Search

CN-115661931-B - Operation and maintenance platform management system and method based on semantic track analysis

CN115661931BCN 115661931 BCN115661931 BCN 115661931BCN-115661931-B

Abstract

The invention discloses an operation and maintenance platform management system and method based on semantic track analysis, which belong to the technical field of semantic track analysis and comprise an operation and maintenance platform server and a plurality of terminal operation and maintenance devices, wherein basic semantic track data and computer control operation instructions are associated to construct a semantic track analysis model, an exclusive semantic track data database is constructed for each operation and maintenance person at the operation and maintenance platform server, and the semantic track analysis model is optimized based on the exclusive semantic track data database to form different exclusive semantic track analysis models for different operation and maintenance persons.

Inventors

  • JI WEI
  • LIU YUN
  • YU JING
  • HUANG TAOWEI
  • DI JIA

Assignees

  • 浙江省邮电工程建设有限公司

Dates

Publication Date
20260508
Application Date
20221027

Claims (8)

  1. 1. The operation and maintenance platform management system based on semantic track analysis is characterized by comprising an operation and maintenance platform server and a plurality of terminal operation and maintenance devices connected with the operation and maintenance platform server; The operation and maintenance platform server is used for pre-constructing a semantic track database and a computer control operation instruction database for storing computer control operation instructions, and associating semantic track data with the computer control operation instructions to construct a semantic analysis model; Constructing a database of stored semantic trajectories, comprising: Step 1, starting from the characteristics of the semantic track, constructing an abstract conceptual model of the semantic track, wherein the abstract conceptual model comprises tracks, track fragments, stay, travel chains and activities; step2, formalizing definition is carried out on the semantic track; step 3, supporting space data types and developing source expansibility, so that semantic track data types and other PostgreSQL original data types are used in a database; step 4, realizing operation related to the type of the semantic track data, realizing effective storage of the track data, and constructing a semantic track storage database; the step 1 comprises the following steps: Step 1.1, acquiring an original image, and segmenting a dynamic object in the original image by utilizing an image semantic segmentation technology in deep learning to obtain a semantic image containing pixel-level semantic information of the dynamic object; Step 1.2, extracting candidate points from the original image, removing dynamic region candidate points according to the semantic image obtained in the step 1, and only keeping static region candidate points; Step 1.3, estimating the pose of the camera by adopting a pyramid model fused with image semantic information based on the reserved candidate points; Step 1.4, optimizing the pose of the key frame based on sliding window optimization and combining with image semantic information, and further constructing an abstract conceptual model of the semantic track; The terminal operation and maintenance equipment is used for uploading semantic track data of operation and maintenance personnel in current operation to the operation and maintenance platform server to bring the semantic track data into an exclusive semantic track analysis model of the operation and maintenance personnel in current operation, acquiring a computer control operation instruction with the association degree of a keyword larger than a preset threshold value, ordering based on the association degree of the computer control operation instruction to create a basic computer control operation instruction list, and transmitting the basic computer control operation instruction list to the terminal operation and maintenance equipment, wherein the terminal operation and maintenance equipment executes the computer control operation instruction selected by the operation and maintenance personnel in current operation from the basic computer control operation instruction list, and uploads the computer control operation instruction selected by the operation and maintenance personnel in current operation to the operation and maintenance platform server to optimize the exclusive semantic track analysis model of the operation and maintenance personnel in current operation, so that the association degree between the keyword and the selected computer control operation instruction is improved.
  2. 2. The operation and maintenance platform management system based on semantic trajectory analysis according to claim 1, wherein the step 1.3 specifically comprises the following steps: Step 1.31 for Key Frames Scaling the image with a scaling factor of 0.5 to obtain a relative image respectively Resolution ratio Will be Constructing an image pyramid with the three zoomed images in sequence from low resolution to high resolution, and recording a k layer image of the image pyramid as Extracting static region candidate points of each layer of the pyramid in the step 2; step 1.32 for the subsequent frame Semantic images corresponding to the same Constructing an image pyramid similar to the image pyramid in the step 1.31, and respectively marking the k-th layer image of the image pyramid as And ; Step 1.33 for Single candidate point in (a) Calculate its projection to the image The photometric error formed above: Wherein, the Is that At the position of The projection point on the upper surface of the lens, And Respectively, are images And Is used for the exposure time of the substrate, , , , Is the photometric transfer function parameter of the image, Is composed of A set of 8 points in total of points and surrounding neighboring points, Is a weight factor that is used to determine the weight of the object, Is the Huber norm; Step 1.34, for each candidate Point According to it in Mid-projection point Calculates a reject label Determining whether the projection residual of the point is removed: ; step 1.35, will The projection residual errors of the points in the pyramid are accumulated, and the residual errors projected to the dynamic area are removed, so that the projection residual error sum of the k layer of the pyramid is obtained: ; Step 1.36 to at Layer relative pose optimization results As the initial value of the optimization, gauss Newton method is utilized for the optimization Optimizing to obtain key frame And subsequent frames Relative pose between ; Step 1.37, repeating steps 1.33-1.36 for all layers of the pyramid from top to bottom, and finally obtaining the key frame And subsequent frames Relative pose between 。
