Full-dimensional monitoring method for pipeline corrosion under thermal insulation layer by infrared thermal imaging

Abstract

The present application relates to the technical field of pipeline detection, in particular to a method for infrared thermal imaging full-dimension monitoring of pipeline corrosion under thermal insulation layer, which obtains temperature distribution under thermal insulation layer, screens measuring points with consistent direction, extracts boundary continuous area to calculate thermal probability, identifies abnormal measuring points and extracts time sequence features, marks response delay measuring points based on model prediction, and generates CUI risk area division results. In the present application, by spatial sequence subdivision of pipeline surface temperature distribution data and heat flow direction trend judgment, a temperature difference direction consistency identification mechanism is constructed, a measuring point set with clear spatial heat conduction characteristics is obtained through continuity screening, further taking local extreme value of gradient change rate as the identification basis, combining spatial density and change intensity to construct thermal anomaly probability evaluation parameters, preliminary identification of abnormal points is realized, time sequence temperature parameters are introduced to construct multi-dimensional dynamic feature set, and multi-step temperature prediction is completed through normalized combination calculation after model training.

Description

Technical Field

[0001] This invention relates to the field of pipeline inspection technology, and in particular to a method for full-dimensional monitoring of pipeline corrosion under insulation layers using infrared thermal imaging. Background Technology

[0002] The field of pipeline inspection technology involves the inspection and evaluation of the structural status and operational safety of industrial pipelines during their service life. This includes determining defects, corrosion, cracks, and deposits on the outer and inner walls of the pipelines, and monitoring and preventing pipeline safety under different environmental conditions.

[0003] Among them, the infrared thermal imaging full-dimensional monitoring method for pipeline corrosion under insulation layer refers to the method of detecting possible metal corrosion under the insulation layer when the pipeline is covered with an insulation layer. It usually uses infrared thermal imaging equipment to collect the temperature distribution information of the pipeline surface. By combining the heat flow conduction characteristics formed by external heating or natural temperature difference with multi-angle and multi-directional thermal imaging observation, the thermal distribution data of the outer surface of the pipeline is obtained in a comprehensive manner.

[0004] Infrared thermal imaging information of pipelines collected without removing the insulation layer only reflects the external distribution of heat conduction and lacks the ability to dynamically analyze the internal thermal change trend. Existing technologies rely on observing thermal distribution images from different angles, which can cover the surface area but cannot accurately distinguish which locations have continuous thermal conduction anomalies. They also lack quantitative judgment of the time-series temperature rise trajectory and fail to distinguish the essential difference between abnormal temperature response and systemic temperature fluctuations, leading to the risk of false alarms in thermal interference environments. For example, in areas exposed to short-term sunlight or with airflow interference, the temperature distribution image may appear uneven, but no corrosion may actually occur. Due to the lack of temperature change trend comparison and spatial aggregation structure analysis, existing methods are difficult to accurately identify corrosion hazards and are prone to missing areas with strong thermal conduction delay under the insulation layer, thus causing the problem of local corrosion development not being detected in time. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and propose a method for full-dimensional monitoring of pipeline corrosion under insulation layers using infrared thermal imaging.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for full-dimensional monitoring of pipeline corrosion under insulation layer using infrared thermal imaging, comprising the following steps:

[0007] S1: Obtain the surface temperature distribution data of the pipe under the insulation layer, divide the surface temperature distribution data into equally spaced temperature measurement points, determine the consistency of the temperature change direction corresponding to each temperature measurement point, and filter the temperature measurement points with the same direction.

[0008] S2: Obtain the spatial location information and gradient change rate peak value of the temperature measuring points with consistent direction, identify the boundary continuous temperature measuring points with continuous gradient change between adjacent temperature measuring points, and calculate the boundary thermal probability of the boundary continuous temperature measuring points.

[0009] S3: Screen the abnormal response temperature measurement points on the outer wall of the pipe under the insulation layer whose boundary thermal probability exceeds the abnormal probability threshold, and extract the abnormal temperature time series feature dataset of the abnormal response temperature measurement points.

[0010] S4: Input the abnormal temperature time series feature dataset into the Long Short-Term Memory (LSTM) network model to perform multi-step temperature prediction of abnormal response temperature measurement points and mark the response delay temperature measurement points.

[0011] S5: Based on the spatial distribution of the response delay temperature measurement points, extract the measurement point cluster area on the outer wall of the pipe under the insulation layer, and generate the CUI risk area division result based on the proportion of the number of response delay temperature measurement points in the measurement point cluster area to the total number of temperature measurement points.

[0012] As a further aspect of the present invention, the step of obtaining the directional temperature measuring point specifically includes:

[0013] S111: Obtain the temperature distribution data of the pipe surface under the insulation layer, divide it into equally spaced temperature measurement points according to the spatial coordinates, establish a corresponding temperature value recording point at each temperature measurement point, and construct a temperature measurement point distribution array.

[0014] S112: For each temperature measuring point in the temperature measuring point distribution array, call the temperature value corresponding to the temperature measuring point and the temperature value of the adjacent temperature measuring point, calculate the direction of change of temperature difference in each direction, and determine whether the change trend of adjacent directions is the same. Mark the temperature measuring points with consistent directions as satisfying the direction consistency condition, and obtain the direction consistency identification mark set.

[0015] S113: Identify adjacent and continuous temperature measurement point chains that satisfy the direction consistency in the set of direction consistency identification markers, remove discrete measurement points with discontinuous direction change trends, and obtain temperature measurement points with consistent direction.

[0016] As a further aspect of the present invention, the step of obtaining the boundary thermal probability specifically includes:

[0017] S211: Obtain the spatial location information and corresponding gradient change rate of each temperature measuring point in the temperature measuring points with consistent direction, extract the location where the gradient change rate reaches the local maximum value as the representative value of the gradient abrupt change point, and obtain the gradient change rate peak set;

[0018] S212: Based on the set of peak gradient change rates, identify adjacent temperature measurement points with consistent gradient change rates that are greater than the boundary gradient threshold and whose positions are continuous, and mark the identified adjacent temperature measurement points with consistent gradient change rates as boundary continuous temperature measurement points, thereby generating a set of boundary continuous temperature measurement points.

[0019] S213: Calculate the proportional relationship between the density of each boundary continuous temperature measuring point in the set of boundary continuous temperature measuring points in the circumferential direction of the outer wall of the pipe under the insulation layer and the peak value of the gradient change rate by normalized weighted calculation method, and obtain the boundary thermal probability of the boundary continuous temperature measuring points.

[0020] As a further aspect of the present invention, the steps for obtaining the abnormal temperature time series feature dataset are as follows:

[0021] S311: Compare the boundary thermal probability corresponding to all continuous boundary temperature measurement points with the preset anomaly probability threshold point by point, filter out temperature measurement points whose boundary thermal probability exceeds the anomaly probability threshold, mark them as abnormal response temperature measurement points, and generate anomaly response temperature measurement point identification results.

