Optimal selection method and system for process data of spraying and winding polyethylene thermal insulation pipe

Abstract

The invention provides a polyethylene thermal insulation pipe spraying and winding process data optimization method and system, and the method comprises the steps: putting a processing pipe into a processing device, setting a processing region, obtaining real-time processing images of the processing pipe in a spraying region and a winding region during spraying and winding, and carrying out the extraction of ROI processing images, carrying out image splicing operation to obtain a real-time spliced processing image, obtaining definition data, carrying out judgment operation based on the definition data, if a difference value is smaller than 0, obtaining an optimal control gain by a proportional controller based on root locus optimization of a processing transfer function, and carrying out adjustment based on the optimal control gain; the real-time processing image is obtained again, related steps are executed, when the difference value is larger than or equal to 0, spraying evaluation data and winding evaluation data are obtained based on the real-time splicing processing image, real-time optimization adjustment is conducted on the processing device, better process data can be obtained through clearer images, and the precision of the subsequent processing process is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of pipeline processing, and more specifically, particularly relates to a method and system for optimizing process data of a sprayed and wound polyethylene insulation pipe. Background Technique

[0002] Sprayed and wound polyethylene insulation pipes are widely used in scenarios such as central heating, oil and gas transportation, and chemical medium transportation due to their good heat insulation performance, corrosion resistance, and long service life. Whether it is the heating pipe network in northern cities in winter or long-distance oil and gas transportation pipelines, polyethylene insulation pipes play an important role in reducing heat or medium loss and ensuring safe and stable transportation. Their quality directly affects the energy transportation efficiency and the operation cost of the pipe network.

[0003] Currently, the production process of sprayed and wound polyethylene insulation pipes mainly relies on fixed image shooting and fixed parameter settings. During the spraying or winding process, problems with poor accuracy often occur in the subsequent process due to image deviation during processing. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention provides a method and system for optimizing process data of a sprayed and wound polyethylene insulation pipe to solve the technical problem that in the prior art, the traditional processing process of insulation pipes often has low accuracy due to image deviation during processing.

[0005] The purpose and effect of a method and system for optimizing process data of a sprayed and wound polyethylene insulation pipe of the present invention are achieved by the following specific technical means: A method for optimizing process data of a sprayed and wound polyethylene insulation pipe includes the following steps: S1: Place the processing pipe into the processing device, and the processing device processes the processing pipe according to the corresponding processing technology; S2: Set the processing area, and the processing area is divided into a spraying area and a winding area; S3: Based on the processing device, obtain the real-time processing images of the processing pipe during spraying and winding operations in the spraying area and the winding area, extract the ROI processing images from the real-time processing images, and perform image stitching operations on the ROI processing images to obtain real-time stitched processing images; S4: Calculate the clarity of the real-time stitched processing image based on the gradient method, obtain clarity data, subtract the clarity data from the preset clarity threshold, and use the obtained difference as feedback data to feedback to the proportional controller and perform a judgment operation; S41: If the difference is less than 0, the proportional controller obtains the optimal control gain based on the root locus optimization of the processing transfer function. The optimal control gain includes the optimal exposure time and the processing tube processing speed. The proportional controller adjusts according to the optimal control gain and executes step S3; S5: When the difference is greater than or equal to 0, spray evaluation data and winding evaluation data are obtained based on the real-time stitched processing image, and the process data of the processing device is optimized and adjusted in real time based on the spray evaluation data and the winding evaluation data.

[0006] As a further solution of the present invention, if the difference is less than 0, the proportional controller obtains the optimal control gain based on the root locus optimization of the processing transfer function. The optimal control gain includes the optimal exposure time and the processing tube processing speed. The proportional controller adjusts according to the optimal control gain, including: Obtain the exposure time and the processing tube processing speed, import the exposure time and the processing tube processing speed as inputs into the proportional controller, establish a basic model for each input and analyze the coupling relationship, obtain the relative gain matrix, confirm the optimal control structure according to the relative gain matrix, and add a feedforward compensator to cancel the coupling effect on the optimal control structure. A multi-input model is established based on the optimal control structure; Establish a processing transfer function based on the multi-input model, draw the root locus diagram of the processing transfer function, and obtain multiple processing poles. Based on the multiple processing poles, obtain the dominant processing pole. The dominant processing pole represents the processing pole that is closest to the imaginary axis and has no zero points nearby in the root locus diagram; Simplify the processing transfer function based on the dominant processing pole, calculate the processing steady-state error of the simplified processing transfer function, obtain the optimal control gain according to the processing steady-state error. The optimal control gain includes the optimal exposure time and the processing tube processing speed. The proportional controller adjusts according to the optimal control gain.

[0007] As a further solution of the present invention, the method further includes: Obtain the blur threshold and the brightness threshold based on the preset clarity threshold, and obtain the blur and brightness of the real-time stitched processing image through the proportional controller; If the blur is greater than the blur threshold and the brightness is greater than the brightness threshold, the proportional controller reduces the processing tube speed; If the blur is less than the blur threshold and the brightness is less than the brightness threshold, the proportional controller increases the exposure time; If the blur is greater than the blur threshold and the brightness is less than the brightness threshold, the proportional controller reduces the processing tube speed and increases the exposure time at the same time.

