next step of the operation. Simultaneously, the data is stored, and through supervised learning, a digital twin simulation model is formed to form a database mapping relationship. The accuracy of the model is improved through continuous verification and learning. It can gradually optimize the position of cutting the binding straps according to the actual operation process, thereby improving the unbundling efficiency and reducing working time. Attached Figure Description

[0030] Appendix Figure 1 This is a flowchart of the present invention;

[0031] In the attached diagram: 1. Body, 2. Cylinder, 3. Crossbar, 4. Connecting rod, 5. Sector plate, 6. Gear seat, 7. Clamping claw; Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this invention clearer, the following description will be provided in conjunction with the appendix of this invention. Figure 1 The present invention will be described in more detail below.

[0033] This invention provides a method for removing transversely bundled steel strips from steel coils. The apparatus includes a visual scanning system, an image processing module, a laser displacement sensor, a debundling assembly, and a debundling assembly. The debundling and debundling assemblies are respectively installed on the foundations on both sides of the debundling station. The debundling assembly includes a robotic arm and pneumatic shears, with the pneumatic shears installed at the end of the robotic arm and capable of free movement in three-dimensional space. The debundling assembly includes a robotic arm and a pneumatic debundling clamp, with the pneumatic debundling clamp installed at the end of the robotic arm and capable of free movement in three-dimensional space. The visual scanning system includes at least two camera mechanisms, a supplementary lighting mechanism, and a control component for controlling the camera mechanisms to capture images. The control component controls the supplementary lighting mechanism to automatically activate during image capture. The visual scanning system also includes at least two laser scanning mechanisms. The laser scanning mechanisms acquire three-dimensional coordinate data of the surface of the object being measured at the debundling station, and after denoising, registration, and stitching as point cloud data, establish and store a three-dimensional model. The image processing module includes an image preprocessing unit, an image feature extraction unit, and an image analysis unit. The image preprocessing unit receives multiple images acquired by the visual scanning system, compares the images after preprocessing, and selects the image with the most obvious image features as the local image. The image feature extraction unit receives the local image, reduces image noise to achieve image smoothing, and then performs convolution operations on the image in the horizontal and vertical directions to obtain the gradient intensity and gradient direction of each pixel. For each pixel, based on its gradient direction, it checks whether it is a local gradient maxima in that direction. If not, the pixel is suppressed. An edge threshold is set, and pixels with gradient intensities greater than the edge threshold are identified as edge pixels. The edge pixels are connected to form a complete edge. The image analysis unit classifies and defines the edge-enclosed regions using pre-input region classification information and fits the recognized image to a stereo model.

[0034] Specifically, the steps include the following:

[0035] S1. The steel coil is placed at the unbundling station using a lifting device. A three-dimensional model is created using a visual scanning system, and the three-dimensional model is placed in a pre-established three-dimensional spatial coordinate system to determine the precise location of the steel coil.

[0036] S2. Image recognition is performed through the image processing module. Based on the pre-input region classification information, the edge-enclosed region is classified and defined, and the recognized image is fitted with the 3D model to determine the specific spatial coordinates of the steel coil, strapping, and strapping buckle.

[0037] S3. After determining the spatial coordinates, the robotic arm drives the pneumatic scissors to move to the position of the strapping on the same side, and the robotic arm drives the pneumatic strapping clamp to move to the position of the strapping on the same side and clamp the strapping.

[0038] S4. The pneumatic scissors start to cut the strapping. At the same time, the laser displacement sensor captures the status of the strapping and determines whether the strapping has been cut. If it is cut, the robotic arm drives the pneumatic strapping clamp to move away from the steel coil, so that the strapping is pulled out from inside the steel coil.

[0039] S5. The camera mechanism captures the movement trajectory of the end of the cable tie after it is cut in real time, and analyzes and stores the movement trajectory in a three-dimensional coordinate set.

