Intelligent detection system and method for scale layer of buried sewage pipeline based on Internet of Things

Abstract

The invention relates to the technical field of intelligent detection, and discloses a buried sewage pipeline scale layer intelligent detection system and method based on the Internet of Things, and the method comprises the steps: collecting the optical albedo and acoustic backscattering data of the section of a target pipeline at each annular angle; a cross-correlation value and time delay of the optical deviation value and the acoustic deviation value are calculated in a post-event time window, and an optical recovery time constant and an acoustic recovery time constant are obtained through fitting of an exponential decay model; the optical positive deviation area and the acoustic negative deviation area are determined, the cross-correlation value and the time delay are combined, then strength parameters are obtained through calculation, and circumferential connectivity sorting is carried out according to the strength parameters to locate candidate scale layer girdles; obtaining a material dominant data set corresponding to the deviation cause material dominant through combined screening; and in response to the material dominant data set, splicing the candidate scale layer annular belt of each section along the axial direction of the pipeline according to the strength parameters to obtain a scale layer belt center track.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent detection, more specifically, it relates to a buried sewage pipeline scale layer intelligent detection system and method based on Internet of Things. BACKGROUND

[0002] The urban drainage pipe network is mostly made of concrete or reinforced concrete structure, and is in a high-humidity, low-flow and sulfur-containing environment for a long time. Sulfate-reducing bacteria in sewage can generate hydrogen sulfide gas, which is oxidized to sulfuric acid by sulfur-oxidizing bacteria in the condensate film in the top arch area, and the local pH value can be reduced to below 2, which can easily cause acid erosion damage to the concrete. In this process, calcium hydroxide in the concrete reacts with sulfuric acid to form gypsum, which further reacts with aluminum to form ettringite, thereby forming a secondary mineral zone with a layered distribution at the top arch. Gypsum and ettringite are both porous structures that can contain and retain liquid water in their pores. The formation of these minerals makes the top arch surface appear as a white, loose and wet band. Its optical performance is high albedo and clear texture, so it can be easily identified under laser scanning. However, in acoustic detection, the pore water changes the acoustic impedance of the interface after filling, which significantly reduces the reflection energy of the sound wave, resulting in a dark signal.

[0003] When the pipeline is short-time full due to heavy rain or pump station scheduling, the secondary mineral zone of the top arch will be infiltrated by water. After the water recedes, the surface quickly recovers to a dry and bright white appearance, so the optical detection result quickly returns to the baseline state. However, the pore structure of the secondary mineral zone can cause water to be retained in the material for a long time, and even if the pipeline water level has returned to normal, the acoustic signal will still be continuously attenuated due to local water saturation. This phenomenon of different recovery of optical and acoustic signals can easily be misinterpreted as equipment calibration error, flow disturbance or gain adjustment problem during detection, thereby masking the true existence and development of the secondary mineral zone. The distortion of the detection result caused by this phenomenon seriously affects the accurate determination of the corrosion state of the buried sewage pipeline and the maintenance decision. SUMMARY

[0004] The present application provides a buried sewage pipeline scale layer intelligent detection system and method based on Internet of Things, which solves the technical problems raised in the background art.

[0005] In a first aspect, a buried sewage pipeline scale layer intelligent detection method based on Internet of Things, comprising:

[0006] determining a pre-time window and a post-time window;

[0007] collecting optical albedo and acoustic backscatter data of the cross section of the target pipeline at each ring angle;

[0008] the optical deviation value and the acoustic deviation value are normalized based on the median and the median absolute deviation of the optical albedo and the acoustic backscatter data within a pre-time window;

[0009] the cross-correlation value and the time delay of the optical deviation value and the acoustic deviation value are calculated within a post-time window, and the optical recovery time constant and the acoustic recovery time constant are fitted by an exponential decay model;

[0010] the optical positive deviation area and the acoustic negative deviation area are determined, and the cross-correlation value and the time delay are combined to calculate the intensity parameter, and the candidate scale ring belt is positioned according to the intensity parameter for annular connectivity sorting;

[0011] the material dominant data set corresponding to the material dominant deviation is obtained through joint screening;

[0012] in response to the material dominant data set, the candidate scale ring belt of each section is spliced according to the intensity parameter along the axial direction of the pipeline to obtain the scale band center trajectory.

[0013] Further, the pre-time window and the post-time window are determined, including:

[0014] the end time of the short flooding event or the high filling event in the target pipeline is determined;

[0015] the pre-time window and the post-time window are segmented with the end time as the center; wherein the pre-time window and the post-time window completely cover the preset time period.

[0016] Further, the optical albedo and acoustic backscatter data of the section of the target pipeline at each annular angle are collected, including:

[0017] at a fixed time interval within the preset time period, and at the annular angle the optical albedo data and the acoustic backscatter original amplitude of each section are collected;

[0018] the annular angle the corresponding compensation coefficient is calculated, and the product of the compensation coefficient and the acoustic backscatter original amplitude is taken as the acoustic backscatter data;

[0019] wherein, the calculation process of the compensation coefficient is as follows:

[0020] the acoustic round trip time of the annular angle i is determined ;

[0021] the compensation coefficient of the annular angle i is calculated according to the acoustic round trip time ; wherein, represents the medium sound speed, including the sewage medium sound speed or the air medium sound speed, represents the calibration gain of the sonar receiving link.

[0022] Further, the optical deviation value and the acoustic deviation value are obtained by normalizing the median value of the optical albedo data and the median absolute deviation of the acoustic backscatter data in the pre-time window, including:

[0023] In the pre-time window, the median value of the optical albedo data is taken as the optical baseline median value;

[0024] The median value of the absolute difference between the optical albedo data and the optical baseline median value is multiplied by a normal distribution constant to obtain an optical robust scale;

[0025] In the pre-time window, the median value of the acoustic backscatter data is taken as the acoustic baseline median value;

[0026] The median value of the absolute difference between the acoustic backscatter data and the acoustic baseline median value is multiplied by a normal distribution constant to obtain an acoustic robust scale;

[0027] The ratio of the difference between the optical albedo data at time k and the optical baseline median value at the ring angle i to the optical robust scale is taken as the corresponding optical deviation value;

[0028] The ratio of the difference between the acoustic backscatter data at time k and the acoustic baseline median value at the ring angle i to the acoustic robust scale is taken as the corresponding acoustic deviation value.

