Strip steel tail positioning method and device, electronic equipment, storage medium and coiling control system

Abstract

The invention provides a strip steel tail positioning method and device, electronic equipment, a storage medium and a coiling control system, and the method comprises the steps: inputting an obtained real-time steel coil image into a pre-trained recognition segmentation model, and obtaining a segmentation feature image and the position information, confidence and channel weight vector of at least one bounding box in the real-time steel coil image; determining a band tail contour position in the real-time steel coil image based on the segmentation feature image and a channel weight vector of a target bounding box when determining that all bounding boxes have the target bounding box where the band steel tail is located based on the position information and the confidence coefficient of each bounding box, and determining a residual coiling angle based on the band tail contour position; the rotating angle reflecting module reflects a rotating angle required by a winding drum of a coiling machine to obtain a standard to-be-bundled steel coil. When the tail of the strip steel appears in the real-time steel coil image, the residual rotation angle of the winding drum is calculated in real time, so that a standard to-be-bundled steel coil can be obtained, and the tail of the strip steel does not need to be manually concerned and positioned in the whole process.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and in particular, to a method, device, electronic device, storage medium, and coiling control system for positioning the tail of a strip steel. Background Art

[0002] The hot rolling process is one of the very important and key steps in the steel production process. Among them, the main function of the hot rolling coiler is to bend the hot rolled strip steel rolled by the rolling mill into a dense hot rolled coil to improve the transportation efficiency. And the coiling effect of the coiler directly determines the quality of the steel coil.

[0003] During the coiling process, it is necessary to accurately position the tail of the strip steel to ensure that the strip tail can fit the steel coil after coiling. If the tail of the strip steel is not coiled to the appropriate position, it is very easy to cause loose coiling, affecting the quality of the finished steel coil. In addition, untimely stopping of coiling may also cause the entire production line to stop. Therefore, the accurate positioning of the strip steel tail is crucial.

[0004] Currently, most hot rolling plants still rely on manual positioning of the strip steel tail. This method is not only time-consuming and laborious, but also difficult to guarantee the accuracy. Summary of the Invention

[0005] The purpose of the present invention is to provide a method, device, electronic device, storage medium, and coiling control system for positioning the tail of a strip steel to improve the problems existing in the prior art.

[0006] The embodiments of the present invention can be implemented as follows: In a first aspect, the present invention provides a method for positioning the tail of a strip steel, including: Obtaining a real-time steel coil image of the current steel coil, and inputting the real-time steel coil image into a pre-trained recognition and segmentation model to obtain an output result; the output result includes a segmentation feature image, as well as the position information, confidence level, and channel weight vector of at least one bounding box in the real-time steel coil image; Based on the position information and confidence level of each bounding box, determining whether there is a target bounding box among all the bounding boxes; If it exists, based on the segmentation feature image and the channel weight vector of the target bounding box, determining the position of the strip tail contour in the real-time steel coil image; the area where the target bounding box is located in the real-time steel coil image includes the tail of the strip steel; Based on the position of the strip tail contour, determining the remaining coiling angle of the current steel coil, and the remaining coiling angle reflects the rotation angle required for the coiler drum from the current steel coil to become a standard steel coil to be strapped.

[0007] In a second aspect, an embodiment of the present invention provides a device for positioning the tail of a strip steel, including: A model processing module, configured to: obtain a real-time steel coil image of the current steel coil, and input the real-time steel coil image into a pre-trained recognition and segmentation model to obtain an output result; the output result includes a segmentation feature image and position information, confidence, and channel weight vectors of at least one bounding box in the real-time steel coil image. A screening module, configured to determine whether there is a target bounding box among all the bounding boxes based on the position information and confidence of each bounding box. A contour determination module, configured to, when the target bounding box exists, determine the position of the strip tail contour in the real-time steel coil image based on the segmentation feature image and the channel weight vector of the target bounding box; the area where the target bounding box is located in the real-time steel coil image includes the strip tail. An angle calculation module, configured to determine the remaining coiling angle of the current steel coil based on the position of the strip tail contour, where the remaining coiling angle reflects the rotation angle required for the reel of the coiling machine to change the current steel coil into a standard steel coil to be bundled.

[0008] In a third aspect, an embodiment of the present invention provides an electronic device, including: a memory and a processor, where the memory stores a software program, and when the electronic device runs, the processor executes the software program to implement the method for positioning the strip tail as described in the first aspect.

[0009] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for positioning the strip tail as described in the first aspect is implemented.

[0010] In a fifth aspect, an embodiment of the present invention provides a coiling control system, where the coiling control system includes: At least one coiling machine and a camera configured for each coiling machine, where the camera is used to collect a real-time steel coil image during the coiling process of the corresponding coiling machine. An image processing device, configured to receive the real-time steel coil image transmitted by the camera and execute the method for positioning the strip tail as described in the first aspect to analyze the received real-time steel coil image to obtain the remaining coiling angle of the coiling machine.

[0011] Compared with the prior art, the embodiment of the present invention provides a method, a device, an electronic device, a storage medium and a coiling control system for positioning the tail of a strip steel. First, a real-time steel coil image of the current steel coil is obtained, and the real-time steel coil image is input into a pre-trained recognition and segmentation model to obtain a segmented feature image, as well as the position information, confidence level and channel weight vector of at least one bounding box in the real-time steel coil image. Then, based on the position information and confidence level of each bounding box, it is determined whether there is a target bounding box among all the bounding boxes. If there is, based on the segmented feature image and the channel weight vector of the target bounding box, the position of the strip tail contour is determined in the real-time steel coil image. The area where the target bounding box is located in the real-time steel coil image includes the tail of the strip steel. Finally, based on the position of the strip tail contour, the remaining coiling angle of the current steel coil is determined. Since the remaining coiling angle reflects the rotation angle required for the coiler mandrel to change from the current steel coil to a standard steel coil to be strapped, the present invention calculates the remaining rotation angle of the mandrel in real time when the tail of the strip steel appears in the real-time steel coil image, so as to obtain a standard steel coil to be strapped, without the need for manual full-time attention to positioning the tail of the strip steel. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0013] Figure 1 FIG. 1 is one of the schematic flowcharts of a method for positioning the tail of a strip steel provided by an embodiment of the present invention.

