A method for rapid laying of a kilometer-scale continuous pipeline

Abstract

The application discloses a kind of kilometer-level continuous pipeline's quick laying construction method, comprising: according to the contact indentation topography evaluation pipe and whether the contact of godet is converted into continuous fitting by local sliding, to identify the godet friction increasing section;For godet friction increasing section, the motion trajectory of pixel point is extracted from the continuous video stream of the pipe surface, the displacement path of surface texture is tracked frame by frame, and the drag trajectory atlas is generated;According to the drag trajectory atlas, the relative sliding trajectory of the pipe surface relative to the godet is analyzed, whether the pipe surface keeps continuous fitting rolling is evaluated, the section where the pipe is in continuous rolling state is identified, and the rolling smooth identification is obtained;Extract the tension data of the pipe into slot traction in the rolling smooth identification corresponding laying section, and identify the laying section with strong traction tension classification in combination with the tension mean value under stable rolling working condition in historical construction data.

Description

Technical Field

[0001] This invention relates to the field of next-generation information technology, and in particular to a rapid construction method for laying kilometer-long continuous pipelines. Background Technology

[0002] The rapid laying of kilometer-long continuous pipelines is a crucial step in modern large-scale engineering construction, directly determining the overall efficiency of energy transmission and infrastructure development. During the long-distance towing of pipelines into trenches, controlling the resistance between the pipeline and the support system is key to ensuring smooth construction. According to conventional mechanics, increased surface roughness of the support rollers means a higher coefficient of friction, inevitably leading to increased pulling force. Therefore, in engineering practice, highly polished or smooth-surfaced support rollers are generally preferred to reduce resistance. However, in actual long-distance towing, when the support roller surface has a shallow groove structure, the indentation distribution on the contact area with the pipeline's outer wall has a directional guiding effect, causing the pipeline to form a localized interlocking effect with the support roller under pressure, thereby inhibiting relative sliding and promoting synchronous rolling. Specifically, when the pipeline travels on the support roller, as the friction on the roller surface increases, a specific indentation distribution is generated on the contact surface between the pipeline and the support roller. This change in indentation transforms the localized sliding friction that would easily occur on a smooth support roller into continuous rolling as the pipeline rotates with the support roller. For example, during actual towing, if the support rollers are too smooth, the pipe will often slip on them, producing severe surface abrasion marks. In this case, the traction machine needs to overcome enormous sliding resistance. However, if the friction is appropriately increased, the pipe will engage with the support rollers and roll smoothly. Although the surface roughness increases and local friction intensifies, the overall towing resistance decreases, and surface damage to the pipe is significantly reduced. Due to the lack of accurate identification of this anomaly between increased friction and decreased resistance, it is difficult to accurately determine when the pipe is in a sliding state and when it is rolling during construction, leading to frequent fluctuations in traction force and easy damage to the pipe surface. How to accurately capture the changes in contact marks and abrasion state on the pipe surface, and dynamically determine the rolling stability of the pipe to identify local slippage, has become a key issue in achieving low-resistance, damage-free laying of kilometers of pipeline. Summary of the Invention

[0003] In view of the shortcomings of the prior art, the purpose of this invention is to provide a rapid construction method for laying kilometer-level continuous pipelines, which solves the above-mentioned problems in the prior art.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] A rapid construction method for laying kilometer-long continuous pipelines includes: during the pipeline entry and lifting process, acquiring an initial contact image of the support surface of a shallow-grooved roller, simultaneously acquiring a continuous video stream of the pipeline surface and pipeline entry traction data; extracting the indentation contour from the indentation edge of the initial contact image, separating the indentation boundary, and forming a contact indentation morphology; identifying the roller friction-increased sections based on the contact indentation morphology; for the roller friction-increased sections, extracting the motion trajectory of pixels from the continuous video stream of the pipeline surface, tracking the surface texture displacement path, and generating a drag trajectory map; and analyzing the relative sliding trajectory of the pipeline surface relative to the roller based on the drag trajectory map. Obtain a rolling stability indicator; extract the tension data of the laying section corresponding to the rolling stability indicator from the traction tension data of the pipeline entering the trench, and combine it with the average tension of stable rolling conditions in historical construction data to identify the laying section with relatively strong traction tension; compare the laying section with relatively strong traction tension with the section in the drag trajectory map where the surface texture displacement is greater than the roller rolling displacement to obtain the local slip segment; extract the sliding image corresponding to the local slip segment from the continuous video stream of the pipeline surface, smooth the motion trajectory in the sliding image, and combine it with the contact indentation morphology and the rolling stability indicator to obtain the recognition result of reduced surface drag.

[0006] Furthermore, extracting the indentation contour from the indentation edge of the initial contact image, separating the indentation boundary, and forming the contact indentation morphology includes: using the Canny edge detection operator to extract the gray-level gradient abrupt change points in the indentation region as indentation edge pixels, connecting the indentation edge pixels point by point to form a closed indentation contour, separating the independent indentation boundary, using structured light projection to obtain the height difference from the bottom of the indentation to the original surface of the pipe outer wall as the contact depth value in the indentation region, and superimposing the contact depth value on the indentation contour to obtain the contact indentation morphology.

[0007] Furthermore, the identification of the roller friction increase section based on the contact indentation morphology includes: dividing the roller circumference and pipeline laying direction into square statistical units; counting the pixels covered by the indentation outline and normalizing them by area to obtain the indentation distribution density value; splicing along the pipeline laying mileage to obtain the indentation density distribution sequence; retrieving the indentation distribution density reference values ​​under sparse and dense contact conditions in historical construction data as the dividing threshold; merging the continuous mileage range where the density value of the statistical unit is higher than the dividing threshold to identify the roller friction increase section.

[0008] Furthermore, for the section where the roller friction increases, the motion trajectory of pixels is extracted from the continuous video stream on the pipe surface, and the displacement path of the surface texture is tracked to generate a drag trajectory map. This includes: extracting a video frame sequence with the same timestamp as the section where the roller friction increases; detecting pixels with significant grayscale gradients at welds, anti-corrosion coating particles, and random scratches on the pipe surface as tracking seed points; solving the instantaneous displacement vector of the seed points between adjacent video frames; tracking the displacement path frame by frame to obtain a continuous displacement trajectory cluster; determining the synchronous rolling displacement of the roller based on the roller rotation speed and the inter-frame time interval; dividing the projection component of the displacement trajectory along the pipe laying direction by the synchronous rolling displacement of the roller to obtain the rolling component; dividing the projection component perpendicular to the laying direction by the synchronous rolling displacement of the roller to obtain the sliding component; and rasterizing the rolling component and the sliding component along the pipe laying mileage and time to obtain the drag trajectory map.

[0009] Furthermore, based on the analysis of the drag trajectory map, the relative sliding trajectory of the pipe surface relative to the support roller is obtained to obtain a rolling stability indicator. This includes: reading the rolling component and sliding component in the drag trajectory map grid by grid along the pipe laying mileage; taking the difference between the rolling component and the sliding component as the relative sliding amplitude; concatenating them to obtain the relative sliding trajectory; retrieving the upper limit of the relative sliding amplitude under continuous contact rolling conditions as the contact judgment threshold; merging the grids of continuous mileage where the relative sliding amplitude is lower than the contact judgment threshold; and assigning the rolling stability indicator.

[0010] Furthermore, the tension data of the laying section corresponding to the rolling stability mark in the pipeline entry traction tension data is extracted, and combined with the average tension of stable rolling conditions in historical construction data, the laying section with strong traction tension is identified. This includes: dividing the tension data of the laying section corresponding to the rolling stability mark into continuous tension segments along the pipeline laying mileage according to a preset mileage length, taking the arithmetic mean of each tension segment to obtain the average tension value, retrieving the average tension of stable rolling conditions of the same pipe diameter in historical construction as a reference benchmark for segment-by-segment comparison, and identifying the tension segment that is more than one level higher than the reference benchmark as the laying section with strong traction tension.

[0011] Furthermore, the tension data of the laying section corresponding to the rolling stability mark is divided into continuous tension segments according to the preset mileage length along the pipeline laying mileage. The arithmetic mean of each tension segment is obtained by taking the average tension value. This includes: dividing the continuous tension segments into segments every 100 meters; using a sliding variance scan for the tension data sequence in each tension segment; determining a preset variance interval based on the statistical interval of tension variance under historical construction stable rolling conditions; removing the initial tensioning segment and the final tensioning segment; retaining the tension stability interval whose variance falls within the preset variance interval; taking the arithmetic mean of the tension data in the tension stability interval to obtain the average tension value of the stability interval; using the average tension value of historical construction stable rolling conditions for the same pipe diameter as a reference benchmark, the upper and lower fluctuations are divided into adjacent levels by a preset gap; and identifying the tension segment that is one level higher than the reference benchmark as the laying section with stronger traction tension.