[0008] As a further solution of the present invention, the real-time processing images of the processing pipes in the spraying area and the winding area during spraying work and winding work are obtained based on the processing device, the ROI processing images are extracted from the real-time processing images, and image stitching operations are performed on the ROI processing images to obtain real-time stitched processing images, including: Dynamically locate the processing pipes during spraying work and winding work in the real-time processing images through a lightweight model, and generate initial ROI candidate boxes; Based on the target tracking algorithm, continuously update the positions and sizes of the ROI candidate boxes in each real-time processing image; Based on the ROI candidate boxes, dynamically extract the ROI processing images in each real-time processing image, and perform registration alignment on each ROI processing image; Obtain the overlapping parts between every four registered and aligned ROI processing images, eliminate the overlapping parts based on hybrid fusion, and stitch the four ROI processing images to generate real-time stitched processing images.

[0009] As a further solution of the present invention, the obtaining of the overlapping parts between every four registered and aligned ROI processing images, eliminating the overlapping parts based on hybrid fusion, and stitching the four ROI processing images to generate real-time stitched processing images includes: Obtain the distance distribution weights of the pixel positions in each ROI processing image from the overlapping boundaries, and perform smooth transition on each ROI processing image based on the distance distribution weights; Based on the image pyramid, decompose each ROI processing image that has completed smooth transition into high-frequency parts, medium-frequency parts, and low-frequency parts. The low-frequency part is represented as the most blurred layer, the high-frequency part is represented as the clearest layer, and the medium-frequency part is represented as the medium clarity layer. Perform hierarchical fusion on the high-frequency part, medium-frequency part, and low-frequency part to obtain the real-time stitched processing image to be optimized; Construct an energy function, find the seam path with the smallest difference in the real-time stitched processing image to be optimized based on the energy function, and eliminate the ghosting in the real-time stitched processing image to be optimized according to the seam path with the smallest difference in the hierarchical fusion; Perform global optimization processing on the real-time stitched processing image to be optimized to generate a real-time stitched processing image.

[0010] As a further solution of the present invention, the clarity calculation of the real-time stitched processing image based on the gradient method to obtain clarity data includes: Calculate the clear gradient amplitudes of the X-axis and Y-axis of the 5×5 window in each real-time stitched processing image based on the Sobel operator with a 5×5 kernel. The clear gradient amplitudes are calculated based on the sum of squares, and normalize the clear gradient amplitudes of the X-axis and Y-axis of the 5×5 window in each real-time stitched processing image; The normalization formula is expressed as: ; Wherein, represents the clear gradient amplitude of the X-axis and Y-axis of the 5×5 window after normalization, represents the average value of the clear gradient amplitude within the th window, represents the standard deviation of the clear gradient amplitude within the th window, represents the clear gradient amplitude of the X-axis and Y-axis of the 5×5 window before normalization; Calibrate the process key area, and the process key area is expressed as the area of 80% of the width of the spraying center of the processed pipe and the area of ±10 mm on both sides of the winding joint; Based on the process key area, perform the summation calculation of the clear gradient amplitude to obtain the sum of the clear gradient amplitude, and linearly map the sum of the clear gradient amplitude to the range of 0-100% to obtain the clarity data.

[0011] As a further solution of the present invention, the summation calculation of the clear gradient amplitude based on the process key area to obtain the sum of the clear gradient amplitude includes: After performing the summation calculation of the clear gradient amplitude, obtain the processing blur length based on the exposure time and the processing rotation speed of the processed pipe, and perform noise optimization on the sum of the clear gradient amplitude based on the noise suppression function; The noise suppression function can be expressed as: ; Wherein, represents the sum of the clear gradient amplitude after noise optimization, represents the sum of the clear gradient amplitude, represents the processing blur length.

[0012] As a further solution of the present invention, the obtaining of the spraying evaluation data and the winding evaluation data based on the real-time stitched processing image includes: The real-time stitched processing image includes the real-time stitched processing image of the processed pipe during the spraying operation and the real-time stitched processing image of the processed pipe during the winding operation; Obtain the spraying characteristics of the processed pipe based on the real-time stitched processing image of the processed pipe during the spraying operation, and the spraying characteristics of the processed pipe include defect characteristics and thickness characteristics. Obtain the winding characteristics of the processed pipe based on the real-time stitched processing image of the processed pipe during the winding operation, and the winding characteristics of the processed pipe are expressed as the winding angle deviation. Obtain the spraying evaluation data and the winding evaluation data according to the spraying characteristics of the processed pipe and the winding characteristics of the processed pipe.

[0013] As a further solution of the present invention, obtaining the spraying evaluation data and the winding evaluation data according to the spraying characteristics and the winding characteristics of the processed pipe includes: Establish a scoring model, import the spraying characteristics and the winding characteristics of the processed pipe into the scoring model, and obtain the spraying evaluation data and the winding evaluation data through the scoring model; The spraying evaluation data is expressed as a score for spraying processing, which is between 0 and 100, and the winding evaluation data is expressed as a score for winding processing, which is between 0 and 100.