[0040] The coordinates of the strapping cut position in S3 are used as marker values, and the three-dimensional spatial coordinate set in S5 is used as reference values. A database is constructed by mapping the marker values ​​and reference values ​​one-to-one. For the reference values ​​from the 1st to the Nth unbundling, the range of movement of the strapping end and the time it takes for the strapping to completely detach from the steel coil are compared. The optimal cutting position is selected through weighted calculation, and all data is stored in the database. Machine learning is performed based on the database to identify the mapping relationship between the marker values ​​and reference values, as well as the optimal cutting position, to construct a recognition model. This model continuously optimizes the optimal strapping cut position through supervised learning and is applied to the (N+1)th cutting operation. Data after the (N+1)th cutting is recorded, and calculations and optimizations continue using the above processing method. The image preprocessing unit obtains the local image with the most obvious image features through noise reduction, image visualization enhancement, and lossless image compression. Image visualization enhancement includes contrast enhancement, brightness adjustment, and saturation adjustment. In noise reduction, the contrast and clarity of the image are improved by replacing each pixel value in the image with the average value of its surrounding pixels. The pneumatic strapping clamp includes a body 1, inside which is a cylinder 2. A crossbar 3 is connected to the piston rod end of the cylinder 2. Connecting rods 4 are hinged to both ends of the crossbar 3, and a fan plate 5 is hinged to the other end of the connecting rod 4. The arc-shaped edge of the fan plate 5 has serrations. A slide rail is provided on the upper part of the body 1, and two toothed seats 6 are slidably mounted on the slide rail. Clamping claws 7 are fixedly connected to the upper part of the toothed seats 6, and the two clamping claws are mirror-symmetrical. The serrations mesh with the teeth on the toothed seats 6. An inclined surface is provided on the upper part of the clamping claws 7. In step S4, if the strapping is blocked when being pulled out due to being pressed by the steel coil, an alarm signal is output, and the pneumatic strapping clamp stops working.

[0041] Example 1 is as follows:

[0042] A method for removing transversely bundled steel strips from steel coils, using a device including a vision scanning system, an image processing module, a laser displacement sensor, a debundling assembly, and a debundling assembly. The debundling assembly and the debundling assembly are respectively installed on the foundation on both sides of the debundling station. The debundling assembly is mainly used to cut the bundled strips, and the debundling assembly is mainly used to pull the cut bundled strips out of the inside of the steel coil, thus avoiding the transverse bundled strips remaining stuck inside the steel coil after debundling, which would affect the next step of the operation.

[0043] The unbundling assembly includes a robotic arm and pneumatic scissors. The pneumatic scissors are installed at the end of the robotic arm and can move freely in three-dimensional space. The robotic arm is a robotic arm, which is existing technology and will not be described in detail here. The pneumatic scissors can be laser cutters, which is existing technology and will not be described in detail here.

[0044] The strapping assembly includes a robotic arm and a pneumatic strapping clamp. The pneumatic strapping clamp is installed at the end of the robotic arm and can move freely in three-dimensional space. The robotic arm is a robotic arm, which is existing technology and will not be described in detail here.

[0045] The pneumatic bundling clamp includes a body 1, a cylinder 2 is installed inside the body 1, a crossbar 3 is connected to the piston rod end of the cylinder 2, a connecting rod 4 is connected to both ends of the crossbar 3 in a hinged manner, a fan plate 5 is connected to the other end of the connecting rod 4 in a hinged manner, and serrations are provided on the arc-shaped edge of the fan plate 5.

[0046] The upper part of the main body 1 is provided with a slide rail, on which two toothed seats 6 are slidably arranged. The upper part of the toothed seats 6 is fixedly connected with a clamping claw 7. The two clamping claws are mirror symmetrical. The saw teeth mesh with the teeth on the toothed seats 6. When the piston rod of the cylinder 2 retracts, the crossbar 3 will move downward, and at the same time drive the fan plate 5 to rotate. Since the fan plate 5 meshes with the toothed seats 6, the toothed seats 6 drive the clamping claws 7 to move towards each other. The upper part of the clamping claws 7 is provided with an inclined surface, which makes it convenient for the clamping claws 7 to be inserted into the gap between the strapping and the steel coil. The width of the gap between the clamping claws 7 is the same as the width of the strapping, which can clamp the strapping. Under the drive of the robotic arm, the strapping is pulled out in an anthropomorphic manner.

[0047] The visual scanning system includes at least two camera units, a supplementary lighting mechanism, and a control component for controlling the camera units to capture images. The control component controls the supplementary lighting mechanism to automatically turn on during shooting. The camera units are located at the left front, left rear, right front, and right rear of the unpacking station, respectively, to capture images from multiple angles and obtain accurate image information. The supplementary lighting mechanism includes multiple supplementary lights, which are located next to the camera units to provide sufficient light enhancement and ensure image clarity and brightness.

[0048] The visual scanning system also includes at least two laser scanning mechanisms. The laser scanning mechanisms acquire three-dimensional coordinate data of the surface of the object being measured at the unpacking station. They can accurately acquire high-precision information of irregular structures and establish and store a three-dimensional model after noise reduction, registration and stitching in the form of point cloud data.

[0049] The image processing module includes an image preprocessing unit, an image feature extraction unit, and an image analysis unit;

[0050] The image preprocessing unit receives multiple images acquired by the visual scanning system, compares the multiple images after image preprocessing, and selects the image with the most obvious image features as the local image.