[0014] Figure 2 FIG. 2 is an example diagram of perspective comparison provided by an embodiment of the present invention.

[0015] Figure 3 FIG. 3 is another schematic flowchart of a method for positioning the tail of a strip steel provided by an embodiment of the present invention.

[0016] Figure 4 FIG. 4 is a comparison schematic diagram of a real-time steel coil image and a strip tail mask image provided by an embodiment of the present invention.

[0017] Figure 5 FIG. 5 is a schematic diagram of angle determination provided by an embodiment of the present invention.

[0018] Figure 6 FIG. 6 is a schematic structural diagram of a device for positioning the tail of a strip steel provided by an embodiment of the present invention.

[0019] Figure 7 FIG. 7 is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. Detailed implementation manners

[0020] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. The components of the embodiments of the present invention described and illustrated herein generally can be arranged and designed in a variety of different configurations.

[0021] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0022] It should be noted that: like reference numerals and letters denote like items in the following drawings, and thus, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0023] It should be noted that, without conflict, the features in the embodiments of the present invention can be combined with each other.

[0024] Herein, first, keywords or key terms related to the present invention are introduced: 1. Coiler: An auxiliary equipment in a rolling mill workshop that winds hot-rolled or cold-rolled steel into a coil.

[0025] 2. Coiling: Refers to the process of winding materials (such as strip steel) into a coil.

[0026] 3. ROI (Region of Interest): That is, the region of interest, which refers to a specific region of concern or interest in an image or picture.

[0027] The method for positioning the tail of a strip steel provided by the embodiments of the present invention can be applied to an electronic device. The electronic device can be, but is not limited to, computing devices such as smart phones, personal notebooks, personal computers, servers, and industrial control computers.

[0028] Please refer to Figure 1 , Figure 1 which is a schematic flowchart of a method for positioning the tail of a strip steel provided by the embodiments of the present invention. The method includes the following steps S101 to S104.

[0029] S101. Obtain a real-time coiled steel image of the current coiled steel, and input the real-time coiled steel image into a pre-trained recognition and segmentation model to obtain an output result.

[0030] It can be understood that for a coiler that is performing strip coiling, the image obtained by the camera taking real-time pictures of the steel coil being coiled on the coiler drum is the real-time coiling image.

[0031] Optionally, in an actual strip production workshop, a corresponding camera can be installed for each coiler. For example, ordinary high-definition cameras can be used in cold rolling workshops, while infrared thermal imaging cameras can be used in hot rolling workshops. When installing, the imaging perspective needs to be considered. Ideally, the ideal imaging perspective of the camera is preferably directly facing the front of the coiled steel coil. For example, the ideal imaging perspective can Figure 2 be similar to the imaging perspective of the left image in

[0032] However, due to possible other obstructions or some other reasons in the environment where the coiler is located, the camera cannot be installed at the ideal position with the ideal imaging perspective. Its actual imaging is usually side-facing the front of the coiled steel coil. For example, the actual imaging perspective of the camera may be similar to Figure 2 the imaging perspective of the right image in

[0033] It should be noted that Figure 2 The illustration shown is only for reference of perspective explanation. The embodiments of the present invention do not limit the camera type, the imaging perspective of the actual steel coil image, the imaging distance, the coiling state of the current steel coil, etc.

[0034] In this embodiment, when a coiler starts to work for strip coiling, the corresponding camera can take pictures of the steel coil image at a preset acquisition frequency (for example, between 30 and 60 fps). For each real-time steel coil image collected, it can be input into the recognition and segmentation model for processing. The output results of the recognition and segmentation model include the segmentation feature image and the position information, confidence level, and channel weight vector of at least one bounding box in the real-time steel coil image. The confidence level reflects the probability that the bounding box contains the strip tail.

[0035] Optionally, the recognition and segmentation model can include a detection network and a segmentation network. Input a real-time steel coil image into the recognition and segmentation model: the detection network can identify the strip tail in the real-time steel coil image, and thus output the position information, confidence level, and channel weight vector of at least one bounding box; while the segmentation network can adopt instance segmentation technology to output a multi-channel segmentation feature image.

[0036] S102. Based on the position information and confidence level of each bounding box, determine whether there is a target bounding box among all the bounding boxes.

[0037] Due to environmental interference or interference from other coiling machines in the image field of view, the recognition and segmentation model may identify multiple bounding boxes. In the area of some of these bounding boxes, it may not be the tail of the strip on the current coiling machine, but other objects are misidentified as the tail of the strip or belong to the tail of the strip on other coiling machines in the camera's field of view. Therefore, it is necessary to distinguish from all the bounding boxes whether there is a target bounding box. The area where the target bounding box is located in the real-time steel coil image should include the tail of the strip.

[0038] If it is determined based on the position information and confidence level of each bounding box that there is no target bounding box among all the bounding boxes, it indicates that the coiling machine may have just started coiling or is in the middle stage of coiling, and there is still some time until the coiling ends. At this time, there is no need to execute the subsequent steps S103~S104 to locate the tail of the strip.

[0039] S103: If there is a target bounding box, based on the segmented feature image and the channel weight vector of the target bounding box, determine the position of the strip tail contour in the real-time steel coil image.

