CN114841204A - Pipeline sonar point cloud denoising method and system - Google Patents

Pipeline sonar point cloud denoising method and system Download PDF

Info

Publication number
CN114841204A
CN114841204A CN202210454829.3A CN202210454829A CN114841204A CN 114841204 A CN114841204 A CN 114841204A CN 202210454829 A CN202210454829 A CN 202210454829A CN 114841204 A CN114841204 A CN 114841204A
Authority
CN
China
Prior art keywords
clustering
point cloud
pipeline
sonar
sonar point
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210454829.3A
Other languages
Chinese (zh)
Inventor
刘文黎
骆汉宾
李琛
吴俊豪
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huazhong University of Science and Technology
Original Assignee
Huazhong University of Science and Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huazhong University of Science and Technology filed Critical Huazhong University of Science and Technology
Priority to CN202210454829.3A priority Critical patent/CN114841204A/en
Publication of CN114841204A publication Critical patent/CN114841204A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • G06F2218/04—Denoising
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Signal Processing (AREA)
  • Investigating Or Analyzing Materials By The Use Of Ultrasonic Waves (AREA)

Abstract

The invention discloses a pipeline sonar point cloud denoising method and system, and belongs to the field of drainage pipeline detection. The method comprises the following steps: removing outlier noise from original sonar scanning data in the pipeline to obtain primarily screened sonar point cloud; clustering the primarily screened sonar point clouds to obtain 3 globally optimal clustering center coordinates and clustering radii; and determining high-density noise according to the global optimal clustering center coordinate and the clustering radius at the topmost layer, and removing the high-density noise from the primarily screened sonar point cloud. According to the method, firstly, outlier noise is removed from original sonar scanning data in a pipeline, sonar point cloud after primary screening is obtained, more accurate data are obtained and further processed, then clustering is carried out, 3 global optimal clustering center coordinates and clustering radiuses are obtained, finally, high-density noise is determined according to the global optimal clustering center coordinates and the clustering radiuses at the topmost layer, the high-density noise is removed from the sonar point cloud after primary screening, the accuracy and readability of scanning results are guaranteed, and the processing efficiency of the sonar point cloud data is improved.