[0014] A process data optimization system for spraying and winding polyethylene insulating pipes includes: A processing device, which includes a sensor module, an image module and a processing module; A sensor module, which is used to obtain the process data of the processing device; A processing module, which processes the processed pipe according to the corresponding processing process steps; An image module, which is used to photograph and collect the processed pipe during the processing process, and obtain the real-time processing images of the sprayed area and the wound area of the processed pipe for spraying work and winding work; A processing module, which is used to extract the ROI processing images from the real-time processing images and obtain the real-time spliced processing images; A proportional controller, which can obtain the optimal control gain based on the root locus optimization of the processing transfer function and can be adjusted according to the optimal control gain; An adjustment module, which is used to obtain the spraying evaluation data and the winding evaluation data, and can perform real-time optimization adjustment on the process data of the processing device based on the spraying evaluation data and the winding evaluation data.

[0015] Compared with the prior art, the present invention has the following beneficial effects: First, place the processing pipe into the processing device through step S1, set the processing area through step S2, and then obtain the real-time processing images of the processing pipe in the spraying area and the winding area during the spraying work and the winding work through step S3. Extract the ROI processing images from the real-time processing images and obtain the real-time spliced processing images. Then, perform sharpness calculation through step S4 to obtain sharpness data, subtract the sharpness data from the preset sharpness threshold, and perform a judgment operation on the difference to determine whether it is necessary to execute step S41 to adjust the control gain. If the difference is less than 0, execute step S41 to obtain the optimal control gain through the root locus optimization of the processing transfer function. The proportional controller is adjusted based on the optimal control gain to improve the sharpness of the obtained real-time processing images, so as to optimize and adjust the process data of the subsequent processing device. If the difference is greater than or equal to 0, execute step S5 to obtain the spraying evaluation data and the winding evaluation data based on the real-time spliced processing images, and perform real-time optimization and adjustment of the process data of the processing device according to the spraying evaluation data and the winding evaluation data, thereby improving the accuracy of the subsequent processing process and solving the problem of low accuracy often caused by image deviation during processing in the traditional processing process of thermal insulation pipes. Description of the Drawings

[0016] Figure 1 is a flowchart of the steps of a method for optimizing process data of a spraying and winding polyethylene thermal insulation pipe according to the present invention; Figure 2 is a schematic diagram of a system for optimizing process data of a spraying and winding polyethylene thermal insulation pipe according to the present invention; Figure 3 is a schematic diagram of the processing device in a system for optimizing process data of a spraying and winding polyethylene thermal insulation pipe according to the present invention. Detailed Embodiments

[0017] The following further describes in detail the embodiments of the present invention in conjunction with the drawings and embodiments. The following embodiments are used to illustrate the technical solutions of the present invention, but cannot be used to limit the protection scope of the present invention.

[0018] Embodiment 1: As shown in the attached Figure 1 , Figure 2 , Figure 3 : The present invention provides a method for optimizing process data of a spraying and winding polyethylene thermal insulation pipe, which is applicable to the processing of spraying and winding polyethylene thermal insulation pipes, and includes the following steps: Step S1, place the processing pipe into the processing device, and the processing device processes the processing pipe according to the corresponding processing technology.

[0019] Step S2, set the processing area, and the processing area is divided into a spraying area and a winding area.

[0020] Specifically, the relevant process flows for performing spraying work within the spraying area and the relevant process flows for performing winding work within the winding area.

[0021] Step S3: Based on the processing device, obtain the real-time processing images of the processing pipe during spraying work and winding work within the spraying area and the winding area, extract the ROI processing images from the real-time processing images, and perform image stitching operations on the ROI processing images to obtain the real-time stitched processing images.

[0022] In this embodiment, step S3 includes: Step S3-1: Dynamically locate the processing pipe during spraying work and winding work in the real-time processing image through a lightweight model, and generate an initial ROI candidate box.

[0023] Specifically, lightweight models such as YOLO and MobileNet-SSD can be used to dynamically locate the processing pipe during spraying work and winding work in the real-time processing image and generate an initial ROI candidate box.

[0024] Step S3-2: Continuously update the position and size of the ROI candidate box in each real-time processing image based on the target tracking algorithm.

[0025] Step S3-3: Dynamically extract the ROI processing image in each real-time processing image based on the ROI candidate box, and perform registration alignment on each ROI processing image.

[0026] It should be noted that the core principle of dynamic ROI extraction is to capture the dynamically changing region of interest in the real-time processing image through real-time detection and tracking technologies, such as the processing pipe during spraying work and winding work, and use the spatio-temporal correlation between multiple frames to achieve precise alignment and seamless fusion.