[0012] Furthermore, the paving sections with stronger traction force are compared with the sections in the drag trajectory map where the surface texture displacement is greater than the roller rolling displacement to obtain local slip segments. This includes: reconstructing the surface texture displacement and roller rolling displacement grid by grid along the pipeline paving mileage based on the rolling and sliding components in the drag trajectory map; using the difference between the surface texture displacement and roller rolling displacement as displacement deviation values ​​to obtain a displacement deviation sequence; merging the positions where the displacement deviation value is greater than zero and continuously covers multiple consecutive mileage grids into a displacement leading segment; comparing the displacement leading segment with the paving sections with stronger traction force according to mileage coordinates, and identifying the overlapping positions as the local slip segments.

[0013] Furthermore, a sliding image corresponding to the local slip segment is extracted from the continuous video stream of the pipe surface, and the motion trajectory in the sliding image is extracted smoothly. This includes: according to the pipe laying mileage range and timestamp corresponding to the local slip segment, a corresponding video frame subsequence is extracted from the continuous video stream of the pipe surface and named as a sliding image sequence; the motion trajectory of the texture feature points on the outer wall of the pipe is tracked frame by frame in the sliding image sequence; the motion trajectory is extracted smoothly using cubic spline interpolation; isolated peak displacements are removed to obtain a smooth sliding trajectory.

[0014] Furthermore, the smooth extraction of the motion trajectory using cubic spline interpolation includes: taking the grid-by-grid amplitude of the smooth sliding trajectory along the pipeline mileage as the sliding amplitude value; reading the contact depth and indentation distribution density at the corresponding mileage position of the local slip segment from the contact indentation morphology; determining the sliding-to-roll transition marker; comparing the adjacency relationship between the mileage position of the sliding-to-roll transition marker and the mileage segment of the rolling stability marker to obtain the identification result of reduced surface drag.

[0015] Compared to existing technologies, this invention offers the following advantages: During the pipe lifting process in the trench, by acquiring initial contact images of the support surface of the support roller and continuous video streams of the pipe surface, combined with real-time traction force data, it solves the problem of accurately identifying changes in the contact state between the pipe and the support roller in traditional construction. This invention extracts the contour morphology from the indentation edge, calculates the indentation distribution density per unit area, determines whether the contact has changed from localized slippage to continuous contact, and thus identifies sections with increased friction. For these sections, the surface texture displacement path is tracked frame by frame to generate a drag trajectory map, analyzes the sliding characteristics of the pipe relative to the support roller, and identifies sections in a continuous rolling state. Further, the traction force data for these sections is extracted and graded for strength, compared with the historical average traction force under stable working conditions, to identify laying sections with relatively strong traction. This section is cross-validated with sections where the texture displacement exceeds the rolling displacement of the support roller to accurately locate the position where local slippage of the pipe occurs. Finally, the surface drag resistance change is evaluated by combining the contact indentation morphology. This invention enables dynamic monitoring and accurate identification of abnormal states throughout the entire process of laying kilometer-level pipelines, effectively avoiding construction risks caused by uncontrolled contact conditions, and significantly improving the construction efficiency and safety of long-distance pipeline laying. Attached Figure Description

[0016] Figure 1 This is a flowchart of a rapid laying method for a kilometer-long continuous pipeline according to the present invention.

[0017] Figure 2 This is a schematic diagram of a rapid laying method for a kilometer-level continuous pipeline according to the present invention.

[0018] Figure 3 This is another schematic diagram of a rapid laying method for a kilometer-level continuous pipeline according to the present invention. Detailed Implementation

[0019] The present invention will now be described in detail through specific embodiments:

[0020] like Figures 1-3 This embodiment of a rapid construction method for laying a kilometer-long continuous pipeline may specifically include:

[0021] S101. During the process of lifting the pipe into the trench, an initial contact image is acquired on the support surface of the shallow grooved roller, and a continuous video stream along the laying direction of the pipe surface and the traction force data of the pipe entering the trench are acquired simultaneously. The indentation contour is extracted from the edge of the indentation in the initial contact image, the indentation boundary is separated, and the contact indentation morphology is formed.

[0022] During the pipe insertion and lifting process, an industrial camera is mounted on the support surface of the shallow grooved roller to capture the initial contact image of the roller carrying the pipe at the moment of insertion. A line array camera is mounted along the laying direction to capture a continuous video stream of the pipe surface texture. At the traction end, a tension sensor collects the traction force data of the pipe insertion. The initial contact image, continuous video stream, and traction force data are time-aligned according to a unified timestamp to obtain multi-source acquisition data with a time index. For the initial contact image in the multi-source acquisition data, the Canny edge detection operator is used to extract the gray-level gradient abrupt change points in the indentation area as indentation edge pixels. Points are connected along the indentation edge pixels to form a closed indentation contour. Based on the indentation contour, the shallow groove background texture of the roller and the extrusion marks on the outer wall of the pipe are distinguished, and independent indentation boundaries are separated to obtain a set of indentation boundaries reflecting the contact range between the outer wall of the pipe and the shallow grooved roller. Based on the set of indentation boundaries, the height difference between the bottom of the indentation and the original surface of the outer wall of the pipe is obtained by structured light projection within the indentation area as the contact depth value. The contact depth value is mapped point-to-point with the planar coordinates of the indentation boundary, and the depth value is superimposed on the indentation contour to obtain the contact indentation morphology containing the contact depth.

[0023] Specifically, the shallow-grooved support rollers used during the lifting process of kilometer-long continuous pipelines have a regularly distributed shallow groove structure on their supporting surface. The groove depth is typically in the range of 0.2 to 0.5 millimeters, and the groove direction is perpendicular to the axis of the support roller. Under the combined action of its own weight and traction force, the outer wall of the pipeline is pressed against the support roller, and the shallow groove structure forms an indentation distribution on the outer wall of the pipeline corresponding to the direction of the groove.

[0024] In one embodiment, an industrial camera is mounted at a height of 800–1200 mm above the outer wall of the pipe, directly above the support surface of the shallow grooved roller. The industrial camera uses an area array imaging method, with a sensor resolution of no less than 2048×2048 pixels, equipped with a fixed-focus lens with a focal length of 25mm to 35mm, and a field of view coverage of 300mm×300mm to 400mm×400mm, ensuring that the actual size of the pipe surface corresponding to a single pixel is no greater than 0.2mm. It is positioned directly opposite the contact area between the pipe and the roller, capturing the initial contact image of the roller bearing the pipe at the moment of contact. Simultaneously, a line array camera is mounted 200 to 300 mm behind the roller along the pipe laying direction. The scan lines of the line array camera are perpendicular to the pipe axis, with a line array resolution of no less than 4096 pixels and a line frequency set to 5000 to 8000Hz. As the pipe moves forward, the camera captures the surface texture of the pipe line by line, stitching the images into a continuous video stream. A tension sensor with a range of 0 to 500 kN is connected in series at the end of the traction machine's rope, and the sampling frequency is set to 100 Hz to continuously sample the traction force data of the pipeline entering the trench.

[0025] It should be noted that the three types of acquisition devices—industrial camera, line scan camera, and tension sensor—are triggered by a unified clock source. The initial contact image, continuous video stream, and traction force data are time-aligned using a unified timestamp, with timestamp accuracy matched to the millisecond level, resulting in multi-source acquisition data with a time index. This time index allows subsequent indentation recognition results to correspond to the pipeline mileage position at the same time. The initial contact image in the multi-source acquisition data is processed using the Canny edge detection operator.

[0026] Specifically, the initial contact image is first Gaussian smoothed to suppress oil stain reflection and acquisition noise on the roller surface. Then, the magnitude and direction of the grayscale gradient are calculated along the horizontal and vertical directions. Non-maximum suppression is applied to the magnitude map, retaining only pixels with local maximum values ​​along the gradient direction. Finally, a hysteresis connection is performed using both high and low thresholds to identify pixels with abrupt changes in grayscale gradient as indentation edge pixels. The high threshold is set to the average gradient magnitude plus 1.2 times the standard deviation, with a value range of 80 to 120, and is used to determine strong edge seed points. The low threshold is set to 0.4 times the high threshold, with a value range of 32 to 48, and is used to extend weak edge pixels connected to strong edges along the seed points. When the contrast of the roller surface is low, the high threshold is taken as close to 80; when the indentation edge is clear, the high threshold can be taken as close to 120. The low threshold is adjusted synchronously with the high threshold to maintain a 0.4 times ratio.