Description

Pipeline sonar point cloud denoising method and system
Technical Field
The invention belongs to the field of drainage pipeline detection, and particularly relates to a pipeline sonar point cloud denoising method and system.
Background
Along with the continuous development of human society, the urban gathering effect becomes more obvious, and large cities inevitably generate a large amount of sewage and wastewater due to the increasing use amount of water resources, which puts higher requirements on urban underground drainage pipelines. However, most of the traditional drainage pipe network drawings are paper folders and scanning pieces, and the workload of manual arrangement and filing, drawing retrieval, proofreading, modification and the like is huge; moreover, with the continuous development of cities, the butt joint and matching between a newly-built pipe network project and an already-built pipe network project are disordered, the arrangement condition of underground pipelines is difficult to clean, and the mistaken digging of the existing sewage pipelines can be caused, so that great economic loss is caused. Therefore, the method has obvious practical significance on how to accurately and quickly identify the pipeline information and clear the veins of the urban underground pipelines.
Sonar detection is used as a scanning technology, has the advantages of high sensitivity, strong penetrating power, flexible flaw detection, high efficiency, low cost and the like, can provide accurate data information, and can be combined with CCTV (Closed Circuit Television) to carry out comprehensive inspection on a pipeline, so that the outline of the pipeline at any cross section position of the pipeline can be known. However, noise pollution is inevitably generated in the pipeline detection process, and the current mainstream denoising method mainly comprises filtering denoising, so that the efficiency and the precision of the method need to be improved.
Disclosure of Invention
Aiming at the defects and improvement requirements of the prior art, the invention provides a pipeline sonar point cloud denoising method and system, aiming at ensuring the accuracy and readability of a scanning result and improving the processing efficiency of sonar point cloud data.
To achieve the above object, according to a first aspect of the present invention, there is provided a method for denoising a point cloud of a pipeline sonar, the method comprising:
s1, removing outlier noise from original sonar scanning data in a pipeline to obtain primarily screened sonar point cloud;
s2, clustering the primarily screened sonar point clouds to obtain 3 globally optimal clustering center coordinates and clustering radii;
and S3, determining high-density noise according to the global optimal clustering center coordinate and the clustering radius at the topmost layer, and removing sonar point cloud after primary screening.
Preferably, in step S1, the adaptive DBSCN algorithm is used to remove the outlier noise.
Has the advantages that: aiming at the problems of noise points, non-visual pipeline surface and the like in a sonar scanning result, the adaptive DBSCN algorithm is preferably used for removing outlier noise, and because the connection capability between samples is investigated from the angle of sample density and clustering is continuously expanded on the basis of continuous samples, clusters with different sizes and shapes are simply and flexibly detected, and a final clustering result is obtained.
Preferably, clustering is carried out on the primarily screened sonar point clouds by adopting an improved whale algorithm, and a clustering center is used as a whale position, wherein the improved whale algorithm specifically comprises the following steps:
the spiral birth position updating mechanism of the classical whale algorithm is unchanged, and the contraction surrounding mechanism adopts a self-adaptive weight:
Figure BDA0003618405590000021
Figure BDA0003618405590000022
wherein the content of the first and second substances,
Figure BDA0003618405590000023
indicating the updated position vector and the position vector,
Figure BDA0003618405590000024
represents the current best position vector, omega represents the adaptive weight,
Figure BDA0003618405590000025
representing the vector factor, ω max ,ω min Respectively representing the maximum value and the minimum value of the self-adaptive weight, i representing the current iteration turn, and M representing the maximum iteration number.
Has the advantages that: compared with the problem that other whale algorithms are easy to fall into a local optimal solution, the method improves a WOA weight value adjusting mode, and when whales are close to food, the positions of the whales are changed by introducing a smaller weight value, so that the local searching capacity of the whales is improved. Aiming at the problem of low convergence precision, the improved whale algorithm is preferably used for clustering the primarily screened sonar point clouds, the self-adaptive weight is introduced, so that the local optimization capability of WOA is enhanced, the detailed search around local optimization is realized, the local search capability is improved, and the convergence precision is improved.
Preferably, the objective function of the improved whale algorithm is as follows:
Figure BDA0003618405590000031
Figure BDA0003618405590000032
wherein, the fitness represents the optimization target, n represents the variable number of the clustering center, w ij Weight, x, of ith cluster center representing jth cluster p Representing a point cloud coordinate vector, x ij Represents the cluster center coordinate, | x, to be solved p -x ij I represents x p Euclidean distances to the center of each cluster.
Has the advantages that: the present invention preferably selects the objective function that reflects the convergence accuracy, and the smaller the fit, the higher the convergence accuracy, and the improvement of the local search capability is achieved.
Preferably, the method further comprises:
and S4, determining a fitting section of the pipeline according to the global optimal clustering center coordinate and the clustering radius which are positioned in the middle, and determining a sludge section according to the global optimal clustering center coordinate and the clustering radius which are positioned at the bottommost layer.
Has the beneficial effects that: according to the method, the fitting section of the pipeline and the sludge section are further determined through the steps.
Preferably, if the fitting section of the pipeline has data loss, the complete section of the pipeline is completed according to the pipeline section points obtained by clustering.
Has the advantages that: the invention enables the model to obtain a more accurate fitting result in a proper time by completing the complete pipeline section.
To achieve the above object, according to a second aspect of the present invention, there is provided a pipe sonar point cloud denoising system, comprising: a computer-readable storage medium and a processor;
the computer-readable storage medium is used for storing executable instructions;
the processor is configured to read executable instructions stored in the computer-readable storage medium, and execute the pipe sonar point cloud denoising method according to the first aspect.
Generally, by the above technical solution conceived by the present invention, the following beneficial effects can be obtained:
the invention provides a pipeline sonar point cloud denoising method which comprises the steps of removing outlier noise from original sonar scanning data in a pipeline to obtain primarily screened sonar point cloud, further processing the acquired more accurate data, clustering to obtain 3 globally optimal clustering center coordinates and clustering radii, determining high-density noise according to the globally optimal clustering center coordinates and clustering radii at the top layer, and removing the noise from the primarily screened sonar point cloud. Due to the fact that the characteristic of sonar detection of the drainage pipeline is utilized, accuracy and readability of scanning results are guaranteed, and efficiency of processing sonar point cloud data is improved.