[0027] Specifically, first, a lightweight object detection model, such as YOLO, is used to locate the ROI candidate box, and then the tracking algorithm is used to continuously update its position and shape to ensure temporal consistency. Subsequently, the ROI processing image in each real-time processing image is dynamically extracted based on the ROI candidate box, and geometric registration is performed on the ROI processing image based on feature matching and motion model prediction to eliminate the perspective and displacement differences.

[0028] Step S3-4: Obtain the overlapping part between every four registered and aligned ROI processing images, eliminate the overlapping part based on hybrid fusion, and stitch the four ROI processing images to generate the real-time stitched processing image.

[0029] In this embodiment, step S3-4 includes: Step S3-4-1: Obtain the weights assigned according to the distances between the pixel positions in each ROI processing image and the overlapping boundaries, and perform smooth transition on each ROI processing image based on the weights assigned according to the distances.

[0030] In this embodiment, the method of fading in and fading out is adopted to perform smooth transition on each ROI processing image, and weights are dynamically assigned according to the distances between the pixels in the overlapping area and the boundaries. For example, the linear weights fade from 0 to 1 to smooth the splicing gap. When using linear weights, blurring may occur in areas with complex textures. Therefore, gradient-sensitive weights can be introduced. If the gradient value of a certain pixel is higher than the threshold, its weight can be increased to retain edge details.

[0031] Step S3-4-2: Decompose each ROI processing image that has completed smooth transition into high-frequency part, medium-frequency part and low-frequency part based on the image pyramid. The low-frequency part is represented as the most blurred layer, the high-frequency part is represented as the clearest layer, and the medium-frequency part is represented as the medium-clarity layer. Perform hierarchical fusion on the high-frequency part, medium-frequency part and low-frequency part to obtain the real-time splicing processing image to be optimized.

[0032] In this embodiment, the high-frequency, low-frequency and medium-frequency image information is decomposed through the image pyramid and fused separately to avoid blurring and color difference. For the high-frequency part, that is, the clearest layer, select the pixels with the largest gradient for hierarchical fusion to retain texture details. For the medium-frequency part, that is, the medium-clarity layer, select the method of weighted mixing for hierarchical fusion. For the low-frequency part, that is, the most blurred layer, select the method of average fusion for hierarchical fusion to eliminate illumination differences.

[0033] Step S3-4-3: Construct an energy function, search for the seam path with the smallest difference in the real-time splicing processing image to be optimized based on the energy function, and eliminate the ghosting in the real-time splicing processing image to be optimized according to the seam path with the smallest difference in the hierarchical fusion.

[0034] In this embodiment, by constructing an energy function, use the maximum flow algorithm to search for the seam path with the smallest difference in the splicing area and eliminate the ghosting.

[0035] It can be understood that for the overlapping parts between multiple ROI processed images, a hybrid strategy based on spatial or temporal weights is adopted. For example, when using fade-in and fade-out, weights are dynamically allocated according to the distance between the pixel position and the overlapping boundary. The weight at the boundary is 0.5, and when far from the boundary, the weight gradually changes towards 1 or 0 to achieve a smooth transition. For multi-band fusion, the image is decomposed by Laplacian pyramid, and hierarchical fusion is performed by using average fusion in the low-frequency part (i.e., the most blurred layer) to eliminate illumination differences. In the high-frequency part (i.e., the clearest layer), the maximum gradient value is directly retained to maintain detail clarity. In the middle-frequency part (i.e., the medium clarity layer), for the small misalignments caused by motion blur or registration errors, the optimal seam finding technology can be used to construct an energy function using the graph cut algorithm. The energy function contains constraints such as color difference and gradient continuity, and the path with the smallest difference in the overlapping area of the splice is found as the seam, thus avoiding ghosting or fracture phenomena.

[0036] Step S3-4-4: Perform global optimization processing on the real-time stitching processed image to be optimized to generate a real-time stitching processed image.

[0037] In this embodiment, after the fusion and stitching are completed, it is necessary to ensure the geometric and illumination consistency between the stitching result and the global canvas. Poisson fusion can be used to solve the Poisson equation, combine the gradient field of the ROI processed image with the boundary conditions of the target canvas, and while retaining the details of the original ROI processed image, force the color transition at the stitching boundary to be natural. For the ROI processed images that are frequently updated in a dynamic scene, an incremental canvas update strategy is adopted, and only the changed areas are re-rendered by dirty rectangle marking or motion vector detection, rather than refreshing the entire canvas, reducing the computational load. In addition, an illumination equalization mechanism can be introduced to perform histogram matching or brightness correction on each ROI to make it consistent with the illumination conditions of the global canvas and avoid sudden brightness changes caused by environmental changes.

[0038] Specifically, the solution of the Poisson equation can be expressed as: ; where is the Laplacian operator, represents the gradient field of the ROI processed image, represents the solution of the Poisson equation. When solving the Poisson equation, it is discretized into a linear equation system and solved using the preconditioned conjugate gradient method or the multigrid method.

[0039] Specifically, the illumination equalization mechanism is expressed as extracting the RGB histograms of the ROI processed image and the reference area, mapping through the cumulative distribution function, matching the histogram of the ROI processed image to the reference histogram, restricting the contrast of the ROI processed image with uneven illumination, and performing adaptive histogram equalization.