[0040] S104: Based on the position of the strip tail contour, determine the remaining coiling angle of the current steel coil.

[0041] In this embodiment, if there is a target bounding box, it is necessary to determine the position where the tail of the strip is located in the target bounding box. The present invention indirectly determines the position where the tail of the strip is located by calculating the remaining coiling angle of the current steel coil. The remaining coiling angle directly reflects the rotation angle required for the coiler's reel to change the current steel coil into a standard steel coil to be strapped. For a standard steel coil to be strapped on the coiling machine, the tail of the strip is located at the optimal stop position (i.e., the lowest position of the steel coil on the coiling machine), so that the tension distribution on the steel coil is uniform and it is not easy to unwind.

[0042] Therefore, the calculated remaining coiling angle can be used to prompt the operator that the tail of the strip of the current steel coil is about to reach the optimal stop position. In this way, when the operator sees that the tail of the strip reaches the optimal stop position, the operator can timely control the coiling machine to stop coiling.

[0043] Alternatively, in an automated control workshop, the coiling machine can be automatically controlled. For example, an industrial control device acting as a host computer is responsible for the production of a production line, and it is responsible for when the coiling machine starts coiling and when it stops coiling. In this way, when the industrial control device confirms that the remaining coiling angle of the current steel coil is 0, that is, the tail of the strip has reached the optimal stop position, it can immediately control the coiling machine to end coiling.

[0044] The positioning method for the tail of the strip steel provided by the embodiment of the present invention first obtains the real-time strip steel image of the current steel coil, and inputs the real-time strip steel image into a pre-trained recognition and segmentation model to obtain a segmented feature image, as well as the position information, confidence level, and channel weight vector of at least one bounding box in the real-time strip steel image; then, based on the position information and confidence level of each bounding box, it is determined whether there is a target bounding box among all the bounding boxes; if there is, based on the segmented feature image and the channel weight vector of the target bounding box, the position of the strip tail contour is determined in the real-time strip steel image; the area where the target bounding box is located in the real-time strip steel image includes the tail of the strip steel; finally, based on the position of the strip tail contour, the remaining coiling angle of the current steel coil is determined, and the remaining coiling angle reflects the rotation angle required for the reel of the coiling machine to change the current steel coil into a standard steel coil to be bundled. Therefore, when the tail of the strip steel appears in the real-time strip steel image in the present invention, the remaining rotation angle of the reel is calculated in real time, so that the operator or industrial control equipment can control the coiling machine to stop coiling when the remaining rotation angle is 0°, thereby obtaining a standard and non-loosening steel coil to be bundled, without the need for manual full-time attention to positioning the tail of the strip steel.

[0045] In an optional implementation manner, please refer to Figure 3 , for the step S102 in the above "Based on the position information and confidence level of each bounding box, determine whether there is a target bounding box among all the bounding boxes", it includes sub-steps S1021 to S1025.

[0046] S1021. Determine whether there is only one pending bounding box among all the bounding boxes whose confidence level is greater than or equal to the preset confidence threshold.

[0047] In this embodiment, when there is only one pending bounding box among all the bounding boxes, the following step S1022 is continued; when there are multiple pending bounding boxes among all the bounding boxes, the following step S1024 is continued.

[0048] Optionally, the confidence level can be a value between 0 and 1, and the preset confidence threshold can be but is not limited to values such as 0.8, 0.9, 0.95, etc. This example is only for illustration and is not limited here.

[0049] S1022. If the only pending bounding box is located in the preset ROI region, the pending bounding box is used as the target bounding box.

[0050] In this embodiment, the preset ROI region is the region through which the tail of the strip steel will appear on the collected strip steel image during the rotation of the reel of the coiling machine. Using the preset ROI region, the attention area required for positioning the tail of the strip steel can be focused on the actual area where the tail of the strip steel appears, avoiding other interferences.

[0051] S1023. Determine the pending bounding box with the highest confidence from all the pending bounding boxes.

[0052] In this embodiment, in the real-time steel coil image, the pending bounding box with the highest confidence is most likely to include the true tail of the strip steel.

[0053] S1024. Based on the position information of each pending bounding box, calculate the intersection over union (IoU) between the pending bounding box with the highest confidence and each other pending bounding box respectively.

[0054] In this embodiment, the position information of the pending bounding box may include the horizontal value range and the vertical value range of the bounding box area, or may include the center point coordinates, width, and height of the bounding box area. The intersection over union is IoU (Intersection over Union), which is used to reflect the overlapping degree of two areas.

[0055] For two pending bounding boxes (assuming box A and box B), the method for calculating the intersection over union between them is as follows: Using the respective position information of box A and box B, calculate the area of the overlapping part between box A and box B and the total area covered by box A and box B respectively, and then divide the area of the overlapping part by the total area to obtain the intersection over union of box A and box B.

[0056] S1025. Take the pending bounding boxes in the ROI area among each other pending bounding box with an intersection over union lower than the preset intersection over union threshold and the pending bounding box with the highest confidence as the target bounding boxes.

[0057] In this embodiment, the value range of the intersection over union is 0 to 1, and the larger the value, the higher the overlapping degree of the two boxes. The preset intersection over union threshold can be 0.1 or 0.2, etc., which is not limited here.

[0058] For example, assume there are 4 pending bounding boxes (box 1 to box 4), and box 1 is the pending bounding box with the highest confidence. Therefore, it is necessary to calculate the intersection over union between box 1 and box 2, box 3, and box 4 respectively. If only the intersection over union corresponding to box 2 is lower than the preset intersection over union threshold, then it is necessary to select the target bounding boxes in the ROI area from box 1 and box 2. If neither box 1 nor box 2 is in the ROI area, there is no target bounding box. This example is only for illustration and is not limited here.

