GIS equipment laser three-dimensional point cloud data denoising method

Abstract

The invention discloses a GIS equipment laser three-dimensional point cloud data denoising method, and relates to the field of data denoising, and the method comprises the steps: S1, obtaining, processing and analyzing GIS equipment laser three-dimensional point cloud data, the GIS point cloud data comprising X, Y and Z coordinate information of a Cartesian coordinate system of a three-dimensional object on a sampling curved surface and the sampling curved surface of the three-dimensional object; s2, constructing a de-noising mapping model, wherein the de-noising mapping model comprises a local region feature de-noising layer and an overlapping region feature de-noising layer; and S3, substituting the GIS equipment laser three-dimensional point cloud data into the denoising mapping model to obtain a target denoising point cloud and generate a three-dimensional image. According to the GIS equipment laser three-dimensional point cloud data denoising method provided by the invention, data of irrelevant equipment can be more accurately removed and subjected to noise reduction processing, and the complexity in the operation process is reduced.

Description

Technical Field

[0001] The invention relates to the technical field of noise reduction processing, and in particular to a method for denoising laser three-dimensional point cloud data of a GIS device. Background Art

[0002] The strong uncertainty, volatility and large amount of harmonics in the new power system will cause power equipment to be subjected to more extreme and drastically changing operating conditions, which puts higher requirements on the safe and reliable operation of power equipment. Comprehensive, timely and accurate perception of the status of power equipment is a prerequisite for ensuring the safety of equipment, and it is also a technical bottleneck for realizing the digitalization and intelligence of power equipment.

[0003] Under the current background of new power systems, the digitization and intelligence of GIS equipment are still in their early stages. The existing GIS status monitoring technology cannot fully and accurately grasp the internal status of GIS equipment. The operation and maintenance efficiency of the equipment is difficult to effectively improve, and GIS equipment failures occur frequently. Therefore, it is necessary to carry out research on holographic perception and precise evaluation technology of GIS equipment based on multi-parameter collaboration.

[0004] GIS power equipment is increasingly used in power systems due to its superior performance. Three-dimensional GIS technology can achieve accurate and rapid positioning and management of facilities through three-dimensional modeling of power facilities, and display status information, equipment parameters and other information of power facilities in three-dimensional scenes.

[0005] In the process of establishing GIS three-dimensional models, the acquisition of point cloud data is very important, but the point cloud data acquisition process is often affected by factors such as instrument accuracy, scanning environment, air conditions, external interference, operator experience, etc., resulting in a large number of noise points and abnormal points in the original point cloud data. Point cloud data is a data form that expresses the geometric and physical characteristics of the measured object. When the above problems occur, the obtained point cloud data may not be able to fully reflect the geometric characteristics and local details of the transmission line. Therefore, the quality of the original point cloud obtained cannot fully meet industrial requirements and directly affects the accuracy of the later model reconstruction results.

[0006] Denoising of laser 3D point cloud data from GIS equipment is a complex and important process, and it is necessary to select an appropriate denoising method based on the specific application scenario and data characteristics.

[0007] Therefore, a GIS equipment laser three-dimensional point cloud data denoising method is urgently needed to solve the problem of removing noise in the original point cloud data in a point cloud set with distinct features. Summary of the invention

[0008] To solve the above problems, the present application provides a method for denoising laser three-dimensional point cloud data of GIS equipment, with the goal of improving data accuracy, aiming to improve the precision of later modeling.

[0009] In this application, a method for denoising laser three-dimensional point cloud data of a GIS device is provided, comprising the following steps: S1. Acquire and process the GIS equipment laser three-dimensional point cloud data, wherein the GIS point cloud data includes the X, Y and Z coordinate information of the Cartesian coordinate system of the three-dimensional object on the sampling surface and the sampling surface of the three-dimensional object; S2. Constructing a denoising mapping model, wherein the denoising mapping model includes a local area feature denoising layer and an overlapping area feature denoising layer; S3. Bring the GIS equipment laser three-dimensional point cloud data into the denoising mapping model to obtain the target denoised point cloud and generate a three-dimensional image.

[0010] Preferably, the specific contents of obtaining the GIS equipment laser three-dimensional point cloud data and processing and analyzing it in S1 are: The laser 3D point cloud data of GIS equipment is sorted and analyzed to obtain N noise 3D points and sampling surface S; For a noisy 3D point P i Perform matrix processing to obtain the matrix P=[p1, p2, ..., p N ]∈R 3×N .

[0011] Preferably, the specific content of the local area feature denoising layer in S2 is: P i The calculation is limited to r within the adjacent range K i ∈R N×K , then the local noise 3D point x in the adjacent range K i The local matrix is ​​X i =[x1, x2, ..., x K ]∈R 3×K ; Establish a local reference plane and transform the local noise 3D point x i Projecting onto the local reference plane yields X i ; The orthonormal basis of the local reference plane is W i =[w1,w2]∈R 3×2 ; Relative to W i , p i The coordinates are (s1, s2) T ; A local area feature denoising layer model is established to obtain the projection of Xi on the local reference plane and then obtain the initial denoised point cloud data.