[0029] Further, the cross-correlation value and the time delay of the optical deviation value and the acoustic deviation value are calculated in the post-time window, and the optical recovery time constant and the acoustic recovery time constant are fitted by an exponential decay model, including:

[0030] The optical deviation average value and the acoustic deviation average value corresponding to the calculated optical deviation value and the acoustic deviation value in the pre-time window are determined respectively; and the optical albedo data and the acoustic backscatter data in the post-time window are respectively de-biased based on the optical deviation average value and the acoustic deviation average value to obtain the optical de-bias matrix and the acoustic de-bias matrix ; wherein the i-th column of the r-th row in the optical de-bias matrix represents the optical de-bias value at the ring angle i at the r-th time; and the i-th column of the r-th row in the acoustic de-bias matrix represents the acoustic de-bias value at the ring angle i at the r-th time;

[0031] The discrete lag amount is initialized, and the cross-correlation value of the optical de-bias sequence and the acoustic de-bias sequence is calculated according to the discrete lag amount , as follows:

[0032]

[0033] wherein, denotes the discrete lag amount of the circumferential angle i at time r corresponding cross-correlation value, denotes the post-time window, denotes the optical depolarization value of the optical depolarization matrix at circumferential angle i at time r denotes the average value of the optical depolarization matrix, denotes the acoustic depolarization value of the acoustic depolarization matrix at circumferential angle i at time r denotes the acoustic depolarization value of the acoustic depolarization matrix at circumferential angle i at time r denotes: shifting the acoustic depolarization matrix to the time axis by a discrete lag amount after, still in the post-time window, the average value of the acoustic depolarization value;

[0034] determining the discrete lag amount along which the cross-correlation value is minimized to obtain the time delay;

[0035] least square fitting each optical depolarization value as an optical event post-offset amount by a forward exponential decay model to obtain an optical recovery time constant;

[0036] least square fitting each acoustic depolarization value as an acoustic event post-offset amount by a forward exponential decay model to obtain an acoustic recovery time constant;

[0037] wherein, the forward exponential decay model is specifically as follows:

[0038]

[0039] wherein, denotes the end time of the short-time submergence event or the high-fulfillment event in the target pipeline, denotes the optical recovery time constant or the acoustic recovery time constant to be fitted, denotes the initial offset amplitude, specifically: the ratio of the first parameter in the optical depolarization sequence and the average value of the optical albedo data in the pre-time window, or the ratio of the first parameter in the acoustic depolarization sequence and the average value of the acoustic backscattering data in the pre-time window.

[0040] Further, the optical positive offset area and the acoustic negative offset area are determined, combined with the cross-correlation value and the time delay, and then the intensity parameter is calculated to sort and locate the candidate scale ring according to the intensity parameter, including:

[0041] determining the optical offset value and the acoustic offset value of the circumferential angle i at each time in the post-time window;

[0042] If the optical deviation value of the ring direction angle i at the rth moment is positive, the optical deviation value is merged into the first positive part set; the acoustic deviation value of the ring direction angle i at the rth moment is processed by sign inversion to obtain the corresponding inverted acoustic deviation value; if the inverted acoustic deviation value of the ring direction angle i at the rth moment is positive, the inverted acoustic deviation value is merged into the second positive part set;

[0043] The first positive part set is integrated in the post-time window to obtain an optical positive deviation area;

[0044] The second positive part set is integrated in the post-time window to obtain an acoustic positive deviation area;

[0045] The product of the optical positive deviation area, the acoustic positive deviation area, the time delay and the cross-correlation value corresponding to the time delay is taken as the intensity parameter of the ring direction angle i;

[0046] If the intensity parameters of the ring direction angle i and the ring direction angle i+1 are both greater than 0, the ring direction angle i and the ring direction angle i+1 are merged to form a connected segment;

[0047] The range of the ring direction angle of each connected segment is determined, and the intensity parameters in the connected segment are integrated according to the range of the ring direction angle to obtain an intensity integral;

[0048] The connected segment with the maximum intensity integral is taken as the candidate scale layer ring belt corresponding to the cross section.

[0049] Further, joint screening includes first screening, second screening and third screening;

[0050] The first screening includes: traversing each moment in the post-time window; if all the optical depolarization values corresponding to the moment r in the optical depolarization matrix are positive numbers, and all the acoustic depolarization values corresponding to the moment r in the acoustic depolarization matrix are negative numbers, the optical albedo and acoustic backscattering data corresponding to the rth moment are retained to form a first screening data set;

[0051] The second screening includes: for the qth moment in the first screening data set, after deviating from the time delay on the time axis, still belonging to the post-time window, and the cross-correlation values of all the ring direction angles corresponding to the qth moment are negative values, the optical albedo and acoustic backscattering data corresponding to the qth moment are retained to form a second screening data set;

[0052] The third screening includes: for the ring direction angle i corresponding to the pth moment in the second screening data set, if the acoustic recovery time constant is greater than the optical recovery time constant, the optical albedo and acoustic backscattering data corresponding to the ring direction angle i of the pth moment are retained to form a third screening data set;

[0053] The third screening data set is marked as a material dominant data set.

[0054] Further, in response to the material dominant data set, candidate scale ring belts of each section are spliced according to the intensity parameters in the axial direction of the pipeline to obtain a scale belt center track, including:

[0055] The candidate scale ring belts of each section are determined respectively, and the intensity parameters of each circumferential angle corresponding to the candidate scale ring belts are determined;

[0056] The circumferential center angle of the candidate scale ring belt is calculated based on the intensity parameters, as follows:

[0057] The circumferential center angle is obtained by calculating the ratio of the product of each circumferential angle and the corresponding intensity parameter to the sum of the intensity parameters of each circumferential angle ;

[0058] An optimization function is constructed , as follows:

[0059]

[0060] wherein, the sum of the intensity parameters of each circumferential angle corresponding to the candidate scale ring belt of the s-th section, the circumferential center angle of the s-th section, the angle weight, , the interval distance between the s-th section and the t-th section; The optimization function is solved based on a greedy search to obtain a scale belt center track; wherein the scale belt center track is used for visualizing the scale of the target pipeline.

[0061] In the second aspect, a buried sewage pipeline scale intelligent detection system based on the Internet of Things is applied to any of the buried sewage pipeline scale intelligent detection methods based on the Internet of Things, and includes:

[0062] A time window determination module is configured to determine a pre-time window and a post-time window.

[0063] A multi-modal data acquisition module is configured to acquire optical albedo and acoustic backscatter data of a section of a target pipeline at each circumferential angle.

[0064] A deviation value normalization calculation module is configured to normalize the optical deviation value and the acoustic deviation value based on the median and the median absolute deviation of the optical albedo and the acoustic backscatter data in the pre-time window.

[0065] A deviation value normalization calculation module is configured to normalize the optical deviation value and the acoustic deviation value based on the median and the median absolute deviation of the optical albedo and the acoustic backscatter data in the pre-time window.

[0066] ​A time delay and recovery time constant analysis module is configured to calculate a cross-correlation value and a time delay of the optical deviation value and the acoustic deviation value within a post-time window, and to obtain an optical recovery time constant and an acoustic recovery time constant by fitting an exponential decay model;

[0067] A circumferential candidate scale layer ring belt positioning module is configured to determine an optical positive deviation area and an acoustic negative deviation area, combine the cross-correlation value and the time delay, and calculate an intensity parameter, and position a candidate scale layer ring belt according to the circumferential connectivity sorting of the intensity parameter;

[0068] A material dominant data set screening module is configured to obtain a material dominant data set corresponding to a material dominant deviation cause by joint screening;

[0069] An axial scale layer band center trajectory splicing module is configured to respond to the material dominant data set, splice the candidate scale layer ring belt of each section according to the intensity parameter along the axial direction of the pipeline, and obtain a scale layer band center trajectory.