[0040] Step S4, calculate the sharpness of the real-time stitching and processing image based on the gradient method, obtain the sharpness data, subtract the sharpness data from the preset sharpness threshold, use the obtained difference as feedback data to feedback to the proportional controller, and perform a judgment operation.

[0041] In this embodiment, step S4 includes: Step S4-1, calculate the clear gradient amplitudes of the X-axis and Y-axis of the 5×5 window in each real-time stitching and processing image based on the Sobel operator with a 5×5 kernel. The clear gradient amplitudes are calculated based on the sum of squares, and normalize the clear gradient amplitudes of the X-axis and Y-axis of the 5×5 window in each real-time stitching and processing image.

[0042] The normalization formula is expressed as: ; where, represents the clear gradient amplitudes of the X-axis and Y-axis of the 5×5 window after normalization, represents the mean value of the clear gradient amplitudes in the th window, represents the standard deviation of the clear gradient amplitudes in the th window, represents the clear gradient amplitudes of the X-axis and Y-axis of the 5×5 window before normalization.

[0043] It can be understood that when detecting minute textures, such as the orientation of polyethylene films, the 3×3 kernel is prone to noise interference, and for large-scale features, such as the response to the distribution of spraying particles is insufficient. Therefore, the horizontal and vertical gradients are extended to a 5×5 kernel to enhance the sensitivity to medium-frequency features.

[0044] For example, the horizontal kernel can be expressed as: ; where, represents the horizontal kernel, which enhances the edge contrast by increasing the central weight and suppresses noise by smoothing the distal weights. When calculating with the horizontal kernel, the horizontal kernel is optimized by using separable convolution, and the 5×5 kernel is decomposed into the product of 5×1 and 1×5 vectors, thereby reducing the number of calculations.

[0045] It can be understood that when calculating the clear gradient amplitudes, calculating based on the sum of squares can avoid the open operation of the traditional gradient magnitude and can retain the relative intensity.

[0046] Step S4-2, calibrate the key process area, and the key process area is expressed as the area of 80% of the width of the spraying center of the processing pipe and the area of ±10 mm on both sides of the winding seam.

[0047] It is understandable that only the clarity gradient amplitude of the process critical area can exclude the interference at the pipe end.

[0048] Step S4-3: Calculate the sum of the clarity gradient amplitudes based on the process critical area to obtain the sum of the clarity gradient amplitudes, linearly map the sum of the clarity gradient amplitudes to the range of 0-100% to obtain the clarity data.

[0049] Furthermore, after calculating the sum of the clarity gradient amplitudes, obtain the processing blur length based on the exposure time and the processing pipe rotation speed, and optimize the noise of the sum of the clarity gradient amplitudes based on the noise suppression function.

[0050] The noise suppression function can be expressed as: ; Wherein, represents the sum of the clarity gradient amplitudes after noise optimization, represents the sum of the clarity gradient amplitudes, represents the processing blur length.

[0051] It should be noted that the gradient field quantization calculation performs high-sensitivity gradient analysis on the real-time stitched processing image through an extended 5×5 kernel Sobel operator, captures the microscopic texture changes of the sprayed coating and the wound fibers. Its core principle is to use a weighted gradient kernel to enhance the response ability to intermediate-frequency features, such as the fiber orientation of the polyethylene film and the distribution of spraying particles, calculate the gradient energy through the sum of squares to replace the traditional modulus operation to improve the calculation efficiency, and combine normalization to eliminate the illumination interference, and suppress the periodic texture interference through the noise suppression function to ensure that the quantization result is strongly correlated with the real process quality.

[0052] Step S41: If the difference is less than 0, the proportional controller obtains the optimal control gain based on the root locus optimization of the processing transfer function. The optimal control gain includes the optimal exposure time and the optimal processing pipe rotation speed. The proportional controller adjusts according to the optimal control gain and executes step S3.

[0053] In this embodiment, step S41 includes: Step S41-1: Obtain the exposure time and the processing pipe rotation speed, import the exposure time and the processing pipe rotation speed as inputs into the proportional controller, establish a basic model for each input and analyze the coupling relationship, obtain the relative gain matrix, confirm the optimal control structure according to the relative gain matrix, and add a feedforward compensator to cancel the coupling effect on the optimal control structure, and establish a multi-input model based on the optimal control structure.

[0054] Step S41-2: Based on the multi-input model, establish a processing transfer function, plot the root locus of the processing transfer function, and obtain multiple processing poles. Based on the multiple processing poles, obtain the dominant processing poles, where the dominant processing poles refer to the processing poles that are closest to the imaginary axis and have no zeros nearby in the root locus diagram.

[0055] Step S41-3: Simplify the processing transfer function based on the dominant processing poles, calculate the steady-state error of the simplified processing transfer function, and obtain the optimal control gain according to the processing steady-state error. The optimal control gain includes the optimal exposure time and the rotational speed of the processing tube. The proportional controller is adjusted according to the optimal control gain.