  3. 3. The operation and maintenance platform management system based on semantic trajectory analysis according to claim 2, wherein in step 1.4, the key frame pose is optimized by adopting a sliding window and semantic information, and the operation and maintenance platform management system specifically comprises the following steps: Step 1.41 for Key Frames Single point in (a) It projects to another key frame in the sliding window The photometric error formed above is: Wherein the method comprises the steps of Is that At the position of The projection point on the upper surface of the lens, And Respectively, are images And Is used for the exposure time of the substrate, , , , Is the photometric transfer function parameter of the image, Is composed of A set of 8 points in total of points and surrounding neighboring points, Is a weight factor that is used to determine the weight of the object, Is the Huber norm; step 1.42 for Key Frames Each candidate point According to it in Mid-projection point Calculates a reject label Determining whether the projection residual of the point is removed: wherein, the method comprises the steps of, Representing key frames Is a semantic image of (1); Step 1.43, traversing all key frames in the sliding window, projecting all candidate points in the key frames to other key frames in the window, and counting and accumulating all photometric errors: Where F is the set of all key frames within the sliding window, Is a key frame A set of all the candidate points in the (c), The finger can observe A keyframe set of points; Step 1.44, couple using Gauss Newton method And (3) optimizing to obtain the optimized pose of all the key frames, and completing tracking of the camera motion.
  4. 4. An operation and maintenance management system based on semantic trajectory analysis according to claim 3, wherein the method for assisting operation and maintenance management by semantic trajectory analysis further comprises the steps of: if the number of the computer control operation instructions which are analyzed by the exclusive semantic track analysis model of the currently operated operation and maintenance personnel and have the association degree with the keywords larger than a preset threshold value is 0, sequentially uploading the keywords to exclusive semantic track analysis models corresponding to other operation and maintenance personnel; Acquiring a computer control operation instruction with the keyword association degree larger than a preset threshold value and the highest keyword association degree through a special semantic track analysis model corresponding to other operation and maintenance personnel; Integrating and sequencing the computer control operation instructions obtained by the exclusive semantic track analysis models corresponding to other operation and maintenance personnel to form an alternative computer control operation instruction list, and pushing the alternative computer control operation instruction list to the terminal operation and maintenance equipment for selection of the operation and maintenance personnel currently operating; If the computer operation control instruction is selected by the operation and maintenance personnel currently operating, the terminal operation and maintenance equipment executes the computer control operation instruction selected by the operation and maintenance personnel currently operating from the alternative computer control operation instruction list, and uploads the computer control operation instruction selected by the operation and maintenance personnel currently operating to the operation and maintenance platform server to optimize the exclusive semantic track analysis model of the operation and maintenance personnel currently operating, and adds the association relation between the keyword and the selected computer control operation instruction and improves the association degree; If the computer operation control instruction is not selected by the operation and maintenance personnel currently operated, the terminal operation and maintenance equipment executes the computer control operation instruction searched and selected by the operation and maintenance personnel currently operated, and uploads the computer control operation instruction searched and selected by the operation and maintenance personnel currently operated to the operation and maintenance platform server to optimize the exclusive semantic track analysis model of the operation and maintenance personnel currently operated, and adds the association relation between the keyword and the selected computer control operation instruction and improves the association degree.
  5. 5. The operation and maintenance management system based on semantic trajectory analysis according to claim 4, wherein if a difference between the degree of association of the first-ranked computer control operation instruction list and the degree of association of the second-ranked computer control operation instruction list in the basic computer control operation instruction list is greater than a preset difference, broadcasting the first-ranked computer control operation instruction to the currently operated operation and maintenance personnel and requesting to confirm whether to execute the computer control operation instruction; the terminal operation and maintenance equipment receives the confirmation information input by the operation and maintenance personnel of the current operation through the semantic track data, and the operation and maintenance personnel of the current operation select a computer control operation instruction from the basic computer control operation instruction list, and the terminal operation and maintenance equipment executes the selected computer control operation instruction; If only one computer control operation instruction exists in the basic computer control operation instruction list, the relevance of the computer control operation instruction list with the second relevance rank is regarded as 0.
  6. 6. The operation and maintenance management system based on semantic trajectory analysis according to claim 5, The terminal operation and maintenance equipment transmits login information of operation and maintenance personnel to the operation and maintenance platform server for identity verification, and allows the operation and maintenance personnel to log in after the identity verification is passed; after receiving a semantic track data control instruction of an operation and maintenance person currently operating, the terminal operation and maintenance equipment firstly confirms whether the semantic track data control instruction comes from the corresponding operation and maintenance person through voiceprint recognition.