[0070] The present application has the beneficial effects that: by introducing optical and acoustic dual-mode detection data, combining time window division, deviation value normalization, anti-correlation and time delay analysis, and exponential decay model fitting, the system can effectively overcome the detection distortion problem caused by different synchronization of optical and acoustic signal recovery in the traditional method. The method not only realizes accurate identification of the secondary mineral scale layer in the buried sewage pipeline, but also can perform multi-dimensional screening and splicing based on the intensity parameter and the material dominant data set, and finally form a spatial trajectory visualization result of the scale layer band, thereby greatly improving the accuracy and reliability of the detection, and providing a scientific basis for corrosion state evaluation and maintenance decision of the urban drainage network. BRIEF DESCRIPTION OF DRAWINGS

[0071] Fig. 1 is a module diagram of the present application;

[0072] Fig. 2 is a first flowchart of the present application;

[0073] Fig. 3 is a second flowchart of the present application. DETAILED DESCRIPTION

[0074] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that the discussion of these implementations is merely meant to provide a better understanding of the subject matter described herein and can be changed in function and arrangement without departing from the scope of the present disclosure. Each of the various examples can omit, substitute or add various procedures or components as appropriate or desired. In addition, features described in relation to some examples can also be combined in other examples.

[0075] Example One:

[0076] As Figs. 1-3 shown, the buried sewage pipeline scale layer intelligent detection method based on Internet of Things comprises:

[0077] determining the pre-time window and the post-time window;

[0078] collecting the optical albedo and acoustic backscatter data of the cross section of the target pipeline at each ring angle;

[0079] normalizing the median and median absolute deviation of the optical albedo and acoustic backscatter data in the pre-time window to obtain optical deviation value and acoustic deviation value;

[0080] calculating the cross-correlation value and time delay of the optical deviation value and acoustic deviation value in the post-time window, and fitting to obtain the optical recovery time constant and acoustic recovery time constant through the exponential decay model;

[0081] determining the optical positive deviation area and the acoustic negative deviation area, combining the cross-correlation value and the time delay, and then calculating the intensity parameter to sort and locate the candidate scale ring according to the ring connectivity;

[0082] obtaining the material dominant data set corresponding to the material dominant deviation through joint screening;

[0083] in response to the material dominant data set, splicing the candidate scale ring of each cross section according to the intensity parameter along the pipeline axial direction to obtain the scale layer band center trajectory.

[0084] It needs to be explained that for the gypsum ettringite secondary mineral scale layer formed by biological corrosion in the top arch area of the buried sewage pipeline, the cross-modal time delay characteristics of the porous structure are used to realize accurate detection after the optical signal quickly recovers and the acoustic signal delays to recover. Specifically, first, the pre-event time window and the post-event time window are divided with the end time of the flooding event as the boundary; the optical albedo and acoustic backscattering data of each ring angle of the same pipeline section are collected, and the median and median absolute deviation of the pre-event data are used for normalization to obtain dimensionless and noise-free optical and acoustic deviation values; in the post-event window, the time delay (acoustic lag length) is determined by calculating the cross-correlation of the two types of deviation values, and the optical and acoustic recovery time constants (capture recovery speed difference) are obtained by fitting the exponential decay model; the intensity parameters are calculated by combining the optical positive deviation area (optical persistent high), the acoustic negative deviation area (acoustic persistent low), the cross-correlation value and the time delay; the candidate scale ring belt of ring connection is located; after three screenings (direction consistency, time delay and inverse correlation, recovery time constant order), non-material factors such as equipment fluctuation and water flow interference are excluded, and the data dominated by the scale layer are retained; finally, the candidate ring belt is spliced along the pipeline axis to obtain the continuous scale belt center trajectory. Thus, based on the unique cross-modal time delay difference caused by the porous structure of the scale layer, the precise positioning and tracking of the top arch scale layer which is difficult to identify by traditional methods are realized, and the problem of distorted detection results easily disturbed is solved.

[0085] It needs to be explained that the collection, transmission, processing and application of optical albedo and acoustic backscattering data are deeply integrated with Internet of Things technology to form a closed-loop intelligent detection system of perception→transmission→analysis→decision, as follows:

[0086] The optical albedo data are collected by high-definition optical sensors (such as laser radar, industrial camera) deployed on the pipeline detection robot (or fixed monitoring node), the sensor scans the pipeline section at a preset ring angle interval, and real-time captures the light reflection intensity at different angles; the acoustic backscattering data are obtained by the sonar array carried by the same node, the specific frequency sound wave is emitted and the reflected signal is received, and the original amplitude is obtained after preliminary processing by the hardware circuit. These sensors are integrated into the Internet of Things perception terminal, which has low power consumption characteristics and can be powered by battery or pipeline environment energy harvesting technology, realizing long-term autonomous operation and providing a stable physical basis for data collection.

[0087] The optical and acoustic raw data collected by the perception terminal are uploaded in real time to an edge gateway or a cloud platform through an Internet of Things communication module (such as a low-power wide-area network technology such as NB-IoT, LoRaWAN, etc.). Considering that the closed environment of the buried pipeline may cause signal attenuation, the system uses multi-hop routing or relay nodes to enhance communication, ensuring the integrity and timeliness of the data under complex working conditions. Data encryption and verification mechanisms are embedded during transmission to ensure the security and reliability of Internet of Things data, solving the problems of data lag and low transmission efficiency in traditional manual detection.

[0088] The data processing layer relies on the computing power support of the Internet of Things cloud platform to realize intelligent analysis combining distributed and centralized systems. The platform first preprocesses the uploaded raw data, then calls the preset algorithm module for automated analysis, generating characteristic parameters such as optical deviation and acoustic recovery time constant. At the same time, the platform has edge computing capabilities, which can complete some real-time processing on the terminal side, reducing the cloud computing power pressure, and embodying the technical advantages of Internet of Things cloud-edge collaboration.

[0089] Through the visual interface of the Internet of Things platform, the analysis results are presented to the operation and maintenance personnel in real time, and spatial positioning is realized in combination with the pipeline GIS map. When abnormal expansion of the scale layer is detected, the system automatically triggers the early warning mechanism, notifies relevant personnel through SMS, APP push, etc., forming an intelligent closed loop from data perception to decision execution. Based on the architecture of the Internet of Things, the limitations of traditional pipeline detection relying on manual inspection and long analysis period are broken through, realizing dynamic monitoring, remote diagnosis and precise early warning of the scale layer of buried sewage pipelines, fully embodying the technical integration of the Internet of Things and intelligent detection.

[0090] In an embodiment of the present application, the pre-event time window and the post-event time window are determined, comprising:

[0091] Determining the end time of the short-term flooding event or the high-fulfillment event in the target pipeline;

[0092] In detail, the short-term flooding event refers to the situation that the pipeline is temporarily submerged by water due to backflow caused by heavy rain, etc. The high-fulfillment event refers to the situation that the water level in the pipeline rises sharply for a short time due to pump station scheduling, etc. These two types of events will cause the scale layer (mainly the secondary mineral zone of gypsum ettringite) on the top of the pipeline to be submerged by water, and the end time is the starting point of the recovery period of the scale layer.

[0093] The end time is used as the center to divide the pre-event time window and the post-event time window; wherein the pre-event time window and the post-event time window completely cover the preset time period.

[0094] In detail, the pre-event time window is located before the end time and is used to collect stable state data of the pipeline before the event occurs, including optical albedo and acoustic backscatter data of each circumferential angle. The data in the pre-event time window will be used as a benchmark for subsequent analysis, reflecting the normal performance of the scale layer without interference from the event. The post-event time window is located after the end time and is used to collect recovery process data of the pipeline after the event ends, focusing on capturing the dynamic changes in optical and acoustic responses of the scale layer during the gradual desorption of moisture, especially the time-lag property of the different recovery of the two.