[0051] The image preprocessing unit obtains the local image with the most obvious image features through noise reduction, image visualization enhancement, and lossless image compression. Image visualization enhancement includes contrast enhancement, brightness adjustment, and saturation adjustment. In noise reduction, the contrast and clarity of the image are improved by replacing each pixel value in the image with the average value of its surrounding pixels.

[0052] The image feature extraction unit receives the local image, reduces image noise to achieve image smoothing, and then performs convolution operations on the image in the horizontal and vertical directions to obtain the gradient intensity and gradient direction of each pixel. For each pixel, it checks whether it is a local maximum value of the gradient in that direction based on its gradient direction. If not, the pixel is suppressed.

[0053] Set an edge threshold, identify pixels with gradient strength greater than the edge threshold as edge pixels, and connect the edge pixels to form a complete edge;

[0054] The image analysis unit classifies and defines the edge-enclosed regions using pre-input region classification information, and then fits the recognized image to a stereo model;

[0055] Specifically, the steps include the following:

[0056] S1. The steel coil is placed at the unbundling station using a lifting device. A 3D model is created using a vision scanning system and placed within a pre-established 3D coordinate system to determine the precise location of the steel coil. This 3D coordinate system is either manually or automatically created beforehand and remains unchanged after creation; all subsequent operations are based on this coordinate system to ensure data accuracy. The 3D coordinate system includes X, Y, and Z axes, with the midpoint of the unbundling station as its origin. The 3D model determines the position of the steel coil within the 3D coordinate system, simultaneously determining the positions of the strapping straps and their clips. However, at this stage, the classification of each component is not automatically determined within the model.

[0057] S2. Image recognition is performed through the image processing module. Through image preprocessing and image feature extraction, the features of each region in the image are determined. Based on the pre-input region classification information, the regions enclosed by the edges are classified and defined. For example, based on the pre-defined structural and size features of the strapping, the region is compared with the similar regions after the edges are determined in the image recognition to determine which region is the strapping. The recognized image is then fitted to the 3D model. This process bends the image so that the edges of the image recognition overlap with the raised edges of the 3D model. The feature values ​​of the 3D model are matched one-to-one with the feature values ​​of the image recognition to determine the specific spatial coordinates of the steel coil, strapping, and strapping buckle.

[0058] S3. After determining the spatial coordinate position, the robotic arm drives the pneumatic scissors to move to the position of the strapping on the same side, and the robotic arm drives the pneumatic strapping clamp to move to the position of the strapping on the same side and clamp the strapping. This position is determined by three-dimensional coordinates. During the first operation, this position can be selected manually or automatically by the system.

[0059] S4. The pneumatic scissors start to cut the strapping. At the same time, the laser displacement sensor captures the status of the strapping and determines whether the strapping has been cut. If it is cut, the robotic arm drives the pneumatic strapping clamp to move away from the steel coil, so that the strapping is pulled out from inside the steel coil.

[0060] S5. The camera mechanism captures the movement trajectory of the end of the cable tie after it is cut in real time, and analyzes and stores the movement trajectory in a three-dimensional coordinate set.

[0061] The coordinates of the cut position of the strapping in S3 are used as the marker value, and the three-dimensional spatial coordinate set in S5 is used as the reference value. The marker value and the reference value are matched one-to-one to form a database and form a mapping relationship. The reference values ​​of the first to Nth unbundling are compared with the size of the movement range of the end of the strapping and the length of time the strapping is completely separated from the steel coil. The optimal cutting position is selected through weight calculation, and all data are stored to form a database.

[0062] The weight of the range of movement of the end of the strapping is 0.4, and the weight of the time it takes for the strapping to completely detach from the steel coil is 0.6. After calculating each data, the cutting position corresponding to the minimum value is selected as the optimal cutting position.

[0063] Machine learning is performed based on the database to identify the mapping relationship between the labeled value and the reference value, as well as the optimal cutting position to form an identification model. The identification model continuously optimizes the optimal strapping cut position through supervised learning and applies it to the N+1th cutting operation. The data after the N+1th cutting is recorded, and the calculation and optimization are continued in the above processing method.

[0064] During the placement of the steel coil, the strapping may become trapped between the coil and the base, preventing it from being pulled out. This problem can be resolved by manual adjustment during placement. The system also includes an emergency measure: if the strapping is blocked by the coil, an alarm signal will be output and the pneumatic strapping clamp will stop working to prevent damage to the robotic arm and the clamp. After adjusting the position, the process will be repeated.

