Power Facility Shape Distortion Analysis System

Abstract

The present invention relates to a shape distortion analysis system for electric power facilities, comprising: a distortion detection component, for simultaneously inputting selected multiple visual data into a deep neural network that has been reconstructed after multiple trainings, and running the deep neural network that has been reconstructed after multiple trainings to obtain a state identification outputted by the deep neural network indicating whether the set hardware is distorted in shape; an identification analysis component, connected to the distortion detection component, for sending a first analysis signal when the received state identification indicates that the set hardware is not distorted, otherwise, sending a second analysis signal. Through the present invention, it is possible to use a deep neural network that has been reconstructed after multiple trainings to complete intelligent identification of whether the set hardware is distorted in shape based on multiple basic data, thereby using a high-precision electronic detection mechanism at the pixel level to complete reliable identification of whether the set hardware is distorted in shape.

Description

Technical Field

[0001] The present invention relates to the field of electric power facilities, and in particular to a shape distortion analysis system for electric power facilities. Background Art

[0002] Power appliances refer to various electrical equipment or components, including cables, switches, sockets, meters, wires, electrical accessories, motors, electronic components, etc., whose main function is to receive, distribute or use electrical energy. Power facilities play a very important role in the power supply of modern society. Various equipment and facilities involved in power generation, transmission, transformation, distribution and power consumption are all components of power facilities. The development and operation of power facilities must be taken seriously to ensure people's normal production and life.

[0003] For example, as an important type of electrical appliance, whether the shape of the hardware is distorted determines the quality and performance of the hardware. A large degree of shape distortion can be measured roughly through a simple electronic detection mechanism or a manual measurement mechanism. However, for a slight degree of shape distortion, a high-precision electronic detection mechanism is required to conduct targeted identification of the shape distortion of the hardware. Obviously, the existing technology lacks a corresponding high-precision electronic detection mechanism. Summary of the invention

[0004] In order to solve the technical problems in related fields, the present invention provides a shape distortion analysis system for power facilities, which can use a deep neural network that has performed multiple training reconstruction processes to complete the intelligent identification of whether the set hardware is distorted in shape based on multiple basic data, thereby using a high-precision electronic detection mechanism at the pixel level to complete the reliable identification of whether the set hardware is distorted in shape, thereby improving the reliability and effectiveness of the identification system.

[0005] According to the present invention, a system for analyzing shape distortion of electric power facilities is provided, the system comprising:

[0006] A visual monitor is used to perform a bird's-eye view operation on the set hardware of the power transmission line to obtain a bird's-eye view of the environment where the set hardware is located;

[0007] A front-end mapping device, connected to the visual monitor, is used to perform arithmetic mean filtering on the received on-site overhead picture to obtain and output a corresponding mean filtered image;

[0008] A mid-end mapping device, connected to the front-end mapping device, for performing gamma correction processing on the received mean-filtered image to obtain and output a corresponding corrected image;

[0009] A multi-layer filtering device, connected to the mid-end mapping device and comprising a first filtering unit, a second filtering unit and a third filtering unit, for sequentially performing bilateral filtering, wavelet filtering and homomorphic filtering on the received corrected processed image to obtain and output a corresponding multi-layer filtered image;

[0010] A distortion detection component, comprising a value extraction device, an information analysis device and a network reconstruction device, wherein the information analysis device is connected to the value extraction device and the network reconstruction device respectively, and the distortion detection component is used to detect multiple image values ​​of the setting hardware in the received multi-layer filtered image, and simultaneously perform reconstruction processing based on multiple trainings on the deep neural network, and simultaneously input the multiple image values ​​of the setting hardware in the received multi-layer filtered image and the multiple standard values ​​corresponding to the setting hardware into the deep neural network that has performed the reconstruction processing of multiple trainings, and run the deep neural network that has performed the reconstruction processing of multiple trainings to obtain a state identifier outputted by the setting hardware indicating whether the setting hardware is distorted in shape, the multiple image values ​​of the setting hardware in the received multi-layer filtered image are the coordinate values ​​of each edge pixel point of the image block of the setting hardware in the received multi-layer filtered image, and the multiple standard values ​​corresponding to the setting hardware are the pattern values ​​of the digital representation of the standard overhead pattern corresponding to the setting hardware;