[0095] The preset time period is determined according to the actual working condition of the pipeline and includes the entire event process and a sufficient time period before and after the event. The preset time period needs to ensure that the pre-event window can fully reflect the stable state and the post-event window can completely record the recovery process.

[0096] In an embodiment of the present application, the optical albedo and acoustic backscatter data of the cross section of the target pipeline at each circumferential angle are collected, including:

[0097] In the preset time period, the optical albedo and acoustic backscatter data of each circumferential angle are collected at fixed time intervals. The optical albedo data and acoustic backscatter raw amplitude of each cross section are collected.

[0098] In detail, the data are collected at fixed time intervals to ensure the regularity of the time series, allowing the subsequent analysis to clearly capture the dynamic changes before and after the event. The data are collected at circumferential angles to cover the entire circumferential range of the pipeline cross section, ensuring that the optical and acoustic data of each angle of the inner wall of the pipeline are recorded, providing a spatial coordinate benchmark for locating the circumferential distribution of the scale layer.

[0099] The circumferential angle is divided into a plurality of intervals, and the optical albedo and acoustic backscatter data of each interval are collected. The corresponding compensation coefficient is calculated, and the product of the compensation coefficient and the acoustic backscatter raw amplitude is used as the acoustic backscatter data.

[0100] In detail, the acoustic backscatter raw amplitude is affected by the propagation distance and the device gain, and cannot directly reflect the true acoustic properties of the pipeline interface. Therefore, the compensation coefficient must be corrected. The calculation of the compensation coefficient is based on the round-trip time of the sound wave, because the round-trip time of the sound wave is directly related to the propagation distance, and the propagation distance is equal to half of the product of the medium sound speed and the round-trip time. The intensity of the acoustic signal decreases with the square of the propagation distance, so the compensation coefficient needs to include the square term of the propagation distance to offset this attenuation effect.

[0101] In detail, the compensation coefficient is multiplied by the acoustic backscatter raw amplitude to obtain the corrected acoustic backscatter data. Thus, the acoustic data can truly reflect the reflection characteristics of the pipeline interface (including the scale layer) to the sound wave.

[0102] The calculation process of the compensation coefficient is as follows:

[0103] Determination of the sound wave round trip time of the ring angle i ;

[0104] According to the sound wave round trip time, the compensation coefficient of the ring angle i is calculated ; wherein, The medium sound speed includes the sewage medium sound speed or the air medium sound speed, The calibration gain of the sonar receiving link.

[0105] In detail, the medium sound speed distinguishes sewage and air because the sound speed difference is significant, and if not distinguished, the propagation distance calculation deviation will be caused, and then the compensation accuracy is affected. The calibration gain of the sonar receiving link is used to eliminate the influence of the gain fluctuation of the device itself, and to ensure that the acoustic data collected at different times or by different devices are consistent.

[0106] It should be noted that the optical albedo refers to the reflection ability of the inner wall surface of the pipeline to visible light (or laser), which is manifested as the optical signal intensity at a specific ring angle position in detection. The scale layer (mainly gypsum, ettringite secondary mineral zone) in the top arch area has a strong diffuse reflection ability to visible light due to the granular and needle-like microstructure, and has a high optical albedo, which appears bright white; and the optical albedo of the complete concrete matrix is relatively low. By collecting optical albedo data at different times and different ring angles, the optical signal recovery process caused by the evaporation of water on the scale layer surface after the event can be tracked.

[0107] It should be noted that the acoustic backscatter data refers to the reflection intensity of the inner wall surface of the pipeline to the sound wave, which is obtained by collecting the original amplitude of the sonar device and then performing compensation processing (to compensate for the propagation distance attenuation and device gain influence). In the propagation process of the sound wave, the difference in acoustic impedance of the interface determines the reflection intensity: the acoustic impedance difference between the complete concrete and the air (or water) is large, and the acoustic backscatter is strong; and the scale layer is easy to retain water due to the porous structure, and the acoustic impedance difference between the water and the scale layer is small, resulting in weakened sound wave reflection and low acoustic backscatter. By collecting acoustic backscatter data at different times and different ring angles, the acoustic signal delayed recovery process caused by the water held in the pores of the scale layer after the event can be captured.

[0108] In an embodiment of the present application, the optical deviation value and the acoustic deviation value are obtained by normalizing the median and the median absolute deviation of the optical albedo and the acoustic backscatter data in the pre-event time window, including:

[0109] In the pre-event time window, the median of the optical albedo data is taken as the optical baseline median;

[0110] The median of the absolute difference between the optical albedo data and the optical baseline median is multiplied by a normal distribution constant to obtain an optical robust scale;

[0111] In detail, in the pre-event time window, the median of the optical albedo data is selected as the optical baseline median. The median is selected instead of the mean because the median is more robust to outliers in the data, avoiding baseline shifts caused by occasional impurities in the pipeline that reflect light, speckle flickering, and other interference, and better represents the stable optical properties of the pipeline surface before the event. Then, the absolute difference between the optical albedo data and the optical baseline median is calculated, and the median of these absolute differences is taken, multiplied by a normal distribution constant (usually 1.4826, which is derived from the theoretical conversion relationship between the standard deviation and the median absolute deviation in a normal distribution), to obtain the optical robust scale. The optical robust scale is used to quantify the natural fluctuation range of the optical data in the stable state, and its robustness ensures that even if there are a few outliers, the fluctuation scale will not be excessively amplified, thereby ensuring the reliability of subsequent deviation calculations.

[0112] In the pre-event time window, the median of the acoustic backscatter data is taken as the acoustic baseline median;

[0113] The median of the absolute difference between the acoustic backscatter data and the acoustic baseline median is calculated, multiplied by a normal distribution constant, to obtain the acoustic robust scale;

[0114] It should be noted that the acoustic robust scale is obtained in the same way and logic as the optical robust scale.

[0115] The difference between the optical albedo data at time k and the optical baseline median at time k is divided by the optical robust scale to obtain the corresponding optical deviation value.

[0116] The difference between the acoustic backscatter data at time k and the acoustic baseline median at time k is divided by the acoustic robust scale to obtain the corresponding acoustic deviation value.

[0117] In detail, for the optical albedo data at time k and the optical baseline median at time k, the difference between the optical albedo data at time k and the optical baseline median at time k is divided by the optical robust scale to obtain the corresponding optical deviation value. Similarly, the difference between the acoustic backscatter data at time k and the acoustic baseline median at time k is divided by the acoustic robust scale to obtain the corresponding acoustic deviation value. The deviation value reflects the standardized deviation of the data relative to the reference data in the pre-event time window, a positive value indicates that the optical or acoustic signal at that time and angle is higher than the stable baseline, and a negative value indicates that it is lower than the baseline.

[0118] In an embodiment of the present application, the cross-correlation value and time delay of the optical deviation value and the acoustic deviation value are calculated in the post-event time window, and the optical recovery time constant and the acoustic recovery time constant are fitted by an exponential decay model, including:

[0119] The average optical and acoustic deviation values ​​corresponding to the calculated optical and acoustic deviation values ​​within the pre-event time window are determined respectively. Based on these average optical and acoustic deviation values, the optical albedo and acoustic backscattering data within the post-event time window are depolarized to obtain the optical depolarization matrix. Harmony and acoustic depolarization matrix In the optical depolarization matrix, the r-th row and i-th column represents the optical depolarization value of the circumferential angle i at the r-th time; the r-th row and i-th column of the acoustic depolarization matrix represents the acoustic depolarization value of the circumferential angle i at the r-th time.

