CN-121982610-A - Point cloud and video combined power transmission line online monitoring method and system

CN121982610ACN 121982610 ACN121982610 ACN 121982610ACN-121982610-A

Abstract

The invention relates to the technical field of power system automation and image recognition, and discloses a point cloud and video combined power transmission line online monitoring method and system. The method comprises the steps of obtaining three-dimensional point cloud data of an insulator through a laser radar, synchronously collecting multispectral video frames, denoising, registering and reconstructing the point cloud to generate a high-fidelity geometric model, enhancing the video, correcting distortion and aligning space time to form a texture mapping atlas, fusing and constructing a pollution joint characterization model, calculating thickness, volume density and space continuity indexes of a pollution layer, and generating a cleaning priority decision instruction according to the pollution joint characterization model. The system comprises a three-dimensional point cloud acquisition unit, a multispectral video acquisition unit, a point cloud processing unit, a video processing unit, a joint modeling unit, a pollution quantization unit and a decision output unit. According to the invention, three-dimensional accurate quantification of pollution is realized through geometric-optical fusion, and the scientificity and the operation and maintenance efficiency of a cleaning decision are improved.

Inventors

WANG XIAOPENG
ZHOU RUIBO
ZHENG ZEYUN

Assignees

湖南捷讯通电力技术有限公司

Dates

Publication Date: 20260505
Application Date: 20260127

Claims (10)