[0011] an identification analysis component connected to the distortion detection component, and configured to send a first analysis signal when the received status identification indicates that the setting hardware is not distorted, and also configured to send a second analysis signal when the received status identification indicates that the setting hardware is distorted;

[0012] Among them, the distortion detection component is used to detect multiple image values ​​of the setting hardware in the received multi-layer filtered image, and at the same time, the reconstruction processing based on multiple trainings is performed on the deep neural network, including: in the reconstruction processing based on multiple trainings performed on the deep neural network, the number of trainings is positively correlated with the amount of data of the pattern value of the digital representation of the standard overhead pattern corresponding to the setting hardware.

[0013] It can be seen that the present invention has at least the following important invention points:

[0014] First invention point: introducing a picture quality optimization mechanism with a customized structural design including a front-end mapping device, a mid-end mapping device and a multi-layer filter device to sequentially perform arithmetic mean filtering, gamma correction, bilateral filtering, wavelet filtering and homomorphic filtering on the on-site overhead picture of the set hardware, so as to obtain and output a multi-layer filtered image with better picture quality, thereby ensuring the stability and effectiveness of the subsequent intelligent judgment results of hardware shape distortion;

[0015] The second invention point: a distortion detection component with a customized structure design including a numerical extraction device, an information analysis device and a network reconstruction device is used to provide reliable and effective hardware resources for intelligent judgment of whether the shape of the hardware is distorted;

[0016] The third inventive point: Detect multiple image values ​​of the setting hardware in the received multi-layer filtered image, and simultaneously perform reconstruction processing based on multiple trainings on the deep neural network, and simultaneously input the multiple image values ​​of the setting hardware in the received multi-layer filtered image and the multiple standard values ​​corresponding to the setting hardware into the deep neural network that has performed the reconstruction processing of multiple trainings, and run the deep neural network that has performed the reconstruction processing of multiple trainings to obtain its output status identification indicating whether the setting hardware is distorted in shape. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The embodiments of the present invention will be described below with reference to the accompanying drawings, wherein:

[0018] Figure 1 It is a structural block diagram of a power facility shape distortion analysis system according to Embodiment A of the present invention.

[0019] Figure 2 It is a structural block diagram of a power facility shape distortion analysis system according to Embodiment B of the present invention.

[0020] Figure 3 It is a structural block diagram of a power facility shape distortion analysis system according to Embodiment C of the present invention. DETAILED DESCRIPTION

[0021] An embodiment of the electric power facility shape distortion analysis system of the present invention will be described in detail below with reference to the accompanying drawings.

[0022] Figure 1 The structure block diagram of the power facility shape distortion analysis system according to the A embodiment of the present invention is shown, and the system includes:

[0023] A visual monitor is used to perform a bird's-eye view operation on the set hardware of the power transmission line to obtain a bird's-eye view of the environment where the set hardware is located;

[0024] A front-end mapping device, connected to the visual monitor, is used to perform arithmetic mean filtering on the received on-site overhead picture to obtain and output a corresponding mean filtered image;

[0025] A mid-end mapping device, connected to the front-end mapping device, for performing gamma correction processing on the received mean-filtered image to obtain and output a corresponding corrected image;

[0026] A multi-layer filtering device, connected to the mid-end mapping device and comprising a first filtering unit, a second filtering unit and a third filtering unit, for sequentially performing bilateral filtering, wavelet filtering and homomorphic filtering on the received corrected processed image to obtain and output a corresponding multi-layer filtered image;

[0027] Specifically, the multi-layer filtering device is connected to the mid-end mapping device and includes a first filtering unit, a second filtering unit and a third filtering unit, and is used to sequentially perform bilateral filtering processing, wavelet filtering processing and homomorphic filtering processing on the received correction processing image to obtain and output the corresponding multi-layer filtering image, including: different SOC chips can be selected to respectively implement the first filtering unit, the second filtering unit and the third filtering unit;