[0056] In this embodiment, the processing transfer function can be expressed as: ; where , , , are the processing poles, is the gain; Based on the above processing poles, plot the root locus diagram, analyze the movement locus of the processing poles when the gain K changes, determine the possible pole distributions. Assume , , , these four points are the four processing poles of the processing transfer function, and based on these four poles, obtain the dominant processing poles. The dominant processing poles are the poles that are closest to the imaginary axis and have no zeros nearby. Then and are the dominant processing poles, , can be ignored. Simplify the processing transfer function based on the dominant processing poles, simplify the original fourth-order closed-loop transfer function system to a second-order system, and then calculate the optimal control gain according to the steady-state error of its second-order closed-loop transfer function. The proportional controller is adjusted based on the optimal control gain.

[0057] In this embodiment, the specific method for adjusting the control gain is as follows: Obtain the blur threshold and the brightness threshold based on the preset clarity threshold, and obtain the blur and brightness of the real-time stitched processing image through the proportional controller; If the blur is greater than the blur threshold and the brightness is greater than the brightness threshold, the proportional controller reduces the rotational speed of the processing tube; If the blur is less than the blur threshold and the brightness is less than the brightness threshold, the proportional controller increases the exposure time; If the blur is greater than the blur threshold and the brightness is less than the brightness threshold, the proportional controller reduces the rotational speed of the processing tube and increases the exposure time simultaneously.

[0058] In step S5, when the difference is greater than or equal to 0, spray evaluation data and winding evaluation data are obtained based on the real-time spliced processing image, and the process data of the processing device is adjusted in real time based on the spray evaluation data and the winding evaluation data.

[0059] Specifically, the real-time spliced processing image includes a real-time spliced processing image of the processing pipe during spraying work and a real-time spliced processing image of the processing pipe during winding work.

[0060] Specifically, the spraying characteristics of the processing pipe are obtained based on the real-time spliced processing image of the processing pipe during spraying work. The spraying characteristics of the processing pipe include defect characteristics and thickness characteristics. The winding characteristics of the processing pipe are obtained based on the real-time spliced processing image of the processing pipe during winding work. The winding characteristics of the processing pipe are expressed as the winding angle deviation. The spray evaluation data and the winding evaluation data are obtained according to the spraying characteristics of the processing pipe and the winding characteristics of the processing pipe.

[0061] Furthermore, a scoring model is established. The spraying characteristics of the processing pipe and the winding characteristics of the processing pipe are imported into the scoring model. The spray evaluation data and the winding evaluation data are obtained through the scoring model. The spray evaluation data is expressed as the score of the spraying process, which is between 0 and 100. The winding evaluation data is expressed as the score of the winding process, which is between 0 and 100.

[0062] Furthermore, the scoring model can be expressed as: ; ; Among them, represents the thickness characteristic, represents the number of defect characteristics, represents the winding characteristics of the processing pipe. The process data is adjusted in real time based on the built-in process data in the adjustment module, the spray evaluation data, and the winding evaluation data to improve the accuracy of the process.

[0063] A process data optimization system for spraying and winding polyethylene insulation pipes, which is applicable to the processing of spraying and winding polyethylene insulation pipes, includes: A processing device, which includes a sensor module, an image module, and a processing module; A sensor module, which is used to obtain the process data of the processing device; A processing module, which processes the processing pipe according to the corresponding processing process steps; An image module, which is used to photograph and collect the processing pipe during the processing process to obtain a real-time processing image; A processing module, which is used to obtain real-time local processing images, real-time stitched processing images and clarity data, and can perform judgment operations based on the clarity data; A proportional controller, which can obtain control gain based on the root locus optimization of the processing transfer function and can adjust the control; An adjustment module, which is used to obtain spraying evaluation data and winding evaluation data, and can perform real-time optimization adjustment on the process data of the processing device based on the spraying evaluation data and the winding evaluation data.

[0064] The specific usage mode and function of the first embodiment: First, put the processing pipe into the processing device through step S1, set the processing area through step S2, then obtain the real-time processing images of the processing pipe in the spraying area and the winding area during the spraying work and the winding work through step S3, extract the ROI processing images from the real-time processing images, and obtain the real-time stitched processing images. Then, perform clarity calculation through step S4 to obtain clarity data, subtract the clarity data from the preset clarity threshold, and perform a judgment operation on the difference to determine whether it is necessary to execute step S41 to adjust the control gain. If the difference is less than 0, then execute step S41, obtain the optimal control gain through the root locus optimization of the processing transfer function, and the proportional controller adjusts based on the optimal control gain to improve the clarity of the obtained real-time processing images, so as to optimize and adjust the process data of the subsequent processing device. If the difference is greater than or equal to 0, then execute step S5, obtain the spraying evaluation data and the winding evaluation data based on the real-time stitched processing images, and perform real-time optimization adjustment on the process data of the processing device according to the spraying evaluation data and the winding evaluation data, thereby improving the accuracy of the subsequent processing process and solving the problem of low accuracy often caused by image deviation during processing in the traditional processing process of thermal insulation pipes.