Drawings
FIG. 1 is a flow chart of a pipeline sonar point cloud denoising method provided by the invention;
FIG. 2 is a schematic diagram of the pipeline sonar point cloud data provided by the embodiment of the present invention;
FIG. 3 is a schematic diagram of removing outlier noise points by the adaptive DBSCAN method provided by the present invention;
FIG. 4 is a schematic diagram of a noise point removal process provided by the present invention;
FIG. 5 is a schematic view of a fitting of sludge fouling lines provided by the present invention;
FIG. 6 is a diagram of the cluster optimization results obtained by applying the improved whale algorithm provided by the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is 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 of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
As shown in FIG. 1, the invention provides a pipeline sonar point cloud denoising method, which comprises the following steps:
(1) removing outlier noise from original sonar scanning data in the pipeline to obtain sonar point cloud after primary screening.
Through sonar scanning robot scan on the spot, obtain the interior original sonar scanning data of pipeline as shown in FIG. 2, it includes: external outliers, internal outliers, noise points, and sludge points.
Further, the obtained spatial coordinate data is subjected to formatting processing to realize the standardization of the data.
Aiming at the problems that the sonar scanning result has noise points, the pipeline surface is not visual and the like, each noise point is initialized to be a vector, a point cloud main body and outliers are identified by using a self-adaptive DBSCAN algorithm, the initial screening of data is realized, the outliers in the data are removed preliminarily, and the specific principle is shown in FIG. 3.
(2) Setting K clustering centers, taking coordinates of the clustering centers and the clustering radius as optimized variables of an improved whale algorithm, and initializing the improved whale algorithm to obtain initial optimal values of the coordinates of the clustering centers and the clustering radius according to primarily screened sonar data.
Further, the Euclidean distance from each data vector to each cluster center is calculated through a formula (1), a formula (2) is an optimization objective function, and the fitness value is updated through a contraction surrounding mechanism and a spiral position updating mechanism of an improved whale algorithm to obtain an optimal solution, so that the initial optimal values of the coordinates and the cluster radii of the K cluster centers are obtained.
Figure BDA0003618405590000051
Figure BDA0003618405590000052
Figure BDA0003618405590000053
Wherein x is p For loaded point cloud coordinate vectors, x ij The cluster center coordinate to be solved; the fitness is an optimization objective function, k is the number of clustering centers, and n is the number of variables of the clustering centers; w is a ij The value is determined by equation (3).
(3) And calculating the distance from each data point to the cluster center based on an improved whale algorithm, and iterating the optimal position of each cluster center by utilizing a contraction surrounding mechanism and spiral position updating.
And obtaining an optimal clustering solution through an improved whale optimization algorithm, wherein a mathematical model of the improved whale optimization algorithm is shown as a formula (4), the mathematical model comprises two search modes of circular shrinkage and spiral rising, and the probabilities of the two search modes are equal. When p is more than or equal to 0.5, adopting a spiral ascending position updating mechanism; otherwise, a shrink wrapping mechanism with adaptive weights is adopted.
Figure BDA0003618405590000061
Figure BDA0003618405590000062
Wherein the content of the first and second substances,
Figure BDA0003618405590000063
indicating the updated position vector and the position vector,
Figure BDA0003618405590000064
represents the current best position vector, omega represents the adaptive weight,
Figure BDA0003618405590000065
representing the vector factor, ω max ,ω min Respectively representing the maximum value and the minimum value of the self-adaptive weight, i representing the current iteration turn, and M representing the maximum iteration number.
p is [0,1 ]]T represents the number of iterations, b is a constant whose value determines the shape of the spiral update, l is [ -1,1 [ -1]Any random number in between.
Figure BDA0003618405590000066
And
Figure BDA0003618405590000067
in order to be a vector factor, the vector factor,
Figure BDA0003618405590000068
the distance between the ith whale and the prey,
Figure BDA0003618405590000069
for the current best position vector to be used,
Figure BDA00036184055900000610
is a vector representation of the cluster center position.
When the model adopts a circular contraction mechanism, as whales have two search forms, the contraction search or random search of the whales can be simulated by the value of | A |, wherein the | A | value is expressed by a formula
Figure BDA00036184055900000611
It is decided that,
Figure BDA00036184055900000612
t is the current iteration number, t max In order to be the maximum number of iterations,
Figure BDA00036184055900000613
is [0,1 ]]Random vector betweenAmount of the compound (A). When | A | < 1,
Figure BDA00036184055900000614
wherein
Figure BDA00036184055900000615
In order to enclose the step size,
Figure BDA00036184055900000616
when | A |>When the pressure of the mixture is 1, the pressure is lower,
Figure BDA00036184055900000617
wherein
Figure BDA00036184055900000618
(4) And continuously iterating through the whale algorithm flow according to the initially set maximum iteration times until the iteration times meet the requirements and are the globally optimal solution, stopping the algorithm, and outputting the globally optimal clustering center coordinates and clustering radius values.
Further, in the step (4), by means of an improved whale algorithm iterative operation rule, the coordinates of the clustering centers and the clustering radius values are continuously updated until the iteration times meet requirements and the overall optimal solution is obtained, the algorithm is stopped, and the optimal clustering result of each sonar scanning point is obtained according to the optimal spatial positions and the optimal clustering radii of the K clustering centers obtained through calculation.
Sampling the point cloud data from which the outliers are removed to obtain a more accurate point cloud sample, as shown in fig. 4.
(5) Deleting noise points according to the calculated global optimal clustering center coordinates and clustering radius values to obtain a pipeline and sludge siltation optimal clustering fitting curve; and the optimized fitting curve is combined to complete the complete pipeline section, so that a more accurate fitting result can be obtained by the model in a proper time.
Further, in the step (5), the noise points and the clustering blocks of the best fitting curve are respectively given, the noise points are deleted, and the complete pipeline section is completed, so that the model can obtain a more accurate fitting result in a proper time.
For the pipe interior points, the fouling points are identified to obtain a fouling fitting curve, as shown in fig. 5.
Substituting the primarily screened point cloud vectors into a whale algorithm to obtain globally optimal clustering center coordinates and clustering radius, and finally realizing noise point identification, pipeline fitting curves and sedimentation fitting curves as shown in FIG. 6.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims (7)