[0120] In detail, there may be slight systematic shifts in the optical and acoustic deviation values ​​within the pre-event time window. These shifts originate from natural fluctuations in the baseline data rather than the true post-event response, and retaining them would interfere with subsequent related analyses. Therefore, it is necessary to first calculate the average optical deviation value and the average acoustic deviation value within the pre-event window. Based on these two averages, the optical albedo data and acoustic backscattering data within the post-event time window are corrected respectively, i.e., by subtracting the corresponding average value from each data point, to obtain the optical depolarization matrix and the acoustic depolarization matrix. The r-th row and i-th column of the optical depolarization matrix directly corresponds to the optical depolarization value of the circumferential angle i at the r-th time, and the same applies to the acoustic depolarization matrix.

[0121] Initialize discrete hysteresis And based on discrete hysteresis Computational optical depolarization sequence Harmony and acoustic debiasing sequence The cross-correlation values ​​are as follows:

[0122]

[0123] in, This represents the discrete hysteresis of the circumferential angle i at time r. The corresponding cross-correlation values, Indicates a subsequent time window. This represents the optical debiasing value of the optical debiasing matrix at time r and the circumferential angle i. This represents the average value of the optical debiasing matrix. Indicates the acoustic debiasing sequence at time... The acoustic debiasing value of the circumferential angle i. Represents: Shifting the acoustic debiasing matrix backward along the time axis by the discrete hysteresis. Afterwards, the average value of the acoustic debiasing value is still within the subsequent time window;

[0124] In detail, the discrete lag represents the time offset of the acoustic signal relative to the optical signal. By initializing different discrete lags, the cross-correlation value of the optical and acoustic de-biased sequences at the lag is calculated, which can quantify the degree of synchronization of the two types of signals. In the formula for calculating the cross-correlation value, the numerator is the sum of the products of the deviations of the two types of de-biased values from their respective means, and the denominator is the square root product of the sum of the squares of the deviations of the two types of de-biased values from the mean. This form can standardize the correlation degree to between -1 and 1. The acoustic de-biased sequence at time r minus the value of the lag needs to ensure that it is still within the post-event time window, and the corresponding mean value, which is the mean value of the acoustic de-biased matrix after shifting the lag within the window, is used to eliminate the overall offset of the shifted sequence.

[0125] The discrete lag is determined along the correlation value to be minimized to obtain the final value of the time delay.

[0126] In detail, since the scale layer area recovers rapidly after the event, and the acoustic signal remains low due to the water held in the pores, the two types of signals will show significant anti-correlation, and this anti-correlation will reach the strongest at a certain lag, i.e., the cross-correlation value is the smallest. Therefore, along the value range of the discrete lag, the lag that minimizes the cross-correlation value is found, which is the time delay. The time delay refers to the time length of the acoustic signal recovery lagging behind the optical signal, which directly reflects the water retention effect caused by the porous structure of the scale layer.

[0127] Each optical de-biased value is taken as the post-event deviation of the optical signal, and a forward exponential decay model is used for least squares fitting to obtain the optical recovery time constant.

[0128] Each acoustic de-biased value is taken as the post-event deviation of the acoustic signal, and a forward exponential decay model is used for least squares fitting to obtain the acoustic recovery time constant.

[0129] wherein the forward exponential decay model is as follows:

[0130]

[0131] wherein, represents the end time of a short submergence event or a high fill event in the target pipeline, represents the optical or acoustic recovery time constant to be fitted, represents the initial deviation amplitude, which is specifically the ratio of the first parameter in the optical de-biased sequence to the average value of the optical albedo data in the pre-event time window, or the ratio of the first parameter in the acoustic de-biased sequence to the average value of the acoustic backscatter data in the pre-event time window.

[0132] ​In detail, after the event ends, the optical and acoustic deviation values gradually return to the baseline data over time, and the return process conforms to the exponential decay law, that is, the initial deviation amplitude is the largest, and then gradually decreases over time. Each optical deviation value is taken as the optical post-event deviation, and a forward exponential decay model is used for least squares fitting to obtain the optical recovery time constant, which quantifies the speed of the optical signal returning to the baseline. Similarly, each acoustic deviation value is taken as the acoustic post-event deviation, and the acoustic recovery time constant obtained by fitting quantifies the speed of the acoustic signal returning to the baseline.

[0133] In detail, in the forward exponential decay model, the event end time is the starting point of time, and the optical recovery time constant or the acoustic recovery time constant to be fitted is the core parameter representing the decay speed, and the smaller the value is, the faster the recovery is. The initial deviation amplitude, that is, the parameter in the model, is determined by the ratio of the first parameter of the optical deviation sequence to the average value of the optical albedo data in the pre-window, or the ratio of the first parameter of the acoustic deviation sequence to the average value of the acoustic backscatter data in the pre-window, and is used to reflect the initial intensity of the signal deviation from the baseline at the end of the event.

[0134] In detail, for the scale layer region, the optical recovery time constant is small due to the rapid evaporation of surface moisture, and the acoustic recovery time constant is significantly larger than the optical recovery time constant due to the slow desorption of water in the pores. The difference in time constants is a manifestation of the porous structure of the scale layer.

[0135] In an embodiment of the present application, the optical positive deviation area and the acoustic negative deviation area are determined, and the intensity parameter is calculated in combination with the cross-correlation value and the time delay, and the candidate scale layer annulus is located according to the annular connectivity sorting of the intensity parameter, which includes:

[0136] The optical deviation value and the acoustic deviation value of the ring direction angle i at each time in the post-time window are determined.

[0137] It should be noted that the optical deviation value and the acoustic deviation value of the ring direction angle i at each time in the post-time window are obtained in the same way as the optical deviation value and the acoustic deviation value of the ring direction angle i at each time in the pre-time window.

[0138] If the optical deviation value of the ring direction angle i at the rth time is positive, the optical deviation value is merged into the first positive set; the acoustic deviation value of the ring direction angle i at the rth time is symbolically inverted to obtain the corresponding inverted acoustic deviation value; if the inverted acoustic deviation value of the ring direction angle i at the rth time is positive, the inverted acoustic deviation value is merged into the second positive set.

[0139] In detail, for the optical deviation value, the optical property of the scale layer is that the optical albedo is continuously higher than the baseline after the event, so only the part of the circumferential angle i at the r-th moment whose optical deviation value is positive is retained, and is merged into a first positive part set. The positive value directly corresponds to the scale layer property of high optical deviation, and the negative value is excluded because it does not meet the characteristics. For the acoustic deviation value, the acoustic property of the scale layer is that the backscattering is continuously lower than the baseline after the event, so the acoustic deviation value is first processed by sign inversion to convert the negative deviation into an inverted acoustic deviation value; then the part that is positive after inversion is retained, and is merged into a second positive part set. Thus, the property of low acoustic is converted into a positive and accumulative quantity.