[0022] S312: Based on the abnormal response temperature measurement point identification results, extract the temperature value sequence of each abnormal response temperature measurement point in a continuous time frame, calculate the temperature change between adjacent time frames, and obtain the total temperature rise value of the time series corresponding to the abnormal response temperature measurement point to obtain the abnormal response temperature measurement point temperature change parameter group.

[0023] S313: Based on the temperature change parameter group of the abnormal response temperature measurement point, combine and archive the temperature change amount, temperature change rate of adjacent time frames and cumulative temperature rise value corresponding to each abnormal response temperature measurement point to obtain an abnormal temperature time series feature dataset.

[0024] As a further aspect of the present invention, the step of obtaining the response delay temperature measurement point specifically includes:

[0025] S411: Input the abnormal temperature time series feature dataset into the Long Short-Term Memory (LSTM) network model, call the temperature change, temperature change rate and cumulative temperature rise value corresponding to each abnormal response temperature measurement point, perform multi-step temperature prediction calculation, and obtain the predicted temperature sequence of abnormal response temperature measurement points.

[0026] S412: Call the predicted temperature sequence of the abnormal response temperature measurement point respectively, call the actual temperature value sequence in the historical observation record that is the same as the prediction time period, and calculate the corresponding difference between the predicted value of each abnormal response temperature measurement point and the actual value in the historical record according to the time step to obtain the prediction residual sequence of the abnormal response temperature measurement point.

[0027] S413: Count the number of consecutive positive or negative time steps in each abnormal response temperature measurement point in the predicted residual sequence of the abnormal response temperature measurement point, and determine whether the duration of the time step exceeds the response delay judgment threshold. If the condition is met, mark the corresponding abnormal response temperature measurement point as a response delay state and obtain the response delay temperature measurement point.

[0028] As a further aspect of the present invention, the steps for obtaining the CUI risk area division results are specifically as follows:

[0029] S511: Based on the spatial coordinate information of the response delay temperature measurement points, locate the distribution area on the outer wall of the pipe under the insulation layer, identify the spatial clustering trend between adjacent response delay temperature measurement points, extract interconnected measurement point areas by combining the position continuity rule, and generate a group of clustered positions of response delay temperature measurement points.

[0030] S512: Based on the group of temperature measurement points clustered together, extract the clustered areas of measurement points with spatial connectivity within the outer wall of the pipe under the insulation layer, and calculate the proportion of the number of temperature measurement points in each clustered area to the total number of temperature measurement points, thereby obtaining the proportion of temperature measurement points with response delay.

[0031] S513: Based on the proportion of response delay temperature measurement points, combined with the location boundary and area information of the corresponding spatial cluster area, construct a spatial risk distribution structure covering the outer wall of the pipe under the insulation layer, and mark the area where the response delay temperature measurement points are densely distributed as the CUI risk area, and obtain the CUI risk area division result.

[0032] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0033] In this invention, a temperature difference direction consistency identification mechanism is constructed by subdividing the spatial sequence of pipeline surface temperature distribution data and judging the trend of heat flow direction. A set of measuring points with clear spatial thermal conduction characteristics is obtained through continuous screening. Further, based on the local extrema of the gradient change rate, thermal anomaly probability assessment parameters are constructed by combining spatial density and change intensity to achieve preliminary identification of anomaly points. Then, time series temperature parameters are introduced to construct a multi-dimensional dynamic feature set. Multi-step temperature prediction is completed by normalized combination calculation after model training. The prediction results are compared with historical data to form a residual sequence. The thermal response hysteresis behavior is judged by the persistence of deviation, and connectivity is identified based on the spatial aggregation trend of this behavior to complete the extraction of delayed aggregation regions. The risk distribution level is constructed by calculating the proportion of response measuring points in the aggregation area. Finally, the corrosion possibility area under the insulation layer is marked, effectively realizing the continuous transformation of thermal distribution data into corrosion risk area judgment, improving the spatial accuracy and temporal resolution of corrosion hazard identification, enhancing the ability to identify high-hysteresis heat conduction areas, reducing the misjudgment rate and the risk of omission, and improving the pertinence and applicability of the assessment results. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the main steps of the present invention;

[0035] Figure 2 This is a flowchart of step S1 of the present invention;

[0036] Figure 3 This is a flowchart of step S2 of the present invention;

[0037] Figure 4 This is a flowchart of step S3 of the present invention;

[0038] Figure 5 This is a flowchart of step S4 of the present invention;

[0039] Figure 6 This is a flowchart of step S5 of the present invention. Detailed Implementation

[0040] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0041] Please see Figure 1 This invention provides a technical solution: a method for full-dimensional monitoring of pipeline corrosion under insulation layer using infrared thermal imaging, comprising the following steps:

[0042] S1: Obtain the surface temperature distribution data of the pipe under the insulation layer, divide the surface temperature distribution data into equally spaced temperature measurement points, determine the consistency of the temperature change direction corresponding to each temperature measurement point, and filter the temperature measurement points with the same direction.

[0043] S2: Obtain the spatial location information and gradient change rate peak of temperature measurement points with consistent orientation, identify the boundary continuous temperature measurement points with continuous gradient change between adjacent temperature measurement points, and calculate the boundary thermal probability of the boundary continuous temperature measurement points.

[0044] S3: Screening abnormal response temperature measurement points on the outer wall of the pipe under the insulation layer where the boundary thermal probability exceeds the abnormal probability threshold, and extracting the abnormal temperature time series feature dataset of the abnormal response temperature measurement points.

[0045] S4: Input the abnormal temperature time series feature dataset into the Long Short-Term Memory (LSTM) network model to perform multi-step temperature prediction of abnormal response temperature measurement points and mark the response delay temperature measurement points.

[0046] S5: Based on the spatial distribution of response delay temperature measurement points, extract the measurement point cluster area on the outer wall of the pipe under the insulation layer, and generate the CUI risk area division result according to the proportion of the number of response delay temperature measurement points in the cluster area to the total number of temperature measurement points.

[0047] Please see Figure 2 The specific steps for obtaining temperature measurement points with consistent direction are as follows:

[0048] S111: Obtain the temperature distribution data of the pipe surface under the insulation layer, divide it into equally spaced temperature measurement points according to the spatial coordinates, establish a corresponding temperature value recording point at each temperature measurement point, and construct a temperature measurement point distribution array.