[0065] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wired (such as infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that contains one or more collections of available media. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0066] It should be understood that the term “and / or” in this document is merely a description of the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone. Here, A and B can be singular or plural. In addition, the character “ / ” in this document generally represents an “or” relationship between the associated objects before and after, but it may also represent an “and / or” relationship, which can be specifically understood with reference to the context.

[0067] It should be understood that in the embodiments of the present invention, the magnitudes of the serial numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0068] The above-described embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for optimizing process data of a sprayed and wound polyethylene insulation pipe, characterized in that, It includes the following steps: S1: Place the processing pipe into the processing device, and the processing device processes the processing pipe according to the corresponding processing technology; S2: Set the processing area, and the processing area is divided into a spraying area and a winding area; S3: Based on the processing device, obtain the real-time processing images of the processing pipe in the spraying area and the winding area during the spraying work and the winding work, extract the ROI processing images from the real-time processing images, and perform an image stitching operation on the ROI processing images to obtain the real-time stitched processing images; S4: Calculate the clarity of the real-time stitched processing images based on the gradient method, obtain the clarity data, subtract the clarity data from the preset clarity threshold, use the obtained difference as the feedback data to feedback to the proportional controller, and perform a judgment operation; S41: If the difference is less than 0, the proportional controller obtains the optimal control gain based on the root locus optimization of the processing transfer function. The optimal control gain includes the optimal exposure time and the optimal processing pipe processing speed. The proportional controller adjusts according to the optimal control gain and executes step S3; S5: When the difference is greater than or equal to 0, obtain the spraying evaluation data and the winding evaluation data based on the real-time stitched processing images, and perform real-time optimization adjustment on the process data of the processing device based on the spraying evaluation data and the winding evaluation data.

2. The preferred method for process data of a spray-wound polyethylene insulation pipe according to claim 1, characterized in that, If the difference is less than 0, the proportional controller obtains the optimal control gain based on the root locus optimization of the processing transfer function. The optimal control gain includes the optimal exposure time and the optimal processing pipe processing speed. The proportional controller adjusts according to the optimal control gain, including: Obtain the exposure time and the processing pipe processing speed, import the exposure time and the processing pipe processing speed as inputs into the proportional controller, establish a basic model for each input and analyze the coupling relationship, obtain the relative gain matrix, confirm the optimal control structure according to the relative gain matrix, and add a feedforward compensator to cancel the coupling effect on the optimal control structure. Establish a multi-input model based on the optimal control structure; Establish a processing transfer function based on the multi-input model, draw the root locus diagram of the processing transfer function, and obtain multiple processing poles. Obtain the dominant processing pole based on the multiple processing poles. The dominant processing pole represents the processing pole that is closest to the imaginary axis and has no zero points nearby in the root locus diagram; Simplify the processing transfer function based on the dominant processing pole, calculate the processing steady-state error of the simplified processing transfer function, obtain the optimal control gain according to the processing steady-state error. The optimal control gain includes the optimal exposure time and the optimal processing pipe processing speed. The proportional controller adjusts according to the optimal control gain.

3. A method for optimizing process data of a sprayed and wound polyethylene insulation pipe according to claim 2, characterized in that, The method further includes: Obtain the blurriness threshold and the brightness threshold based on the preset clarity threshold, and obtain the blurriness and brightness of the real-time stitched processing images through the proportional controller; If the blurriness is greater than the blurriness threshold and the brightness is greater than the brightness threshold, the proportional controller reduces the processing pipe speed; If the blurriness is less than the blurriness threshold and the brightness is less than the brightness threshold, the proportional controller increases the exposure time; If the ambiguity is greater than the ambiguity threshold and the brightness is less than the brightness threshold, the proportional controller reduces the rotational speed of the processing pipe while increasing the exposure time.

4. The preferred method for process data of a spray-wound polyethylene insulation pipe according to claim 1, characterized in that Based on the processing device, the real-time processing images of the processing pipe during spraying work and winding work in the spraying area and the winding area are obtained. The ROI processing images are extracted from the real-time processing images, and the ROI processing images are subjected to image stitching operations to obtain real-time stitched processing images, including: Dynamically locate the processing pipe during spraying work and winding work in the real-time processing image through a lightweight model, and generate an initial ROI candidate box; Based on the target tracking algorithm, continuously update the position and size of the ROI candidate box in each real-time processing image; Dynamically extract the ROI processing image in each real-time processing image based on the ROI candidate box, and perform registration alignment on each ROI processing image; Obtain the overlapping parts between every four registered and aligned ROI processing images, eliminate the overlapping parts based on hybrid fusion, and stitch the four ROI processing images to generate a real-time stitched processing image.

