Search

CN-121997527-A - Method for establishing inspection model of box-type transformer

CN121997527ACN 121997527 ACN121997527 ACN 121997527ACN-121997527-A

Abstract

The invention provides a method for establishing a box-type transformer inspection model, which relates to the technical field of data processing, and comprises the steps of controlling inspection equipment to acquire information of target equipment to obtain inspection data; the method comprises the steps of carrying out element identification on inspection data to obtain an environment element, carrying out multidimensional screening on the environment element to obtain a static reference element, generating an environment model based on the static reference element, obtaining the position relation between a target element corresponding to target equipment and the static reference element, obtaining the target orientation of the target element based on the position relation and the reference orientation of the static reference element, carrying out model fusion on a standard box model corresponding to the target element and the environment model based on the target orientation to generate an inspection model, and carrying out model update on the inspection model according to the target element when the image comparison result of the standard box model and the target element is inconsistent, so as to obtain an updated inspection model, and carrying out high-efficiency model construction according to automatically acquired data to improve the accuracy of the inspection model.

Inventors

  • QIU ZHEN
  • ZHAO BIN

Assignees

  • 长沙理工大学

Dates

Publication Date
20260508
Application Date
20251204

Claims (13)

  1. 1. The method for establishing the inspection model of the box-type transformer is characterized by comprising the following steps of: receiving the inspection coordinates to control the inspection equipment to acquire information of the target equipment, so as to obtain inspection data; Performing element identification on the inspection data to obtain an environment element, performing multidimensional screening on the environment element to obtain a static reference element, and generating an environment model based on the static reference element; acquiring a position relation between a target element corresponding to the target device and a static reference element, and acquiring a target orientation of the target element based on the position relation and a reference orientation of the static reference element; Carrying out model fusion on a standard box model corresponding to the target element and an environment model based on the target orientation to generate a patrol model; And comparing the standard box type model with a target element, and updating the model of the inspection model according to the target element when the comparison result is inconsistent, so as to obtain an updated inspection model.
  2. 2. The method of claim 1, wherein the receiving the inspection coordinates controls the inspection device to collect information on the target device to obtain inspection data, comprising: Determining a top surface contour center point of the target equipment based on the inspection coordinates, and obtaining an initial overlooking acquisition point of the target equipment according to the center point and a preset acquisition height; Controlling the inspection equipment to move to an initial overlooking acquisition point, and when determining that the target contour of the complete target equipment does not exist in overlooking data corresponding to the inspection equipment, adjusting the height of the inspection equipment to obtain a reference acquisition height; Taking overlooking shooting on the target equipment based on the reference acquisition height to obtain overlooking reference video; And determining a side-looking acquisition point of the inspection equipment according to the overlooking reference video, and obtaining a side-looking reference video based on the side-looking acquisition point, wherein the inspection data comprise the overlooking reference video and the side-looking reference video.
  3. 3. The method of claim 2, wherein determining a side view acquisition point of the inspection device from the top view reference video, obtaining a side view reference video based on the side view acquisition point, comprises: controlling the inspection equipment to move to the corresponding contour vertex angle according to any contour vertex angle of the target contour of the target equipment in the overlooking reference video, and determining a height contour line corresponding to the corresponding contour vertex angle of the target equipment; Determining a side view acquisition point based on a preset side view distance and the height profile, and adjusting the distance of the inspection equipment to obtain a reference acquisition distance when determining the complete boundary profile of the non-target equipment in the side view data at the side view acquisition point; And performing side-looking shooting on the target equipment based on the reference acquisition interval to obtain a side-looking reference video.
  4. 4. The method of claim 3, wherein the performing element recognition on the inspection data to obtain an environmental element, and performing multidimensional screening on the environmental element to obtain a static reference element, comprises: performing element identification on the inspection data, taking an image element corresponding to target equipment as a target element, and taking the rest image elements except the target element in the inspection data as environment elements; counting the environmental elements to obtain an environmental element set, and carrying out static identification on the inspection data based on a preset judging time length to obtain a first reference element; and acquiring physical parameters of the first reference element according to the inspection image in the inspection data, and screening the first reference element based on the physical parameters to obtain a static reference element, wherein the physical parameters comprise physical height, physical width and physical length.
  5. 5. The method as recited in claim 4, further comprising: when the static reference elements are not found in the environment data, enlarging and adjusting the standard acquisition height and the standard acquisition interval to obtain current inspection data; Repeating the step of judging the static reference element based on the current inspection data until the environment element in the current inspection data has the static reference element.