[0049] To acquire the surface temperature distribution data of the pipe under the insulation layer, the area is uniformly divided into equally spaced temperature measurement points based on spatial coordinates. A corresponding temperature value recording point is established at each measurement point, constructing a temperature measurement point distribution array. To obtain this data, a portion of the insulation layer needs to be removed, or thermal imaging images need to be acquired non-destructively using infrared imaging. The outer wall surface of the pipe is scanned and measured using thermal imaging equipment. A coordinate unfolding method is used to map the three-dimensional outer wall of the pipe into a two-dimensional coordinate plane. The circumferential position is unfolded to the Y direction (i.e., the pipe's circumference), and the axial position corresponds to the X direction (i.e., the pipe's length). To represent spatial layers or thickness, the Z direction can be introduced to stack multiple time-series images or different imaging angles, forming a complete data volume structure. The temperature measurement points are divided into uniform grids with a standard spacing of 1cm × 1cm in both the X and Y directions. After flattening the cylindrical outer wall, a regular two-dimensional measurement point array is formed, with each intersection point representing the temperature measurement point location. Through the mapping between the thermal image coordinates and the actual physical coordinates, the pixel coordinates extracted from the image are converted into corresponding spatial coordinates, establishing the correspondence between the measurement points and their actual locations. Each recording point corresponds to a specific physical location and its temperature value. The temperature value at the time of acquisition is recorded for each measuring point and kept synchronized through time stamps, forming a multi-dimensional array structure that includes spatial location, time series, and temperature data. To improve the accuracy of temperature measurement, multiple measurements are taken at each measuring point, and the arithmetic mean is calculated. Outliers with deviations exceeding a preset fluctuation range (such as ±0.5℃) are removed, ultimately forming a uniformly distributed and stable array of temperature measuring points.

[0050] S112: For each temperature measuring point in the temperature measuring point distribution array, call the temperature value corresponding to the temperature measuring point and the temperature value of the adjacent temperature measuring points, calculate the direction of change of temperature difference in each direction, and determine whether the change trend of adjacent directions is the same. Mark the temperature measuring points with consistent directions as satisfying the direction consistency condition, and obtain the direction consistency identification mark set.

[0051] For each temperature measuring point in the temperature measuring point distribution array, the system retrieves the corresponding temperature value of the measuring point and the temperature values ​​of adjacent temperature measuring points, calculates the direction of temperature difference change in each direction, and determines whether the change trends in adjacent directions are the same. Temperature measuring points with consistent directions are marked as meeting the direction consistency condition, resulting in a direction consistency identification tag set. In specific operations, the temperature value of the central measuring point is first read, and then the temperature values ​​of the eight surrounding measuring points (up, down, left, right, and diagonal) are retrieved. The temperature difference is calculated for each, and its positive or negative value is determined to determine the heat flow direction. The heat flow trend labels in the eight directions are recorded. When a measuring point exhibits a consistent temperature change trend (such as decreasing or increasing temperature) in any three or more adjacent directions and the temperature difference exceeds a set threshold (such as 0.8℃), the measuring point can be marked as meeting the direction consistency condition. In actual situations, there may be complex structures where some directions have positive trends and some have negative trends. For such cases, the system allows groups of adjacent directions with consistent trends to be regarded as a direction consistency subset, and as long as any one of these groups meets the judgment condition, the central measuring point can be marked as a direction consistency measuring point. If multiple directional consistency subsets exist for the same central measurement point, they will be recorded in parallel and merged uniformly in subsequent processing. Since each measurement point will be referenced multiple times in the neighborhood analysis, some measurement points may appear in multiple directional consistency label sets. To address this duplicate labeling phenomenon, a deduplication and priority retention mechanism is adopted in subsequent processing. This includes prioritizing the selection of the most representative chain group path based on the directional continuity length and temperature difference variation amplitude of the chain group to which the measurement point belongs, and merging duplicate label points during the clustering chain building process to avoid duplicate counting or directional label conflicts, thereby ensuring that the final selected set of directional consistency measurement points has uniqueness and consistency with physical heat conduction logic.

[0052] S113: Identify the chain of adjacent and continuous temperature measurement points that meet the direction consistency in the identification mark set, remove discrete measurement points with discontinuous direction change trends, and obtain temperature measurement points with consistent direction.

[0053] After constructing the directional consistency identification mark set, the measurement points in the mark set need to be further subjected to spatial structure identification to extract continuous measurement point chains with clear heat conduction path characteristics. Specifically, starting with each measurement point marked as directionally consistent, a recursive search is performed according to its direction label, following the same heat flow direction, to find the next adjacent measurement point with the same direction label and a temperature difference that still meets the threshold condition, thus constructing a continuous directionally consistent path. This recursive process requires maintaining the actual physical connectivity between measurement points in space; that is, any two adjacent measurement points in a chain must be within their eight-neighborhood, and the distance between them must not exceed twice the standard measurement point spacing. During the recursive process, if a change in the direction label is detected, or the ΔT value is lower than the set threshold, the chain connection is terminated. This path tracing operation is performed on each measurement point to identify directionally consistent paths with a length of not less than three consecutive measurement points, which are called directionally consistent measurement point chains.

[0054] Meanwhile, measurement points that are marked as directionally consistent in the marker set but cannot form a continuous chain that meets the length requirement in any directional path are considered discrete measurement points and should be removed. The criteria for judging discrete points are: in any direction, there are fewer than 3 consecutive adjacent directionally consistent measurement points, or there are directional breaks or reversals in heat flow trends, indicating a lack of spatial heat conduction continuity and inability to form an effective heat conduction structure; therefore, they are not retained in subsequent calculations. In addition, to prevent measurement points from repeatedly participating in multiple paths due to the intersection of multiple chain groups, conflict detection is performed on all formed chain groups. Chain group paths with longer continuous lengths or higher average temperature difference intensity are prioritized for retention, and other overlapping measurement points can be merged according to the directional consistency scoring strategy to ensure that each measurement point belongs to only one valid chain group.

[0055] Please see Figure 3 The specific steps for obtaining the boundary thermal probability are as follows:

[0056] S211: Obtain the spatial location information and corresponding gradient change rate of each temperature measuring point in the temperature measuring points with consistent orientation, extract the location where the gradient change rate reaches the local maximum value as the representative value of the gradient abrupt change point, and obtain the gradient change rate peak set;

[0057] To obtain the spatial location information of temperature measurement points with consistent orientation, it is necessary to combine the mapping relationship between the pixel coordinates extracted from the thermal imaging image and the layout of the measurement points in physical space, and establish a two-dimensional unfolded coordinate matrix of the measurement points in the physical coordinate system to ensure that each temperature measurement point has an identifiable location attribute. The "adjacent measurement points" mentioned here refer to the selected temperature measurement points with consistent orientation, not all original measurement points, to ensure the continuity of gradient analysis and the rationality of thermal conductivity. Based on this, the gradient change rate of each measurement point is calculated. Specifically, taking the measurement point as the center, the ratio of the temperature difference to the directional distance between it and the adjacent measurement points in the four diagonal directions (up, down, left, right, and four sides) is calculated, and the maximum value is selected as the representative gradient change rate of that measurement point for sensitive identification of local drastic temperature change trends. To further reduce the impact of measurement fluctuations, based on the previous temperature data being averaged from multiple measurements, the "iterative averaging" introduced here refers to the time-series averaging of the gradient change rate of the same measurement point at multiple time sections, eliminating occasional interference and improving the stability of gradient abrupt change judgment. After calculating the gradient rate of change, local maxima are determined for temperature measurement points with consistent orientations across all directions. If the gradient rate of change at a given measurement point is higher than that of all eight neighboring measurement points and exceeds 20% of the average gradient rate of change of its neighboring measurement points, then that point is marked as a gradient abrupt change point and recorded as a local extremum. The "20% threshold" set here is based on the characteristics of changes in normal heat flow decay regions and abrupt change regions extracted from field measurement data: in well-insulated areas, the temperature gradient is usually around 1.0℃ / cm, while in corroded or thermally resisted abrupt change regions, the gradient often rises to 1.2℃ / cm or higher. Therefore, using the neighborhood mean as a baseline and setting a 20% amplitude as the abrupt change judgment threshold can effectively distinguish between stable heat flow and abnormal abrupt change regions. Finally, all local extrema points that meet the above conditions are summarized to form a gradient rate of change peak set.