[0028] A distortion detection component, comprising a value extraction device, an information analysis device and a network reconstruction device, wherein the information analysis device is connected to the value extraction device and the network reconstruction device respectively, and the distortion detection component is used to detect multiple image values ​​of the setting hardware in the received multi-layer filtered image, and simultaneously perform reconstruction processing based on multiple trainings on the deep neural network, and simultaneously input the multiple image values ​​of the setting hardware in the received multi-layer filtered image and the multiple standard values ​​corresponding to the setting hardware into the deep neural network that has performed the reconstruction processing of multiple trainings, and run the deep neural network that has performed the reconstruction processing of multiple trainings to obtain a state identifier outputted by the setting hardware indicating whether the setting hardware is distorted in shape, the multiple image values ​​of the setting hardware in the received multi-layer filtered image are the coordinate values ​​of each edge pixel point of the image block of the setting hardware in the received multi-layer filtered image, and the multiple standard values ​​corresponding to the setting hardware are the pattern values ​​of the digital representation of the standard overhead pattern corresponding to the setting hardware;

[0029] an identification analysis component connected to the distortion detection component, and configured to send a first analysis signal when the received status identification indicates that the setting hardware is not distorted, and also configured to send a second analysis signal when the received status identification indicates that the setting hardware is distorted;

[0030] The distortion detection component is used to detect multiple image values ​​of the set hardware in the received multi-layer filtered image, and the reconstruction processing based on multiple trainings is performed on the deep neural network, including: in the reconstruction processing based on multiple trainings performed on the deep neural network, the number of trainings is positively correlated with the data volume of the pattern value of the digital representation of the standard overhead pattern corresponding to the set hardware;

[0031] The steps of simultaneously inputting multiple image values ​​of the setting hardware in the received multi-layer filtered image and multiple standard values ​​corresponding to the setting hardware into a deep neural network that has been reconstructed after multiple trainings, and running the deep neural network that has been reconstructed after multiple trainings to obtain a state identifier outputted by the deep neural network indicating whether the setting hardware is distorted in shape comprises: the state identifier outputted by the deep neural network indicating whether the setting hardware is distorted in shape is a binary value;

[0032] Among them, multiple image values ​​of the setting hardware in the received multi-layer filtered image and multiple standard values ​​corresponding to the setting hardware are simultaneously input into the deep neural network that has completed the reconstruction processing of multiple trainings, and the deep neural network that has completed the reconstruction processing of multiple trainings is run to obtain its output status identification indicating whether the setting hardware is distorted in shape, which also includes: the multiple image values ​​of the setting hardware in the received multi-layer filtered image and the multiple standard values ​​corresponding to the setting hardware are all in the form of binary values.

[0033] Figure 2 It is a structural block diagram of a power facility shape distortion analysis system according to Embodiment B of the present invention.

[0034] Compared to the A embodiment, the power facility shape distortion analysis system shown in the B embodiment of the present invention may also include:

[0035] a timing server, connected to the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, respectively, and used to provide the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit with respective timing services required by the respective filtering units;

[0036] The timing server is connected to the distortion detection component, the first filter unit, the second filter unit and the third filter unit respectively, and is used to provide the distortion detection component, the first filter unit, the second filter unit and the third filter unit with the timing services required by each of them, including: the timing server uses the same timing clock to provide the distortion detection component, the first filter unit, the second filter unit and the third filter unit with the timing services required by each of them;

[0037] Among them, the timing server uses the same timing clock to provide the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit with the timing services they need respectively, including: the timing server uses the same timing clock to provide the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit with reference clock signals of various changing waveforms of the same timing clock.

[0038] Figure 3 It is a structural block diagram of a power facility shape distortion analysis system according to Embodiment C of the present invention.