  6. 6. The method of claim 4, wherein generating an environmental model based on the static reference element comprises: performing model construction according to the environmental elements to obtain an environmental sub-model, and acquiring the relative position relation among the environmental elements based on the inspection data; And carrying out model positioning on each environment sub-model based on the relative position relation to obtain an environment model.
  7. 7. The method of claim 4, wherein the performing static identification on the inspection data based on the preset determination time length to obtain the first reference element includes: Intercepting video segments of the inspection data based on a preset judging time length to obtain a comparison video segment; Sequentially identifying the outline of the environmental element in each video frame in the comparison video band according to the environmental element set to obtain the element outline of each environmental element corresponding to each video frame; acquiring contour coordinates of pixel points corresponding to the element contours to obtain a plurality of contour coordinate sets corresponding to each video frame; And comparing the contour coordinate sets corresponding to the same environmental element in each video frame to obtain a comparison result, and determining a first reference element based on the comparison result.
  8. 8. The method of claim 7, wherein comparing the set of contour coordinates corresponding to the same environmental element in each video frame to obtain a comparison result, and determining the first reference element based on the comparison result comprises: When the contour coordinate sets are determined to be identical, a comparison consistent result is generated, and when the contour coordinate sets are determined to be different, a comparison inconsistent result is generated, wherein the comparison result comprises a comparison consistent result and a comparison inconsistent result; based on the inconsistent comparison result, taking the corresponding environment element as a non-reference element; and determining the comparison consistent result, and taking the corresponding environment element as a first reference element.
  9. 9. The method of claim 8, wherein the obtaining the physical parameter of the first reference element from the inspection image in the inspection data comprises: according to the target model of the target equipment, corresponding physical parameters are called, and the pixel length of the target equipment and the contour length and the contour width of the first reference element are obtained based on the overlooking image of the overlooking reference video; obtaining a conversion ratio based on the physical length of the target equipment and the pixel length, and determining the physical length and the physical width corresponding to the first reference element according to the contour length, the contour width and the conversion ratio; Respectively obtaining contour heights corresponding to target equipment and a first reference element according to the side view image of the side view reference video; Determining the interval distance between each first reference element and the target element based on the overlook image, and determining the first distance between the inspection equipment and the first reference element according to the interval distance and the base acquisition interval; And obtaining the physical height corresponding to the first reference element according to the first distance, the contour height and the physical height corresponding to the target device and the contour height corresponding to the first reference element.
  10. 10. The method of claim 9, wherein the filtering the first reference element based on the physical parameter to obtain the static reference element comprises: comparing the physical length in the physical parameters of the first reference element with a length threshold value to obtain a first comparison result; comparing the physical height with a height threshold to obtain a second comparison result, and comparing the physical width with a width threshold to obtain a third comparison result; And when any comparison result is determined to be larger than a preset threshold value, taking the corresponding first reference element as a static reference element, wherein the preset threshold value comprises a length threshold value, a height threshold value and a width threshold value.
  11. 11. The method of claim 10, wherein the deriving the target orientation of the target element based on the positional relationship and the reference orientation of the static reference element comprises: determining the position relation between the target element and each static reference element based on the inspection data; Comparing the physical length and the physical width corresponding to the static reference elements in the overlook image, and selecting the orientation of the boundary contour corresponding to the maximum value as the reference orientation of the corresponding static reference elements; Determining a corresponding boundary contour as a reference contour based on the reference orientation, and taking a long-side contour corresponding to a target element in the overlook image as a comparison contour; and obtaining a boundary included angle according to the reference contour and the comparison contour, and obtaining a target orientation corresponding to the target element based on the reference orientation and the boundary included angle.
  12. 12. The method of claim 11, wherein the model fusing the standard box model corresponding to the target element with the environment model based on the target orientation to generate the inspection model comprises: performing model selection on a preset model library based on the target model to obtain a standard box model, and positioning the standard box model into an environment model to obtain an initial inspection model; and carrying out azimuth adjustment on the standard box type model in the initial inspection model according to the target orientation to obtain an inspection model.
  13. 13. The method of claim 12, wherein the comparing the standard box model with the target element to determine that the comparison result is inconsistent, updating the inspection model according to the target element to obtain an updated inspection model comprises: obtaining a standard box image based on a standard box model, carrying out image recognition on the target elements, and taking the image elements connected with the standard box image in the target elements as connection elements; and generating a connection model according to the connection elements, and updating the standard box type model in the inspection model based on the connection model to obtain an updated inspection model.