[0027] In one embodiment, the pixels at the edge of the indentation are fitted to a continuous curve using a Hough transform, and then simplified into a closed contour using the Douglas-Peucker algorithm, with a simplification threshold set to 0.5 pixels. It should be noted that the surface texture of the shallow grooved roller itself also appears as periodic stripes in the initial contact image, with its contour extending along the groove direction and its depth not changing with pipe compression, while the pressure marks on the pipe outer wall appear as locally closed elliptical or arc-shaped contours. The closure degree C and aspect ratio R are calculated for each extracted contour, where the closure degree C is equal to the ratio of the distance between the first and last endpoints of the contour to the perimeter of the contour, and the aspect ratio R is equal to the ratio of the long side to the short side of the circumscribed rectangle of the contour. When C is less than 0.1 and 1.2 is less than R and less than 5, it is determined to be an indentation contour; otherwise, it is determined to be roller surface texture. This discrimination criterion separates independent indentation boundaries, resulting in a set of indentation boundaries reflecting the contact range between the pipe outer wall and the shallow grooved roller.

[0028] It is understood that the set of indentation boundaries only describes the two-dimensional distribution of indentations on the projection plane of the roller support surface, and cannot reflect the depth undulations of the indentations. Furthermore, based on the set of indentation boundaries, contact depth values ​​are obtained within the indentation area using structured light projection.

[0029] For example, a structured light projector projects equally spaced parallel stripe gratings onto the indentation area. The stripes are reflected by the outer wall of the pipe and the surface of the support roller and then received by the structured light camera. The phase difference of the stripes at the bottom of the indentation is offset relative to the original surface outside the indentation. The height difference from the bottom of the indentation to the original surface of the outer wall of the pipe is calculated based on the phase offset, and the height difference is used as the contact depth value at that point.

[0030] Specifically, a four-step phase-shifting method is used to obtain the phase distribution of the indentation area. The phase value of each pixel is calculated using the arctangent function. The difference between the phase value at the bottom of the indentation and the phase value on the original surface is the phase offset Δφ. Before use, the system needs to be calibrated. A standard step of known height is placed at the measurement position, and the phase offset corresponding to different heights is obtained to establish a mapping relationship between phase and height, thus obtaining the system sensitivity coefficient K. Its value typically ranges from 0.05 to 0.2 mm per radian, with the specific value depending on the baseline distance between the projector and the camera and the period of the projection grating. During measurement, the phase offset Δφ is multiplied by the sensitivity coefficient K to obtain the height difference h, calculated as h = K × Δφ. For example, when the sensitivity coefficient K is calibrated to 0.1 mm per radian, and the measured phase offset at a point is 1.5 radians, the contact depth at that point is 0.15 mm.

[0031] Preferably, the structured light projector and the industrial camera share the same bracket, and its projection area corresponds to the same physical plane coordinate system as the imaging field of view of the initial contact image. A point-to-point mapping is performed between the contact depth value and the planar coordinates of the indentation boundary. Each pixel within the indentation boundary is assigned a corresponding contact depth value, and the depth values ​​are superimposed on the indentation contour to obtain a contact indentation morphology that includes the contact depth. This contact indentation morphology simultaneously conveys the planar position, contour direction, and depth distribution of the indentation, serving as a basis for subsequent determination of the contact state between the pipe and the shallow grooved support roller.

[0032] S102. Evaluate whether the contact between the pipe and the support roller changes from localized sliding friction to continuous contact based on the contact indentation morphology, so as to identify the section where the friction of the support roller increases.

[0033] Based on the contact indentation morphology, square statistical units of uniform size are divided along the circumference of the support roller and the pipeline laying direction. Within each statistical unit, the pixels covered by the indentation outline are counted, and the area of ​​the statistical unit is normalized to obtain the indentation distribution density value. The indentation distribution density values ​​of adjacent statistical units are spliced ​​along the pipeline laying mileage to obtain a continuous indentation density distribution sequence extending along the laying direction. For the indentation density distribution sequence, reference values ​​of indentation distribution density under sparse slip-friction conditions and dense adhesion conditions in historical construction data are retrieved as boundary thresholds. The indentation density distribution sequence is compared unit by unit. If the density values ​​of multiple consecutive statistical units in the indentation density distribution sequence are higher than the boundary threshold and the indentation direction is consistent with the shallow groove direction, it is determined that the pipeline-support roller contact at that statistical unit has changed from local slip-friction to continuous adhesion, thus obtaining an adhesion status indicator. Based on the bonding status identifier, the statistical units of adjacent bonding statuses are merged along the pipeline laying mileage, and isolated bonding units with a length shorter than the preset mileage are removed. The pipeline mileage coordinates are marked on the merged continuous bonding mileage intervals to identify the laying mileage range where the surface friction of the support roller increases, thus obtaining the section where the support roller friction increases.

[0034] It is understood that the morphology of the contact indentation reflects the planar position, contour direction, and depth distribution of the indentation within the contact area between the pipe outer wall and the shallow grooved support roller. During the pipe insertion and lifting process, when the pipe and support roller are in a state of partial sliding, the indentation is sparsely and discretely distributed within the contact area; when the two transition from sliding to continuous contact, the indentation is densely arranged along the shallow groove direction, and the number of indentations per unit area increases significantly.

[0035] Specifically, square statistical units of uniform size are divided along the circumference of the support roller and the pipeline laying direction. The side length of the square statistical unit is an integer multiple of the shallow groove period, so that each statistical unit can completely contain the indentation marks corresponding to several grooves. Within each statistical unit, the pixels falling into the set of indentation boundaries are counted to obtain the total number of pixels covered by the indentation outline within the statistical unit. In one embodiment, the total number of pixels is divided by the area of ​​the statistical unit to obtain the indentation distribution density value of the statistical unit. The indentation distribution density value characterizes the density of indentations per unit area; the higher the value, the more continuous and close the contact between the outer wall of the pipeline and the shallow groove support roller at that statistical unit. The indentation distribution density values ​​of adjacent statistical units are spliced ​​together along the pipeline laying mileage to obtain an indentation density distribution sequence that extends continuously along the laying direction.

[0036] It should be noted that the interpretation of the indentation density distribution sequence depends on a pre-established boundary threshold.

[0037] In one possible implementation, reference values ​​for the indentation distribution density are retrieved from historical construction data for two typical working conditions with the same pipe diameter and shallow grooved rollers of the same type: one is a sparse slip-friction condition, where the pipe slips significantly on the roller, and indentations appear sporadically only at a few groove locations, resulting in a low density reference value; the other is a dense fit condition, where the pipe and roller roll synchronously, and indentations are regularly distributed along all grooves, resulting in a high density reference value. The median value of the density reference values ​​for these two conditions is taken as the dividing threshold, denoted as Td, serving as the boundary for distinguishing between the slip-friction state and the fit state. Furthermore, the indentation density distribution sequence is compared unit by unit with the boundary threshold Td. If the density values ​​of multiple consecutive statistical units in the indentation density distribution sequence are all higher than Td, and the deviation between the main direction of the indentation contour and the shallow groove direction within the statistical unit is within a preset angle range, then it is determined that the contact between the pipe and the support roller at the statistical unit has changed from local sliding to continuous contact, and a contact state identifier is assigned to it; otherwise, the sliding state is retained. The calculation of the main direction of the indentation adopts the principal component analysis method: extract the coordinate set of all indentation contour pixels within the statistical unit, calculate the covariance matrix of the coordinate set, and find the eigenvector direction corresponding to the largest eigenvalue, which is the main direction θm. The difference between θm and the known shallow groove direction θg is compared. When |θm−θg|≤15 degrees, it is determined that the direction is consistent, which meets the direction condition of the contact state, and is used to exclude false high-density areas caused by oil stains or foreign objects on the pipe surface.

[0038] For example, in the case of a gas transmission steel pipe with a diameter of 610 mm entering the groove on a shallow grooved support roller, the side length of the statistical unit is 6 mm, the number of statistical units for continuous comparison is 5, and the preset angle range is within 15 degrees, which serves as a set of implementable parameters for determining the fit state.

[0039] Preferably, based on the bonding status indicator, statistical units of adjacent bonding status are merged along the pipeline laying mileage, and statistical units that are adjacent in position and all have bonding status indicators are connected in series to form a continuous bonding mileage interval.

[0040] In one embodiment, the length of each segment of the merged continuous bonding mileage interval is checked, and isolated bonding units shorter than a preset mileage L0 are discarded. The preset mileage L0 is the minimum effective mileage for the shallow grooved roller to continuously support the pipe. Physically, it represents the shortest axial contact length required for the roller and pipe to form stable frictional contact, and its value is determined based on the geometric relationship between the roller groove width and the pipe diameter, typically ranging from 0.5 meters to 2 meters. For the remaining continuous bonding mileage intervals after discarding the discarded units, pipe mileage coordinates are marked based on the pipe's traction travel position in the groove, identifying these intervals as the paving mileage range where the friction between the roller surface and the pipe outer wall is relatively increased, thus obtaining the roller friction increase section. This roller friction increase section is presented as a continuous segment along the kilometer-level paving mileage.