  7. 7. An operation and maintenance management method based on semantic track analysis, which realizes the operation and maintenance management system based on semantic track analysis as set forth in any one of claims 1 to 6, and is characterized by comprising the following steps: A basic semantic track data database for storing basic semantic track data and a computer control operation instruction database for storing computer control operation instructions are pre-built in an operation and maintenance platform server, and the basic semantic track data and the computer control operation instructions are associated to construct a semantic track analysis model; constructing an exclusive semantic track data database aiming at each operation and maintenance person at the operation and maintenance platform server, and optimizing the semantic track analysis model based on the exclusive semantic track data database so as to form different exclusive semantic track analysis models aiming at different operation and maintenance persons; The terminal operation and maintenance equipment converts semantic track data control instructions of operation and maintenance personnel currently operated into words, extracts keywords, uploads the keywords to an operation and maintenance platform server to be brought into a dedicated semantic track analysis model of the operation and maintenance personnel currently operated, acquires computer control operation instructions with the association degree with the keywords being larger than a preset threshold value, and performs sorting based on the association degree of the computer control operation instructions to create a basic computer control operation instruction list and transmits the basic computer control operation instruction list to the terminal operation and maintenance equipment; and the terminal operation and maintenance equipment executes a computer control operation instruction selected by the operation and maintenance personnel of the current operation from the basic computer control operation instruction list, uploads the computer control operation instruction selected by the operation and maintenance personnel of the current operation to an operation and maintenance platform server, optimizes a dedicated semantic track analysis model of the operation and maintenance personnel of the current operation, and improves the association degree between the keywords and the selected computer control operation instruction.
  8. 8. The operation and maintenance management method based on semantic trajectory analysis according to claim 7, further comprising the steps of: if the number of the computer control operation instructions which are analyzed by the exclusive semantic track analysis model of the currently operated operation and maintenance personnel and have the association degree with the keywords larger than a preset threshold value is 0, sequentially uploading the keywords to exclusive semantic track analysis models corresponding to other operation and maintenance personnel; Acquiring a computer control operation instruction with the keyword association degree larger than a preset threshold value and the highest keyword association degree through a special semantic track analysis model corresponding to other operation and maintenance personnel; Integrating and sequencing the computer control operation instructions obtained by the exclusive semantic track analysis models corresponding to other operation and maintenance personnel to form an alternative computer control operation instruction list, and pushing the alternative computer control operation instruction list to terminal operation and maintenance equipment for selection of the operation and maintenance personnel currently operating; If the computer operation control instruction is selected by the operation and maintenance personnel currently operating, the terminal operation and maintenance equipment executes the computer control operation instruction selected by the operation and maintenance personnel currently operating from the alternative computer control operation instruction list, uploads the computer control operation instruction selected by the operation and maintenance personnel currently operating to an operation and maintenance platform server to optimize a dedicated semantic track analysis model of the operation and maintenance personnel currently operating, adds the association relation between the keywords and the selected computer control operation instruction and improves the association degree; if the computer operation control instruction is not selected by the operation and maintenance personnel currently operated, the terminal operation and maintenance equipment executes the computer control operation instruction searched and selected by the operation and maintenance personnel currently operated, and uploads the computer control operation instruction searched and selected by the operation and maintenance personnel currently operated to an operation and maintenance platform server to optimize a dedicated semantic track analysis model of the operation and maintenance personnel currently operated, and adds the association relation between the keywords and the selected computer control operation instruction and improves the association degree; If the difference value between the relevance of the computer control operation instruction list with the first relevance rank and the relevance of the computer control operation instruction list with the second relevance rank in the basic computer control operation instruction list is larger than a preset difference value, broadcasting the computer control operation instruction with the first relevance rank to the operation and maintenance personnel currently operating and requesting to confirm whether to execute the computer control operation instruction; the terminal operation and maintenance equipment receives the confirmation information input by the operation and maintenance personnel of the current operation through the semantic track data, and the operation and maintenance personnel of the current operation select a computer control operation instruction from the basic computer control operation instruction list, and the terminal operation and maintenance equipment executes the selected computer control operation instruction; If only one computer control operation instruction exists in the basic computer control operation instruction list, the relevance of the computer control operation instruction list with the second relevance rank is regarded as 0.