[0058] S212: Based on the set of peak gradient change rates, identify adjacent temperature measurement points with consistent gradient change rates that are greater than the boundary gradient threshold and whose positions are continuous, and mark the identified adjacent temperature measurement points with consistent gradient change rates as boundary continuous temperature measurement points, generating a set of boundary continuous temperature measurement points.

[0059] Starting from each measurement point in the peak gradient rate of change, a path expansion operation is performed sequentially in eight directions. Path expansion is only performed along the direction consistent with the heat flow direction of the starting measurement point. This means that all measurement points on the expanded path must have the same direction label as the starting point and must themselves be labeled as directionally consistent temperature measurement points. On each directional path, the gradient rate of change between each point and the previous measurement point is checked to see if it exceeds a set boundary gradient threshold. Simultaneously, the spatial distance between these measurement points is checked to see if it is continuous, i.e., the physical distance between them does not exceed twice the standard measurement point spacing. When three or more measurement points in a certain direction simultaneously meet both of these conditions, it indicates a stable high gradient abrupt change trend on the path. In this case, all directionally consistent temperature measurement points in the path are marked as boundary continuous temperature measurement points. The direction consistency requirement involves three levels of requirements: first, the starting point itself must be a directionally consistent temperature measurement point; second, all measurement points traversed during path expansion must also be directionally consistent temperature measurement points; and third, the path direction must be consistent with the heat flow direction of the starting point, with no direction jumps or reverse reversals allowed. The continuity condition refers to the dual requirements of continuous gradient value exceeding the threshold and spatial position continuity mentioned above. Ensure that the extracted boundary continuous temperature measurement points form a heat conduction path in physical space that is guided by the same heat flow direction, has a significant abrupt change in thermal gradient, and is continuously located.

[0060] S213: The boundary thermal probability of the boundary continuous temperature measurement points is obtained by calculating the proportional relationship between the density of each boundary continuous temperature measurement point in the circumferential direction of the outer wall of the pipe under the insulation layer and the peak value of the gradient change rate through normalized weighted calculation method.

[0061] To obtain the boundary thermal probability of continuous temperature measurement points at the boundary using a normalized weighted calculation method, two key variables must first be defined: the density of these measurement points along the circumference of the pipe's outer wall and the peak value of the gradient rate of change corresponding to each measurement point. In practice, the density is determined by counting the number of measurement points marked as continuous temperature points per unit circumferential length. A unit length of 10 cm is set, and a cylindrical mapping is constructed on the thermal image. The circumference is unfolded into a one-dimensional linear coordinate system, and segments are sequentially divided. The number of measurement points in each segment is recorded, and the ratio of the number of measurement points in a segment to the maximum number of measurement points in that segment is used as the normalized density weight. Next, the peak value of the gradient rate of change of all boundary measurement points within the corresponding segment is extracted, its average value is taken, and divided by the maximum gradient rate of change of all continuous temperature measurement points at the boundary to obtain the normalized gradient intensity coefficient. Finally, the boundary thermal probability of each measurement point or segment is calculated using the following weighted formula: ;in, : indicates the first Boundary thermal probability of the segment; : indicates the first The number of continuous temperature measurement points at the boundary of the segment; : Represents the maximum number of consecutive temperature measurement points at the boundaries of all segments; : indicates the first Peak value of the average gradient change rate at continuous temperature measurement points at the boundary of the segment; : Represents the peak value of the maximum gradient rate of change at all continuous temperature measurement points at the boundary.

[0062] In this formula, the two components normalize the distribution density and gradient intensity of the measurement points, respectively, so that different sections have a unified standard when comparing them. The product reflects the possibility that the region has both high density and high gradient jump, and the final boundary thermal probability value is obtained. Between 0 and 1, it can be used for ranking and classifying the intensity of thermal anomalies, serving as an important quantitative basis for identifying risk areas. In actual calculations, data extraction relies on thermal imaging image analysis modules, coordinate expansion and transformation tools, and numerical processing tools such as Python scripts or nested functions in Excel to achieve batch calculations. All parameters come from the collected coordinate-temperature dataset and its second-order calculation results, without involving external estimates or subjective human settings. Only in the normalization stage is a comparison benchmark set, with the current sample's maximum value used as the denominator, thus ensuring that the proportional expression has internal data consistency.

[0063] Suppose a section of pipe under insulation is subjected to thermal imaging measurement. This section is unfolded into cylindrical coordinates and divided into 5 equal-length segments, each 10cm long, with a total circumference of 50cm. The analysis objective is to calculate the boundary thermal probability of the 3rd segment. .

[0064] Obtain the number of continuous temperature measurement points at the boundary After the image coordinates are expanded, continuous temperature measurement points are marked at the boundaries: Segment 1: 8 points, Segment 2: 12 points, Segment 3: 20 points, Segment 4: 10 points, Segment 5: 15 points. Therefore: , ;

[0065] Obtain the average gradient rate of change of continuous temperature measurement points at the boundary. The rate of change of all continuous boundary measurement points in the third segment is as follows (unit: ℃ / cm): 1.25, 1.30, 1.20, 1.10, 1.40, 1.35, 1.28, 1.22, 1.18, 1.26 (a total of 10 sample points).

[0066] Calculate the average: ;

[0067] Assume the peak value of the maximum gradient rate of change at all continuous temperature measurement points at the boundary: ;

[0068] Substitute into the formula to calculate Substitute the above values ​​into:

[0069] ;

[0070] The result indicates that the overall assessment of the third segment in terms of boundary thermal risk is 0.8957.