[0041] S103. For the section where the friction of the roller increases, the motion trajectory of the pixels is extracted from the continuous video stream on the surface of the pipe, and the displacement path of the surface texture is tracked frame by frame to generate a drag trajectory map.

[0042] For the pipeline mileage range corresponding to the increased friction section of the support roller, a video frame sequence with timestamps consistent with this mileage range is extracted from the continuous video stream of the pipeline surface. The video frame sequence is then processed for grayscale conversion and histogram equalization. Shi-Tomasi corner detection is used to select pixels with significant grayscale gradients at weld seams, anti-corrosion coating particles, and random scratches on the pipeline surface as tracking seed points, resulting in a set of seed points with stable texture features. For each seed point in the seed point set, the Lucas-Kanade optical flow method is used to solve for the instantaneous displacement vector of the seed point between adjacent video frames. The displacement path of the seed point in the pipeline laying direction is traced frame by frame along the video frame sequence. Broken or out-of-bounds displacement paths are reconnected using the nearest neighbor principle, resulting in a continuous displacement trajectory cluster of the seed point. Based on the continuous displacement trajectory cluster, the synchronous rolling displacement of the roller is determined according to the roller rotation speed and the inter-frame time interval. The rolling component is obtained by dividing the projection component of each displacement trajectory along the pipeline laying direction by the synchronous rolling displacement of the roller. The sliding component is obtained by dividing the projection component perpendicular to the laying direction by the synchronous rolling displacement of the roller. The rolling component and the sliding component are rasterized and distributed along the two dimensions of pipeline laying mileage and time to obtain a dragging trajectory map reflecting the change of surface friction.

[0043] Understandably, within the paving mileage corresponding to the aforementioned section of increased roller friction, the relationship between the pipe and the shallow grooved roller has shifted from sparse slippage to continuous contact. However, contact does not necessarily equate to synchronous rolling. Within this section, the outer wall of the pipe may still experience localized slippage relative to the roller, which needs to be identified using visual motion information from the pipe surface.

[0044] Specifically, a line scan camera pre-installed along the laying direction of the pipe during the pipe entry and lifting process has captured a continuous video stream of the pipe surface texture. For the pipe laying mileage range corresponding to the increased friction section of the support rollers, based on the correspondence between video frame timestamps and pipe mileage coordinates, a sequence of video frames with timestamps consistent with that mileage range is extracted from the continuous video stream of the pipe surface.

[0045] It should be noted that the captured video frame sequence was affected by uneven lighting at the construction site, reflections from the anti-corrosion coating, and dust interference during the acquisition process, resulting in an overall dark grayscale distribution or local overexposure.

[0046] In one possible implementation, the video frame sequence is grayscaled frame by frame, and then histogram equalization is applied to make the grayscale histogram of each frame tend to be uniformly distributed across the entire grayscale range, which facilitates the subsequent identification of pixels with significant grayscale gradients on the outer wall texture of the pipe. Further, Shi-Tomasi corner detection is used to scan the equalized video frames frame by frame. The core of Shi-Tomasi corner detection lies in constructing a 2x2 matrix of grayscale gradients in the neighborhood of each candidate pixel, where the neighborhood size is set to a 3×3 to 7×7 pixel window, preferably a 5×5 pixel window to balance computational efficiency and feature stability. Two eigenvalues ​​λ1 and λ2 of this matrix are calculated, and the smaller eigenvalue min(λ1, λ2) is taken as the corner response value. Only when the response value is higher than a preset response threshold of 0.01 is the candidate pixel determined to be a corner. On the texture of the outer wall of the pipe, the gray-scale gradients at the weld edge, anti-corrosion coating particles, and random scratches are large in both the horizontal and vertical directions, and the corresponding small feature values ​​are significant. Therefore, they are preferentially selected as tracking seed points to obtain a set of seed points with stable texture features.

[0047] Preferably, the upper limit of the number of tracking seed points in each frame is limited according to the unit field of view area of ​​the pipeline.

[0048] Specifically, 8 to 15 seed points are set per square meter of the field of view to avoid excessively dense corner points, which would lead to excessive computational load for subsequent optical flow tracing. To ensure that the small displacement assumption of the Lucas-Kanade optical flow method holds, the displacement between adjacent frames is adjusted to ensure that it does not exceed 10 pixels.

[0049] Specifically, the required frame rate f is calculated based on the pipe traction speed v and the camera resolution R, where f ≥ v × R ÷ 10, v is the traction speed in meters per second, R is the horizontal resolution in pixels per meter, and f is the frame rate in frames per second. In one embodiment, for each seed point in the seed point set, the instantaneous displacement vector of the seed point is solved between adjacent video frames using the Lucas-Kanade optical flow method. The Lucas-Kanade optical flow method, based on the assumptions of constant brightness and small displacement, assumes that the grayscale value of the same seed point remains unchanged between adjacent frames. A window is selected around the seed point, with its size adaptively adjusted according to the image texture density. The initial window side length is set to 15 to 21 pixels. When the gradient change within the window is less than a threshold of 5, the window is expanded to 31 pixels; when the displacement between adjacent frames is greater than 8 pixels, the window is shrunk to 11 pixels. An overdetermined system of equations is established for the brightness changes of all pixels within the window. The least squares method is used to solve for the horizontal displacement component u and the vertical displacement component v of the seed point. The instantaneous displacement vector is denoted as v, which is equal to the combination of u and v, where u corresponds to the displacement component in the pipe laying direction and v corresponds to the displacement component perpendicular to the pipe laying direction. The instantaneous displacement vectors of various seed points are concatenated frame by frame along the video frame sequence to obtain the displacement path of the seed point in the pipe laying direction. During actual acquisition, some seed points may experience displacement path breaks or cross boundaries due to occlusion on the pipe surface or textures detaching from the field of view. To address this, the nearest neighbor principle is used to reconnect the frames before and after the break based on the minimum spatial distance, thus obtaining a continuous displacement trajectory cluster of the seed points.

[0050] Under the ideal condition of synchronous rolling of the pipe and the support roller, the distance that any point on the outer wall of the pipe moves along the laying direction per unit time should be equal to the displacement corresponding to the proportion of the arc length that the support roller spreads on the outer wall of the pipe in one revolution.

[0051] Specifically, based on the roller rotation speed ω collected by the speed sensor during the corresponding time period of the increased roller friction section and the time interval Δt between adjacent video frames, the synchronous rolling displacement of the roller is denoted as s = ω × R × Δt, where R is the radius of the roller support surface in meters; ω is the roller angular velocity in rad / s; and Δt is the interval between adjacent video frames in seconds. According to the continuous displacement trajectory cluster, the projection components along the pipeline laying direction and perpendicular to the pipeline laying direction of each displacement trajectory are calculated. The rolling component Kr is obtained by dividing the projection component along the pipeline laying direction by the synchronous rolling displacement s of the roller, and the sliding component Ks is obtained by dividing the projection component perpendicular to the pipeline laying direction by the synchronous rolling displacement s of the roller. Set the judgment threshold: when Kr is in the range of 0.95 to 1.05 and Ks is less than 0.1, the judgment seed point rolls synchronously with the pipe and the support roller; when Kr deviates from the range of 0.95 to 1.05 or Ks is greater than or equal to 0.1, the judgment seed point shifts laterally or slips forward relative to the support roller.

[0052] For example, the rolling component Kr is defined as the ratio of the projected component of the pipeline laying direction to the synchronous rolling displacement of the support roller, and the sliding component Ks is the ratio of the lateral displacement component of the pipeline to the synchronous rolling displacement of the support roller. The pipeline laying mileage is divided into mileage grids according to a preset mileage step, and the video frame sequence time axis is divided into time grids according to a preset time step. The selection of the mileage step Δd and the time step Δt is determined based on the pipeline laying speed v, satisfying Δd=v×Δt. When the laying speed is 5 meters per second, the mileage step is 0.5 meters and the time step is 0.1 seconds, ensuring that the pipeline surface area covered by each grid unit and the time window have sufficient seed point samples. For all seed points falling within each mileage minus time grid cell, when the number of seed points is not less than 3, the mean values ​​of Kr and Ks are calculated respectively, and the mean values ​​of Kr and Ks are written into the grid array as binary attributes of the grid cell; when the number of seed points is less than 3, the attribute value of the grid cell is filled by weighted interpolation of adjacent grid cells, with the weight coefficient being inversely proportional to the spatial distance. The grid array is distributed along the two dimensions of pipeline laying mileage and time to form a drag trajectory map that extends continuously along the kilometer-level laying mileage. Each grid cell in the drag trajectory map carries two types of values: rolling component Kr and sliding component Ks, reflecting the degree of rolling synchronization and lateral slippage of the pipeline surface relative to the shallow grooved roller at the mileage minus time position, providing a two-dimensional gridded kinematic description for subsequent determination of the continuity of the pipeline rolling state.