[0140] The first positive part set is integrated in the post-event time window to obtain an optical positive deviation area;

[0141] The second positive part set is integrated in the post-event time window to obtain an acoustic positive deviation area;

[0142] In detail, the first positive part set is integrated in the post-event time window to obtain an optical positive deviation area, which quantifies the time accumulation effect of the optical signal continuously higher than the baseline, and the larger the area, the more significant the property of high optical deviation. The second positive part set is integrated in the same way to obtain an acoustic positive deviation area, which is essentially a time accumulation effect of the acoustic signal continuously lower than the baseline (converted into a positive area by sign inversion), and the larger the area, the more significant the property of low acoustic. The optical positive deviation area and the acoustic positive deviation area jointly reflect the persistence of the scale layer properties in the optical and acoustic aspects after the event.

[0143] The product of the optical positive deviation area, the acoustic positive deviation area, the time delay, and the cross-correlation value corresponding to the time delay is taken as the intensity parameter of the circumferential angle i;

[0144] In detail, the intensity parameter is obtained by multiplying the optical positive deviation area, the acoustic positive deviation area, the time delay, and the corresponding cross-correlation value. Among them, the optical and acoustic positive deviation areas represent the degree of significance of the two types of properties respectively; the time delay reflects the lag length of the acoustic relative to the optical, and the more obvious the lag, the greater the contribution of this factor; the cross-correlation value (the anti-correlation intensity because the scale layer property is anti-correlated) is used to strengthen the reverse synchronization of the two types of signals. The intensity parameter formed by the multiplication of the four can comprehensively measure the degree to which a certain circumferential angle meets the scale layer cross-modal characteristics, and the larger the value, the higher the possibility that the angle belongs to the scale layer.

[0145] If the intensity parameters of the circumferential angle i and the circumferential angle i+1 are both greater than 0, then the circumferential angle i and the circumferential angle i+1 are merged to form a connected segment;

[0146] The range of the circumferential angle of each connected segment is determined, and the intensity parameter in the connected segment is integrated with respect to the range of the circumferential angle to obtain an intensity integral;

[0147] The connected section with the maximum intensity integral is taken as a candidate scale layer girdle corresponding to the section.

[0148] In detail, since the scale layer is a secondary mineral girdle continuously distributed in the crown region of the pipeline, its spatial form must be annularly continuous. Therefore, if the intensity parameters of adjacent annular angles i and i+1 are both greater than 0, it indicates that both of them have the property of scale layer, and they are combined into a connected section. After determining the annular angle range of each connected section, the intensity integral is obtained by integrating the intensity parameters in the section, which represents the overall intensity of the scale layer characteristics in the connected section. The connected section with the maximum intensity integral is taken as a candidate scale layer girdle corresponding to the section, because it is most consistent with the properties of scale layer representation, significant and continuous distribution.

[0149] In an embodiment of the present application, the joint screening includes a first screening, a second screening and a third screening.

[0150] The first screening includes: traversing each time in the post-event time window; if all the optical depolarization values corresponding to time r in the optical depolarization matrix are positive numbers, and all the acoustic depolarization values corresponding to time r in the acoustic depolarization matrix are negative numbers, then the optical albedo and acoustic backscatter data corresponding to the rth time are retained to form a first screening data set;

[0151] In detail, the first screening focuses on the overall consistency of the signal direction. Traversing each time in the post-event time window, it is required that all the optical depolarization values of the annular angle in the optical depolarization matrix are positive numbers, and all the acoustic depolarization values of the annular angle in the acoustic depolarization matrix are negative numbers. This selection constraint is derived from the inherent properties of the scale layer: the optical albedo of the event after the scale layer is continuously higher than the baseline (positive deviation), and the acoustic backscatter is continuously lower than the baseline (negative deviation). If all the annular angles at a certain time satisfy this condition, it indicates that the data at this time is not disturbed by device gain mutation, uneven illumination, etc. (such disturbances often lead to overall deviation from the expected direction), and the overall direction is consistent with the properties of the scale layer. The first screening data set formed thereby preliminarily ensures the reliability of the signal direction.

[0152] The second screening includes: after time delay on the time axis, the qth time in the first screening data set still belongs to the post-event time window, and the cross-correlation values of all the annular angles corresponding to the qth time are negative values, then the optical albedo and acoustic backscatter data corresponding to the qth time are retained to form a second screening data set;

[0153] In detail, the second screening strengthens the anti-correlation and time lag characteristics of the cross-modal. Based on the result of the first screening, if the data at each time is still within the time window after deviating from the time delay (i.e. the lag time of the acoustic relative to the optical), and the cross-correlation value of all ring angles at this time is negative, the data at this time is retained. The deviation from the time delay still within the window ensures the effectiveness of the lag analysis; the negative cross-correlation value indicates that the optical and acoustic signals are anti-correlated (the optical is high and the acoustic is low), which is the core cross-modal property of the scale layer, i.e. due to the water holding in the pores, the optical recovers quickly and the acoustic remains low. Thus, the synchronous fluctuations caused by water flow disturbance (such as the optical and acoustic being high or low at the same time) are excluded, and only the data meeting the anti-correlation characteristics is retained to form the second screening data set.

[0154] The third screening includes: for the acoustic recovery time constant and the optical recovery time constant corresponding to the ring angle i at the pth time in the second screening data set, if the acoustic recovery time constant > the optical recovery time constant, the optical albedo and the acoustic backscattering data corresponding to the ring angle i at the pth time are retained to form a third screening data set.

[0155] The third screening data set is marked as a material dominant data set.

[0156] In detail, the third screening anchors the material specificity of the recovery dynamics. For the time and ring angle retained by the second screening, it is checked whether the acoustic recovery time constant of each ring angle is greater than the optical recovery time constant. The porous structure of the scale layer causes water to be retained in the pores, resulting in continuous acoustic signal decay due to water saturation and slow recovery speed (large time constant); while the optical signal is affected by the evaporation of surface moisture, and the recovery speed is fast (small time constant). This dynamic difference is the key to distinguish the scale layer from non-porous areas such as intact concrete. By requiring the acoustic recovery time constant to be greater than the optical recovery time constant, the interference of non-porous areas (which have no such significant difference in recovery time constant) can be excluded, and the third screening data set formed finally is dominated by the material properties of the scale layer.

[0157] In an embodiment of the present application, in response to the material dominant data set, the candidate scale layer ring belt of each section is spliced according to the intensity parameter along the axial direction of the pipeline to obtain a scale layer belt center track, including:

[0158] The candidate scale layer ring belt of each section is determined respectively, and the intensity parameter of each ring angle corresponding to the candidate scale layer ring belt is determined.

[0159] In detail, candidate scale ring belts of each pipe section are determined respectively, and the candidate scale ring belt is a continuous annular region that is screened and meets the scale layer property; meanwhile, the intensity parameters corresponding to each annular angle in each ring belt are extracted, and the intensity parameters comprehensively consider the optical positive deviation area, the acoustic positive deviation area, the time delay and the cross-correlation value, and quantize the possibility that each annular angle belongs to the scale layer.

[0160] The annular central angle of the candidate scale ring belt is calculated based on the intensity parameter, as follows:

[0161] The ratio of the product of each annular angle and the corresponding intensity parameter to the sum of the intensity parameters of each annular angle is calculated to obtain the annular central angle .