5. The preferred method for process data of a spray-wound polyethylene insulation pipe according to claim 4, characterized in that, The obtaining of the overlapping parts between every four registered and aligned ROI processing images, eliminating the overlapping parts based on hybrid fusion, and stitching the four ROI processing images to generate a real-time stitched processing image includes: Obtain the distance distribution weights of the pixel positions in each ROI processing image from the overlapping boundary, and perform smooth transition on each ROI processing image based on the distance distribution weights; Based on the image pyramid, decompose each ROI processing image that has completed smooth transition into high-frequency part, medium-frequency part, and low-frequency part. The low-frequency part is represented as the most blurred layer, the high-frequency part is represented as the clearest layer, and the medium-frequency part is represented as the medium clarity layer. Perform hierarchical fusion on the high-frequency part, medium-frequency part, and low-frequency part to obtain the real-time stitched processing image to be optimized; Construct an energy function, search for the seam path with the smallest difference in the real-time stitched processing image to be optimized based on the energy function, and eliminate the ghosting in the real-time stitched processing image to be optimized according to the seam path with the smallest difference in the hierarchical fusion; Perform global optimization processing on the real-time stitched processing image to be optimized to generate a real-time stitched processing image.

6. The preferred method for process data of a sprayed and wound polyethylene insulation pipe according to claim 1, characterized in that The calculation of the clarity of the real-time stitched processing image based on the gradient method to obtain clarity data includes: Calculate the clear gradient amplitudes of the X-axis and Y-axis of the 5×5 window in each real-time stitched processing image based on the Sobel operator with a 5×5 kernel. The clear gradient amplitudes are calculated based on the sum of squares, and normalize the clear gradient amplitudes of the X-axis and Y-axis of the 5×5 window in each real-time stitched processing image; The normalization formula is expressed as: ； Among them, represents the clear gradient magnitude of the X-axis and Y-axis of the 5×5 window after normalization, represents the mean value of the clear gradient magnitude within the th window, represents the standard deviation of the clear gradient magnitude within the th window, represents the clear gradient magnitude of the X-axis and Y-axis of the 5×5 window before normalization; Calibrate the process key area, which is represented as the area of 80% of the width of the spraying center of the processing pipe and the area of ±10 mm on both sides of the winding seam; Perform a summation calculation of the clear gradient amplitudes based on the process key area to obtain the sum of clear gradient amplitudes, and linearly map the sum of clear gradient amplitudes to the 0-100% interval to obtain clarity data.

7. A method for optimizing process data of a spray-wound polyethylene insulation pipe according to claim 6, characterized in that, Performing clear gradient amplitude summation calculation based on process key areas to obtain the clear gradient amplitude sum, including: After performing the clear gradient amplitude summation calculation, obtaining the processing blur length based on the exposure time and the processing rotation speed of the processing tube, and performing noise optimization on the clear gradient amplitude sum based on the noise suppression function; The noise suppression function can be expressed as: ； Among them, represents the sum of clear gradient magnitudes after noise optimization, represents the sum of clear gradient magnitudes, represents the machining blur length.

8. A method for optimizing process data of a spray-wound polyethylene insulation pipe according to claim 1, characterized in that, Obtaining the spraying evaluation data and the winding evaluation data based on the real-time stitched processing images, including: The real-time stitched processing images include the real-time stitched processing images of the processing tube during spraying work and the real-time stitched processing images of the processing tube during winding work; Obtaining the spraying characteristics of the processing tube based on the real-time stitched processing images of the processing tube during spraying work, where the spraying characteristics of the processing tube include defect characteristics and thickness characteristics, obtaining the winding characteristics of the processing tube based on the real-time stitched processing images of the processing tube during winding work, and the winding characteristics of the processing tube are expressed as the winding angle deviation, and obtaining the spraying evaluation data and the winding evaluation data according to the spraying characteristics and the winding characteristics of the processing tube.

9. A method for optimizing process data of a sprayed and wound polyethylene insulation pipe according to claim 8, characterized in that, Obtaining the spraying evaluation data and the winding evaluation data according to the spraying characteristics and the winding characteristics of the processing tube, including: Establishing a scoring model, importing the spraying characteristics and the winding characteristics of the processing tube into the scoring model, and obtaining the spraying evaluation data and the winding evaluation data through the scoring model; The spraying evaluation data is expressed as the score of spraying processing, which is between 0 and 100, and the winding evaluation data is expressed as the score of winding processing, which is between 0 and 100.

10. A process data optimization system for spray-wound polyethylene insulation pipes, characterized in that, Including: A processing device, which includes a sensor module, an image module and a processing module; A sensor module, which is used to obtain the process data of the processing device; A processing module, which processes the processing tube according to the corresponding processing process steps; An image module, which is used to photograph and collect the processing tube during the processing process, and obtain the real-time processing images of the processing tube during spraying work and winding work in the spraying area and the winding area; A processing module, which is used to extract the ROI processing images from the real-time processing images and obtain the real-time stitched processing images; A proportional controller, which can obtain the optimal control gain based on the root locus optimization of the processing transfer function and can be adjusted according to the optimal control gain; An adjustment module, which is used to obtain the spraying evaluation data and the winding evaluation data, and can perform real-time optimization adjustment on the process data of the processing device based on the spraying evaluation data and the winding evaluation data.

Images