[0053] S104. Based on the dragging trajectory map, analyze the relative sliding trajectory of the pipe surface with respect to the roller, evaluate whether the pipe surface maintains continuous contact and rolling, identify the sections of the pipe in a continuous rolling state, and obtain a rolling stability indicator.

[0054] Based on the drag trajectory map, the rolling and sliding components are read grid by grid along the pipeline laying mileage. The difference between the sliding and rolling components is used as the relative sliding amplitude, reflecting the degree of deviation between sliding and rolling. The relative sliding amplitudes are sequentially connected along the pipeline laying mileage to obtain the relative sliding trajectory, reflecting the degree of offset of the pipeline surface relative to the shallow grooved roller. For the relative sliding trajectory, the upper limit of the relative sliding amplitude under the pre-established continuous contact rolling condition is retrieved as the contact judgment threshold. Based on the historical construction of stable rolling conditions for the same pipe diameter, a preset fluctuation range between adjacent grids is determined. The relative sliding trajectory is compared grid by grid with the contact judgment threshold. If the relative sliding amplitude of multiple consecutive mileage grids in the relative sliding trajectory is lower than the contact judgment threshold, and the amplitude fluctuation between adjacent grids falls within the preset fluctuation range, then it is determined that the pipeline surface at the mileage grid maintains continuous contact rolling, and a continuous contact judgment result is obtained. Based on the continuous adhesion determination result, the adjacent mileage grids that maintain continuous adhesion and rolling are serialized and merged along the pipeline laying mileage. Isolated adhesion grids with a length shorter than the preset mileage are removed. The merged mileage interval is given a rolling stability mark to obtain the laying mileage section where the pipeline is in a continuous rolling state.

[0055] Understandably, the drag trajectory map presents two types of values—rolling and sliding components—across the dimensions of pipeline laying mileage and time. The rolling component reflects the degree to which the outer wall of the pipeline rotates synchronously with the shallow grooved roller, while the sliding component reflects the degree to which the outer wall of the pipeline shifts laterally relative to the shallow grooved roller. Only when the rolling component is close to 1 and the sliding component is close to 0 is the pipeline and the shallow grooved roller in a truly continuous rolling contact state.

[0056] It should be noted that relying solely on either the rolling component or the sliding component is insufficient to fully characterize the motion coordination between the pipe surface and the shallow grooved roller; the two must be considered in conjunction.

[0057] Specifically, based on the drag trajectory map, data is read grid by grid along the pipeline laying mileage, within each mileage time grid cell. The rolling component Kr is defined as the ratio of the pipeline laying direction projection component to the synchronous rolling displacement of the support roller, ranging from 0 to 1. A value closer to 1 indicates a higher degree of synchronous rotation of the pipeline outer wall with the support roller. The sliding component Ks is defined as the ratio of the pipeline lateral offset to the support roller outer diameter, also ranging from 0 to 1. A larger value indicates a higher degree of lateral offset of the pipeline outer wall relative to the support roller. Ks-Kr is used as the relative sliding amplitude reflecting the degree of sliding deviation from rolling. A smaller amplitude value indicates that the pipeline surface movement at that grid cell is more biased towards synchronous rolling; a larger value indicates that there is a significant lateral offset or forward slippage of the pipeline surface relative to the shallow grooved support roller at that grid cell. The relative sliding amplitudes are sequentially concatenated along the pipeline laying mileage to obtain the relative sliding trajectory reflecting the degree of offset of the pipeline surface relative to the support roller. The relative sliding trajectory is then smoothed by applying a sliding mean based on the pipeline laying mileage, with a window length of 5 mileage grid cells, resulting in the smoothed relative sliding trajectory. To determine the relative sliding trajectory, a pre-established fit judgment threshold Tb is retrieved. The initial determination method for this threshold is as follows: 10 pipeline laying videos with a total length of no less than 500 meters are selected. Three experts with over 5 years of construction experience independently mark synchronous rolling sections. The relative sliding amplitude of the overlapping sections marked by the three experts is taken, and its mean plus twice the standard deviation is calculated as the initial threshold. This threshold is then dynamically updated according to the 95th percentile based on accumulated historical data. For DN1200 pipelines, this threshold is typically 0.15. Furthermore, a relative sliding amplitude below Tb for a single grid is insufficient to determine continuous fit rolling; the fluctuation of the relative sliding amplitude between adjacent grids must also be examined. Based on the difference in relative sliding amplitude between adjacent grids under stable rolling conditions for the same pipe diameter in historical construction, its standard deviation σ is calculated, and 3σ is used as the preset fluctuation range Wb. For DN1200 pipelines, this fluctuation range is typically 0.08. The relative sliding trajectory is compared grid by grid with Tb, and the absolute value of the difference in relative sliding amplitude between adjacent grids is compared with Wb. If the relative sliding amplitude of five or more consecutive mileage grids in the relative sliding trajectory is lower than Tb, and the absolute value of the amplitude difference between adjacent grids is less than Wb, then the pipe surface at that mileage grid is determined to maintain continuous contact and rolling. If the relative sliding amplitude of any grid is higher than Tb or the absolute value of the amplitude difference between adjacent grids exceeds Wb, then a discontinuous contact is determined to occur at that grid. The above comparison is performed grid by grid along the relative sliding trajectory to obtain the continuous contact determination result.

[0058] For example, in the case of a gas transmission steel pipe with a diameter of 610 mm entering the groove on a shallow grooved roller, the adhesion judgment threshold Tb is 0.15, the preset fluctuation range Wb is 0.05, and the number of mileage grids for continuous comparison is 5, which are used as a set of implementable parameters for continuous adhesion rolling judgment.

[0059] Preferably, based on the continuous fit determination result, the adjacent mileage grids that maintain continuous fit rolling are connected and merged along the pipeline laying mileage, and the grids that are adjacent in position and all have continuous fit determination results are connected to form a continuous fit mileage interval.

[0060] In one embodiment, the continuous bonding mileage interval is checked segment by segment for its coverage mileage length, and isolated bonding grids shorter than a preset mileage are discarded. The preset mileage is the minimum effective mileage under which the shallow grooved rollers continuously support the pipeline, avoiding false bonding judgments caused by instantaneous texture mismatch. A rolling stability indicator is assigned to the merged mileage interval. This rolling stability indicator carries the start and end pipeline coordinates and continuous bonding attributes of the mileage interval, identifying the laying mileage section where the pipeline is in a continuous rolling state. The laying mileage sections are distributed in segments along the kilometer-level laying mileage, reflecting the actual operating range of the pipeline maintaining synchronous rolling under the support of the shallow grooved rollers.

[0061] S105. Extract the tension data of the laying section corresponding to the rolling stability mark from the traction tension data of the pipeline entering the trench, and identify the laying section with strong traction tension by combining the average tension under stable rolling conditions in historical construction data.

[0062] Extract the tension data of the laying section corresponding to the rolling stability indicator from the traction tension data of the pipeline entry trench. Divide the pipeline laying mileage into continuous tension segments according to a preset mileage length. Take the arithmetic mean of the tension data in each tension segment to obtain the average tension value of each tension segment. Serially concatenate the average tension values ​​along the pipeline laying mileage to obtain a sequence of average tension values ​​distributed along the laying mileage. Based on the sequence of average tension values, retrieve the average tension value under stable rolling conditions of the same pipe diameter from historical construction data as a reference benchmark. Compare the average tension value of each tension segment with the reference benchmark segment by segment to classify the strength. If the average tension value is more than one level higher than the reference benchmark, the traction tension of that tension segment is determined to be relatively strong, and the laying segment with relatively strong traction tension is obtained.

[0063] It is understood that the laying mileage section corresponding to the rolling stability mark represents the working range in which the pipeline and the shallow grooved roller are in a continuous rolling state, and the comparison of the strength of traction force data is comparable within this range.