[0162] In detail, the sum of the products of each annular angle and the corresponding intensity parameter is divided by the total sum of the intensity parameters to obtain the annular central angle. The annular central angle is a central angle weighted by the intensity parameter, and the annular angle with a higher intensity parameter contributes more to the central angle, which ensures that the central angle can accurately point to the core region of the scale ring belt of the section, avoids the interference of the low-intensity region at the edge, and is more consistent with the geometric center of the actual scale layer.

[0163] An optimization function is constructed , as follows:

[0164]

[0165] wherein, the sum of the intensity parameters of each annular angle corresponding to the candidate scale ring belt of the s-th section, the annular central angle of the s-th section, the angle weight, , the interval distance between the s-th section and the t-th section, In detail, the optimization function includes two parts: the first part is the sum of the intensity parameter sums of the candidate ring belts of the sections, and the purpose is to maximize the overall intensity accumulation to ensure that the spliced trajectory contains as many high-intensity regions as possible, and these regions are the parts with the most significant scale layer properties; the second part is the sum of the absolute values of the differences between the annular central angles of adjacent sections multiplied by the angle weight, and the purpose is to minimize the changes in the central angles of adjacent sections to ensure the axial continuity of the trajectory. The value of the angle weight is determined by the interval distance between the s-th section and the t-th section, so that the weight of spatial smoothing matches the sampling density, and the loss of high-intensity regions due to excessive emphasis on continuity or the breaking of the trajectory due to excessive emphasis on intensity is avoided.

[0166]

[0167] ​​​The scale layer band center trajectory is obtained by solving the optimization function based on the greedy search, and the scale layer band center trajectory is used for visualizing the scale layer of the target pipeline.

[0168] In detail, the greedy search selects a ring direction center angle that maximizes the optimization function. In each decision step, the current section with high intensity is preferentially retained to maximize the first part of the sum, and the change in the center angle of the previous section is ensured to be minimal to meet the smoothing constraint of the second part. The greedy search can efficiently obtain a continuous, smooth, and concentrated trajectory, i.e., the scale layer band center trajectory. The scale layer band center trajectory directly presents the distribution of the scale layer in the axial direction of the pipeline and can be directly used for visual display, thereby providing precise spatial positioning basis for pipeline maintenance. That is, the area through which the trajectory passes is the concentrated area of the scale layer that needs to be focused on and repaired, and the continuity and intensity distribution of the trajectory fully meet the properties of the scale layer growing continuously along the axial direction of the pipeline.

[0169] Embodiment two:

[0170] The buried sewage pipeline scale layer intelligent detection system based on the Internet of Things is applied to any one of the buried sewage pipeline scale layer intelligent detection methods based on the Internet of Things, and includes:

[0171] The time window determination module is configured to determine the pre-time window and the post-time window.

[0172] The multi-modal data acquisition module is configured to acquire optical albedo and acoustic backscatter data of the section of the target pipeline at each ring direction angle.

[0173] The deviation value normalization calculation module is configured to normalize the median and the median absolute deviation of the optical albedo and the acoustic backscatter data in the pre-time window to obtain optical deviation values and acoustic deviation values.

[0174] The time lag and recovery time constant analysis module is configured to calculate the cross-correlation value and the time delay of the optical deviation values and the acoustic deviation values in the post-time window, and fit the optical recovery time constant and the acoustic recovery time constant through an exponential decay model.

[0175] The ring direction candidate scale layer ring band positioning module is configured to determine the optical positive deviation area and the acoustic negative deviation area, combine the cross-correlation value and the time delay, and then calculate the intensity parameter to position the candidate scale layer ring band according to the ring direction connectivity sorting.

[0176] The material dominant data set screening module is configured to obtain a material dominant data set corresponding to the material dominant cause by joint screening.

[0177] The axial scale layer band center trajectory splicing module is configured to splice the candidate scale layer ring band of each section according to the intensity parameter along the axial direction of the pipeline in response to the material dominant data set, and obtain the scale layer band center trajectory.

[0178] The above describes the embodiments of the present embodiments, but the present embodiments are not limited to the specific embodiments described above, which are only illustrative but not restrictive, and those skilled in the art can make many forms under the inspiration of the present embodiments, which all belong to the protection of the present embodiments.

Claims

1. A buried sewage pipeline scale layer intelligent detection method based on the Internet of Things, characterized in that, The method comprises the following steps: determining a pre-time window and a post-time window; collecting optical albedo and acoustic backscatter data of each cross section of the target pipeline at each circumferential angle; normalizing the median and median absolute deviation of the optical albedo and acoustic backscatter data in the pre-time window to obtain optical deviation values and acoustic deviation values; calculating the cross-correlation value and time delay of the optical deviation values and acoustic deviation values in the post-time window, and fitting an exponential decay model to obtain optical recovery time constant and acoustic recovery time constant; determining the optical positive deviation area and the acoustic negative deviation area, combining the cross-correlation value and the time delay, and then calculating the intensity parameter to sort and locate the candidate scale ring according to the intensity parameter; obtaining the material dominant data set corresponding to the material dominant cause through joint screening; in response to the material dominant data set, splicing the candidate scale ring of each cross section according to the intensity parameter to obtain the scale band center trajectory along the pipeline axis. 2.The Internet of Things based intelligent detection method for scale layer of buried sewage pipeline according to claim 1, characterized in that, The method comprises the following steps: determining the end time of a short flooding event or a high filling event in the target pipeline; dividing the end time into a pre-time window and a post-time window; wherein the pre-time window and the post-time window completely cover a preset time period. 3.The Internet of Things based intelligent detection method for scale layer of buried sewage pipeline according to claim 2, characterized in that, The method comprises the following steps: at fixed time intervals and in ring angles Collecting optical reflectance data and acoustic backscatter raw amplitude for each section; to the circumferential angle corresponding compensation coefficients are calculated, and the product of the compensation coefficients and the acoustic backscatter raw amplitude is taken as the acoustic backscatter data; The calculation process of the compensation coefficient is as follows: Determining the acoustic round trip time of the ringdown angle i ; The compensation coefficient of the ring direction angle i is calculated according to the sound wave round trip time ; wherein, represents the medium sound speed, the medium sound speed including the sewage medium sound speed or the air medium sound speed, represents the calibration gain of the sonar receiving link.