[0039] Compared to the A embodiment, the power facility shape distortion analysis system shown in the C embodiment of the present invention may further include:

[0040] a parallel data bus, connected to the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit respectively, and used to provide a parallel data communication link between any two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit;

[0041] The parallel data bus is connected to the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit respectively, and is used to provide a parallel data communication link between two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, including: providing a 64-bit parallel data communication link between two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit;

[0042] The parallel data bus is respectively connected to the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, and is used to provide a parallel data communication link between two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, including: providing a parallel data communication link between two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit as a 32-bit parallel data communication link;

[0043] The parallel data bus is connected to the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit respectively, and is used to provide a parallel data communication link between two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, including: providing a parallel data communication link between two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit as a 16-bit parallel data communication link;

[0044] And wherein, a parallel data bus is respectively connected to the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, and is used to provide a parallel data communication link between two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, and also includes: the distance between the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit is less than or equal to a set distance threshold.

[0045] In addition, in the electric power facility shape distortion analysis system, multiple image values ​​of the setting hardware in the received multi-layer filtered image and multiple standard values ​​corresponding to the setting hardware are simultaneously input into the deep neural network that has completed the reconstruction processing of multiple trainings, and the deep neural network that has completed the reconstruction processing of multiple trainings is run to obtain the output status identification indicating whether the setting hardware is distorted in shape, including: using MATLAB toolbox to realize the simulation and test of the data processing process of simultaneously inputting the multiple image values ​​of the setting hardware in the received multi-layer filtered image and multiple standard values ​​corresponding to the setting hardware into the deep neural network that has completed the reconstruction processing of multiple trainings, and running the deep neural network that has completed the reconstruction processing of multiple trainings to obtain the output status identification indicating whether the setting hardware is distorted in shape.

[0046] For example, the MATLAB toolbox is used to implement the simulation and testing of a data processing process in which multiple image values ​​of the setting hardware in the received multi-layer filtered image and multiple standard values ​​corresponding to the setting hardware are simultaneously input into a deep neural network that has completed multiple training reconstruction processes, and the deep neural network that has completed multiple training reconstruction processes is run to obtain a status identification of whether the setting hardware is distorted in shape, which is output by the deep neural network. The process includes: selecting and using a programmable logic device to implement the simulation and testing of a data processing process in which multiple image values ​​of the setting hardware in the received multi-layer filtered image and multiple standard values ​​corresponding to the setting hardware are simultaneously input into a deep neural network that has completed multiple training reconstruction processes, and the deep neural network that has completed multiple training reconstruction processes is run to obtain a status identification of whether the setting hardware is distorted in shape, which is output by the deep neural network.

[0047] The electric power facility shape distortion analysis system of the present invention solves the technical problem of the lack of a high-precision identification mechanism for whether the hardware is distorted in shape in the prior art. By introducing a deep neural network that has performed multiple training reconstruction processes, the system completes the intelligent identification of whether the hardware is distorted in shape based on multiple visual basic data, thereby improving the accuracy and intelligence level of the identification mechanism for the shape distortion of the hardware.

[0048] Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art will appreciate that various modifications and substitutions may be made thereto without departing from the spirit and scope of the invention as set forth in the appended claims.

Claims

1. A power facility shape distortion analysis system, characterized in that: The system comprises: A visual monitor is used to perform a bird's-eye view operation on the set hardware of the power transmission line to obtain a bird's-eye view of the environment where the set hardware is located; A front-end mapping device, connected to the visual monitor, is used to perform arithmetic mean filtering on the received on-site overhead picture to obtain and output a corresponding mean filtered image; A mid-end mapping device, connected to the front-end mapping device, for performing gamma correction processing on the received mean-filtered image to obtain and output a corresponding corrected image; A multi-layer filtering device, connected to the mid-end mapping device and comprising a first filtering unit, a second filtering unit and a third filtering unit, for sequentially performing bilateral filtering, wavelet filtering and homomorphic filtering on the received corrected processed image to obtain and output a corresponding multi-layer filtered image; A distortion detection component, including a value extraction device, an information analysis device and a network reconstruction device, wherein the information analysis device is connected to the value extraction device and the network reconstruction device respectively, and the distortion detection component is used to detect multiple image values ​​of the setting hardware in the multi-layer filtered image, and simultaneously perform reconstruction processing based on multiple trainings on the deep neural network, and simultaneously input the multiple image values ​​of the setting hardware in the multi-layer filtered image and the multiple standard values ​​corresponding to the setting hardware into the deep neural network that has performed the reconstruction processing of multiple trainings, and run the deep neural network that has performed the reconstruction processing of multiple trainings to obtain a state identifier outputted by the setting hardware indicating whether the setting hardware is distorted in shape, the multiple image values ​​of the setting hardware in the multi-layer filtered image are the coordinate values ​​of each edge pixel point of the image block of the setting hardware in the multi-layer filtered image, and the multiple standard values ​​corresponding to the setting hardware are the pattern values ​​of the digital representation of the standard overhead pattern corresponding to the setting hardware; an identification analysis component connected to the distortion detection component, and configured to send a first analysis signal when the received status identification indicates that the setting hardware is not distorted, and also configured to send a second analysis signal when the received status identification indicates that the setting hardware is distorted; Among them, the distortion detection component is used to detect multiple image values ​​of the set hardware in the multi-layer filtered image, and at the same time, the reconstruction processing based on multiple trainings is performed on the deep neural network, including: in the reconstruction processing based on multiple trainings performed on the deep neural network, the number of trainings is positively correlated with the amount of data of the pattern value of the digital representation of the standard overhead pattern corresponding to the set hardware.