[0064] Specifically, a tension sensor connected in series at the end of the traction machine's rope records the traction force data of the pipeline entering the trench at fixed sampling intervals, and this data is synchronized with the pipeline laying mileage coordinates by timestamp. The tension data corresponding to the rolling stability marker in the laying section is extracted from the pipeline entry traction force data, and continuous tension segments are divided along the pipeline laying mileage according to a preset mileage length. In one embodiment, the preset mileage length is 100 meters, dividing the kilometer-level laying mileage into several adjacent tension segments. The arithmetic mean of the tension data within each tension segment is taken to obtain the average tension value of each tension segment. These average tension values ​​are then sequentially concatenated along the pipeline laying mileage to obtain a sequence of average tension values ​​distributed along the laying mileage.

[0065] It should be noted that the historical construction data contains the average tension value of pipes of the same diameter under stable rolling conditions on shallow grooved rollers. This average tension value reflects the conventional traction level required by the traction machine when the pipe is continuously rolling. Based on the aforementioned average tension value sequence, the average tension value of the same diameter under stable rolling conditions in the historical construction data is retrieved as a reference benchmark. The average tension value of each tension segment is then compared with the reference benchmark segment by segment to classify the strength.

[0066] Preferably, the reference benchmark is divided into three levels: weak, average, and strong, with the level interval being 10% of the reference benchmark. If the average tension is more than one level higher than the reference benchmark, the traction tension of that tension segment is determined to be strong, thus obtaining a paving section with a strong traction tension.

[0067] Retrieve the traction force data of the paving section corresponding to the rolling stability mark, divide the continuous traction segment along the kilometer-level mileage into segments of 100 meters each, extract the average traction force of the stable interval of each traction segment, and compare it with the average traction force of the same pipe diameter under stable rolling conditions in historical construction segment by segment. Identify the traction segment with an average value more than one level higher, and identify the paving section with a relatively strong traction force classification.

[0068] The traction force data of the laying section corresponding to the rolling stability mark is retrieved. The laying mileage of the kilometer-level pipeline is divided into continuous tension segments every 100 meters. The tension data sequence in each tension segment is scanned using sliding variance. A preset variance interval is determined based on the statistical interval of tension variance under the historical construction stable rolling conditions. The initial tensioning segment and the final tension easing segment are removed. The tension stability intervals whose variance falls within the preset variance interval are retained. The arithmetic mean of the tension data in the tension stability interval is taken to obtain the average tension value of the stability interval of each tension segment. Based on the average tension value of the stable interval of each tension segment, the average tension value of the same pipe diameter under stable rolling conditions in historical construction is retrieved as a reference benchmark. The reference benchmark is divided into adjacent levels by a preset interval. The average tension value of each stable interval is compared with the reference benchmark segment by segment. If the average tension value of the stable interval is more than one level higher than the reference benchmark, the tension segment is determined to be a tension segment with an average value more than one level higher, and the traction tension is identified as a laying section with relatively strong traction tension.

[0069] Understandably, during the traction process of the pipeline entering the trench, the initial segment of each 100-meter tension segment corresponds to the tensioning process of the traction machine from static to stable tension, while the final segment corresponds to the tensioning process when entering the next tension segment. The tension values ​​of the two segments fluctuate drastically and need to be discarded before the average value is calculated.

[0070] Specifically, the traction force data of the laying section corresponding to the rolling stability indicator is retrieved. The laying mileage of the kilometer-long pipeline is divided into continuous tension segments every 100 meters. A sliding variance scan is applied to the tension data sequence within each tension segment. The sliding variance scan uses a fixed-length sliding window to slide sequentially along the tension data sequence, calculating the variance of the tension value within each window. In one embodiment, the length of the sliding window is taken to cover the data length of 10 seconds of sampling. The statistical interval of tension variance under the same pipe diameter and stable rolling conditions in historical construction is retrieved as a preset variance interval. Windows with sliding variance values ​​higher than the upper limit of the preset variance interval are identified as initial tensioning segments or final tensioning segments and are removed. Windows whose variance falls within the preset variance interval are retained and spliced ​​together to form a stable tension interval. The arithmetic mean of the tension data within the stable tension interval is calculated to obtain the average stable tension value of each tension segment.

[0071] Preferably, based on the average tensile force in the stable interval of each tensile segment, the average tensile force under stable rolling conditions of the same pipe diameter during historical construction is retrieved as a reference benchmark. The reference benchmark is then divided into adjacent ranges by a preset gap, where the preset gap is 10% of the reference benchmark. The reference benchmark is then divided upwards into equal range, slightly stronger range one, and slightly stronger range two, and downwards into equal range, slightly weaker range one, and slightly weaker range two.

[0072] In one embodiment, the average tensile force of each stable interval is compared segment by segment with the reference benchmark. If the average tensile force of a stable interval is more than one level higher than the reference benchmark, i.e., it falls into the range of being one or two levels stronger, then the tensile force segment is determined to be a tensile force segment with an average value more than one level higher. All tensile force segments with an average value more than one level higher are summarized along the kilometer-level pipeline laying mileage to identify laying sections with relatively strong traction tensile force classification.

[0073] S106. The laying sections with relatively strong traction force are compared with the sections in the drag trajectory map where the surface texture displacement is greater than the rolling displacement of the support roller. The overlapping positions of the two types of sections show a tendency for local slippage of the pipe surface, and the sections where local slippage occurs are identified to obtain the local slippage sections.

[0074] Based on the drag trajectory map, the surface texture displacement and the roller rolling displacement are reconstructed grid by grid along the pipeline laying mileage, according to the rolling and sliding components within the mileage-time grid unit. The difference between the surface texture displacement and the roller rolling displacement is used as the displacement deviation value of the mileage grid. The displacement deviation values ​​are sequentially concatenated along the pipeline laying mileage to obtain a displacement deviation sequence reflecting the degree to which the surface texture leads the roller rolling. For the displacement deviation sequence, it is determined grid by grid whether the displacement deviation value is greater than zero. If the displacement deviation value is greater than zero and continuously covers multiple consecutive mileage grids, then the mileage grid is determined to be a displacement leading grid where the surface texture displacement is greater than the roller rolling displacement. Adjacent displacement leading grids are concatenated and merged along the pipeline laying mileage to obtain the displacement leading section. Based on the displacement leading section and the traction tension graded as strong, the pipeline laying mileage coordinates are compared segment by segment to determine the overlapping position of the displacement leading section and the traction tension graded as strong in the pipeline laying mileage. It is confirmed that the overlapping position shows a local slippage tendency of the pipeline surface relative to the shallow groove support roller, and the section of the pipeline where local slippage occurs is identified, thus obtaining the local slippage section.

[0075] Understandably, the aforementioned drag trajectory map carries both rolling and sliding components across the pipeline laying mileage and time dimensions, while the aforementioned laying sections with relatively strong traction force classifications reflect an abnormal distribution of traction levels along the pipeline laying mileage. Determining the tendency for localized pipeline slippage requires correlating the kinematic texture advance phenomenon with the mechanically strong traction phenomenon on the pipeline laying mileage coordinates.

[0076] It should be noted that relying solely on the displacement lead phenomenon in the drag trajectory map is insufficient to rule out occasional false lead caused by noise introduced by pipe surface texture identification; similarly, relying solely on the phenomenon of excessive traction force cannot distinguish whether the excessive traction force originates from the relative slippage between the pipe and the shallow grooved roller, or from the additional resistance at the pipe bend or inlet. Only when both types of phenomena occur simultaneously on the mileage coordinates can a local slippage tendency of the pipe surface relative to the shallow grooved roller be confirmed.

[0077] Specifically, based on the drag trajectory map, the rolling component Kr and sliding component Ks are read grid by grid along the pipeline laying mileage, within each mile-time grid unit. Since the rolling component Kr is the ratio of the pipeline laying direction projection component to the synchronous rolling displacement of the support roller, the rolling component Kr in this grid is multiplied by the synchronous rolling displacement of the support roller to obtain the surface texture displacement. Based on the synchronous rolling displacement of the support roller s = ω × R × Δt, the support roller rotation speed ω, the support roller support surface radius R, and the time interval Δt between adjacent video frames are multiplied to obtain the support roller rolling displacement. Further, the surface texture displacement and the support roller rolling displacement are subtracted one by one along the mileage grid to obtain the displacement deviation value of the mileage grid. In one embodiment, the displacement deviation values ​​are concatenated along the pipeline laying mileage at a preset mileage step size to obtain a displacement deviation sequence reflecting the degree to which the surface texture precedes the rolling of the support roller. A displacement deviation value greater than zero indicates that the amount of movement of the outer wall of the pipeline along the laying direction at that mileage grid exceeds the rolling follow-up amount of the support roller surface, indicating kinematic signs of forward slippage of the pipeline surface relative to the shallow grooved support roller. For the displacement deviation sequence, the displacement deviation value is checked grid by grid to determine whether it is greater than zero.