[0071] The core objective of the entire calculation process is to identify potential thermal anomaly regions on the outer wall of the pipe beneath the insulation layer using a quantitative approach and to assess their risk level. The process first divides the outer wall of the pipe into extended regions and counts the number of continuous temperature measurement points identified as boundary points within each segment, reflecting the spatial density distribution of temperature abrupt changes in that region. Second, it extracts the temperature gradient change rate of these measurement points and calculates their average value to measure the severity of the thermal abrupt change. In the calculation, the relative weight of density is obtained by comparing the number of measurement points with the largest number in each segment, and the intensity weight is obtained by comparing the average gradient change rate with the largest gradient change rate in the entire segment. The product of these two weights is the boundary thermal probability of the region, representing the likelihood that the region possesses both concentrated measurement points and severe temperature differences. The reason for using a product approach is that the density of the measurement point distribution reflects the spatial scale of the thermal anomaly, while the gradient change rate reflects its severity; combining the two can fully express the true characteristics of the thermal instability risk region. Considering only one of these factors might lead to misjudgments of occasional strong gradients or large-area slight changes.

[0072] Please see Figure 4 The specific steps for obtaining the abnormal temperature time series feature dataset are as follows:

[0073] S311: Compare the boundary thermal probability corresponding to all continuous boundary temperature measurement points with the preset anomaly probability threshold point by point, filter out temperature measurement points whose boundary thermal probability exceeds the anomaly probability threshold, mark them as abnormal response temperature measurement points, and generate anomaly response temperature measurement point identification results.

[0074] When identifying anomalies at all continuous boundary temperature measurement points, the boundary thermal probability value corresponding to each measurement point needs to be compared one by one with a pre-set anomaly probability threshold. This threshold should be calibrated based on typical thermal anomaly cases in historical data, typically selecting a lower probability limit that stably reflects the thermal anomaly region as a benchmark. For example, after multiple experiments and analysis of actual thermal failure cases, it was found that areas with boundary thermal probabilities higher than the 75th percentile mostly correspond to areas with higher thermal failure risk. Therefore, this 75th percentile value can be used as the anomaly probability threshold. All measurement points with thermal probability values ​​exceeding this threshold are considered response points of thermal anomalies and are identified as anomaly response temperature measurement points. During the judgment process, each boundary thermal probability needs to be retrieved and compared point by point, and the calculated values ​​corresponding to each measurement point are called. If the value is greater than the set threshold, a marker is added, and its spatial coordinates and timestamp are recorded. Finally, all measurement points that meet the conditions are summarized to generate anomaly response temperature measurement point identification results, forming a complete spatial distribution information set containing anomaly measurement points.

[0075] S312: Based on the abnormal response temperature measurement point identification results, extract the temperature value sequence of each abnormal response temperature measurement point in a continuous time frame, calculate the temperature change between adjacent time frames, and obtain the total temperature rise value of the time series corresponding to the abnormal response temperature measurement point to obtain the abnormal response temperature measurement point temperature change parameter group.

[0076] After identifying the abnormal response temperature measurement points, the time-series temperature data for each marked measurement point needs to be extracted during continuous monitoring. This involves acquiring the temperature value sequence corresponding to the measurement point across multiple consecutive sampling times, typically with a time interval set between 1 and 5 seconds, determined based on the rate of temperature change under actual operating conditions to ensure sufficiently detailed temperature change trends are captured. Based on this, the temperature change between adjacent time points is calculated sequentially, i.e., the temperature at the current moment minus the temperature at the previous moment, yielding the temperature difference value at each time step. The sign of the temperature difference value reflects the direction of temperature change; a positive value indicates a temperature increase, and a negative value indicates a temperature decrease.

[0077] Meanwhile, the temperature differences at all time points (regardless of whether they are positive or negative) will be summed sequentially to obtain the total temperature rise value of that measuring point over the entire monitoring period. This value is used to determine whether there is a trend of continuous temperature increase or instantaneous change. For example, at a certain anomaly point, if the temperature shows an increasing trend for five consecutive time periods and the cumulative temperature rise exceeds 5 degrees Celsius, it indicates that there is a stable and significant thermal change at that point.

[0078] During the processing, abnormal fluctuations in temperature values ​​also need to be removed. For example, if the temperature rises abnormally for a short period of time but then falls rapidly, it should be smoothed using a moving average or rate of change judgment mechanism, or the data should be marked as invalid. Finally, a complete set of abnormal response temperature change parameters is constructed, including the amount of temperature change (including the direction of change), the time point, and the cumulative temperature rise value.

[0079] S313: Based on the temperature change parameter group of the abnormal response temperature measurement point, the temperature change amount corresponding to each abnormal response temperature measurement point, the temperature change rate of adjacent time frames and the cumulative temperature rise value are combined and archived to obtain the abnormal temperature time series feature dataset.

[0080] Based on the temperature change parameter set, the dynamic temperature characteristics of each abnormal response temperature measurement point need to be structurally combined to establish an abnormal temperature time-series feature dataset for subsequent analysis. During the construction process, three core parameters corresponding to each measurement point need to be extracted: the amount of temperature change between time periods, the rate of temperature change between every two time points, and the cumulative temperature rise from the start of anomaly identification to the current time point. These parameters are archived and stored in chronological order. The data structure records the temperature change interval, rate gradient, and cumulative trend within each time period, forming a basic information set for abnormal behavior modeling and trend prediction. This feature dataset can be used to identify typical abnormal evolution patterns, such as continuous gradual increases, periodic fluctuations, or sudden jumps. Structurally, it is stored in a multi-field format divided by time segments, with each record corresponding to a time period and containing multiple indicators such as temperature difference, rate, and cumulative change.

[0081] Please see Figure 5 The specific steps for obtaining the response delay temperature measurement point are as follows:

[0082] S411: Input the abnormal temperature time series feature dataset into the Long Short-Term Memory (LSTM) network model, call the temperature change, temperature change rate and cumulative temperature rise value corresponding to each abnormal response temperature measurement point, perform multi-step temperature prediction calculation, and obtain the predicted temperature sequence of abnormal response temperature measurement points.

[0083] To accurately predict the future state of temperature measurement points with abnormal responses, this solution employs a Long Short-Term Memory (LSTM) network model as the prediction tool. It combines key input variables from the abnormal temperature time-series feature dataset with a time-related weight structure fitted by the model for multi-step temperature prediction. In practical engineering, the LSTM model captures short-term fluctuations and long-term trends in the temperature sequence through memory units and gating mechanisms, dynamically mapping input features to historical change structures. However, during deployment, especially in inference, the complex internal structure of the LSTM model can be approximated through formulas, allowing the influence relationships between various input features to be presented numerically. This simplifies the model prediction into a concrete, executable temperature calculation formula, as follows:

[0084] ;

[0085] This formula is based on the structural reasoning ability and time series feature recognition logic within the LSTM model. Specifically, it approximates the model output as a weighted prediction term based on the current temperature, superimposed with a term consisting of three normalized temperature features. In the formula, This represents the predicted temperature value for the next moment, in degrees Celsius, and is the final output. This represents the measured temperature value at the current moment, obtained through on-site thermal imaging or temperature sensing. This represents the temperature change between the current moment and the previous moment, i.e. It is used to reflect short-term temperature change trends; and These represent the average and standard deviation of historical temperature changes, respectively, and serve as the reference benchmarks required for standard normalization of this feature. The data source is the historical observation records of the corresponding measuring points under long-term normal operating conditions. The rate of temperature change at the current moment, representing the temperature increment per unit time, is expressed by... The value is calculated by dividing by the sampling time interval, reflecting the intensity of instantaneous change; its corresponding normalization parameter is... and These are the mean and standard deviation of the historical rates, respectively. This represents the cumulative temperature rise, which is the sum of temperature changes over all consecutive moments starting from when an abnormal response is identified at a certain measuring point. It is used to reflect the cumulative effect of temperature rise in the medium to long term, and its corresponding normalized parameter is... and , respectively, are the mean and standard deviation of the historical cumulative temperature rise, obtained by statistically analyzing the historical cumulative temperature rise over a specified period and calculating the mean and standard deviation respectively. , and These are weighted coefficients corresponding to the three types of normalized features. These coefficients are not manually set, but are automatically adjusted and learned during the LSTM model training phase by the error minimization principle, reflecting the degree of influence of each type of feature in the final temperature evolution.