[0065] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for removing the transversely bundled steel strips of a steel coil, characterized in that, The device used includes a vision scanning system, an image processing module, a laser displacement sensor, a debundling assembly and a debundling assembly, which are respectively installed on the foundation on both sides of the debundling station; The unbundling assembly includes a robotic arm and pneumatic scissors. The pneumatic scissors are mounted at the end of the robotic arm and can move freely in three-dimensional space. The bundling assembly includes a robotic arm and a pneumatic bundling clamp, wherein the pneumatic bundling clamp is mounted at the end of the robotic arm and can move freely in three-dimensional space; The visual scanning system includes at least two camera mechanisms, a supplementary lighting mechanism, and a control component for controlling the camera mechanisms to capture images. The control component controls the supplementary lighting mechanism to automatically turn on during image capture. The visual scanning system also includes at least two laser scanning mechanisms. The laser scanning mechanisms acquire three-dimensional coordinate data of the surface of the object being measured at the unpacking station, and establish and store a three-dimensional model after noise reduction, registration and stitching in the form of point cloud data. The image processing module includes an image preprocessing unit, an image feature extraction unit, and an image analysis unit; The image preprocessing unit receives multiple images acquired by the visual scanning system, compares the multiple images after image preprocessing, and selects the image with the most obvious image features as the local image. The image feature extraction unit receives the local image, reduces image noise to achieve image smoothing, and then performs convolution operations on the image in the horizontal and vertical directions to obtain the gradient intensity and gradient direction of each pixel. For each pixel, based on its gradient direction, it checks whether it is a local maximum value of the gradient in that direction. If not, the pixel is suppressed. An edge threshold is set, and pixels with gradient intensity greater than the edge threshold are identified as edge pixels. The edge pixels are then connected to form a complete edge. The image analysis unit classifies and defines the edge-enclosed regions using pre-input region classification information, and then fits the recognized image to a stereo model; Specifically, the steps include the following: S1. The steel coil is placed at the unbundling station using a lifting device. A three-dimensional model is created using a visual scanning system, and the three-dimensional model is placed in a pre-established three-dimensional spatial coordinate system to determine the precise location of the steel coil. S2. Image recognition is performed through the image processing module. Based on the pre-input region classification information, the edge-enclosed region is classified and defined, and the recognized image is fitted with the 3D model to determine the specific spatial coordinates of the steel coil, strapping, and strapping buckle. S3. After determining the spatial coordinates, the robotic arm drives the pneumatic scissors to move to the position of the strapping on the same side, and the robotic arm drives the pneumatic strapping clamp to move to the position of the strapping on the same side and clamp the strapping. S4. The pneumatic scissors start to cut the strapping. At the same time, the laser displacement sensor captures the status of the strapping and determines whether the strapping has been cut. If it is cut, the robotic arm drives the pneumatic strapping clamp to move away from the steel coil, so that the strapping is pulled out from inside the steel coil. S5. The camera mechanism captures the movement trajectory of the end of the strapping after it is cut in real time, and analyzes and stores the movement trajectory in a three-dimensional space coordinate set. The coordinates of the cutting position of the strapping in S3 are used as the marker value, and the three-dimensional spatial coordinate set in S5 is used as the reference value. The marker value and the reference value are matched one-to-one to form a database and form a mapping relationship. The reference values ​​of the first to Nth unbundling are compared with the size of the movement range of the end of the strapping and the length of time the strapping is completely separated from the steel coil. The optimal cutting position is selected through weight calculation, and all data are stored to form a database. Machine learning is performed based on the database to identify the mapping relationship between the labeled value and the reference value, as well as the optimal cutting position, to form an identification model. The identification model continuously optimizes the optimal strapping cut position through supervised learning and is applied to the N+1th cutting operation. The data after the N+1th cutting is recorded, and the calculation and optimization are continued in the above processing method.

2. The method of claim 1, wherein: The image preprocessing unit obtains the local image with the most obvious image features through noise reduction, image visualization enhancement, and lossless image compression. The image visualization enhancement includes contrast enhancement, brightness adjustment, and saturation adjustment.

3. The method of claim 2, wherein: In noise reduction, the contrast and sharpness of an image are improved by replacing each pixel value in the image with the average value of its surrounding pixels.

4. The method of claim 1, wherein: The pneumatic bundling clamp includes a body, inside which is a cylinder. The piston rod of the cylinder is connected to a crossbar, and the two ends of the crossbar are connected to a connecting rod in a hinged manner. The other end of the connecting rod is connected to a fan plate in a hinged manner, and the arc-shaped edge of the fan plate is provided with serrations. The upper part of the body is provided with a slide rail, and two tooth seats are slidably arranged on the slide rail. A clamping claw is fixedly connected to the upper part of the tooth seat, and the two clamping claws are mirror symmetrical; the saw teeth mesh with the teeth on the tooth seats.

5. The method of claim 4, wherein: The upper part of the gripper is provided with an inclined surface.

6. The method of claim 1, wherein: In S4, if the strapping is blocked when it is pulled out due to being pressed down by the steel coil, an alarm signal is output and the pneumatic strapping clamp stops working.

Images