[0078] Preferably, a continuous lead length threshold is set for the displacement deviation sequence, which is the length covering 5 mileage grids. If the displacement deviation value is greater than zero and continuously covers multiple consecutive mileage grids while satisfying the continuous lead length threshold, then the mileage grid is determined to be a displacement lead grid where the surface texture displacement is greater than the roller rolling displacement; otherwise, scattered single-grid lead is considered as acquisition noise and discarded. Adjacent displacement lead grids are serially connected and merged along the pipeline mileage to obtain displacement lead sections.

[0079] It is understood that the displacement advance section is distributed in continuous segments along the pipeline laying mileage, and each segment is marked with the start and end coordinates of the pipeline mileage.

[0080] In one embodiment, the start and end coordinates of the mileage of the section with advanced displacement and the start and end coordinates of the mileage of the section with relatively strong traction force are imported into the same pipeline laying mileage coordinate axis. The mileage coordinates are compared segment by segment to find the intersection of the two types of sections on the mileage coordinate axis. If both types of sections exist simultaneously within a certain pipeline laying mileage range, and the intersection length is higher than a preset overlap length threshold, then the mileage range is determined as the position where the two types of sections overlap on the pipeline laying mileage. If the two types of sections only exhibit advanced displacement without relatively strong traction force, or only exhibit relatively strong traction force without advanced displacement, then they are not considered as overlapping positions.

[0081] For example, the preset overlap length threshold is set to 30 meters, serving as an actionable parameter for determining overlap positions along a kilometer-level laying mileage. For all overlap positions, it is confirmed that there is a tendency for local slippage where the pipe surface slides relative to the shallow grooved rollers. These overlap positions are then connected in series along the pipeline laying mileage to identify the sections where local slippage occurs, thus obtaining the local slippage segments. These local slippage segments represent the actual operational locations where the pipe surface slides relative to the shallow grooved rollers, in the form of mileage intervals.

[0082] S107. Extract the sliding image corresponding to the local sliding segment from the continuous video stream of the pipe surface, smoothly extract the motion trajectory in the sliding image, and combine the contact indentation morphology with the rolling smoothness mark to evaluate whether the drag resistance of the pipe surface has decreased, and obtain the recognition result of reduced surface drag.

[0083] Based on the pipeline laying mileage range and timestamp corresponding to the local slip segment, a video frame subsequence corresponding to the mileage range is extracted from the continuous video stream of the pipeline surface. The video frame subsequence is named the sliding image sequence. The motion trajectory of the texture feature points on the outer wall of the pipeline is tracked frame by frame using the Lucas-Kanade optical flow method. The tracked motion trajectory is smoothed by cubic spline interpolation to remove isolated peak displacements, thus obtaining a smooth sliding trajectory that reflects the actual sliding behavior of the pipeline surface relative to the shallow grooved support roller. Based on the smooth sliding trajectory, the grid-by-grid amplitude of the smooth sliding trajectory is taken as the sliding amplitude value along the pipeline laying mileage. The contact depth and indentation distribution density at the corresponding mileage position of the local slip segment are read from the contact indentation morphology. The decreasing trend of the sliding amplitude value with mileage is compared with the increasing trends of the contact depth and indentation distribution density at the same mileage. If the sliding amplitude value decreases with mileage while the contact depth and indentation distribution density increase synchronously with mileage, it is determined that the pipeline surface within the local slip segment has transitioned from sliding behavior to rolling behavior, resulting in a sliding-to-rolling transition indicator. Based on the sliding-to-rolling transition indicator and the rolling stability indicator, the adjacency relationship between the mileage position of the sliding-to-rolling transition indicator and the mileage segment of the rolling stability indicator is compared along the pipeline laying mileage. If the mileage position of the sliding-to-rolling transition indicator is adjacent to the mileage segment of the rolling stability indicator, it is determined that the drag resistance of the pipeline surface decreases with the surface behavior transition, resulting in a reduced surface drag identification result.

[0084] It is understandable that the actual location of relative sliding between the pipe surface and the shallow grooved roller has been confirmed along the pipeline laying mileage. However, relative sliding does not equate to a continuous increase in drag resistance. The anomalous phenomenon of increased friction but decreased resistance mentioned in the background occurs precisely during the transition of the pipeline from sliding behavior to rolling behavior. Therefore, it is necessary to refine and extract the evolution of sliding behavior within the local sliding segment.

[0085] Specifically, based on the pipeline laying mileage range and timestamp corresponding to the local slip segment, a video frame subsequence corresponding to that mileage range is extracted from the continuous video stream on the pipeline surface, and this video frame subsequence is named a sliding image sequence. The sliding image sequence is strictly aligned with the entry time window of the local slip segment on the time axis, and only covers the pipeline laying mileage range where the local slip segment is located on the spatial axis, discarding video frames from other mileage ranges to reduce irrelevant texture interference.

[0086] It should be noted that the aforementioned drag trajectory mapping has already used the Lucas-Kanade optical flow method for tracking on continuous video streams, but the tracking of sliding image sequences here needs to serve the extraction of sliding behavior details, and the two operations are different. In one embodiment, the motion trajectory of the texture feature points on the outer wall of the pipe is tracked frame by frame using the Lucas-Kanade optical flow method, where the tracking window size is set to 15×15 pixels and the feature point spacing is set to 5 pixels, which has a higher spatial resolution than the 31×31 pixel window and 20 pixel spacing used in drag trajectory mapping. The texture feature points are taken from the weld edges, anti-corrosion coating particles, and random scratches on the outer wall of the pipe. The motion trajectory is formed by connecting the instantaneous displacement vectors between adjacent video frames, where each instantaneous displacement vector is denoted as d=(u,v), where u is the displacement component along the pipe laying direction and v is the displacement component perpendicular to the pipe laying direction. Furthermore, due to construction site vibration and video acquisition frame rate fluctuations, the original motion trajectory has sporadic peak displacement points. The motion trajectory is smoothly extracted using cubic spline interpolation. Compared to Gaussian filtering and Kalman filtering, cubic spline interpolation ensures the continuity of the curve at nodes without altering the original data point positions. The cubic spline interpolation constructs cubic polynomial segments between adjacent data points on the motion trajectory, constraining the continuity of the first and second derivatives of adjacent cubic polynomial segments at the nodes, resulting in a curve that smoothly transitions along the pipeline laying mileage. The vertical distance from each original displacement point to the fitted curve is calculated, and isolated peak displacement points with deviations exceeding 0.2 pixels are removed, yielding a smooth sliding trajectory reflecting the actual sliding behavior of the pipeline surface relative to the shallow grooved roller.

[0087] Understandably, during the transition from sliding to rolling contact on the pipe surface, the amount of sliding between the pipe's outer wall and the shallow grooved support roller gradually decreases. Simultaneously, the engagement between the pipe's outer wall and the shallow grooves of the support roller deepens, and the indentation per unit area increases. Based on the smooth sliding trajectory, the amplitude of each grid along the pipe laying mileage is taken as the sliding amplitude value. This sliding amplitude value reflects the combined magnitude of the lateral displacement and forward slippage of the pipe's outer wall relative to the shallow grooved support roller at that mileage grid. Simultaneously, the contact depth and indentation distribution density values ​​at the corresponding mileage positions of the local slippage segments are read from the contact indentation morphology.

[0088] Specifically, the sliding amplitude value, the contact depth value, and the indentation distribution density value are imported into the same pipeline laying mileage coordinate axis, and the evolution direction of the three is observed grid by grid along the pipeline laying mileage. If the sliding amplitude value decreases with mileage, while the contact depth and the indentation distribution density increase synchronously with mileage, and the correlation of the three trends on the pipeline laying mileage is higher than a preset correlation threshold, then it is determined that the pipeline surface in the local sliding segment has changed from sliding behavior to rolling behavior, and a sliding to rolling conversion mark is assigned to that mileage position; if only the sliding amplitude value decreases while the contact depth and the indentation distribution density do not increase significantly, it is determined to be an occasional sliding convergence, and the sliding to rolling conversion mark is not assigned.

[0089] For example, in the case of a gas transmission steel pipe with a diameter of 610 mm entering the groove on a shallow grooved roller, the preset deviation threshold is set to 0.2 pixels and the preset correlation threshold is set to 0.7, which serve as a set of implementable parameters for extracting details of sliding behavior and determining behavior transformation.

[0090] Preferably, based on the sliding-to-roll transition marker and the rolling stability marker obtained in the preceding steps, the adjacency relationship between the mileage of the sliding-to-roll transition marker and the mileage segment of the rolling stability marker is compared along the pipeline laying mileage. The adjacency relationship comparison starts from the mileage of the sliding-to-roll transition marker and extends forward along the pipeline laying direction by a preset adjacent mileage, examining whether the extended range falls within the mileage segment of the rolling stability marker.