[0012] Preferably, the expression of the local area feature denoising layer model is: X i =Pri =w1S i +n i 1 T +O i +E i ; Among them, S i =[s1, s2, ..., s K ]∈R 2×K is a matrix whose j columns represent x j For the local reference plane, O i =[o1, o2, ..., o K ]∈R 3×K The columns of a matrix represent local distribution points with large deviations, i.e., outliers. i is a Gaussian noise component with bounded energy; Preferably, the specific content of the overlapping area feature denoising layer in S2 is: A rank-2 prior model is applied to all neighborhoods of the initial denoised point cloud data to form a collaborative estimate of the redundancy existing in the overlapping areas of the local regions; Solve the prior model to get the denoised point cloud , that is, the target denoised point cloud.

[0013] Preferably, the expression of the prior model is: ; Among them, L i is the local data matrix The rank-2 approximation of is a data fitting parameter that controls the trade-off between denoising “strength” and data fidelity. Each diagonal matrix Ω i ∈R K×K .

[0014] Solve the prior model to get the denoised point cloud The specific content is: The prior model solving algorithm is preset, and the input is the noise point cloud P, the size of the local neighborhood K, and the data fitting parameters And calculate Pˆ.

[0015] In summary, compared with the traditional denoising technology, the denoising method of the laser three-dimensional point cloud data of a GIS device of the present invention constructs a denoising mapping model. The denoising mapping model includes a local area feature denoising layer and an overlapping area feature denoising layer. The spatial information of the north object is multi-directionally removed to achieve the effect of accurately and effectively characterizing the spatial information of the measured object.

[0016] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A method step diagram of a method for denoising laser three-dimensional point cloud data of a GIS device according to the present invention; Figure 2 A point cloud data diagram before denoising of a method for denoising laser three-dimensional point cloud data of a GIS device according to the present invention; Figure 3 This is a denoised point cloud data diagram of a method for denoising laser three-dimensional point cloud data of a GIS device according to the present invention. DETAILED DESCRIPTION

[0018] The technical solution of the present invention is further described below by means of the accompanying drawings and embodiments. It should be noted that unless otherwise specifically stated, the relative arrangement of components and steps, numerical expressions and numerical values ​​described in these embodiments do not limit the scope of the present application.

[0019] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the present application, its application, or uses.

[0020] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0021] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0022] Unless otherwise defined, technical or scientific terms used in the present invention shall have the common meanings understood by one having ordinary skills in the field to which the present invention belongs.

[0023] 3D GIS technology is a new technology that combines geographic information system and 3D visualization technology. In the field of power, 3D GIS technology can help power companies achieve 3D modeling, visualization analysis and spatial data analysis of power facilities and power grid systems. This paper studies the data removal and noise reduction processing technology of irrelevant devices in laser point cloud data suitable for GIS.

[0024] like Figure 1 As shown, the present invention provides a method for denoising laser three-dimensional point cloud data of a GIS device, which is specifically as follows: S1, obtaining laser three-dimensional point cloud data of a GIS device and processing and analyzing the data, wherein the GIS point cloud data includes X, Y and Z coordinate information of a three-dimensional object in a Cartesian coordinate system on a sampling surface and a sampling surface of the three-dimensional object; Furthermore, the specific contents of obtaining the GIS equipment laser three-dimensional point cloud data and processing and analyzing it in S1 are as follows: The laser 3D point cloud data of GIS equipment is sorted and analyzed to obtain N noise 3D points and sampling surface S; For a noisy 3D point P i Perform matrix processing to obtain the matrix P=[p1, p2, ..., p N ]∈R 3×N .

[0025] S2. Constructing a denoising mapping model, wherein the denoising mapping model includes a local area feature denoising layer and an overlapping area feature denoising layer; Furthermore, the specific content of the local area feature denoising layer in S2 is: P i The calculation is limited to r within the adjacent range K i ∈R N×K , then the local noise 3D point x in the adjacent range K i The local matrix is ​​X i =[x1, x2, ..., x K ]∈R 3×K ; Establish a local reference plane and transform the local noise 3D point x i Projecting onto the local reference plane yields X i ; The orthonormal basis of the local reference plane is W i =[w1,w2]∈R 3×2 ; The normal vector of the standard orthogonal basis to the local plane is m i =w1×w2; Relative to W i , p i The coordinates are (s1, s2) T ; A local area feature denoising layer model is established to obtain the projection of Xi on the local reference plane and then obtain the initial denoised point cloud data.