Description

Method for establishing inspection model of box-type transformer Technical Field The invention relates to a data processing technology, in particular to a method for establishing a box-type transformer inspection model. Background As a key node in modern power distribution networks, box transformers operate stably directly with respect to the reliability and safety of the power supply. With the development of intelligent power grids and unmanned inspection technologies, an unmanned aerial vehicle or an inspection robot carrying a high-definition camera is utilized to automatically collect data of a box-type transformer, and a three-dimensional digital inspection model of the box-type transformer is built, so that the intelligent power grid and the unmanned aerial vehicle have become an important means for realizing state evaluation, fault early warning and operation and maintenance management modernization. The digital inspection model can intuitively reproduce equipment and the surrounding environment thereof, provides remote and visual decision support for operation and maintenance personnel, and greatly improves inspection efficiency and safety. However, in the existing box-type transformer inspection model construction technology, most methods rely on directly performing three-dimensional reconstruction on acquired images or point cloud data, and the method is extremely easy to interfere with a field complex environment. For example, reflection, fouling or shielding of the device surface caused by illumination changes can make it difficult for the image recognition algorithm to accurately extract the complete contour of the target device, thereby causing failure in model construction or serious degradation in accuracy. Therefore, how to perform efficient model construction according to the automatically collected data and improve the accuracy of the inspection model becomes a problem to be solved. Disclosure of Invention The invention provides a method for establishing a box-type transformer inspection model, which can be used for carrying out high-efficiency model construction according to automatically acquired data and improving the accuracy of the inspection model. The first aspect of the invention provides a method for establishing a box-type transformer inspection model, which comprises the following steps: receiving the inspection coordinates to control the inspection equipment to acquire information of the target equipment, so as to obtain inspection data; Performing element identification on the inspection data to obtain an environment element, performing multidimensional screening on the environment element to obtain a static reference element, and generating an environment model based on the static reference element; acquiring a position relation between a target element corresponding to the target device and a static reference element, and acquiring a target orientation of the target element based on the position relation and a reference orientation of the static reference element; Carrying out model fusion on a standard box model corresponding to the target element and an environment model based on the target orientation to generate a patrol model; And comparing the standard box type model with a target element, and updating the model of the inspection model according to the target element when the comparison result is inconsistent, so as to obtain an updated inspection model. Optionally, in one possible implementation manner of the first aspect, the receiving patrol coordinates controls the patrol equipment to collect information on the target equipment, so as to obtain patrol data, and the method includes: Determining a top surface contour center point of the target equipment based on the inspection coordinates, and obtaining an initial overlooking acquisition point of the target equipment according to the center point and a preset acquisition height; Controlling the inspection equipment to move to an initial overlooking acquisition point, and when determining that the target contour of the complete target equipment does not exist in overlooking data corresponding to the inspection equipment, adjusting the height of the inspection equipment to obtain a reference acquisition height; Taking overlooking shooting on the target equipment based on the reference acquisition height to obtain overlooking reference video; And determining a side-looking acquisition point of the inspection equipment according to the overlooking reference video, and obtaining a side-looking reference video based on the side-looking acquisition point, wherein the inspection data comprise the overlooking reference video and the side-looking reference video. Optionally, in a possible implementation manner of the first aspect, the determining a side view acquisition point of the inspection device according to the top view reference video, and obtaining the side view reference video based on the side view acquisition point