[0086] The "current moment" and "previous moment" are defined based on the sampling time interval. Typically, the sampling time interval is set to 1 to 5 seconds, depending on the rate of temperature change. The time difference between each sampling interval is called a "time step," and each time step corresponds one-to-one with the actual time interval (such as per second or per minute). For example, if temperature data is collected every 1 second, then each time step represents a 1-second interval. During prediction, the LSTM model makes predictions based on the current time step (i.e., the temperature value at the current sampling moment) and the temperature change at the previous moment.

[0087] This formula expresses the next-time temperature output by LSTM during the prediction phase, based on the model training results and the joint modeling of the three features mentioned above with the current temperature. It is suitable for progressively sliding prediction scenarios with the current time as the base point. By iteratively using this formula, multi-step temperature prediction can be achieved, that is, using the current prediction result as the basis for the next input, and sequentially deriving the temperature sequence of consecutive time points.

[0088] Assuming the current time is the 10th second, the target for prediction is the temperature at the 11th second. Current temperature (10th second): ℃, sampling interval: 1 second; current temperature change: ℃; Current rate of temperature change: ℃ / s; Current cumulative temperature rise (total temperature rise since the start of the abnormal response): ℃; historical average of temperature change ℃; Standard deviation of temperature change ℃; historical average rate of temperature change ℃ / s; Standard deviation of temperature change rate ℃ / s; historical average cumulative temperature rise ℃; cumulative temperature rise standard deviation ℃; weighting coefficients obtained from model training , , .

[0089] Normalization of temperature change: ;

[0090] Normalization of the rate of temperature change: ;

[0091] Cumulative temperature rise normalization: ;

[0092] Substitute into the formula to calculate the predicted temperature:

[0093] ;

[0094] The predicted temperature at the 11th second is approximately 52.07℃, a significant increase from the current 48.0℃. This indicates that the abnormal response measurement point is continuously heating up and exhibits a strong thermal anomaly trend, consistent with the previous judgment logic.

[0095] S412: Call the predicted temperature sequence of the abnormal response temperature measurement point respectively, call the actual temperature value sequence in the historical observation record that is the same as the prediction time period, and calculate the corresponding difference between the predicted value of each abnormal response temperature measurement point and the actual value in the historical record according to the time step to obtain the prediction residual sequence of the abnormal response temperature measurement point.

[0096] After outputting the predicted temperature sequence for each abnormal response temperature measurement point, the measured temperature sequence that strictly corresponds to the predicted time period needs to be extracted from the historical observation records. The two must be perfectly aligned in time, maintaining a time step correspondence. Based on this, the difference between the predicted and actual temperature values ​​is calculated according to each time step; that is, the predicted value at each time step is subtracted from the measured value at the corresponding time, yielding the predicted residual sequence for that measurement point within that time period. Historical observation records are collected from the abnormal response temperature measurement points during the normal continuous monitoring phase before they are identified as abnormal. This process automatically records the temperature values ​​of the measurement points using thermal imaging equipment or temperature sensors at a set sampling interval. The sampling interval can be set to once per second or once every 5 seconds, depending on the actual operating conditions. The collection time range generally covers data from at least several minutes before the prediction begins, to fully reflect the temperature change background of the measurement points. The recorded content includes the measurement point number, collection time, temperature value, and corresponding spatial coordinates, arranged in chronological order to form a time series data, which serves as the basis for subsequent comparison with the predicted sequence. When accessing historical records, the system traces back the corresponding time period from the start time of the predicted sequence, extracting the measured temperature values ​​at the corresponding time points to ensure precise time step matching. The difference calculation process is executed point-by-point, comparing predicted and measured values ​​at the same time points to obtain the residual value for each step. To ensure analytical accuracy, records with missing time points or abnormal data formats are removed from the historical data, retaining only continuous segments with complete time and valid data for analysis. The resulting residual sequence can be used to reflect the difference between the predicted trend and actual temperature changes.

[0097] S413: Count the number of consecutive positive or negative time steps in each abnormal response temperature measurement point in the prediction residual sequence of abnormal response temperature measurement points, and determine whether the duration of the time step exceeds the response delay judgment threshold. If the condition is met, mark the corresponding abnormal response temperature measurement point as a response delay state and obtain the response delay temperature measurement point.

[0098] After obtaining the predicted residual sequence, statistical analysis needs to be performed on the time periods with consecutive deviations in the same direction to identify whether the measuring point has response delay characteristics. The specific judgment rule is as follows: taking each abnormal response temperature measuring point as a unit, count the number of consecutive positive or negative time steps in its residual sequence. If there are time periods with consecutive positive deviations (i.e., the predicted value is consistently higher than the actual value) or consecutive negative deviations (i.e., the predicted value is consistently lower than the actual value), and the length of this time period exceeds the set response delay judgment threshold, then the measuring point is considered to have a lag in actual temperature performance even when the predicted temperature has risen or fallen, and is thus marked as a response delay temperature measuring point. The threshold setting principle should be based on the typical thermal inertia response time of the measuring point. For example, based on the response delay time statistically determined by the pipe material, fluid temperature response characteristics, and heat conduction path, if the actual change delay after a temperature abrupt change under normal conditions does not exceed 5 seconds, then the judgment threshold can be set to 5 consecutive time steps (with a sampling period of 1 second). That is, only when the deviation direction appears consecutively 5 times or more is it marked as a delayed response. The judgment process also needs to exclude the influence of short-term fluctuation errors. For example, if there is a single directional interruption followed by continuous offset, a sliding window method should be used for inclusion processing. Finally, the measurement points that meet the response delay judgment criteria are summarized into a response delay temperature measurement point set.

[0099] Please see Figure 6 The specific steps for obtaining the CUI risk area classification results are as follows:

[0100] S511: Based on the spatial coordinate information of the response delay temperature measurement points, locate the distribution area on the outer wall of the pipe under the insulation layer, identify the spatial clustering trend between adjacent response delay temperature measurement points, extract the interconnected measurement point areas by combining the position continuity rules, and generate a group of clustered positions of response delay temperature measurement points.