[0026] Furthermore, the expression of the local area feature denoising layer model is: X i =Pr i =w1S i +n i 1 T +O i +E i ; Among them, S i =[s1, s2, ..., s K ]∈R 2×K is a matrix whose j columns represent x j For the local reference plane, O i=[o1, o2, ..., o K ]∈R 3×K The columns of a matrix represent local distribution points with large deviations, i.e., outliers. i is a Gaussian noise component with bounded energy; Furthermore, the specific content of the overlapping area feature denoising layer in S2 is: A rank-2 prior model is applied to all neighborhoods of the initial denoised point cloud data to form a collaborative estimate of the redundancy existing in the overlapping areas of the local regions; Solve the prior model to get the denoised point cloud , that is, the target denoised point cloud.

[0027] Preferably, the expression of the prior model is: ; Among them, L i is the local data matrix , τ is a data fitting parameter that controls the trade-off between denoising “strength” and data fidelity. Each diagonal matrix Ω i ∈R K×K .

[0028] S3. Bring the GIS equipment laser three-dimensional point cloud data into the denoising mapping model to obtain the target denoised point cloud and generate a three-dimensional image.

[0029] Solve the prior model to get the denoised point cloud The specific content is: Input the noise point cloud P, the size of the local neighborhood K, the data fitting parameter τ and calculate .

[0030] like Figure 2 and Figure 3 As shown, Figure 2 is the point cloud data before denoising, Figure 3 This is the point cloud data after denoising. The data before denoising has more noise and fuzzy features. The point cloud data before denoising contains a lot of noise, which may be introduced by factors such as sensor errors and environmental interference. Due to the existence of noise, important geometric features in the point cloud are obscured and difficult to identify clearly.

[0031] The data characteristics after denoising are reduced noise and enhanced features. After being processed by the model of the present invention, the noise in the point cloud is significantly reduced, and the data is cleaner. The denoised point cloud retains more geometric features, making the subsequent 3D reconstruction and analysis more accurate.

[0032] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for denoising laser three-dimensional point cloud data of GIS equipment, characterized in that: The following steps are involved: S1. Acquire and process the GIS equipment laser three-dimensional point cloud data, wherein the GIS point cloud data includes the X, Y and Z coordinate information of the Cartesian coordinate system of the three-dimensional object on the sampling surface and the sampling surface of the three-dimensional object; S2. Constructing a denoising mapping model, wherein the denoising mapping model includes a local area feature denoising layer and an overlapping area feature denoising layer; S3. Bring the GIS equipment laser three-dimensional point cloud data into the denoising mapping model to obtain the target denoised point cloud and generate a three-dimensional image.

2. A GIS equipment laser three-dimensional point cloud data denoising method according to claim 1, characterized in that: The specific contents of obtaining and processing the GIS equipment laser 3D point cloud data in S1 are as follows: The laser 3D point cloud data of GIS equipment is sorted and analyzed to obtain N noise 3D points and sampling surface S; For a noisy 3D point P i Perform matrix processing to obtain the matrix P=[p1, p2, ..., p N ]∈R 3×N .

3. A GIS equipment laser three-dimensional point cloud data denoising method according to claim 1, characterized in that: The specific content of the local area feature denoising layer in S2 is: P i The calculation is limited to r within the adjacent range K i ∈R N×K , then the local noise 3D point x in the adjacent range K i The local matrix is ​​X i =[x1, x2, ..., x K ]∈R 3×K ; Establish a local reference plane and transform the local noise 3D point x i Projecting onto the local reference plane yields X i ; The orthonormal basis of the local reference plane is W i =[w1,w2]∈R 3×2 ; Relative to W i , p i The coordinates are (s1, s2) T ; Establish a local area feature denoising layer model and obtain X i The projection on the local reference plane then obtains the initial denoised point cloud data.

4. A method for denoising laser three-dimensional point cloud data of GIS equipment according to claim 3, characterized in that: The expression of the local area feature denoising layer model is: X i =Pr i =w1S i +n i 1 T +O i +E i ; Among them, S i =[s1,s2,…,s K ]∈R 2×K is a matrix whose j columns represent x j For the local reference plane, O i =[o1, o2, ..., o K ]∈R 3×K The columns of a matrix represent local distribution points with large deviations, i.e., outliers. i is a Gaussian noise component with bounded energy.

5. A method for denoising laser three-dimensional point cloud data of GIS equipment according to claim 1, characterized in that: The specific content of the overlapping area feature denoising layer in S2 is: A rank-2 prior model is applied to all neighborhoods of the initial denoised point cloud data to form a collaborative estimate of the redundancy existing in the overlapping areas of the local regions; Solve the prior model to get the denoised point cloud , that is, the target denoised point cloud.

6. A method for denoising laser three-dimensional point cloud data of GIS equipment according to claim 5, characterized in that: The expression of the prior model is: ； Among them, L i is the local data matrix The rank-2 approximation of is a data fitting parameter that controls the trade-off between denoising "strength" and data fidelity. Each diagonal matrix Ω i ∈R K×K .

7. A method for denoising laser three-dimensional point cloud data of GIS equipment according to claim 5, characterized in that: Solve the prior model to get the denoised point cloud The specific content is: The prior model solving algorithm is preset, and the input is the noise point cloud P, the size of the local neighborhood K, and the data fitting parameters And calculate .

Images