[0059] The recognition and segmentation model may identify multiple bounding boxes from the real-time steel coil image. Through the above steps S1021 - S1025, the target bounding box including the strip tail can be accurately selected from multiple bounding boxes. During the selection process, based on the ROI region, the preset confidence threshold, and the preset intersection over union (IoU) threshold, the misidentified bounding boxes caused by environmental interference and the strip tails on other coiling machines are excluded, ensuring that the strip tail in the final target bounding box is located on the coiling machine corresponding to the camera that captures the real-time steel coil image.

[0060] In an optional implementation manner, for the step of "determining the strip tail contour position in the real-time steel coil image based on the segmentation feature image and the channel weight vector of the target bounding box" in the above step S103, it includes sub-steps S1031 - S1032.

[0061] S1031. Generate a strip tail mask image based on the segmentation feature image and the channel weight vector of the target bounding box.

[0062] In this embodiment, in the strip tail mask image, the pixel values of each pixel point in the area of the strip tail are all 1, presenting as white, while the pixel values of each pixel point in the remaining areas are all 0, presenting as black.

[0063] For example, please refer to Figure 4 , Figure 4 the real-time steel coil image on the left in Figure 4 shows a steel coil being coiled (the coiling machine and the reel are not shown in Figure 4 ), and its corresponding strip tail mask image is as shown on the right in

[0064] This example is only for illustration and is not limited here.

[0065]

[0066] Among them, the segmentation feature image is a multi-channel image, and the channel weight vector of the target bounding box includes the weight values of each channel of the segmentation feature image. Therefore, the sub-steps of step S1031 may include S1031 - S1033: S1031. Use each weight coefficient in the channel weight vector of the target bounding box to perform weighted processing on each channel of the segmentation feature image to obtain an initial mask image.

[0065] In this embodiment, multiplying the channel weight vector of the target bounding box with the segmentation feature image can achieve the channel weighted processing of the segmentation feature image, thereby obtaining a single-channel initial mask image.

[0066] For example, the channel weight vector of the target bounding box may include 32 weight values, which reflect the importance of the images of 32 channels of the segmentation feature image for forming the mask of the strip tail. The rule of the segmentation feature image may be 32×160×160, and the specification of the obtained initial mask image is 160×160. This example is only for illustration, and the embodiments of the present invention do not limit the number of channels and the size of the segmentation feature image.

[0067] S10312. Normalize the pixel value of each pixel point in the initial mask image to obtain an intermediate mask image.

[0068] In this embodiment, a normalization function (such as the sigmoid function or the Tanh function) can be used to map the pixel value of each pixel point in the initial mask image to a probability value between [0,1], and the probability value reflects the probability that the pixel point belongs to the strip tail.

[0069] S10313. After setting the pixel value of each pixel point outside the target bounding box area in the intermediate mask image to 0, perform binarization processing on the pixel value of each pixel point in the target bounding box area in the intermediate mask image to obtain a tail mask image.

[0070] In this embodiment, since only the position of the target bounding box needs to be concerned, the pixel value of each pixel point outside the target bounding box area in the intermediate mask image can be directly set to 0, and then the pixel value of each pixel point in the target bounding box area in the intermediate mask image is subjected to binarization processing to obtain a tail mask image.

[0071] Optionally, the binarization processing can adopt a threshold method. For a pixel point in the target bounding box area in the intermediate mask image, the formula for binarization is:

[0072] is the pixel value after binarization at the position, is the pixel value before binarization processing at the position, is the preset binarization threshold.

[0073] S1032. Extract the tail contour position of the strip tail in the real-time steel coil image from the tail mask image.

[0074] In this embodiment, the obtained tail contour position may include the pixel coordinates of each contour point of the strip tail in the real-time steel coil image.

[0075] Optionally, the method for extracting the tail contour position from the tail-masked image may include, but is not limited to: a contour extraction method based on edge detection, a contour extraction method based on regions, an extraction model trained by a deep learning algorithm, etc.

[0076] In an alternative implementation, for the process of "determining the remaining coiling angle of the current steel coil based on the tail contour position" in step S104 above, the following sub-steps S1041 to S1044 may be included.

[0077] S1041. Use a preset transformation matrix to perform angle correction processing on the real-time steel coil image, obtaining the corrected image and the pixel coordinates of each corrected contour point in the corrected image.

[0078] In this embodiment, the preset transformation matrix may represent the perspective conversion relationship between the actual imaging perspective of the camera and the ideal imaging perspective, and the ideal imaging perspective is the perspective directly facing the front of the steel coil on the coiler.

[0079] If the installation position of the camera is not directly facing the front of the current steel coil but is side-facing the front of the steel coil, then there will be perspective distortion in the real-time steel coil image captured by the camera. Therefore, it is necessary to perform angle correction on the real-time steel coil image so that the perfect distortion-free front of the steel coil can be presented in the image.

[0080] Optionally, in order to reduce the processing amount, only the image area of the current steel coil in the real-time steel coil image may be corrected. That is, the sub-steps of step S1041 may include S10411 to S10412: S10411. Based on the preset steel coil position information, determine the rectangular area where the current steel coil is located from the real-time steel coil image.

[0081] In this embodiment, the preset steel coil position information may be the coordinates of two diagonal vertices of the rectangular area where the current steel coil is located, or the center point coordinates, width, and height.

[0082] Among them, during the installation and debugging process of the camera, after determining the perspective conversion relationship between the actual installation position imaging perspective and the ideal imaging perspective of the camera, a frame of test steel coil image may be collected, and then the test steel coil image is converted into a directly facing steel coil image using the perspective conversion relationship, and then the steel coil position information is determined from the test steel coil image and the directly facing steel coil image.