Description

Operation and maintenance platform management system and method based on semantic track analysis Technical Field The invention belongs to the technical field of semantic trajectory analysis, and particularly relates to an operation and maintenance platform management system and method based on semantic trajectory analysis. Background The control of the traditional semantic track recognition is usually based on a large amount of data for model training, and has good recognition and control effects on common universal words. And the professional fields with higher safety requirements such as operation and maintenance management of the power equipment generally specify professional operation and maintenance personnel to execute operation and maintenance tasks, the number of the professional operation and maintenance personnel is small, and the semantic habit of each personnel and the issuing habit of the control instruction have larger differences. The traditional semantic track recognition control mode is relatively low in adaptation degree under the scene, and personalized setting is required for operation and maintenance personnel, so that the convenience of the operation and maintenance personnel in controlling an operation and maintenance system is improved. Disclosure of Invention The invention aims to provide an operation and maintenance platform management system and method based on semantic track analysis aiming at the defects of the background technology so as to solve the problems in the background technology. In order to achieve the above purpose, the present invention provides the following technical solutions: An operation and maintenance platform management system based on semantic track analysis comprises an operation and maintenance platform server and a plurality of terminal operation and maintenance devices connected with the operation and maintenance platform server; The operation and maintenance platform server is used for pre-constructing a semantic track database and a computer control operation instruction database for storing computer control operation instructions, and associating semantic track data with the computer control operation instructions to construct a semantic analysis model; The terminal operation and maintenance equipment is used for uploading semantic track data of operation and maintenance personnel in current operation to the operation and maintenance platform server to bring the semantic track data into an exclusive semantic track analysis model of the operation and maintenance personnel in current operation, acquiring a computer control operation instruction with the association degree of a keyword larger than a preset threshold value, ordering based on the association degree of the computer control operation instruction to create a basic computer control operation instruction list, and transmitting the basic computer control operation instruction list to the terminal operation and maintenance equipment, wherein the terminal operation and maintenance equipment executes the computer control operation instruction selected by the operation and maintenance personnel in current operation from the basic computer control operation instruction list, and uploads the computer control operation instruction selected by the operation and maintenance personnel in current operation to the operation and maintenance platform server to optimize the exclusive semantic track analysis model of the operation and maintenance personnel in current operation, so that the association degree between the keyword and the selected computer control operation instruction is improved. As a further preferable scheme of the operation and maintenance platform management system based on semantic track analysis, the invention constructs a database for storing semantic tracks, and specifically comprises the following steps of; Step 1, an abstract conceptual model of the semantic track is constructed from the characteristics of the semantic track, wherein the abstract conceptual model comprises tracks, track fragments, stay, travel chains and activities. Step2, formalizing definition is carried out on the semantic track; step 3, supporting space data types and developing source expansibility, so that semantic track data types and other PostgreSQL original data types are used in a database; and 4, realizing the operation related to the type of the semantic track data, realizing the effective storage of the track data, and constructing a semantic track database. As a further preferable scheme of the operation and maintenance platform management system based on semantic trajectory analysis, the step 1 specifically comprises the following steps: Step 1.1, acquiring an original image, and segmenting a dynamic object in the original image by utilizing an image semantic segmentation technology in deep learning to obtain a semantic image containing pixel-level semantic information of the dynamic object; Step 1.2,