[0091] In one embodiment, the preset adjacent mileage is determined based on the pipe outer diameter D and the roller spacing L, with a value ranging from 3 to 5 times the roller spacing. When the roller spacing is 12 to 15 meters, the preset adjacent mileage is 50 meters. If the mileage section where the sliding to rolling transition marker is located extends along the laying direction for the preset adjacent mileage and then enters the mileage section where the rolling stability marker is located, it is determined that the pipe surface has entered a continuous rolling state after experiencing local slippage. Correspondingly, the dragging resistance between the pipe and the shallow grooved roller decreases from sliding friction to rolling friction, resulting in a reduced surface dragging resistance identification result. The reduced surface dragging resistance identification result includes the starting mileage station number, the ending mileage station number, the time of transition, and the time of rolling stability, forming a record of sections where roller friction increases but traction dragging resistance decreases during the laying of a kilometer-level pipeline.

[0092] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for rapid laying of kilometer-scale continuous pipelines, characterized in that, The method includes: during the pipe entry and lifting process, acquiring an initial contact image of the support surface of the shallow grooved roller, simultaneously acquiring a continuous video stream of the pipe surface and data on the traction force of the pipe entering the groove; extracting the indentation contour from the indentation edge of the initial contact image, separating the indentation boundary, and forming a contact indentation morphology; identifying the roller friction increase section based on the contact indentation morphology; for the roller friction increase section, extracting the motion trajectory of pixels from the continuous video stream of the pipe surface, tracking the surface texture displacement path, and generating a drag trajectory map; analyzing the relative sliding trajectory of the pipe surface relative to the roller based on the drag trajectory map to obtain a rolling stability standard. The process involves: extracting the tension data of the laying section corresponding to the rolling stability indicator from the traction tension data of the pipeline entering the trench; combining the average tension of stable rolling conditions in historical construction data to identify laying sections with relatively strong traction tension; comparing the laying sections with relatively strong traction tension with the sections in the drag trajectory map where the surface texture displacement is greater than the roller rolling displacement to obtain local slip segments; extracting the sliding image corresponding to the local slip segment from the continuous video stream of the pipeline surface; smoothing the motion trajectory in the sliding image; and combining the contact indentation morphology with the rolling stability indicator to obtain the recognition result of reduced surface drag.

2. The method according to claim 1, characterized in that, The step of extracting the indentation contour from the indentation edge of the initial contact image, separating the indentation boundary, and forming the contact indentation morphology includes: using the Canny edge detection operator to extract the gray-level gradient abrupt change points in the indentation region as indentation edge pixels, connecting the indentation edge pixels point by point to form a closed indentation contour, separating the independent indentation boundary, using structured light projection to obtain the height difference from the bottom of the indentation to the original surface of the outer wall of the pipe as the contact depth value in the indentation region, and superimposing the contact depth value on the indentation contour to obtain the contact indentation morphology.

3. The method according to claim 1, characterized in that, The step of identifying the section of increased roller friction based on the contact indentation morphology includes: dividing the roller circumference and the pipeline laying direction into square statistical units; counting the pixels covered by the indentation outline and normalizing them by area to obtain the indentation distribution density value; splicing the indentation density distribution sequence along the pipeline laying mileage; retrieving the reference values ​​of indentation distribution density under sparse sliding friction conditions and dense contact conditions from historical construction data as the dividing threshold; merging the continuous mileage range where the density value of the statistical unit is higher than the dividing threshold; and identifying the section of increased roller friction.

4. The method according to claim 1, characterized in that, The process for extracting the motion trajectory of pixels from the continuous video stream on the pipe surface for the section where the roller friction increases, tracking the displacement path of the surface texture, and generating a drag trajectory map includes: extracting a video frame sequence with the same timestamp as the section where the roller friction increases; detecting pixels with significant grayscale gradients at welds, anti-corrosion coating particles, and random scratches on the pipe surface as tracking seed points; solving for the instantaneous displacement vector of the seed points between adjacent video frames, and tracking the displacement path frame by frame to obtain a continuous displacement trajectory cluster; determining the synchronous rolling displacement of the roller based on the roller rotation speed and the inter-frame time interval; dividing the projection component of the displacement trajectory along the pipe laying direction by the synchronous rolling displacement of the roller to obtain the rolling component, and dividing the projection component perpendicular to the laying direction by the synchronous rolling displacement of the roller to obtain the sliding component; and rasterizing the rolling component and the sliding component along the pipe laying mileage and time to obtain the drag trajectory map.

5. The method according to claim 1, characterized in that, The step of analyzing the relative sliding trajectory of the pipe surface relative to the support roller based on the dragging trajectory map to obtain a rolling stability indicator includes: reading the rolling component and sliding component of the dragging trajectory map grid by grid along the pipe laying mileage; taking the difference between the rolling component and the sliding component as the relative sliding amplitude; concatenating them to obtain the relative sliding trajectory; retrieving the upper limit of the relative sliding amplitude under continuous contact rolling conditions as the contact determination threshold; merging the grids of continuous mileage where the relative sliding amplitude is lower than the contact determination threshold; and assigning the rolling stability indicator.

6. The method according to claim 1, characterized in that, The step of extracting the tension data of the rolling stability marker corresponding to the laying section from the pipeline entry traction tension data, and combining it with the average tension of stable rolling conditions in historical construction data, to identify laying sections with relatively strong traction tension includes: dividing the tension data of the laying section corresponding to the rolling stability marker into continuous tension segments along the pipeline laying mileage according to a preset mileage length; taking the arithmetic mean of each tension segment to obtain the average tension value; retrieving the average tension value of stable rolling conditions of the same pipe diameter in historical construction as a reference benchmark for segment-by-segment comparison; and identifying the tension segment that exceeds the reference benchmark by more than one level as the laying section with relatively strong traction tension.

7. The method according to claim 6, characterized in that, The tensile data of the laying section corresponding to the rolling stability marker along the pipeline is divided into continuous tensile segments according to a preset mileage. The arithmetic mean of each tensile segment is obtained by taking the tensile mean. This includes: dividing the continuous tensile segments into segments every 100 meters; using a sliding variance scan on the tensile data sequence within each tensile segment; determining a preset variance interval based on the tensile variance statistical interval under historical stable rolling conditions; removing the initial tensioning segment and the final tensioning segment; retaining the tensile stability interval whose variance falls within the preset variance interval; taking the arithmetic mean of the tensile data within the tensile stability interval to obtain the tensile mean of the stability interval; using the tensile mean of the historical stable rolling conditions of the same pipe diameter as a reference benchmark, the upper and lower fluctuations are divided into adjacent levels by a preset level; and identifying tensile segments that are one level higher than the reference benchmark as laying sections with relatively strong traction tensile force.

8. The method according to claim 1, characterized in that, The step of comparing the paving sections with stronger traction force classifications with the sections in the drag trajectory map where the surface texture displacement is greater than the roller rolling displacement to obtain local slip segments includes: reconstructing the surface texture displacement and roller rolling displacement grid by grid along the pipeline paving mileage based on the rolling and sliding components in the drag trajectory map; concatenating the difference between the surface texture displacement and roller rolling displacement as displacement deviation values ​​to obtain a displacement deviation sequence; merging the positions where the displacement deviation value is greater than zero and continuously covers multiple consecutive mileage grids into a displacement leading segment; comparing the displacement leading segment with the paving sections with stronger traction force classifications according to mileage coordinates, and identifying the overlapping positions as the local slip segments.

9. The method according to claim 1, characterized in that, The step of extracting the sliding image corresponding to the local slip segment from the continuous video stream of the pipe surface and smoothing the motion trajectory in the sliding image includes: extracting the corresponding video frame subsequence from the continuous video stream of the pipe surface according to the pipe laying mileage range and timestamp corresponding to the local slip segment, naming it a sliding image sequence; tracking the motion trajectory of the texture feature points on the outer wall of the pipe frame by frame in the sliding image sequence; smoothing the motion trajectory using cubic spline interpolation; removing isolated peak displacements; and obtaining a smooth sliding trajectory.

10. The method according to claim 9, characterized in that, The process of smoothing the motion trajectory using cubic spline interpolation includes: taking the grid-by-grid amplitude of the smooth sliding trajectory along the pipeline mileage as the sliding amplitude value; reading the contact depth and indentation distribution density at the corresponding mileage position of the local slip segment from the contact indentation morphology; determining the sliding-to-roll transition marker; comparing the adjacency relationship between the mileage position of the sliding-to-roll transition marker and the mileage segment of the rolling stability marker to obtain the identification result of reduced surface drag.

Images