2. The power facility shape distortion analysis system according to claim 1, characterized in that: Inputting multiple image values ​​of the set hardware in the multi-layer filtered image and multiple standard values ​​corresponding to the set hardware into the deep neural network that has been reconstructed after multiple trainings at the same time, and running the deep neural network that has been reconstructed after multiple trainings to obtain the state identifier outputted by the deep neural network indicating whether the set hardware is distorted in shape includes: the state identifier outputted by the deep neural network indicating whether the set hardware is distorted in shape is a binary value; Among them, multiple image values ​​of the setting hardware in the multi-layer filtered image and multiple standard values ​​corresponding to the setting hardware are simultaneously input into the deep neural network that has completed the reconstruction processing of multiple trainings, and the deep neural network that has completed the reconstruction processing of multiple trainings is run to obtain its output status identification indicating whether the setting hardware is distorted in shape, which also includes: the multiple image values ​​of the setting hardware in the multi-layer filtered image and the multiple standard values ​​corresponding to the setting hardware are all in the form of binary values.

3. The electric power facility shape distortion analysis system according to claim 2, characterized in that: The system further comprises: a timing server, connected to the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, respectively, and used to provide the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit with respective timing services required by the respective filtering units; Among them, the timing server is respectively connected to the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, and is used to provide the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit with their respective required timing services, including: the timing server uses the same timing clock to provide the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit with their respective required timing services.

4. The electric power facility shape distortion analysis system according to claim 3, characterized in that: The timing server uses the same timing clock to provide the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit with the timing services they need, including: the timing server uses the same timing clock to provide the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit with reference clock signals of various changing waveforms of the same timing clock.

5. The electric power facility shape distortion analysis system according to claim 3, characterized in that: The system further comprises: A parallel data bus is connected to the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit respectively, and is used to provide a parallel data communication link between any two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit.

6. The electric power facility shape distortion analysis system according to claim 5, characterized in that: A parallel data bus is respectively connected to the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, and is used to provide a parallel data communication link between two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, including: providing a parallel data communication link between two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit is a 64-bit parallel data communication link.

7. The electric power facility shape distortion analysis system according to claim 5, characterized in that: A parallel data bus is respectively connected to the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, and is used to provide a parallel data communication link between two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, including: providing a parallel data communication link between two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit is a 32-bit parallel data communication link.

8. The electric power facility shape distortion analysis system according to claim 5, characterized in that: A parallel data bus is respectively connected to the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, and is used to provide a parallel data communication link between two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, including: providing a parallel data communication link between two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit is a 16-bit parallel data communication link.

9. The electric power facility shape distortion analysis system according to claim 5, characterized in that: A parallel data bus is respectively connected to the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit, and is used to provide a parallel data communication link between two devices of the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit. It also includes: the distance between the distortion detection component, the first filtering unit, the second filtering unit and the third filtering unit is less than or equal to a set distance threshold.

Images