1. A pipeline sonar point cloud denoising method is characterized by comprising the following steps:
s1, removing outlier noise from original sonar scanning data in a pipeline to obtain primarily screened sonar point cloud;
s2, clustering the primarily screened sonar point clouds to obtain 3 globally optimal clustering center coordinates and clustering radii;
and S3, determining high-density noise according to the global optimal clustering center coordinate and the clustering radius at the topmost layer, and removing sonar point cloud after primary screening.
2. The method of claim 1, wherein in step S1, the adaptive DBSCN algorithm is used to remove the outlier noise.
3. The method according to claim 1, wherein clustering is performed on the primarily screened sonar point clouds by using an improved whale algorithm, and a clustering center is used as a whale position, wherein the improved whale algorithm is as follows:
the spiral birth position updating mechanism of the classical whale algorithm is unchanged, and the contraction surrounding mechanism adopts a self-adaptive weight:
Figure FDA0003618405580000011
Figure FDA0003618405580000012
wherein, the first and the second end of the pipe are connected with each other,
Figure FDA0003618405580000013
indicating the updated position vector and the position vector,
Figure FDA0003618405580000014
represents the current best position vector, omega represents the adaptive weight,
Figure FDA0003618405580000015
representing the vector factor, ω max ,ω min Respectively representing the maximum value and the minimum value of the self-adaptive weight, i representing the current iteration turn, and M representing the maximum iteration number.
4. A method as claimed in claim 3, wherein the objective function of the modified whale algorithm is as follows:
Figure FDA0003618405580000016
Figure FDA0003618405580000017
wherein, the fitness represents the optimization target, n represents the variable number of the clustering center, w ij Weight, x, of ith cluster center representing jth cluster p Representing a point cloud coordinate vector, x ij Represents the cluster center coordinate, | x, to be solved p -x ij I represents x p Euclidean distances to the center of each cluster.
5. The method of any of claims 1 to 4, further comprising:
and S4, determining a fitting section of the pipeline according to the global optimal clustering center coordinate and the clustering radius which are positioned in the middle, and determining a sludge section according to the global optimal clustering center coordinate and the clustering radius which are positioned at the bottommost layer.
6. The method of claim 5, wherein if there is a data loss in the pipeline fit section, the complete pipeline section is completed based on the clustered pipeline section points.
7. A pipeline sonar point cloud denoising system is characterized by comprising: a computer-readable storage medium and a processor;
the computer-readable storage medium is used for storing executable instructions;
the processor is used for reading executable instructions stored in the computer-readable storage medium and executing the pipe sonar point cloud denoising method according to any one of claims 1 to 6.
CN202210454829.3A 2022-04-27 2022-04-27 Pipeline sonar point cloud denoising method and system Pending CN114841204A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210454829.3A CN114841204A (en) 2022-04-27 2022-04-27 Pipeline sonar point cloud denoising method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210454829.3A CN114841204A (en) 2022-04-27 2022-04-27 Pipeline sonar point cloud denoising method and system