1. The method for online monitoring of the power transmission line by combining the point cloud and the video is characterized by comprising the following steps of: Acquiring three-dimensional point cloud data of an insulator chain and hardware fitting of the insulator chain through a laser radar sensor deployed on a transmission tower; acquiring a multispectral video frame sequence under a corresponding visual angle through a synchronously triggered visible light and ultraviolet imaging sensor; Denoising, registering and surface reconstructing the three-dimensional point cloud data to generate a high-fidelity insulator three-dimensional geometric model; Performing image enhancement, distortion correction and space-time alignment on the multispectral video frame sequence to generate a texture mapping atlas strictly corresponding to the three-dimensional geometric model; Projecting the texture mapping atlas to the surface of the three-dimensional geometric model to construct a pollution joint characterization model fusing geometric structures and optical reflection characteristics; Calculating the thickness, the volume density and the space continuity index of the pollution layer in each region of the insulator based on the pollution joint characterization model; generating an insulator cleaning priority grade according to the thickness, the volume density and the space continuity index of the pollution layer, and outputting a cleaning decision instruction.
2. The method for online monitoring of a power transmission line by combining point cloud and video according to claim 1, wherein the denoising, registering and surface reconstruction processing are performed on the three-dimensional point cloud data to generate a high-fidelity three-dimensional geometrical model of an insulator, and the method comprises the following steps: adopting a statistical outlier removal algorithm to remove abnormal points caused by atmospheric disturbance or flying birds; performing initial coarse registration on the multi-view point cloud segments by using a feature descriptor FPFH, and performing registration by using an iterative nearest point algorithm; and a poisson surface reconstruction algorithm is adopted to generate a closed and watertight triangular grid model, and the geometric details of the connection part of the edge of the umbrella skirt and the steel feet are reserved.
3. The method for on-line monitoring of a power transmission line by combining point cloud and video according to claim 2, wherein performing image enhancement, distortion correction and space-time alignment processing on the multispectral video frame sequence to generate a texture map set strictly corresponding to a three-dimensional geometric model comprises: performing illumination equalization processing based on Retinex theory on the visible light image to inhibit local overexposure or underexposure caused by backlight or strong glare; Performing background noise suppression and hot spot enhancement treatment on the ultraviolet image, and improving the signal-to-noise ratio of the weak corona discharge signal; performing radial and tangential distortion correction on all images by using a camera internal reference matrix and a lens distortion coefficient which are calibrated in advance; And establishing a one-to-one mapping relation between the corrected image pixel coordinate system and the point cloud world coordinate system through known sensor external parameters, so as to realize pixel-level space alignment.
4. The method for online monitoring of a power transmission line by combining point cloud and video according to claim 3, wherein projecting the texture mapping atlas onto the three-dimensional geometric model surface, constructing a pollution joint characterization model integrating geometric structures and optical reflection characteristics, comprises: Binding corresponding visible light reflectivity values and ultraviolet radiation intensity values on each triangular surface patch of the three-dimensional geometric model; Defining a pollution existence area as a surface area which simultaneously meets the following 2 conditions that the visible light reflectivity is smaller than a preset threshold value and the ultraviolet radiation intensity is 3 times of standard deviation larger than the background noise mean value; And further identifying geometric deformation characteristics of the region judged to be polluted according to the point cloud normal vector change rate and the surface curvature gradient.
5. The method for online monitoring of a power transmission line by combining point cloud and video according to claim 4, wherein calculating the thickness, volume density and spatial continuity index of the pollution layer in each region of the insulator based on the pollution joint characterization model comprises: The thickness of the pollution layer is obtained by comparing Euclidean distance between the current point cloud surface and a clean state reference model, and the reference model is obtained by training point cloud data in a historical pollution-free period; The volume density is defined as the average thickness of the dirt layer on the unit surface area, and integral calculation is carried out according to the umbrella skirt partition; The space continuity index is determined by calculating the number of connected components of the pollution area and the area ratio of the maximum connected area.
6. The method for online monitoring of a power transmission line by combining point cloud and video according to claim 5, wherein generating an insulator cleaning priority score according to the thickness, volume density and spatial continuity index of the pollution layer, and outputting a cleaning decision instruction comprises: setting a cleaning priority scoring function The method comprises the following steps: ; in order to normalize the thickness of the film, In order to normalize the bulk density, Is a spatial continuity index; Marking the insulator as "emergency cleaning" when the score is greater than 0.7, marking the insulator as "planned cleaning" when the score is between 0.4 and 0.7, and marking the insulator as "no cleaning" when the score is less than 0.4; The cleaning decision instruction comprises an insulator number, a position of a tower, a pollution distribution thermodynamic diagram, a suggested cleaning mode and a predicted pollution growth trend prediction.
7. The method for online monitoring of a power transmission line by combining point cloud and video according to claim 6, wherein the clean state reference model is stored in a pollution reference database, and the pollution reference database comprises three-dimensional geometric models recorded by cleaning a model insulator in a factory state, a post-rain cleaning state and a history cleaning state and is used for dynamically updating the calculation reference of the current pollution thickness.
8. The method for online monitoring of a power transmission line by combining point cloud and video according to claim 7, wherein the pollution distribution thermodynamic diagram represents the thickness of a pollution layer in color gradation, the suggested cleaning mode comprises high-pressure water flushing, mechanical scrubbing or chemical cleaning, and the predicted pollution growth trend prediction is generated based on a historical pollution data fitting index growth model.
9. The utility model provides a transmission line on-line monitoring system that point cloud and video combine which characterized in that includes: the three-dimensional point cloud acquisition unit is used for acquiring three-dimensional point cloud data of the insulator string and hardware fitting thereof through a laser radar sensor deployed on the transmission tower; The multispectral video acquisition unit is used for acquiring multispectral video frame sequences under corresponding visual angles through synchronously triggered visible light and ultraviolet imaging sensors; The point cloud processing unit is used for denoising, registering and surface reconstruction processing the three-dimensional point cloud data to generate a high-fidelity insulator three-dimensional geometric model; the video processing unit is used for carrying out image enhancement, distortion correction and space-time alignment processing on the multispectral video frame sequence to generate a texture mapping atlas strictly corresponding to the three-dimensional geometric model; the joint modeling unit is used for projecting the texture mapping atlas to the surface of the three-dimensional geometric model to construct a pollution joint characterization model fusing geometric structures and optical reflection characteristics; the pollution quantification unit is used for calculating the thickness, the volume density and the space continuity index of the pollution layer in each region of the insulator based on the pollution joint characterization model; the decision output unit is used for generating an insulator cleaning priority grade according to the thickness, the volume density and the space continuity index of the pollution layer and outputting a cleaning decision instruction.
10. The online monitoring system of the power transmission line combining point cloud and video according to claim 9, wherein when the joint modeling unit performs texture-geometry fusion, a barycentric coordinate interpolation method is adopted to map image pixel attributes to triangular mesh vertices, and surface rendering is performed through a Phong illumination model so as to verify visual consistency of a pollution area; And the decision output unit uploads a cleaning decision instruction to the provincial power transmission monitoring center through the power dispatching data network and synchronously pushes the cleaning decision instruction to the mobile terminal of the operation and maintenance personnel, and the instruction format accords with the IEC61850 communication standard.

Description

Point cloud and video combined power transmission line online monitoring method and system Technical Field The invention belongs to the technical field of power system automation and image recognition, and particularly relates to a point cloud and video combined power transmission line online monitoring method and system. Background With the deep promotion of smart power grid construction, on-line monitoring of the operation state of key equipment such as insulators of a power transmission line becomes an important link for guaranteeing the safety and stability of a power system. The current mainstream monitoring means rely on visible light or infrared video images for pollution assessment, and the core principle is that the pollution degree is indirectly inferred by analyzing two-dimensional visual characteristics such as color, gray scale, coverage area and the like of a pollution area. However, such methods are essentially limited by planar imaging mechanisms, and cannot obtain three-dimensional spatial distribution information of the contamination on the surface of the insulator, and particularly, it is difficult to quantify the actual thickness of the contamination layer and the local stacking morphology thereof. Because the influence of the pollution on the insulation performance is not only dependent on the apparent area, but also closely related to the physical thickness, density and adhesion uniformity, the cleaning decision is very easy to be too early or late only by using a two-dimensional image, so that the operation and maintenance resource waste is possibly caused, and the flashover accident is possibly caused by the missing judgment of thick-layer pollution. The three-dimensional perception technology based on the point cloud provides a new idea for solving the problems. The laser radar or the structured light scanning can acquire high-precision three-dimensional point cloud data of the surface of the insulator, and can reflect geometric deformation caused by pollution deposition theoretically. The existing point cloud processing method focuses on geometric reconstruction or simple curvature analysis, and lacks effective depicting capability on complex surface topological structure changes induced by pollution. For example, the filling of the dirt may cause the disappearance of the tiny holes existing originally or form new concave areas, and the change of the topological characteristics contains key information of dirt distribution, but the conventional point cloud algorithm is difficult to extract the quantitative index with discriminant property from the key information. In the prior art, the inherent defects of single information dimension and one-sided feature expression exist in both a pure visual method and a pure point cloud method. Although the video analysis can capture abundant texture and color depth information, but lacks space thickness, the point cloud has three-dimensional geometric advantages, but is difficult to directly characterize the material properties and optical characteristics of the pollution. The two are used independently, and complete cognition of the dirty state cannot be constructed. Especially under complex meteorological conditions or long-term operation aging scenes, the pollution presents heterogeneous, multilayer, mixed state and other complex forms, the limitation of single-mode data is further amplified, and the existing monitoring system is seriously insufficient in the aspects of pollution thickness quantification, risk level judgment, cleaning opportunity decision and the like. Therefore, an online monitoring method for depth fusion video texture semantics and a point cloud topological structure is needed to realize high-precision, three-dimensional and quantifiable intelligent evaluation of the pollution state of the insulator of the power transmission line. Disclosure of Invention The invention provides a point cloud and video combined power transmission line online monitoring method and system, and aims to solve the technical problems that the existing image method can only evaluate the color and the area of pollution and cannot quantify the thickness and the spatial distribution of three-dimensional pollution, so that the cleaning decision is inaccurate. According to the invention, by fusing high-precision three-dimensional point cloud data with a multispectral video image, a three-dimensional geometric model and material attribute combined representation of the surface pollution of the insulator are constructed, the accurate quantification of the thickness, volume, space coverage density and attachment form of the pollution layer is realized, and a cleaning priority decision instruction with physical interpretability is generated based on the accurate quantification. The invention provides a point cloud and video combined power transmission line online monitoring method, which comprises the following steps: Acquiring three-dimensional point clou