[0101] First, the spatial coordinates of the response delay temperature measurement points are mapped to their actual physical locations on the outer wall of the pipe beneath the insulation layer. This process is achieved by establishing a mapping relationship between image coordinates and physical coordinates, converting the temperature measurement points from image coordinates to the actual coordinates of the pipe's outer surface. Next, the spatial connectivity between adjacent measurement points is determined. The maximum allowable distance between measurement points is generally set to twice the standard sampling interval. If the distance between two measurement points does not exceed this maximum distance, they are considered potentially connected. Furthermore, the temperature change directions of the two measurement points must be consistent for them to be considered spatially connected. Specifically, if the direction of the temperature difference change between two measurement points is consistent within eight possible directions (up / down, left / right, and four diagonal directions), then the two measurement points are considered spatially connected. By recursively searching for all connectable measurement points, multiple response delay measurement points are grouped into the same cluster. To exclude isolated points or pseudo-clusters, a minimum connectivity threshold is set, requiring a region to contain at least three consecutive connected measurement points to be considered to have effective spatial clustering. Each valid cluster is assigned a unique number, forming a cluster location group for the response delay temperature measurement points.

[0102] S512: Based on the cluster location group of response delay temperature measuring points, extract the cluster area of ​​measuring points with spatial connectivity within the outer wall of the pipe under the insulation layer, and calculate the proportion of the number of response delay temperature measuring points in each cluster area to the total number of temperature measuring points, and obtain the proportion value of response delay temperature measuring points.

[0103] In the extraction and analysis of clustered locations of response delay temperature measurement points, the spatial occupancy of the clustered area is first obtained by calculating the boundary of the region where each measurement point is located. Specifically, for all response delay measurement points within each clustered area, the boundary envelope of the region is determined based on their spatial coordinates, generating a minimum closed polygon that accurately represents the physical spatial extent of the region. For example, if a clustered area contains 5 response delay measurement points, their spatial coordinates are used to calculate the minimum enclosing polygon of the region, assuming the area of ​​this polygon is 10 square meters. Next, within this boundary envelope, the total number of all temperature measurement points and the number of response delay temperature measurement points are counted. Assuming that in a specific clustered area, the total number of temperature measurement points is 100, of which 20 are response delay measurement points, then the proportion of response delay measurement points is 20 / 100 = 0.2, meaning that the proportion of response delay temperature measurement points in this clustered area is 20%. The calculation of this percentage is based on standard grid mapping. By defining a grid within the clustered region, using the total number of measurement points within that region as the denominator and the number of response delay temperature measurement points as the numerator, the percentage of response delay temperature measurement points within that region is finally obtained.

[0104] The threshold setting is typically determined based on the statistical relationship between the clustering distribution and defect location in historical pipeline corrosion events. For example, historical data might show that when the proportion of response delay measurement points exceeds 50%, the corrosion risk in the pipeline area increases significantly. Therefore, a threshold of 0.5 is set; if the proportion of response delay measurement points in a region exceeds 50%, the region is considered a "high response delay clustering region." For instance, in another clustering region with 60 response delay measurement points out of a total of 100 points, the proportion is 60 / 100 = 0.6, indicating that this region belongs to a high response delay clustering region and may have a high corrosion risk. Therefore, this method can effectively identify response delay regions of varying severity.

[0105] S513: Based on the proportion of response delay temperature measurement points, combined with the location boundary and area information of the corresponding spatial clustering area, construct a spatial risk distribution structure covering the outer wall of the pipe under the insulation layer, and mark the area where response delay temperature measurement points are densely distributed as CUI risk area, and obtain the CUI risk area division result.

[0106] Based on the proportion of temperature measurement points with response delay and the spatial boundary information of each clustered area on the outer wall of the pipeline under the insulation layer, a spatial risk distribution structure needs to be constructed to identify potential corrosion risks. This structure divides the pipeline outer wall area into multiple sub-segments, embedding the corresponding clustered area proportion value within each segment, and using color coding or hierarchical labels to identify different risk levels. Clustered areas with proportion values ​​exceeding a set threshold are marked as CUI risk zones, i.e., areas where corrosion under the insulation layer may occur. During the construction process, the area of ​​the clustered areas also needs to be calculated, i.e., the area is calculated using the coordinate boundaries of the envelope polygon to obtain the specific coverage area of ​​each risk zone, thereby establishing a risk level map covering the entire outer wall of the pipeline.

[0107] This method uses infrared thermal imaging to acquire temperature distribution data on the pipe surface and combines spatial and temporal series analysis to accurately determine the corrosion situation under the insulation layer. First, by analyzing the consistency of temperature change direction, temperature measuring points with stable heat flow characteristics are screened out, and their spatial location and gradient change information are further extracted to identify areas of abrupt thermal gradient changes. Next, by comparing the thermal probabilities of these areas, abnormal temperature response points are identified, and a long short-term memory network is used for multi-step temperature prediction to identify measuring points with response delays. These measuring points reflect areas where corrosion problems may exist. Finally, by analyzing the spatial distribution of the response delay measuring points, high-risk clustering areas are identified, and CUI (Corrosion Under Insulation) risk areas are generated based on the degree of measuring point clustering.

[0108] CUI (Conditional Environmental Intake) risk areas refer to regions on the pipe surface beneath the insulation layer that exhibit significant temperature anomalies and response delays, identified through infrared thermal imaging monitoring. Temperature changes in these areas reflect abnormal heat conduction behavior on the pipe surface, potentially indicating internal corrosion issues. Specifically, CUI risk areas encompass the following levels of risk:

[0109] Early corrosion risk: These areas exhibit mild temperature anomalies, possibly manifesting as small temperature changes, but a certain decrease in thermal conductivity already exists, suggesting that minor corrosion or thermal resistance buildup may be beginning to appear on the pipe surface. These areas may be early signs of corrosion and require regular monitoring and early intervention to prevent further deterioration.

[0110] Moderate corrosion risk: These areas exhibit significant temperature variations, characterized by large temperature fluctuations and marked differences from the surrounding areas. This may indicate more severe corrosion or uneven thermal conductivity due to moisture buildup under the insulation layer, suggesting that corrosion within the pipeline has progressed to a moderate level. Further inspection and intervention are needed in these areas to prevent further corrosion.

[0111] Severe Corrosion Risk: In these areas, drastic temperature anomalies are often accompanied by significant thermal gradient changes or temperature rises, indicating severely impaired heat flow conduction, most likely due to heat flow delays caused by severe corrosion within the pipeline. Corrosion in these areas has reached a relatively severe level, potentially leading to reduced pipeline strength or leaks, necessitating repair or replacement.

[0112] These CUI risk areas correspond to different levels of corrosion risk. Through infrared thermal imaging technology and time series analysis, dynamic monitoring and precise location of pipeline corrosion status can be achieved.