[0083] S10412. Use the preset transformation matrix to perform position correction on the pixel coordinates of each pixel point within the rectangular area, obtaining the corrected image and the pixel coordinates of each corrected contour point in the corrected image.

[0084] In this embodiment, the preset transformation matrix M may be a 3×3 matrix. The formula for position correction is as follows: ,in

[0085] Then, we can get the pixel The pixel coordinates obtained after position correction are:

[0086] S1042. Obtain the pixel coordinates of the preset center point of the steel coil in the corrected image.

[0087] In this embodiment, the center point of the steel coil can be calibrated during the installation and debugging process of the camera.

[0088] S1043, calculating the interval distance between the center point of the steel coil and each corrected contour point, and taking the corrected contour point with the longest interval distance as the tail reference point of the strip tail.

[0089] In this embodiment, in the corrected image, all corrected contour points can form a contour line with a tail.

[0090] For example, see Figure 5 , Figure 5 The real-time steel coil image on the left side of the middle can be obtained after the perspective correction processing. Figure 5 The right side of the figure shows the steel coil outline in the corrected image. Figure 5 In the figure, line segment AB is the tail contour line of the strip tail under the ideal viewing angle.

[0091] Figure 5 The above is only an example and is not intended to be limiting.

[0092] S1044. Determine the remaining coiling angle of the current steel coil using the pixel coordinates of the tape tail reference point and the pixel coordinates of the center point of the steel coil.

[0093] In this embodiment, the imaging main optical axis corresponding to the steel coil portion in the corrected image coincides with the axis of the roll. Since the roll is cylindrical, the axis is also called the high line.

[0094] Optionally, the sub-steps of step S1044 may include S10441~S10443.

[0095] S10441. Based on the pixel coordinates of the tape tail reference point and the pixel coordinates of the center point of the steel coil, calculate the slope of the reference connecting line between the tape tail reference point and the center point of the steel coil.

[0096] In this embodiment, combined with Figure 5, the pixel coordinates of the center point P of the steel coil can be , the pixel coordinates of the reference point A at the tail of the strip can be , the reference connection line between the reference point A at the tail of the strip and the center point P of the steel coil has a slope of:

[0097] S10442. Obtain the slope of the horizontal line passing through the center point of the steel coil and parallel to the horizontal axis of the pixel coordinate system of the corrected image.

[0098] In this embodiment, in combination with Figure 5 , the horizontal line passing through the center point of the steel coil and parallel to the horizontal axis (i.e., the X-axis) of the pixel coordinate system XOY of the corrected image is , The slope of is denoted as = 0.

[0099] S10443. Based on the slope of the reference connection line and the slope of the horizontal line, calculate the angle between the reference connection line and the stop reference line to obtain the remaining winding angle of the current steel coil.

[0100] In this embodiment, the stop reference line is the straight line passing through the preset stop point at the tail of the strip and the center point of the steel coil in the corrected image, and the distance between the center point of the steel coil and the stop point at the tail of the strip reflects the maximum radius of the steel coil that the coiler can wind. In Figure 5 , the stop point at the tail of the strip is point D, Figure 5 For the steel coil in, continue to wind, the tail of the strip gradually adheres to the steel coil. When point A coincides with point B, or point A winds to directly above point D, stop winding at this time, and a standard steel coil to be bundled can be obtained, which is not easy to unwind.

[0101] In combination with Figure 5 , the reference connection line has a slope of , the horizontal line has a slope of , then the angle between the reference connection line and the horizontal line is is

[0102]

[0103] In Figure 5 , since the reference point A at the tail of the strip is below the horizontal line , that is, is a positive number, so the calculated is also a positive value. At this time, the remaining winding angle of the current steel coil is: .

[0104] If the trailing reference point A is located above the horizontal line , that is is negative, then the calculated is also negative. At this time, the remaining coiling angle of the current steel coil is: .

[0105] Optionally, in the manual control workshop, the operator's habit may be to determine whether to control the coiler to stop coiling by judging whether the rotated angle of the strip tail reaches 180°. Then, the rotated angle of the strip tail can also be calculated.

[0106] Therefore, combining Figure 5 , when the trailing reference point A is located below the horizontal line , the rotated angle of the strip tail is: ; when the trailing reference point A is located above the horizontal line , the rotated angle of the strip tail is: .

[0107] In another alternative implementation, after performing the above S1043, combining Figure 5 , it is also possible to directly calculate the remaining coiling angle and / or the rotated angle based directly on the pixel coordinates of the steel coil center point P, the strip tail stop point D, and the trailing reference point A: (1) Based on the pixel coordinates of the steel coil center point P and the strip tail stop point D respectively, determine the stop reference vector pointing from the steel coil center point P to the strip tail stop point D; (2) Based on the pixel coordinates of the steel coil center point P and the trailing reference point A respectively, determine the trailing rotation vector pointing from the steel coil center point P to the trailing reference point A; (3) Calculate the angle between the stop reference vector and the trailing rotation vector to obtain the remaining coiling angle . The calculation formula of is as follows:

[0108] Next, the rotated angle of the strip tail can be determined as:

[0109] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: The present invention uses an identification and segmentation model, which can identify at least one bounding box including the strip tail in the real-time steel coil image and output a segmentation feature image corresponding to the real-time steel coil image. In order to ensure the positioning accuracy, it is also necessary to accurately screen out the target bounding box including the strip tail from multiple bounding boxes, and during the screening process, based on the ROI region, a preset confidence threshold, and a preset intersection over union threshold, false recognition bounding boxes caused by environmental interference and strip tails on other coiling machines are excluded, ensuring that the strip tail in the final target bounding box is on the coiling machine corresponding to the camera that collects the real-time steel coil image; The present invention uses the segmentation feature image output by the identification and segmentation model and the channel weight vector of the target bounding box to first generate a tail mask image, and then extracts each contour point included in the tail contour position from the tail mask image, realizing accurate extraction of the tail contour; The present invention uses a preset transformation matrix to perform angle correction processing on the real-time steel coil image, and can also only correct the rectangular area where the current steel coil is located to reduce the processing amount and speed up the processing speed, so that in the obtained corrected image, the steel coil area is the image area under the ideal perspective, and a tail contour line composed of the corrected contour points of each contour point in the corrected image is obtained for subsequent angle calculation; Before calculating the angle, the present invention needs to first determine the tail reference point in the corrected image, and then combine the preset center point of the steel coil to calculate the remaining coiling angle or the rotated angle of the strip tail. The obtained accurate remaining coiling angle or rotated angle can accurately reflect the position of the strip tail, which helps the operator or industrial control equipment to timely control the coiling machine to stop coiling based on the remaining coiling angle or the rotated angle, so as to obtain a standard steel coil to be bundled, and it is not easy to loosen the coil.

[0110] In order to execute the corresponding steps in the above method embodiments and various possible implementation manners, the following gives an implementation manner of a positioning device for the strip tail.

[0111] Please refer to Figure 6 , Figure 6 which shows a schematic structural diagram of a positioning device for the strip tail provided by an embodiment of the present invention. The positioning device 200 for the strip tail includes: a model processing module 210, a screening module 220, a contour determination module 230, and an angle calculation module 240.

[0112] The model processing module 210 is configured to: acquire a real-time steel coil image of the current steel coil, and input the real-time steel coil image into a pre-trained identification and segmentation model to obtain an output result; the output result includes a segmentation feature image, as well as the position information, confidence level, and channel weight vector of at least one bounding box in the real-time steel coil image; A screening module 220, configured to determine whether there is a target bounding box among all the bounding boxes based on the position information and confidence of each bounding box; A contour determination module 230, configured to determine the position of the strip tail contour in the real-time steel coil image based on the segmentation feature image and the channel weight vector of the target bounding box when there is a target bounding box; the area where the target bounding box is located in the real-time steel coil image includes the strip tail; An angle calculation module 240, configured to determine the remaining coiling angle of the current steel coil based on the position of the strip tail contour, and the remaining coiling angle reflects the rotation angle required for the reel of the coiling machine to change the current steel coil into a standard steel coil to be bundled.

[0113] Optionally, when the screening module 220 is configured to determine whether there is a target bounding box among all the bounding boxes based on the position information and confidence of each bounding box, it may specifically be configured to: if there is only one pending bounding box with a confidence greater than or equal to a preset confidence threshold among all the bounding boxes and the pending bounding box is located in a preset ROI region, then use the pending bounding box as the target bounding box; if there are multiple pending bounding boxes with a confidence greater than or equal to the preset confidence threshold among all the bounding boxes, then determine the pending bounding box with the highest confidence from all the pending bounding boxes, calculate the intersection over union (IoU) between the pending bounding box with the highest confidence and each other pending bounding box respectively, and use the pending bounding boxes located in the ROI region among each other pending bounding box with an IoU lower than the preset IoU threshold and the pending bounding box with the highest confidence as the target bounding boxes.

[0114] Optionally, when the contour determination module 230 is configured to determine the position of the strip tail contour in the real-time steel coil image based on the segmentation feature image and the channel weight vector of the target bounding box, it may specifically be configured to: generate a strip tail mask image based on the segmentation feature image and the channel weight vector of the target bounding box; extract the position of the strip tail contour of the strip tail in the real-time steel coil image from the strip tail mask image.

[0115] Optionally, when the contour determination module 230 is configured to generate a strip tail mask image based on the segmentation feature image and the channel weight vector of the target bounding box, it may specifically be configured to: perform weighted processing on each channel of the segmentation feature image using each weight coefficient in the channel weight vector of the target bounding box to obtain an initial mask image; perform normalization processing on the pixel value of each pixel point in the initial mask image to obtain an intermediate mask image; after setting the pixel value of each pixel point outside the region where the target bounding box is located in the intermediate mask image to 0, perform binarization processing on the pixel value of each pixel point in the region where the target bounding box is located in the intermediate mask image to obtain a strip tail mask image.

[0116] Optionally, the tail contour position includes the pixel coordinates of each contour point of the strip tail in the real-time coil image. The angle calculation module 240, in the process of determining the remaining coiling angle of the current coil based on the tail contour position, can specifically be used to: perform angle correction processing on the real-time coil image using a preset transformation matrix to obtain a corrected image and the pixel coordinates of each corrected contour point in the corrected image; obtain the pixel coordinates of the preset coil center point in the corrected image; calculate the distance between the coil center point and each corrected contour point, and use the corrected contour point with the farthest distance as the tail reference point of the strip tail; determine the remaining coiling angle of the current coil using the pixel coordinates of the tail reference point and the pixel coordinates of the coil center point.

[0117] Optionally, the real-time coil image is obtained by the camera taking a real-time picture of the coil being wound on the reel of the coiling machine. The angle calculation module 240, in the process of performing angle correction processing on the real-time coil image using a preset transformation matrix to obtain a corrected image and the pixel coordinates of each corrected contour point in the corrected image, can specifically be used to: determine the rectangular area where the current coil is located from the real-time coil image based on the preset coil position information; perform position correction on the pixel coordinates of each pixel point in the rectangular area using the preset transformation matrix to obtain a corrected image and the pixel coordinates of each corrected contour point in the corrected image; where the preset transformation matrix represents the perspective conversion relationship between the actual imaging perspective of the camera and the ideal imaging perspective, and the ideal imaging perspective is the perspective facing the front of the coil on the coiling machine.

[0118] Optionally, the main optical axis corresponding to the coil part in the corrected image coincides with the axis of the reel. The angle calculation module 240, in the process of determining the remaining coiling angle of the current coil using the pixel coordinates of the tail reference point and the pixel coordinates of the coil center point, can specifically be used to: calculate the slope of the reference connection line between the tail reference point and the coil center point based on the pixel coordinates of the tail reference point and the pixel coordinates of the coil center point; obtain the slope of the horizontal line passing through the coil center point and parallel to the horizontal axis of the pixel coordinate system of the corrected image; calculate the angle between the reference connection line and the stop reference line based on the slope of the reference connection line and the slope of the horizontal line to obtain the remaining coiling angle of the current coil, and the stop reference line is the line passing through the preset strip tail stop point and the coil center point in the corrected image.

[0119] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the positioning device 200 of the strip tail described above can refer to the corresponding process in the foregoing method embodiment, and will not be elaborated here.

[0120] Please refer to Figure 6 , Figure 7A schematic structural diagram of an electronic device provided by an embodiment of the present invention. The electronic device 300 includes a processor 310, a memory 320, and a bus 330. The processor 310 is connected to the memory 320 through the bus 330.

[0121] The memory 320 can be used to store software programs. For example, the software program corresponding to the strip tail positioning device 200 provided by the embodiment of the present invention. The processor 310 executes various functional applications and data processing by running the software program stored in the memory 320 to implement the strip tail positioning method provided by the embodiment of the present invention.

[0122] Among them, the memory 320 can be, but is not limited to: RAM (Random Access Memory, random access memory), ROM (Read Only Memory, read-only memory), FLASH (flash memory), PROM (Programmable Read-Only Memory, programmable read-only memory), EPROM (Erasable Programmable Read-Only Memory, erasable read-only memory), EEPROM (Electric Erasable Programmable Read-Only Memory, electrically erasable read-only memory), etc.

[0123] The processor 310 can be an integrated circuit chip with signal processing capabilities. The processor 310 can be a general-purpose processor, including: CPU (Central Processing Unit, central processor), NP (Network Processor, network processor), SoC (System on Chip, system-on-chip), etc.; it can also be: DSP (Digital Signal Processing, digital signal processor), ASIC (Application Specific Integrated Circuit, application-specific integrated circuit), FPGA (Field-Programmable Gate Array, field-programmable gate array) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0124] It can be understood that Figure 7 The structure shown is only schematic. The electronic device 300 may further include more or fewer components than those shown Figure 7 in the figure, or have a different configuration from that shown Figure 7 in the figure. Figure 7 Each component shown in the figure can be implemented by hardware, software, or a combination thereof.

[0125] An embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is run by a processor, the positioning method of the strip tail disclosed in the above embodiment is implemented. The computer-readable storage medium may be, but is not limited to: various media such as USB flash drives, mobile hard disks, ROM, RAM, PROM, EPROM, EEPROM, FLASH magnetic disks, or optical discs that can store program codes.

[0126] An embodiment of the present invention also provides a coiling control system. The coiling control system includes: an image processing device, at least one coiler, and a camera configured for each coiler. The image processing device is communicatively connected to the camera. The camera is used to collect real-time steel coil images during the coiling process of the corresponding coiler. The image processing device is used to receive the real-time steel coil images transmitted by the camera and execute the above positioning method of the strip tail to analyze the received real-time steel coil images to obtain the remaining coiling angle of the coiler.

[0127] In summary, an embodiment of the present invention provides a positioning method, device, electronic device, storage medium, and coiling control system for the strip tail. First, a real-time steel coil image of the current steel coil is obtained, and the real-time steel coil image is input into a pre-trained recognition and segmentation model to obtain a segmented feature image, the position information, confidence, and channel weight vector of at least one bounding box in the real-time steel coil image. Then, based on the position information and confidence of each bounding box, it is determined whether there is a target bounding box among all the bounding boxes. If so, based on the segmented feature image and the channel weight vector of the target bounding box, the position of the strip tail contour is determined in the real-time steel coil image. The area where the target bounding box is located in the real-time steel coil image includes the strip tail. Finally, based on the position of the strip tail contour, the remaining coiling angle of the current steel coil is determined. Since the remaining coiling angle reflects the rotation angle required for the coiler drum to change the current steel coil into a standard steel coil to be strapped, the present invention calculates the remaining rotation angle of the drum in real time when the strip tail appears in the real-time steel coil image, so as to obtain a standard steel coil to be strapped without the need for manual full-time attention to the positioning of the strip tail.

[0128] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A positioning method for the tail of a strip steel, characterized in that, Including: Obtain a real-time coil image of the current steel coil, and input the real-time coil image into a pre-trained recognition and segmentation model to obtain an output result; The output result includes a segmentation feature image, as well as position information, confidence, and channel weight vectors of at least one bounding box in the real-time coil image; Based on the position information and confidence of each bounding box, determine whether there is a target bounding box among all the bounding boxes; If there is, based on the segmentation feature image and the channel weight vector of the target bounding box, determine the position of the strip tail contour in the real-time coil image; the area where the target bounding box is located in the real-time coil image includes the strip tail; Based on the position of the strip tail contour, determine the remaining coiling angle of the current steel coil, and the remaining coiling angle reflects the rotation angle required for the reel of the coiler to change the current steel coil into a standard coil to be strapped.

2. The positioning method of the strip tail according to claim 1, characterized in that, The step of determining whether there is a target bounding box among all the bounding boxes based on the position information and confidence of each bounding box includes: If there is only one pending bounding box with a confidence greater than or equal to a preset confidence threshold among all the bounding boxes and the pending bounding box is located in a preset ROI region, then use the pending bounding box as the target bounding box; If there are multiple pending bounding boxes with a confidence greater than or equal to the preset confidence threshold among all the bounding boxes, then determine the pending bounding box with the highest confidence from all the pending bounding boxes, calculate the intersection over union between the pending bounding box with the highest confidence and each other pending bounding box respectively, and use the pending bounding box located in the ROI region among each other pending bounding box with an intersection over union lower than the preset intersection over union threshold and the pending bounding box with the highest confidence as the target bounding box.

3. The positioning method of the strip tail according to claim 1, characterized in that The step of determining the position of the strip tail contour in the real-time coil image based on the segmentation feature image and the channel weight vector of the target bounding box includes: Generate a strip tail mask image based on the segmentation feature image and the channel weight vector of the target bounding box; Extract the position of the strip tail contour of the strip tail in the real-time coil image from the strip tail mask image.

4. The positioning method of the strip tail according to claim 3, characterized in that The step of generating a strip tail mask image based on the segmentation feature image and the channel weight vector of the target bounding box includes: Use each weight coefficient in the channel weight vector of the target bounding box to perform weighted processing on each channel of the segmentation feature image to obtain an initial mask image; Normalize the pixel value of each pixel point in the initial mask image to obtain an intermediate mask image; After setting the pixel value of each pixel point outside the region where the target bounding box is located in the intermediate mask image to 0, perform binarization processing on the pixel value of each pixel point in the region where the target bounding box is located in the intermediate mask image to obtain the strip tail mask image.

5. The positioning method of the strip tail according to claim 1, wherein The position of the strip tail contour includes the pixel coordinates of each contour point of the strip tail in the real-time coil image; The step of determining the remaining coiling angle of the current steel coil based on the position of the strip tail contour includes: Using a preset transformation matrix, perform angle correction processing on the real-time steel coil image to obtain a corrected image and the pixel coordinates of each corrected contour point in the corrected image; Obtain the pixel coordinates of the preset center point of the steel coil in the corrected image; Calculate the distance between the center point of the steel coil and each corrected contour point, and use the corrected contour point with the farthest distance as the tail reference point of the strip steel tail; Use the pixel coordinates of the tail reference point and the pixel coordinates of the center point of the steel coil to determine the remaining coiling angle of the current steel coil.

6. The positioning method of the strip tail according to claim 5, characterized in that The real-time steel coil image is obtained by the camera taking a real-time picture of the steel coil being coiled on the reel of the coiling machine; The step of using a preset transformation matrix to perform angle correction processing on the real-time steel coil image to obtain a corrected image and the pixel coordinates of each corrected contour point in the corrected image includes: Based on the preset position information of the steel coil, determine the rectangular area where the current steel coil is located from the real-time steel coil image; Use the preset transformation matrix to perform position correction on the pixel coordinates of each pixel point in the rectangular area to obtain the corrected image and the pixel coordinates of each corrected contour point in the corrected image; wherein, the preset transformation matrix represents the perspective conversion relationship between the actual imaging perspective and the ideal imaging perspective of the camera, and the ideal imaging perspective is the perspective directly facing the front of the steel coil on the coiling machine.

7. The positioning method for the strip tail according to claim 5, characterized in that, The imaging principal axis corresponding to the steel coil part in the corrected image coincides with the axis of the reel; The step of using the pixel coordinates of the tail reference point and the pixel coordinates of the center point of the steel coil to determine the remaining coiling angle of the current steel coil includes: Based on the pixel coordinates of the tail reference point and the pixel coordinates of the center point of the steel coil, calculate the slope of the reference connection line between the tail reference point and the center point of the steel coil; Obtain the slope of the horizontal line passing through the center point of the steel coil and parallel to the horizontal axis of the pixel coordinate system of the corrected image; Based on the slope of the reference connection line and the slope of the horizontal line, calculate the angle between the reference connection line and the stop reference line to obtain the remaining coiling angle of the current steel coil, and the stop reference line is the line passing through the preset strip tail stop point and the center point of the steel coil in the corrected image.

8. A positioning device for the tail of a strip steel, characterized in that, Includes: A model processing module for: obtaining a real-time steel coil image of the current steel coil and inputting the real-time steel coil image into a pre-trained recognition and segmentation model to obtain an output result; the output result includes a segmentation feature image and the position information, confidence level, and channel weight vector of at least one bounding box in the real-time steel coil image; A screening module for determining whether there is a target bounding box among all the bounding boxes based on the position information and confidence level of each bounding box; A contour determination module for, when there is a target bounding box, determining the position of the strip tail contour in the real-time steel coil image based on the segmentation feature image and the channel weight vector of the target bounding box; the area where the target bounding box is located in the real-time steel coil image includes the strip steel tail; An angle calculation module, configured to determine a remaining coiling angle of the current steel coil based on the position of the tail profile, where the remaining coiling angle reflects the rotation angle required for the reel of the coiler to change the current steel coil into a standard steel coil to be strapped.

9. An electronic device, characterized in that, Comprising: A memory and a processor, where the memory stores a software program, and when the electronic device runs, the processor executes the software program to implement the positioning method of the strip tail according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the positioning method of the strip tail according to any one of claims 1-7 is implemented.

11. A coiling control system, characterized in that, The system coiling control system includes: At least one coiler and a camera configured for each coiler, where the camera is used to collect real-time steel coil images during the coiling process of the corresponding coiler; An image processing device, configured to receive the real-time steel coil images transmitted by the camera and execute the positioning method of the strip tail according to any one of claims 1-7 to analyze the received real-time steel coil images to obtain the remaining coiling angle of the coiler.

Images