4. The Internet of Things-based intelligent detection method for the scale layer of a buried sewage pipeline according to claim 3, characterized in that, The method comprises the following steps: In the pre-time window, the median of the optical albedo data is taken as the optical baseline median; The median of the absolute difference between the optical albedo data and the optical baseline median is multiplied by a normal distribution constant to obtain an optical robust scale; In the pre-time window, the median of the acoustic backscatter data is taken as the acoustic baseline median; The median of the absolute difference between the acoustic backscatter data and the acoustic baseline median is multiplied by a normal distribution constant to obtain an acoustic robust scale; The difference between the optical albedo data at time k and circumferential angle i and the optical baseline median is divided by the optical robust scale to obtain the corresponding optical deviation value; The difference between the acoustic backscatter data at time k and circumferential angle i and the acoustic baseline median is divided by the acoustic robust scale to obtain the corresponding acoustic deviation value. 5.The Internet of Things based intelligent detection method for scale layer of buried sewage pipeline according to claim 4, characterized in that, The method comprises the following steps: respectively determine optical deviation average values and acoustic deviation average values corresponding to the calculated optical deviation values and acoustic deviation values in the pre-time window; and perform a de-biasing process on the optical albedo data and acoustic backscattering data in the post-time window based on the optical deviation average values and the acoustic deviation average values, to obtain an optical de-biasing matrix and an acoustic de-biasing matrix ; wherein the ith element in the rth row of the optical de-biasing matrix represents an optical de-biasing value of the circumferential angle i at the rth moment; and the ith element in the rth row of the acoustic de-biasing matrix represents an acoustic de-biasing value of the circumferential angle i at the rth moment. Initialize discrete hysteresis And based on discrete hysteresis Computational optical depolarization sequence Harmony and acoustic debiasing sequence The cross-correlation values ​​are as follows: wherein, denotes the discrete lag amount of the ring-down angle i at time r the corresponding cross-correlation value, denotes the post-time window, denotes the optical depolarization value of the optical depolarization matrix at the ring-down angle i at time r denotes the average value of the optical depolarization matrix, denotes the acoustic depolarization value of the acoustic depolarization matrix at the ring-down angle i at time r denotes the acoustic depolarization value of the acoustic depolarization matrix at the ring-down angle i at time r denotes: the acoustic depolarization matrix is shifted backward along the time axis by a discrete lag amount the average value of the acoustic depolarization value still in the post-time window after Determine the discrete lag by minimizing the cross-correlation value. The final value is used to obtain the time delay; Each optical deviation value is taken as the post-deviation amount of an optical event, and a least square fitting is performed through a forward exponential decay model to obtain the optical recovery time constant; Each acoustic deviation value is taken as the post-deviation amount of an acoustic event, and a least square fitting is performed through a forward exponential decay model to obtain the acoustic recovery time constant; The forward exponential decay model is as follows: wherein, denotes the end time of a short-term flooding event or a high-filling event within the target pipe, denotes the optical or acoustic recovery time constant to be fitted, denotes the initial deviation amplitude, in particular the ratio of the first parameter in the optical depolarization sequence and the average of the optical albedo data of the preceding time window, or the ratio of the first parameter in the acoustic depolarization sequence and the average of the acoustic backscatter data of the preceding time window. 6.The Internet of Things based intelligent detection method for scale layer of buried sewage pipeline according to claim 5, characterized in that, Determine the optical positive deviation area and the acoustic negative deviation area, combine the cross-correlation value and the time delay, and then calculate the intensity parameter. The candidate scale ring belt of the cross-section is located according to the intensity parameter, including: Determine the optical deviation value and the acoustic deviation value of the ring angle i at each time in the post-time window; If the optical deviation value of the ring angle i at the rth time is positive, the optical deviation value is merged into the first positive part set; the acoustic deviation value of the ring angle i at the rth time is processed by sign inversion to obtain the corresponding inverted acoustic deviation value; if the inverted acoustic deviation value of the ring angle i at the rth time is positive, the inverted acoustic deviation value is merged into the second positive part set; Integrate the first positive part set in the post-time window to obtain the optical positive deviation area; Integrate the second positive part set in the post-time window to obtain the acoustic positive deviation area; The product of the optical positive deviation area, the acoustic positive deviation area, the time delay and the cross-correlation value corresponding to the time delay is taken as the intensity parameter of the ring angle i; If the intensity parameters of the ring angle i and the ring angle i+1 are both greater than 0, the ring angle i and the ring angle i+1 are merged to form a connected segment; Determine the range of the ring angle of each connected segment, and integrate the intensity parameter in the connected segment according to the range of the ring angle to obtain the intensity integral; The connected segment with the maximum intensity integral is taken as the candidate scale ring belt corresponding to the cross-section. 7.The Internet of Things based intelligent detection method for scale layer of buried sewage pipeline according to claim 6, characterized in that, Joint screening, including first screening, second screening and third screening; The first screening includes: traversing each time in the post-time window; if all optical depolarization values corresponding to time r in the optical depolarization matrix are positive numbers, and all acoustic depolarization values corresponding to time r in the acoustic depolarization matrix are negative numbers, the optical albedo and acoustic backscatter data corresponding to the rth time are retained to form a first screening data set; The second screening includes: for the qth time in the first screening data set, after deviating from the time delay in the time axis, still belonging to the post-time window, and the cross-correlation value of all ring angles corresponding to the qth time is negative, the optical albedo and acoustic backscatter data corresponding to the qth time are retained to form a second screening data set; The third screening includes: for the pth time in the second screening data set, the acoustic recovery time constant and the optical recovery time constant corresponding to the ring angle i, if the acoustic recovery time constant> optical recovery time constant, the optical albedo and acoustic backscatter data corresponding to the ring angle i of the pth time are retained to form a third screening data set; The third screening data set is marked as the material dominant data set. 8.The Internet of Things based intelligent detection method for scale layer of buried sewage pipeline according to claim 7, characterized in that, In response to the material dominant data set, the candidate scale ring belt of each cross-section is spliced according to the intensity parameter along the pipeline axis to obtain the scale band center trajectory, including: Determine the candidate scale ring belt of each cross-section, and determine the intensity parameter of each ring angle corresponding to the candidate scale ring belt; Based on the intensity parameter, the ring center angle of the candidate scale ring belt is calculated, as follows: calculating the ratio of the product of each azimuthal angle and the corresponding intensity parameter to the sum of the intensity parameters for each azimuthal angle, to obtain the azimuthal central angle ; Constructing an optimization function As follows: in, This represents the sum of the intensity parameters for each circumferential angle corresponding to the candidate scale layer annulus at the s-th cross-section. Indicates the first The circumferential central angle of each cross section, Indicates angle weight, , Represents the s-th cross section and the s-th cross section. The interval between each cross section; Based on the greedy search to solve the optimization function, the scale band center trajectory is obtained; wherein the scale band center trajectory is used to visualize the scale of the target pipeline.

9. The buried sewage pipeline scale layer intelligent detection system based on the Internet of Things is applied to the buried sewage pipeline scale layer intelligent detection method based on the Internet of Things in any one of claims 1-8, characterized in that, Including: The time window determination module is configured to determine a pre-time window and a post-time window; The multi-modal data acquisition module is configured to acquire optical albedo and acoustic backscatter data of a cross section of the target pipeline at each circumferential angle; The deviation value normalization calculation module is configured to normalize the optical albedo and acoustic backscatter data based on the median and median absolute deviation of the optical albedo and acoustic backscatter data in the pre-time window to obtain optical deviation values and acoustic deviation values; The time lag and recovery time constant analysis module is configured to calculate cross-correlation values and time delays of the optical deviation values and the acoustic deviation values in the post-time window, and fit the optical recovery time constant and the acoustic recovery time constant through an exponential decay model; The circumferential candidate scale layer ring belt positioning module is configured to determine optical positive deviation areas and acoustic negative deviation areas, combine the cross-correlation values and the time delays, and then calculate the intensity parameters to position the candidate scale layer ring belts according to the circumferential connectivity sorting of the intensity parameters; The material dominant data set screening module is configured to obtain a material dominant data set corresponding to a material dominant deviation cause through joint screening; The axial scale layer band center trajectory splicing module is configured to splice the candidate scale layer ring belts of each cross section along the pipeline axis according to the intensity parameters to obtain a scale layer band center trajectory in response to the material dominant data set.

Images