[0113] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for full-dimensional monitoring of pipeline corrosion under insulation layer using infrared thermal imaging, characterized in that, Includes the following steps: S1: Obtain the surface temperature distribution data of the pipe under the insulation layer, divide the surface temperature distribution data into equally spaced temperature measurement points, determine the consistency of the temperature change direction corresponding to each temperature measurement point, and filter the temperature measurement points with the same direction. S2: Obtain the spatial location information and gradient change rate peak value of the temperature measuring points with consistent direction, identify the boundary continuous temperature measuring points with continuous gradient change between adjacent temperature measuring points, and calculate the boundary thermal probability of the boundary continuous temperature measuring points. S3: Screen the abnormal response temperature measurement points on the outer wall of the pipe under the insulation layer whose boundary thermal probability exceeds the abnormal probability threshold, and extract the abnormal temperature time series feature dataset of the abnormal response temperature measurement points. S4: Input the abnormal temperature time series feature dataset into the Long Short-Term Memory (LSTM) network model to perform multi-step temperature prediction of abnormal response temperature measurement points and mark the response delay temperature measurement points. S5: Based on the spatial distribution of the response delay temperature measurement points, extract the measurement point cluster area on the outer wall of the pipe under the insulation layer, and generate the CUI risk area division result based on the proportion of the number of response delay temperature measurement points in the measurement point cluster area to the total number of temperature measurement points.

2. The method for full-dimensional monitoring of pipeline corrosion under insulation layer using infrared thermal imaging according to claim 1, characterized in that, The specific steps for obtaining the temperature measurement points with the same direction are as follows: S111: Obtain the temperature distribution data of the pipe surface under the insulation layer, divide it into equally spaced temperature measurement points according to the spatial coordinates, establish a corresponding temperature value recording point at each temperature measurement point, and construct a temperature measurement point distribution array. S112: For each temperature measuring point in the temperature measuring point distribution array, call the temperature value corresponding to the temperature measuring point and the temperature value of the adjacent temperature measuring point, calculate the direction of change of temperature difference in each direction, and determine whether the change trend of adjacent directions is the same. Mark the temperature measuring points with consistent directions as satisfying the direction consistency condition, and obtain the direction consistency identification mark set. S113: Identify adjacent and continuous temperature measurement point chains that satisfy the direction consistency in the set of direction consistency identification markers, remove discrete measurement points with discontinuous direction change trends, and obtain temperature measurement points with consistent direction.

3. The method for full-dimensional monitoring of pipeline corrosion under insulation layer using infrared thermal imaging according to claim 2, characterized in that, The specific steps for obtaining the boundary thermal probability are as follows: S211: Obtain the spatial location information and corresponding gradient change rate of each temperature measuring point in the temperature measuring points with consistent direction, extract the location where the gradient change rate reaches the local maximum value as the representative value of the gradient abrupt change point, and obtain the gradient change rate peak set; S212: Based on the set of peak gradient change rates, identify adjacent temperature measurement points with consistent gradient change rates that are greater than the boundary gradient threshold and whose positions are continuous, and mark the identified adjacent temperature measurement points with consistent gradient change rates as boundary continuous temperature measurement points, thereby generating a set of boundary continuous temperature measurement points. S213: Calculate the proportional relationship between the density of each boundary continuous temperature measuring point in the set of boundary continuous temperature measuring points in the circumferential direction of the outer wall of the pipe under the insulation layer and the peak value of the gradient change rate by normalized weighted calculation method, and obtain the boundary thermal probability of the boundary continuous temperature measuring points.

4. The method for full-dimensional monitoring of pipeline corrosion under insulation layer using infrared thermal imaging according to claim 3, characterized in that, The specific steps for obtaining the abnormal temperature time-series feature dataset are as follows: S311: Compare the boundary thermal probability corresponding to all continuous boundary temperature measurement points with the preset anomaly probability threshold point by point, filter out temperature measurement points whose boundary thermal probability exceeds the anomaly probability threshold, mark them as abnormal response temperature measurement points, and generate anomaly response temperature measurement point identification results. S312: Based on the abnormal response temperature measurement point identification results, extract the temperature value sequence of each abnormal response temperature measurement point in a continuous time frame, calculate the temperature change between adjacent time frames, and obtain the total temperature rise value of the time series corresponding to the abnormal response temperature measurement point to obtain the abnormal response temperature measurement point temperature change parameter group. S313: Based on the temperature change parameter group of the abnormal response temperature measurement point, combine and archive the temperature change amount, temperature change rate of adjacent time frames and cumulative temperature rise value corresponding to each abnormal response temperature measurement point to obtain an abnormal temperature time series feature dataset.

5. The method for full-dimensional monitoring of pipeline corrosion under insulation layer using infrared thermal imaging according to claim 4, characterized in that, The specific steps for obtaining the response delay temperature measurement point are as follows: S411: Input the abnormal temperature time series feature dataset into the Long Short-Term Memory (LSTM) network model, call the temperature change, temperature change rate and cumulative temperature rise value corresponding to each abnormal response temperature measurement point, perform multi-step temperature prediction calculation, and obtain the predicted temperature sequence of abnormal response temperature measurement points. S412: Call the predicted temperature sequence of the abnormal response temperature measurement point respectively, call the actual temperature value sequence in the historical observation record that is the same as the prediction time period, and calculate the corresponding difference between the predicted value of each abnormal response temperature measurement point and the actual value in the historical record according to the time step to obtain the prediction residual sequence of the abnormal response temperature measurement point. S413: Count the number of consecutive positive or negative time steps in each abnormal response temperature measurement point in the predicted residual sequence of the abnormal response temperature measurement point, and determine whether the duration of the time step exceeds the response delay judgment threshold. If the condition is met, mark the corresponding abnormal response temperature measurement point as a response delay state and obtain the response delay temperature measurement point.

6. The method for full-dimensional monitoring of pipeline corrosion under insulation layer using infrared thermal imaging according to claim 5, characterized in that, The specific steps for obtaining the CUI risk area classification results are as follows: S511: Based on the spatial coordinate information of the response delay temperature measurement points, locate the distribution area on the outer wall of the pipe under the insulation layer, identify the spatial clustering trend between adjacent response delay temperature measurement points, extract interconnected measurement point areas by combining the position continuity rule, and generate a group of clustered positions of response delay temperature measurement points. S512: Based on the group of temperature measurement points clustered together, extract the clustered areas of measurement points with spatial connectivity within the outer wall of the pipe under the insulation layer, and calculate the proportion of the number of temperature measurement points in each clustered area to the total number of temperature measurement points, thereby obtaining the proportion of temperature measurement points with response delay. S513: Based on the proportion of response delay temperature measurement points, combined with the location boundary and area information of the corresponding spatial cluster area, construct a spatial risk distribution structure covering the outer wall of the pipe under the insulation layer, and mark the area where the response delay temperature measurement points are densely distributed as the CUI risk area, and obtain the CUI risk area division result.

Images