Publications (1)

Publication Number Publication Date
CN114841204A true CN114841204A (en) 2022-08-02

Family

ID=82568346

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210454829.3A Pending CN114841204A (en) 2022-04-27 2022-04-27 Pipeline sonar point cloud denoising method and system

Country Status (1)

Country Link
CN (1) CN114841204A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115932864A (en) * 2023-02-24 2023-04-07 深圳市博铭维技术股份有限公司 Pipeline defect detection method and pipeline defect detection device

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115932864A (en) * 2023-02-24 2023-04-07 深圳市博铭维技术股份有限公司 Pipeline defect detection method and pipeline defect detection device

Similar Documents

Publication Publication Date Title
CN106599129B (en) A kind of multi-beam point cloud data denoising method for taking lineament into account
US10885352B2 (en) Method, apparatus, and device for determining lane line on road
Olson et al. Automatic target recognition by matching oriented edge pixels
CN111091095B (en) Method for detecting ship target in remote sensing image
CN111008959A (en) Grading ring defect detection method, device, medium and equipment based on aerial image
CN112884886B (en) Three-dimensional point cloud pipeline extraction and modeling method capable of adaptively searching radius
CN116704137B (en) Reverse modeling method for point cloud deep learning of offshore oil drilling platform
CN114841204A (en) Pipeline sonar point cloud denoising method and system
Liu et al. Whale optimization algorithm-based point cloud data processing method for sewer pipeline inspection
CN108765445B (en) Lung trachea segmentation method and device
CN114417951A (en) Unsupervised machine learning-based automatic subdivision optimization method for ocean unstructured grid
CN111445515A (en) Underground cylinder target radius estimation method and system based on feature fusion network
CN115546116A (en) Method and system for extracting and calculating spacing of discontinuous surface of fully-covered rock mass
CN110647647B (en) Closed graph similarity searching method based on time sequence complexity difference
CN112150497A (en) Local activation method and system based on binary neural network
CN112241676A (en) Method for automatically identifying terrain sundries
CN115098717A (en) Three-dimensional model retrieval method and device, electronic equipment and storage medium
WO2024125434A1 (en) Regional-consistency-based building principal angle correction method
CN111340145B (en) Point cloud data classification method and device and classification equipment
CN113111923A (en) Water supply network leakage detection and positioning method based on one-dimensional migration learning convolutional neural network integrated model
CN116597313A (en) Ship optical image wake detection method based on improved YOLOv7
CN115170528A (en) Pavement defect detection method, system, equipment and storage medium
CN114332533A (en) Landslide image identification method and system based on DenseNet
CN111145241A (en) Ellipse feature detection method and device
Zhao et al. Structural shape grammars used in intelligent generation design of discrete structures

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination