CN113036906A

CN113036906A - Automatic inspection robot of power dispatching system

Info

Publication number: CN113036906A
Application number: CN202011638418.7A
Authority: CN
Inventors: 张海永; 高承贵
Original assignee: Nanjing Taisi De Intelligent Electric Co ltd
Current assignee: Nanjing Taisi De Intelligent Electric Co ltd
Priority date: 2020-12-31
Filing date: 2020-12-31
Publication date: 2021-06-25
Anticipated expiration: 2040-12-31
Also published as: CN113036906B

Abstract

The invention discloses an automatic inspection robot of a power dispatching system, which comprises an inspection robot host, a switch and a workstation, wherein the inspection robot host is used for setting inspection test parameters, issuing an inspection command to an inspection client of the workstation, and receiving and displaying an inspection test result; setting polling test parameters including the name of the transformer substation to be polled, a signal set to be polled, a signal threshold and a polling period, and judging whether signals in the signal set to be polled of the transformer substation to be polled are abnormal or not and polling interval time of the polling period through the signal threshold; the switch is used for routing inspection robot host and workstation networking for data transmission; the workstation is used for deploying the power dispatching system and the inspection client, receiving and executing the inspection command, and judging and returning an inspection result. The invention can reduce the manual inspection burden of the staff of the power dispatching master station, improve the inspection work efficiency and improve the accuracy of point and picture association.

Description

Automatic inspection robot of power dispatching system

Technical Field

The invention belongs to the technical field of automatic inspection of power dispatching systems, and particularly relates to an automatic inspection robot of a power dispatching system.

Background

With the continuous popularization of unattended substations, whether each stage of scheduling center can accurately monitor the governed substation in real time is a precondition for safe and stable operation of a power grid, and remote transmission area point table information of a scheduling master station and a plant station end needs to be checked to verify the instantaneity of signals and the consistency of data, so that the scheduling center can be ensured to correctly monitor the substation.

And the transformer substation signal inspection of the dispatching side of the power dispatching system basically adopts a manual verification mode. In the method, a dispatching master station worker contacts a substation end by telephone and manually corrects a real-time alarm window and each wiring sub-diagram picture information of a dispatching master station system. The method brings great workload to dispatching workers, is low in efficiency, is easy to generate careless omission, and brings hidden danger to the safety of a power grid.

Therefore, a more intelligent, more efficient and safer dispatching side inspection robot needs to be researched, the burden of workers is reduced, the inspection work efficiency is improved, and the accuracy of point and picture association is improved.

Disclosure of Invention

The technical problem to be solved by the invention is as follows: the utility model provides an automatic robot of patrolling and examining of power dispatching system to solve the problem that exists among the prior art.

The technical scheme adopted by the invention is as follows: the utility model provides a robot is patrolled and examined in electric power dispatching system's automation, includes patrols and examines robot host computer, switch and workstation, wherein:

the inspection robot host is used for setting inspection test parameters, issuing an inspection command to an inspection client of the workstation, and receiving and displaying an inspection test result;

the method comprises the steps that a polling test parameter is set, wherein the polling test parameter comprises a name of a to-be-polled substation, a signal set to be polled, a signal threshold and a polling period, the name of the polling substation represents a target substation to be polled, the signal set to be polled represents a signal set which is set and selected by a user and needs to be polled under the substation to be polled, the signal set comprises a remote measuring signal and a remote signaling signal, the signal threshold represents a threshold which is compared with an actual polling signal value in a polling process, and if the actual polling signal value is larger than the signal threshold, a polling robot judges that the signal is abnormal; if the actual inspection signal value is less than or equal to the signal threshold, the inspection robot judges that the signal is normal, and the inspection cycle represents the time interval from the next inspection after the inspection robot inspects all signals in the signal set to be inspected at one time; setting a polling period to facilitate automatic periodic polling;

the switch is used for routing inspection robot host and workstation networking for data transmission;

the workstation is provided with a power dispatching system and an inspection client, receives and executes an inspection command, judges and returns an inspection result.

The inspection robot host issues an inspection command to the robot client through TCP communication;

the method comprises the steps that a patrol client of a workstation receives and analyzes a patrol command issued by a patrol robot host, determines the name of a substation to be patrolled, a signal set to be patrolled, a signal threshold and a patrol period, when the patrol client of the workstation patrols, patrol signals in the signal set to be patrolled are sequentially patrolled until all the signals to be patrolled are patrolled, the signals to be patrolled are checked in an alarm window, a power image picture and a front multi-channel picture in a power dispatching system of the workstation, and patrol when the patrol client of the workstation patrols the signals comprises alarm window patrol, power image picture patrol and front multi-channel picture patrol.

The method for realizing the patrol of the alarm window of the patrol signal and the patrol of the preposed multi-channel picture by the patrol client of the workstation comprises the following steps: firstly, screenshot is carried out on an alarm window and a front multi-channel window, then remote signaling and remote measuring information displayed in the alarm window and the front multi-channel window are identified in a character identification mode and are compared with a threshold value of the routing inspection signal, if the identification value of the routing inspection signal exceeds a set signal threshold value, the routing inspection robot judges that the routing inspection signal is abnormal and records the routing inspection signal.

The method for realizing the patrol of the power graphic picture of the patrol signal by the patrol client of the workstation comprises the following steps: firstly, screenshot is carried out on the power graphical interface, the power element, the power analog quantity and the switching value information which are displayed in the power graphical interface are identified in an image identification mode, the power element, the power analog quantity and the switching value information are compared with a threshold value of the measuring point signal, if the identification value of the polling signal exceeds a set signal threshold value, the polling robot judges that the polling signal is abnormal, and records the abnormal polling signal.

The power graphic picture of the power dispatching system comprises wiring diagrams and signal diagrams of all substations in the dispatching system, and the inspection client of the workstation is used for inspecting the power graphic picture of an inspection signal: inquiring which power graphic picture the routing inspection signal belongs to according to the id name of the routing inspection signal, generating a cutting jump track entering the power graphic picture where the routing inspection signal is to be located from the current power graphic picture, inquiring the key _ id attribute of a metadata tag in an svg format file and the key _ id attribute of a GzpGzp, DText, Disconnect, GroundDisconnect and CBreach tag in an inquiring G format file according to known power graphic picture files (svg format and G format), finding the svg format file and the G format file which are the same as the key _ id in the signal parameter to be inspected, determining which power graphic picture the signal to be inspected belongs to, inquiring the xlink of an tag in the svg format file and the poke attribute of an a tag in the inquiring G format file by the routing inspection client of the workstation, acquiring the picture jump relation among the power graphic pictures, and generating the routing jump track entering the power graphic picture where the routing inspection signal is to be located from the shortest power graphic picture The cutting map of (2) jumps to the track.

The path planning method of the map cutting jumping track comprises the following steps:

step 1, analyzing all graphic picture files of a certain transformer substation to be tested to obtain information in all graphic pictures, and forming an XML file RobotGraphs.xml of custom content; XML file robotgraphics, XML contains Graph tag, dst sub tag, home _ page tag; the Graph tag is used for describing certain graphic picture information, the dst sub-tag is used for describing a name of a target picture which can be skipped and coordinates of a skip button which exist in the picture, and the home _ page tag is used for describing a graphic picture file as an initial home page;

step 2, analyzing the signal table and xls file of the check acceptance of the point to be detected to form an XML file RobotItems.xml with customized content, which is used for storing all signal information of the check acceptance of the point to be detected; XML file robotitems, wherein the XML file contains item tags and Graph sub-tags; the item tag is used for describing the point number and name of a certain signal to be detected, and the Graph sub-tag is used for describing the position of the signal in the Graph picture;

step 3, analyzing RobotGraphs.xml of the transformer substation to be tested, and establishing a relation between each picture and a skip target picture stored by a data structure DataA; analyzing a RobotItems.xml file of a to-be-measured transformer substation, and establishing a data structure DataB for storing basic information of each measuring point and pictures to which the measuring point belongs;

step 4, sequentially planning a Graph picture jump track according to a picture set (Graph _1, Graph _2, …, Graph _ n) to which a certain measuring point belongs, which is stored in the data structure DataB;

step 5, setting the distance unit of the path jumping from one graphic picture to another graphic picture to be 1 each time, exhaustively arranging all paths from the current picture to the jumping target picture Graph _1 based on the data structure DataA, and comparing and selecting the graphic picture with the shortest path from all paths to be used as the optimal path to the target picture Graph _ 1;

step 6, driving mouse keys to operate according to the optimal path in the step 5, and obtaining a power graphic picture which jumps to a power wiring graphic picture Graph _ 1;

step 7, based on the DataA, exhaustively analyzing and listing all paths from the current picture Graph _1 to the jump target picture Graph _ 2; comparing and selecting the shortest path from all paths to serve as the optimal path for reaching the target picture Graph _ 2;

and 8, by analogy, finally calculating the shortest path of jumping to all the affiliated pictures (Graph _1, Graph _2, … and Graph _ n) of the measuring point from the current page in sequence.

The inspection client side of the workstation can automatically control a mouse and a keyboard, when an alarm window inspection of an inspection signal, an electric power image inspection and a front multi-channel image inspection are carried out, the alarm window, the electric power image and the front multi-channel image are automatically captured by pressing keys of a keyboard capture shortcut key, the electric power image is skipped to the position of the signal to be inspected by clicking a simulation mouse, the time interval of the automatic capturing of the inspection client side of the workstation is 0.3 second, and 5 images are sequentially captured respectively for the alarm window, the electric power image and the front multi-channel image, wherein 15 images are captured in total.

The inspection robot host receives and displays an inspection result fed back by the inspection client of the workstation, and judges abnormal signals in the inspection process of the inspection client of the workstation, the inspection client of the workstation also feeds back an alarm window, an electric power image picture and a front multichannel picture original screenshot of the abnormal inspection signals to the inspection robot host, the inspection robot host can check the original screenshot of the abnormal inspection signals and provide manual inspection judgment for workers, the inspection robot host can lead out the inspection result into an inspection report in an Excel format, different substations use different Sheet tables, and information in one substation Sheet table comprises the name of the inspected signals, the occurrence time, the signal value and the inspection judgment result (abnormal or normal).

The warning window inspection method comprises the following steps:

step 1, screenshot is carried out on a real-time alarm window of a power dispatching system;

step 2, preprocessing the screenshot of the alarm window, cutting and extracting a window image of the next half of a window of a point-to-point signal refreshed in real time in the screenshot of the alarm window;

step 3, recognizing the lower half-bar window image of the alarm window by adopting text recognition of tesseract, and recognizing text information of each line of the alarm window, wherein the text information comprises signal confirmation state, signal generation time, transformer station name, signal action description and signal text color;

step 4, inquiring whether the image character recognition result of the lower half column of the alarm window has the information which is the same as the point signal description to be aligned: if the identification result has the same information as the point signal to be aligned, the point is aligned correctly by the signal; if the identification result is different from the signal to be aligned, the signal is aligned to the point in a failure mode;

the method for identifying the signal characters and the colors thereof in the alarm window comprises the following steps:

step 1), reading a desktop screenshot file applied by a power dispatching system, calling a cvtColor function in OpenCV to convert the image into a gray-scale image, setting a binarization threshold value as 250, and calling a threshold function of the OpenCV to convert the gray-scale image into a binary image;

step 2), setting the size of an etching kernel to be 2 x 2, and calling an error function of OpenCV to perform etching operation on the binaryzation processed picture;

step 3), carrying out contour searching operation on the picture generated after the corrosion operation in the step 2), reserving two contours with the largest area and the second largest area, calculating the smallest external rectangle of the two contours by using a bounding rectangle function of OpenCV, and obtaining the top left corner vertex (x, y) and the length and width of the two rectangles;

step 4), corresponding the position of the minimum external rectangle to the picture in the step 3), taking the picture in the range of the two rectangles as two upper and lower regions to be identified, taking the vertex y with a small value at the upper left corner as an upper partial region, and taking the rectangle with a large value y as a lower partial region;

step 5), respectively carrying out black pixel longitudinal projection on the determined binary images of the upper partial area and the lower partial area to obtain the number of black pixels in each row and form a longitudinal projection pixel number curve;

step 6), taking the trough position of the longitudinal projection pixel number curve graph as a character segmentation position;

step 7), performing character segmentation on the upper and lower areas according to the character segmentation positions;

step 8), calculating the position of a central line of the segmented character picture, setting a threshold range of white color, and taking the color on the central line which is not in the range of the white threshold as the color of the character of the line;

and 9), calling an identification interface of tesseract to identify the regional characters.

The preposed multi-channel interface patrol method comprises the following steps:

step 1, screenshot is carried out on a front multichannel interface of a power dispatching system;

step 2, preprocessing the preposed multichannel screenshot, cutting and extracting a right window image mainly displaying a point signal in the preposed multichannel interface screenshot;

step 3, segmenting the pre-multi-channel interface preprocessing image by adopting a table analysis algorithm, and identifying the segmented pre-multi-channel interface image by using tesseract character identification to identify character information of each line of signals in the pre-multi-channel interface image, wherein the character information comprises signal names, channel/point numbers and signal values;

step 4, inquiring whether information with the same point number as the point signal to be pointed exists in the preposed multi-channel interface image character recognition result, if the information with the same point number does not exist, comparing the point number of the point signal to be pointed with the maximum point number and the minimum point number in the preposed multi-channel interface image recognition result: A) if the minimum point number in the image recognition result is not more than the point number of the point signal to be pointed and not more than the maximum point number in the image recognition result, the point signal to be pointed is not received in the preposed multi-channel interface, the point pointing fails, and the point pointing work of the next point is started; B) if the point number of the signal to be aligned is less than the minimum point number in the image recognition result, the point number of the signal to be aligned is not displayed in the preposed multi-channel interface, the difference between the point number of the signal to be aligned and the minimum point number in the image recognition result needs to be calculated, the display window of the point table of the preposed multi-channel interface is rolled by the difference distance, the signal to be aligned is displayed in the preposed multi-channel interface, and the step 3 is skipped; C) if the point number of the signal to be pointed is larger than the maximum point number in the image recognition result, the point number of the signal to be pointed is not displayed in the preposed multi-channel interface, the difference between the point number of the signal to be pointed and the maximum point number in the image recognition result needs to be calculated, a point table display window of the preposed multi-channel interface is rolled by the difference distance, the signal to be pointed is displayed in the preposed multi-channel interface, and the step 3 is skipped;

step 5, checking the signal information which has the same point number as the signal to be aligned in the pre-multi-channel interface image character recognition result, and whether other information in the image result is the same as the signal to be aligned: if the identification result is the same information as the point signal to be aligned, the signal aligns correctly; if the identification result is different from the signal to be aligned, the signal is aligned to the point in a failure mode; the table analysis algorithm is: and identifying a table transverse line in the preprocessed pre-multichannel image by adopting opencv, namely, a transverse line of a segmentation record, identifying a table vertical line in the image, namely, a segmentation line of each column, and finding a cell where data in the image is located. And then identifying the text information of the cells in the image by utilizing pyteressact, wherein the text information comprises a signal name, a channel/point number and a signal value.

The automatic graphical picture inspection method for the power dispatching system comprises the following steps:

step 1, screenshot is carried out on a power graphic picture of a power dispatching system;

step 2, preprocessing the screenshot of the power graphic image, cutting a toolbar in the power diagram image, and amplifying the cut power graphic image to enable the coordinate position and the size of a power element displayed in the screenshot of the power graphic image to be the same as those of an original file svg or g of the power graphic image;

step 3, acquiring the coordinate position and the size of the signal to be aligned in the electric power graphic picture according to the original file svg or g file of the electric power graphic picture;

step 4, selecting a position and a size area of a signal to be aligned in the preprocessed electric power graphic picture, and identifying the picture area: the to-be-paired signals comprise remote measuring signals and remote signaling signals, and for the remote measuring signals, tesseract is adopted to identify the text images of the to-be-paired signals and obtain the size of the remote measuring value of the to-be-paired signals; for a remote signaling signal, identifying an image of a signal element to be paired by adopting an image identification method, and acquiring a remote signaling value of the pairing signal;

step 5, checking whether the electric power pattern point signal image identification result is the same as the point signal value: if the identification result is the same as the value of the signal to be aligned, the signal is aligned correctly; if the identification result is different from the value of the signal to be aligned, the signal is failed to align the point.

The image recognition method for the remote signaling signal primitive is used for recognizing the color on the middle point, the vertical center line or the specific area of the primitive to judge the state of the remote signaling signal primitive, and the specific recognition steps are as follows:

for the graphic elements of the circuit breaker in the remote signaling signal, a method for identifying the color of the central point is adopted; if the central point of the circuit breaker primitive is identified as black, judging that the state of the circuit breaker primitive is 'fen'; if the central point of the breaker primitive is identified to be red, judging that the state of the breaker primitive is 'sum';

for the graphic elements of the optical word plate in the remote signaling signal, a method for identifying colors at two ends of a vertical center line is adopted; the vertical central line takes the color of dots from top to bottom, the color of dots from bottom to top, if yellow is taken twice, the state of the light character board primitive is judged to be 'fen'; if the color of the light word board is red in the two times, the state of the primitive of the light word board is judged to be 'combined'; if the colors acquired in the two times are not consistent or the color acquired in the one time is not acquired, reporting that the identification is abnormal;

for the graphic elements of the knife switch in the remote signaling signal, a method for identifying the colors of specific square frame areas in the graphic elements is adopted; if the color is identified in the specific square frame area, judging the state of the knife switch primitive to be 'minute'; if no color, namely black, is recognized in the specific square frame area, the state of the knife switch primitive is judged to be 'closed'.

The invention has the beneficial effects that: compared with the prior art, the invention has the following effects:

1) the invention integrates various intelligent technologies such as character recognition, image recognition technology, keyboard and mouse simulation technology and the like to realize automatic signal routing inspection and picture verification of the power dispatching system, thereby saving labor consumption, greatly improving working efficiency, reducing cost, avoiding the problem that manual point alignment is easy to make mistakes, setting routing inspection period and being capable of automatic periodic routing inspection;

2) the inspection robot can control any workstation computer, and has no interface interaction such as communication and the like with a dispatching system, so that hidden dangers such as network safety and the like do not exist.

Drawings

FIG. 1 is a diagram of an automatic inspection robot system;

FIG. 2 is a flow chart of inspection;

FIG. 3 is a diagram of a table file xls to be measured;

FIG. 4 is a screenshot of the setup of an alarm window configuration description file WCD in the method;

FIG. 5 is a screenshot of an alarm window;

FIG. 6 is a schematic diagram of an automatic point-to-point acceptance flow of a front-mounted multi-channel interface of the power dispatching system;

FIG. 7 is a screenshot of a table file xls of points to be measured;

FIG. 8 is a front-end window parameter configuration screenshot;

FIG. 9 is a front-facing multi-channel interface screenshot;

FIG. 10 is a schematic diagram of an automatic power graphical image inspection process;

FIG. 11 is an xls screenshot of a point table file to be measured for automatic inspection of a power graphic picture;

FIG. 12 is a power graphics screen svg file screenshot;

FIG. 13 is a power graphics screen shot;

FIG. 14 is a flow chart of a path planning for power dispatching system graphic screen switching;

FIG. 15 is a screenshot of a RobotGraphs xml file

Xml file screenshot in robottems.

Detailed Description

The invention is further described below with reference to specific figures and embodiments.

Example 1: as shown in fig. 1, an automatic inspection robot of a power dispatching system comprises an inspection robot host, a switch and a workstation, wherein:

the method comprises the steps that a polling test parameter is set, wherein the polling test parameter comprises a name of a to-be-polled substation, a signal set to be polled, a signal threshold and a polling period, the name of the polling substation represents a target substation to be polled, the signal set to be polled represents a signal set which is set and selected by a user and needs to be polled under the substation to be polled, the signal set comprises a remote measuring signal and a remote signaling signal, the signal threshold represents a threshold which is compared with an actual polling signal value in a polling process, and if the actual polling signal value is larger than the signal threshold, a polling robot judges that the signal is abnormal; if the actual inspection signal value is less than or equal to the signal threshold, the inspection robot judges that the signal is normal, and the inspection cycle represents the time interval from the next inspection after the inspection robot inspects all signals in the signal set to be inspected at one time;

the workstation is used for deploying the power dispatching system and the inspection client, receiving and executing the inspection command, and judging and returning an inspection result.

And the polling robot host issues a polling command to the robot client background program through TCP communication.

The power graphic picture of the power dispatching system comprises wiring diagrams and signal diagrams of all substations in the dispatching system, and the inspection client of the workstation is used for inspecting the power graphic picture of an inspection signal: inquiring which power graphic picture the routing inspection signal belongs to according to the id name of the routing inspection signal, generating a cutting skip track (realized by adopting a planning method of a cutting path) entering the power graphic picture where the signal to be inspected belongs from the current power graphic picture, inquiring the key _ id attribute of a metadata tag in an svg format file and the key _ id attribute of a GzpGzp, DText, Disconnectitor, GroundDisconnectitor and CBreaker tag in the inquiring G format file according to the known power graphic picture files (svg format and G format), finding the svg format file and the G format file which are the same as the key _ id in the signal parameter to be inspected, determining which power graphic picture the signal to be inspected belongs to, inquiring the xlink attribute of an tag in the svg format file by the routing inspection client of the workstation, and acquiring the skip relation between the power graphic ke attributes, and generating a cutting skip track entering the power graphic picture where the signal to be inspected is located from the current power graphic picture according to the principle of the shortest skip path of the power graphic picture.

Example 2: as shown in fig. 2-5, an automatic inspection method for an alarm window of a power dispatching system includes the following steps:

step 2, preprocessing the screenshot of the alarm window, cutting and extracting a window image of the next half of a bar of the point-to-point signal in the screenshot of the alarm window in real time, only cutting the screenshot, removing contents which do not need to be identified, conveniently identifying the signal, having less identification contents and high identification speed;

step 3, recognizing the lower half-bar window image of the alarm window by adopting text recognition of tesseract, and recognizing text information of each line (each signal) of the alarm window, wherein the text information comprises a signal confirmation state, signal generation time, a transformer station name, a signal name, signal action description and signal text color (alarm level);

step 4, inquiring whether the image character recognition result of the lower half column of the alarm window has the information which is the same as the point signal description to be aligned: if the identification result has the same information as the point signal to be aligned, the point is aligned correctly by the signal; if the identification result is different from the signal to be aligned (the name of the transformer substation, the name of the signal and the description of the signal action are different), the signal is aligned to the point in failure.

When the method is used, firstly, the table file xls of the point to be measured is imported, and as shown in FIG. 2, the name of the transformer station, the name of the signal and the sequence of the point table of the point to be measured are obtained. And then configuring an alarm window configuration description file WCD, as shown in FIG. 3, wherein the action description of the alarm window signal is divided into a remote signaling value 0 description and a remote signaling value 1 description, the remote signaling value 0 description in the WCD file is set as 'brake-off' and 'reset', and the remote signaling value 1 description is set as 'brake-on' and 'action'. The signal alarm grades are divided into accidents, abnormity, notification, out-of-limit and deflection, and are related to the color of signal characters in the alarm window. The WCD file may have RGB color parameters set to indicate different signal alarm levels. Through the alarm window configuration description file, the WCD can obtain the action description and the character color of the alarm window signal. The invention refers to the existing manual demand for point, extracts the information to be checked in the point process, expounds and standardizes the information of which measuring points need to be checked in the alarm window; the method does not check the point signals in the alarm window, self-defines the alarm window configuration description file WCD, has universality, and can deal with the point-to-point work of the alarm window signals of different factory dispatching systems;

when the point checking and accepting test is carried out, point signals are sequentially sent out by a plant station end in a fixed period according to a table to be tested, compared with the method for sending the point signals by the plant station end, the automatic inspection method for the alarm window has 2-second delay, and the delay is used for ensuring that the actual telecontrol signal data sent by the transformer station end is correctly sent to the dispatching master station system. The automatic inspection method for the alarm window checks that the fixed time interval of the primary measuring point is 5s, and is the same as the sending interval of the terminal and the point signal of the plant station. The order of the point list checked by the automatic inspection method of the alarm window is the same as the order of the point list sent by the station terminal. The prior art means adopts a point-to-point tester device to test the point-to-point of a telecontrol device and a protection measurement and control device at the side of a transformer substation, and has the problems that: only the point-to-point signals of the substation side are checked, the point-to-point signals in the master station side scheduling system are not checked, and the point-to-point communication link is not complete. The method adopts identification of point-to-point signals in an alarm window, is carried out in a scheduling master station, firstly sends out the point-to-point signals in sequence on a transformer substation side, then sequentially checks the point-to-point signals in the alarm window in a scheduling system on the master station side, and the sent-out signals and the check signals are not related with each other but are synchronous in progress. The method has the advantages that point signals are sequentially sent out at fixed periods by a station end according to a point table to be measured, and the automatic inspection method of the alarm window sequentially carries out the setting of point checking and accepting of signals in the alarm window at the side of a dispatching main station according to the point table to be checked: and checking the integrity of a link of the point-to-point signal, wherein the link comprises a plant station side telecontrol device and a master station side scheduling system.

The screenshot of the alarm window in the step 1 is realized by simulating a keyboard and mouse, the screenshot of the alarm window is realized by automatically controlling the keyboard and mouse to operate, and the screenshot is stored in a regular name (point number, action description and screenshot time) so as to be convenient for later calling and checking, as shown in FIG. 4; and then preprocessing the screenshot of the alarm window, and clipping and extracting the window image of the next half of the window of the alarm window, which refreshes the point-to-point signal in real time. And then, recognizing the image of the lower half column of the alarm window by adopting text recognition of tesseract, and recognizing text information of each line (each signal) of the alarm window, wherein the text information comprises signal confirmation state, signal generation time, transformer station name, signal action description and signal text color (alarm level). The existing method does not check the point signals in the alarm window, and refers to a manual check mode. The screenshot is named regularly, so that the image file is called in a regular format when the image is identified conveniently, and when the point is wrong, the current screenshot file of the alarm window can be quickly searched according to the point number.

step 2), setting the size of a corrosion kernel to be 2 x 2, calling an error function of OpenCV to perform corrosion operation on the binaryzation processed picture, eliminating a small interference area in the picture, and improving the speed of contour searching and processing;

Step 10), setting a character misrecognition adjustment list, wherein the adjustment list is stored in a text file, and the format of each line is as follows:

wrong phrase and correct phrase

Step 11), replacing the wrong phrase with the correct phrase according to the content of the adjustment list, and improving the accuracy rate of character recognition;

and 12), combining the characters and the colors identified in the steps 8) and 9) into a list to be output, and completing the segmentation and processing of the front picture.

The method solves the problem how to extract the attention area of the user on the front interface, and realizes the automatic segmentation adjustment of the area through an image preprocessing technology, a thresholding technology, a connected domain detection technology and a binarization image projection technology;

the invention solves the problems of the character line segmentation and color identification method after extracting the attention area, ensures the accuracy of the character line segmentation by the longitudinal projection segmentation technology of the binary character, and realizes the accuracy of the color identification by filtering and reading the color threshold of the segmentation area.

The invention realizes automatic screen capture of the alarm window picture and automatic identification and judgment of the signal characters in the alarm window picture by a keyboard and mouse simulation technology and a character identification technology, and improves the point-to-point acceptance check efficiency and the intelligent operation level of the alarm window at the master station end.

The alarm window point-to-point method has the following advantages:

1) the invention adopts the character recognition technology to recognize the point-to-point signal information in the alarm window picture, thereby saving the manpower consumption of manpower for people, greatly improving the working efficiency, reducing the cost and avoiding the problem that the manual point-to-point mode is easy to make mistakes;

2) the invention adopts a key mouse simulation technology, can automatically control the key mouse to screen the alarm window picture, and saves the operation of manual screen-capturing;

3) the method self-defines the alarm window configuration description file WCD, has universality, and can deal with the point-to-point work of alarm window signals of scheduling systems of different manufacturers;

4) the inspection method program can be operated on a workstation computer of any scheduling system, does not have interface interaction such as communication and the like with the scheduling system, and does not have hidden dangers such as network safety and the like

Example 3: as shown in fig. 6 to 9, a method for automatically patrolling a front-end multi-channel interface of a power dispatching system includes the following steps:

step 3, segmenting the pre-multi-channel interface preprocessing image by adopting a table analysis algorithm, identifying the segmented pre-multi-channel interface image by using tesseract character identification, and identifying character information of each line of signals in the pre-multi-channel interface image, wherein the character information comprises signal names, channel/point numbers and signal values;

and 4, inquiring whether the information with the same point number as the point signal to be pointed exists in the preposed multi-channel interface image character recognition result. If the same point number information does not exist, comparing the point number of the point signal to be aligned with the maximum point number and the minimum point number in the pre-multichannel interface image recognition result: A) and if the minimum point number in the image recognition result is not more than the point number of the point signal to be point is not more than the maximum point number in the image recognition result, the point signal to be point is not received in the preposed multi-channel interface, the point is failed to be point, and the point aligning work of the next point is started. B) And (3) if the point number of the signal to be aligned is less than the minimum point number in the image recognition result, the point number of the signal to be aligned is not displayed in the preposed multi-channel interface, the difference between the point number of the signal to be aligned and the minimum point number in the image recognition result needs to be calculated, the display window of the point list of the preposed multi-channel interface is rolled by the difference distance, and the signal to be aligned is displayed in the preposed multi-channel interface and jumps to the step 3. C) If the point number of the signal to be pointed is larger than the maximum point number in the image recognition result, the point number of the signal to be pointed is not displayed in the preposed multi-channel interface, the difference between the point number of the signal to be pointed and the maximum point number in the image recognition result needs to be calculated, a point table display window of the preposed multi-channel interface is rolled by the difference distance, the signal to be pointed is displayed in the preposed multi-channel interface, and the step 3 is skipped; in the prior art, factory station side point alignment or manual main station side point alignment is adopted, and a technology for identifying a preposed multi-channel picture is not adopted. The method adopts character recognition technology to recognize the point-to-point signal information in the prepositive multi-channel picture. But the point-to-point signals in the front multichannel are displayed in a table form, a sliding bar needs to be scrolled, and the point-to-point signals are displayed in the current window. In addition, the point signal information of the point in the preposed multi-channel picture is displayed in a multi-moving channel (multiple columns of a corresponding table), and the point is correctly checked until the point is correctly checked;

step 5, checking whether the signal information with the same point number as the signal to be aligned in the pre-multi-channel interface image character recognition result and other information (signal name and signal value) in the image result are the same as the signal to be aligned: if the identification result is the same information as the point signal to be aligned, the signal aligns correctly; if the identification result is not the same as the point aligning signal, the point aligning is failed by the signal.

When the method is used, firstly, a file xls of the table of the point to be measured is imported, and as shown in FIG. 2, the signal name, the signal point number and the sequence of the table of the point to be measured are obtained. And then configuring a remote signaling value description of the preposed multi-channel interface and a maximum allowable value of a remote measuring value error, as shown in figure 3. The signal value of the preposed multi-channel signal is divided into a remote signaling value and a remote measuring value, wherein the remote signaling value is a text description, and the remote measuring value is a specific variable value. In the process of checking the signal values of the preposed multichannel signals, the descriptions of remote signaling as 'minute' and 'sum' need to be configured, and the maximum allowable value of the error percentage of the remote measurement values is set. Remote signaling, telemetry or remote pulse tab paging button coordinates are then configured. The prior art has the problems that: the preposed multi-channel interface can only singly display signal information of remote signaling, remote measuring or remote pulse type, and the solution is as follows: a remote signaling, remote measuring or remote pulse tab paging button coordinate module needs to be arranged, and in the automatic point-to-point checking and accepting process of the front multi-channel interface, a mouse is automatically controlled to click and switch a remote signaling, remote measuring or remote pulse page. The invention refers to the existing requirement of manual point alignment, extracts the information to be checked in the point alignment process, expounds and standardizes the information of which measuring points need to be checked in the preposed multi-channel interface; meanwhile, a remote signaling value is set as a character description, so that a remote signaling character recognition result is conveniently checked; and setting the maximum allowable value of the error percentage of the telemetering value, because the telemetering value in the preposed multichannel interface is a measured value and has an error with the set value of the measuring point.

When the point checking and accepting test is carried out, point signals are sequentially sent out at fixed periods by a plant station end according to a table to be tested, compared with the method for automatically patrolling the multichannel interface, the method for automatically patrolling the point signals by the plant station end has 2-second delay, and the delay has the function of ensuring that the actual telecontrol signal data sent by the transformer station end is correctly sent to the dispatching master station system. The fixed time interval of the primary measuring point is checked to be 5s by the automatic inspection method of the preposed multi-channel interface, and is the same as the sending interval of the end-to-end point signals of the plant station. The order of the point list checked by the automatic inspection method of the preposed multi-channel interface is the same as the order of the point list sent by the station terminal. The prior art means adopts a point-to-point tester device to test the point-to-point of a telecontrol device and a protection measurement and control device at the side of a transformer substation, and has the problems that: only the point-to-point signals of the substation side are checked, the point-to-point signals in the master station side scheduling system are not checked, and the point-to-point communication link is not complete. The method adopts the identification of the point-to-point signals in the preposed multi-channel interface, is carried out in the dispatching master station, firstly sends the point-to-point signals in sequence on the transformer substation side, then sequentially checks the point-to-point signals in the preposed multi-channel interface in the dispatching system of the master station side, and the sent signals and the check signals are not related with each other but are synchronous in progress. The method has the advantages that: and checking the integrity of a link of the point-to-point signal, wherein the link comprises a plant station side telecontrol device and a master station side scheduling system.

The signal information mainly checked by the front multichannel interface point-to-point check acceptance of the power dispatching system is a signal name, a signal point number and a signal value.

When the automatic inspection method for the alarm window starts to identify the point-to-point signals in the alarm window, the screenshot of the alarm window is realized by simulating a keyboard and mouse, and the screenshot of the alarm window is realized by automatically controlling the keyboard and mouse operation, as shown in fig. 4. And then preprocessing the front multichannel screenshot, and clipping and extracting a right window image mainly displaying a point signal in the front multichannel interface screenshot. And then, segmenting the pre-multi-channel interface preprocessing image by adopting a table analysis algorithm, and identifying the segmented pre-multi-channel interface image by using tesseract character identification to identify character information of each line of signals in the pre-multi-channel interface image, wherein the character information comprises signal names, channel/point numbers and signal values. And inquiring whether the information with the same point number as the point signal to be pointed exists in the pre-multi-channel interface image character recognition result. If the same point number information does not exist, comparing the point number of the point signal to be aligned with the maximum point number and the minimum point number in the pre-multichannel interface image recognition result: A) and if the minimum point number in the image recognition result is not more than the point number of the point signal to be point is not more than the maximum point number in the image recognition result, the point signal to be point is not received in the preposed multi-channel interface, the point is failed to be point, and the point aligning work of the next point is started. B) And if the point number of the signal to be pointed is less than the minimum point number in the image recognition result, the point number of the signal to be pointed is not displayed in the preposed multi-channel interface, the difference between the point number of the signal to be pointed and the minimum point number in the image recognition result needs to be calculated, the display window of the preposed multi-channel interface point table is rolled by the difference distance, the signal to be pointed is displayed in the preposed multi-channel interface, and the screenshot recognition of the preposed multi-channel interface is carried out again. C) And if the point number of the signal to be pointed is larger than the maximum point number in the image recognition result, the point number of the signal to be pointed is not displayed in the preposed multi-channel interface, the difference between the point number of the signal to be pointed and the maximum point number in the image recognition result needs to be calculated, the display window of the point table of the preposed multi-channel interface is rolled by the difference distance, the signal to be pointed is displayed in the preposed multi-channel interface, and the screenshot recognition of the preposed multi-channel interface is carried out again. The existing method does not check point signals in the preposed multi-channel interface, and refers to a manual checking mode. The screenshot is named regularly, so that the image file is called in a regular format when the image is identified conveniently, and when the point is wrong, the current preposed multi-channel interface screenshot file can be quickly searched according to the point number.

And finally, checking whether signal information with the same point number as the point signal to be aligned in each channel character recognition result of the preposed multi-channel interface image and other information (signal name and signal value) in the image result are the same as the point signal to be aligned or not: if the identification result of each channel is the same as the information of the signal to be aligned, the signal is aligned correctly; if the identification result of any channel is not the same as the signal to be aligned, the signal is aligned to the point in failure.

The table analysis algorithm is: and identifying a table transverse line in the preprocessed pre-multichannel image by adopting opencv, namely a transverse line of a segmentation record (each row), identifying a table vertical line in the image, namely a segmentation line of each column, and finding a cell where data in the image is located. And then identifying the text information of the cells in the image by utilizing pyteressact, wherein the text information comprises a signal name, a channel/point number and a signal value.

The preposed multi-channel interface point-to-point method has the following advantages:

1) the invention adopts the character recognition technology to recognize the point-to-point signal information in the preposed multiple channels, thereby saving the manpower consumption of manpower for people, greatly improving the working efficiency, reducing the cost and avoiding the problem that the manual point-to-point mode is easy to make mistakes.

2) According to the invention, the table transverse lines in the front multichannel interface are firstly identified, and then the character information of the cells in the image is identified, so that the identification speed of the signal information in the front multichannel interface is increased, and the identification efficiency is improved;

3) the automatic point-aligning method of the preposed multi-channel interface is synchronously performed with the end-aligning signal of a factory station, and a keyboard and mouse simulation technology is adopted, so that the keyboard and mouse can be automatically controlled to perform screenshot of the preposed multi-channel interface and switching of remote signaling, remote measuring and remote pulse paging, the whole-process point-aligning check and acceptance work is automatically performed, and the labor is greatly saved;

4) the invention discloses a patrol method of a dispatching master station end, which verifies the correctness of the whole signal channel from a transformer substation protection device signal to a telemechanical and then to a dispatching master station. The method only obtains the screenshot of the front-end multi-channel interface, does not have interface interaction such as communication and the like with a dispatching system, and does not have hidden dangers such as data network safety and the like.

Example 4: as shown in fig. 10-13, an automatic power graphic picture inspection method includes the following steps:

and 4, selecting a position and a size area of the signal to be aligned in the preprocessed power graphic picture, and identifying the picture area. Dividing the signals to be aligned into remote measuring signals and remote signaling signals, adopting tesseract to identify the images of the signals to be aligned for the remote measuring signals, and obtaining the magnitude of signal remote measuring values in the electric power graph; for remote signaling signals, identifying a signal image to be aligned by adopting an electric power primitive identification method, and acquiring the state and remote measurement values (on or off) of signal elements in the electric power image; the remote measuring signal is a digital numerical value, and the remote signaling signal is a power element graph, so that different identification methods are adopted, and the identification speed is high;

step 5, checking whether the electric power pattern point signal image identification result to be aligned is the same as the point signal value to be aligned: if the identification result is the same as the value of the signal to be aligned, the signal is aligned correctly; if the identification result is different from the value of the signal to be aligned, the signal is failed to align the point.

When the method is used, firstly, the table file xls of the point to be measured is imported, and as shown in FIG. 2, the name of the transformer station, the name of the signal and the sequence of the point table of the point to be measured are obtained.

Then, analyzing an electric power graphic picture svg file or g file of the electric power dispatching system, as shown in fig. 3, and acquiring the position of the signal to be aligned in the electric power graphic picture, the specific steps are as follows:

step 1, inquiring key _ id attribute of metadata sub-label under g label in electric power graphic picture svg file format according to id name of point signal to be aligned, and finding out label with key _ id attribute value same as id name of point signal to be aligned. For the G-format power graphic picture file, inquiring key-id attributes of GzpGzp, DText, disconnect, Grounddisconnect and CBreach labels in the G-format power graphic picture file, and finding out a label with a key-id attribute value which is the same as the id name of the point signal to be aligned;

and 2, acquiring the coordinates of the point signals to be aligned in the electric power graphic picture. For the electric power graphic picture file in the svg file format, under the g label with the same key _ id attribute value, the values of x and y attributes of a text sub-label are searched, wherein x is the horizontal coordinate in the electric power graphic picture, and y is the vertical coordinate in the electric power graphic picture, so that the coordinate of the signal to be aligned in the electric power graphic picture is obtained. For a G file format power graphic picture file, searching values of x and y attributes under GzpGzp, DText, disconnect, Grounddisconnect and CBreach tags with the same keyid attribute value, wherein x is an abscissa in a power graphic picture and y is an ordinate in the power graphic picture, and thus obtaining the coordinate of a signal to be aligned in the power graphic picture;

and step 3, acquiring the size of the picture of the point signal to be aligned in the power graphic picture. For an electric power graphic picture file in an svg file format, when a point signal is a telemetering signal, under a g label with the same key _ id attribute value, finding values of a font-size attribute and a writing-mode attribute of a text sub-label, wherein the font-size is the size of a telemetering signal value character, and the writing-mode is the display digit of a number, so that the picture size of the telemetering signal character is obtained; when the point signal is a remote signaling signal, under the g label with the same key _ id attribute value, the values of the width attribute and the height attribute of the rect sub-label are searched, wherein the width is the width of the signal primitive, and the height is the height of the signal primitive, so that the picture size of the remote signaling signal primitive is obtained. For the G file format power graphic picture file, values of a w attribute and an h attribute are searched under GzpGzp, DText, disconnect, Grounddisconnect and CBreach tags with the same keyid attribute value, wherein w is the width of a signal primitive, and h is the height of the signal primitive, so that the picture size of the signal primitive is obtained.

And then, carrying out consistency check on the graph and picture association project, and finding out the measuring points with original manual association errors in the power graph and picture by checking the keyid attribute and the poke attribute of the measuring points in the SVG file and the XLS file of the point table, wherein the measuring point errors comprise button association errors, repetition, existence of the measuring points in multiple pictures and power interval of measuring point association errors.

The consistency checking method of the graphic picture association project comprises the following steps: and acquiring corresponding association between all point numbers of the transformer substation and the key _ id attribute of the svg power graphic picture file by checking the existing key id. And then finding a signal primitive of key _ id corresponding to the measuring point number in the svg power graphic picture file through the measuring point number in the XLS file of the power transformation station table according to the mapping relation. The svg power graphic picture correlation engineering consistency checking method comprises a picture jump analysis checking method, a measuring point correlation repeatability checking method, a single-page measuring point repeatability checking method and a measuring point attribution checking method, wherein,

the picture jump analysis and check method comprises the following steps: and (5) checking the xlink: href attribute (svg file picture jump path) in the svg power graphic picture file, and acquiring all repeated jump paths in the svg power graphic picture file. If the repeated jump path exists, reminding abnormality;

the measuring point correlation repeatability inspection method comprises the following steps: and checking the number of svg files where key _ id attribute values corresponding to the measuring point numbers are located. 1 remote signaling signal exists in a main wiring diagram (home page) svg file and 1 interval chart svg at most, and if the remote signaling signal exists in more than three svg files, abnormity needs to be reminded;

the method for checking the repeatability of the single page measuring point comprises the following steps: and checking the number of test points with repeated key _ id attribute values in a single svg power graphic picture file. If the key _ id attribute exists at more than 2 measuring points with repeated key _ id attributes, abnormity needs to be reminded;

the method for checking the attribution of the measuring point picture comprises the following steps: and checking and extracting the numerical codes and the English codes in all measuring point names in each svg file (for example, the measuring point name is ' main transformer 220kV second set merging unit GOOSE total alarm ' 1 ', and the numerical codes and the English codes are ' 1-220kV-GOOSE '), and comparing the numerical codes and the English codes in the names of the screen svg files (except the first page picture) to which the measuring points belong (for example, the name of the svg file is ' 220kV Yunlong new 1 main transformer 220kV interval wiring diagram.bay. svg ', and the numerical codes and the English codes are ' 220kV-1-220kV '). If the first 2 numbers (such as '1-220 kV in' 1-220kV-GOOSE ') of the numbers and the English codes in the name of the measuring point are the same as the last 2 numbers (such as' 1-220kV in '220 kV-1-220 kV') of the numbers and the English codes in the name of the svg file of the picture to which the measuring point belongs, judging that the measuring point belongs to the svg file of the picture; if not, judging that the measuring point does not belong to the image svg file, and reminding abnormality.

When the point checking and accepting test is carried out, point signals are sequentially sent out at fixed periods by a plant station end according to a table to be tested, compared with the method for automatically patrolling the power graphic picture, the method for automatically patrolling the power graphic picture has 2-second delay when the point signals are sent out by the plant station end, and the delay is used for ensuring that the actual telecontrol signal data sent by the transformer station end is correctly sent to the dispatching master station system. The fixed time interval of the primary measuring point is checked to be 5s by the automatic power graphic picture inspection method, and is the same as the sending interval of the end-to-end point signals of the plant station. The order of the point list checked by the automatic power graphic picture inspection method is the same as the order of the point list sent by the station terminal. The method is characterized in that point-to-point signals in an alarm window are identified and are transmitted in sequence at a dispatching master station, the point-to-point signals are sequentially transmitted at a transformer substation side, the point-to-point signals are sequentially checked in the alarm window in a dispatching system at the master station side, and the transmitted signals and the check signals are not related with each other but are synchronous in progress. The method has the advantages that: and checking the integrity of a link of the point-to-point signal, wherein the link comprises a plant station side telecontrol device and a master station side scheduling system.

The invention relates to a signal information mainly checked by checking and accepting a graph picture of a power dispatching system, which is a signal name, a signal point number and a signal value.

When the automatic inspection method for the electric power graphic picture starts to identify the point-to-point signals in the electric power graphic picture, the screenshot of the electric power graphic picture is realized by simulating a keyboard and mouse, and the keyboard and mouse are automatically controlled to operate to realize the screenshot of the electric power graphic picture, as shown in fig. 4. The state of the point remote signaling signal is divided into 0 and 1, and the electric power primitive pictures of the remote signaling signal in different states are different, including the difference of shape, color and flashing state. When the image of the electric power graphic primitive picture of the signal is identified, a plurality of graphic pictures need to be continuously intercepted, and the flicker state of the electric power graphic primitive of the point signal needs to be intercepted. The flicker time of the power graphics primitive in the power graphics picture is fixed, 2/3 of the time interval for continuously intercepting the power graphics picture is set as the graphics primitive flicker time interval, the number of the continuously intercepted power graphics picture is 5, and therefore the flicker state of the power graphics primitive of the point signal is obtained. The existing technology has the problems that the graphic elements of the remote signaling signals have flashing states, the information of the graphic elements of the remote signaling signals cannot be confirmed only by identifying one picture, and a plurality of pictures need to be intercepted to obtain the flashing states of the power graphic elements.

And then preprocessing the screenshot of the power graphic picture, cutting a toolbar in the power graphic picture, and amplifying the cut power graphic picture to ensure that the coordinate position and the size of a power element displayed in the screenshot of the power graphic picture are the same as those of an original file svg or g of the power graphic picture.

And then acquiring the coordinate position and the size of the signal to be aligned in the electric power graphic picture according to the original file svg or g file of the electric power graphic picture. And selecting a position and a size area of the point signal to be aligned in the preprocessed electric power graphic picture, and identifying the picture area. Dividing the signals to be aligned into remote measuring signals and remote signaling signals, adopting tesseract to identify the images of the signals to be aligned for the remote measuring signals, and obtaining the magnitude of signal remote measuring values in the electric power graph; for remote signaling signals, an electric power primitive recognition technology is adopted to recognize signal images to be aligned, and the states and remote measurement values (on or off) of signal elements in the electric power images are obtained.

And finally, checking whether the electric power pattern point signal image identification result to be aligned is the same as the point signal value to be aligned: if the identification result is the same as the value of the signal to be aligned, the signal is aligned correctly; if the identification result is different from the value of the signal to be aligned, the signal is failed to align the point.

The image recognition method for the remote signaling signal primitive is used for recognizing the color on the middle point, the vertical center line or the specific area of the primitive to judge the state of the remote signaling signal primitive. The specific identification logic is as follows:

and for the graphic elements of the circuit breaker in the remote signaling signal, a method for identifying the color of the central point is adopted. If the central point of the circuit breaker primitive is identified as black, judging that the state of the circuit breaker primitive is 'fen'; if the central point of the breaker primitive is identified to be red, judging that the state of the breaker primitive is 'sum';

and for the graphic elements of the optical word plate in the remote signaling signal, a method for identifying the colors of two ends of a vertical center line is adopted. The vertical central line takes the color of dots from top to bottom, the color of dots from bottom to top, if yellow is taken twice, the state of the light character board primitive is judged to be 'fen'; if the color of the light word board is red in the two times, the state of the primitive of the light word board is judged to be 'combined'; if the colors acquired in the two times are not consistent or the color acquired in the one time is not acquired, reporting that the identification is abnormal;

and for the graphic elements of the knife switch in the remote signaling signal, a method for identifying the colors of specific square area in the graphic elements is adopted. If the color (not black) is identified in the specific square frame area, judging the state of the knife switch primitive to be 'minute'; if the color, namely black, is not recognized in the specific square frame area, the state of the knife switch primitive is judged to be 'closed';

aiming at the characteristics of the shape and the color of the electric power primitive, the state of the signal is judged by identifying the color of a specific straight line and a specific area in the electric power primitive, the requirement on hardware is low, and the identification efficiency is high.

The automatic power graphic picture point alignment has the following advantages:

the invention adopts the character recognition and image recognition technology to recognize the point-aligning signal information in the electric power graphic picture, thereby saving the manpower consumption of manpower for people, greatly improving the working efficiency, reducing the cost and avoiding the problem that the manual point-aligning mode is easy to make mistakes; the remote measuring signal is a digital value, and the remote signaling signal is a power element graph, so that different identification methods are adopted, and the identification speed is high. Aiming at the characteristics of the shape and the color of the electric power primitive, the state of the signal is judged by identifying the color of a specific straight line and a specific area in the electric power primitive, the requirement on hardware is low, and the identification efficiency is high;

2) the invention adopts a key mouse simulation technology, can automatically control the key mouse to capture the electric power graphic picture, and saves the operation of manual capture;

3) by checking the consistency of the graphic picture association project, the invention finds out the existing error before the point alignment of the graphic picture, thereby improving the accuracy of the point alignment of the power graphic picture;

4) the inspection method program can be operated on any 1 workstation computer of the scheduling system, and has no interface interaction such as communication and the like with the scheduling system, and has no hidden danger such as network safety and the like.

Example 5: as shown in fig. 14 to 16, a path planning method for switching graphic screens of a power dispatching system includes the following steps:

step 1, analyzing all graphic picture files (including SVG format or G format) of a certain transformer substation to be tested, obtaining information (including measuring point information in the picture and a skip button in the picture) in all graphic pictures, and forming an XML file RobotGraphs (XML) of custom content, as shown in FIG. 2; XML file robotgraphics, XML contains Graph tag, dst sub tag, home _ page tag; the Graph tag is used for describing certain graphic picture information, the dst sub-tag is used for describing a name of a target picture which can be skipped and coordinates of a skip button which exist in the picture, and the home _ page tag is used for describing a graphic picture file as an initial home page;

step 2, analyzing the signal table and xls file of the check acceptance of the point to be detected to form an XML file RobotItems.xml with customized content, which is used for storing all signal information of the check acceptance of the point to be detected, as shown in FIG. 3; XML file robotitems, wherein the XML file contains item tags and Graph sub-tags; the item tag is used for describing the point number and name of a certain signal to be detected, and the Graph sub-tag is used for describing the position of the signal in the Graph picture;

The method realizes path planning of switching and jumping from the current picture to all pictures to which a certain measuring point belongs in sequence, and the picture switching speed is high and the efficiency is high by taking the shortest path as a principle.

The invention realizes automatic picture switching jumping without manual intervention by matching with automatic mouse button operation, ensures accurate picture switching browsing, solves the problem that no unified method exists in the switching of measuring point information pictures in a power dispatching system graphic picture, and realizes path planning of all pictures to which a certain measuring point is switched and jumped in sequence from a current picture; the path algorithm of the invention takes the shortest path as a principle, and the picture switching speed is high and the efficiency is high.

The above description is only an embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of changes or substitutions within the technical scope of the present invention, and therefore, the scope of the present invention should be determined by the scope of the claims.

Claims

1. The utility model provides a robot is patrolled and examined in electric power dispatching system's automation which characterized in that: including patrolling and examining robot host computer, switch and workstation, wherein:

2. The automatic inspection robot of the power dispatching system according to claim 1, wherein: the method comprises the steps that a patrol client of a workstation receives and analyzes a patrol command issued by a patrol robot host, determines the name of a substation to be patrolled, a signal set to be patrolled, a signal threshold and a patrol period, when the patrol client of the workstation patrols, patrol signals in the signal set to be patrolled are sequentially patrolled until all the signals to be patrolled are patrolled, the signals to be patrolled are checked in an alarm window, a power image picture and a front multi-channel picture in a power dispatching system of the workstation, and patrol when the patrol client of the workstation patrols the signals comprises alarm window patrol, power image picture patrol and front multi-channel picture patrol.

3. The automatic inspection robot of the power dispatching system according to claim 2, wherein: the method for realizing the patrol of the alarm window of the patrol signal and the patrol of the preposed multi-channel picture by the patrol client of the workstation comprises the following steps: firstly, screenshot is carried out on an alarm window and a front multi-channel window, then remote signaling and remote measuring information displayed in the alarm window and the front multi-channel window are identified by adopting a character identification mode and are compared with a threshold value threshold of the routing inspection signal, if the identification value of the routing inspection signal exceeds a set signal threshold value threshold, the routing inspection robot judges that the routing inspection signal is abnormal and records the routing inspection signal; the method for realizing the patrol of the power graphic picture of the patrol signal by the patrol client of the workstation comprises the following steps: firstly, screenshot is carried out on the power graphical interface, the power element, the power analog quantity and the switching value information which are displayed in the power graphical interface are identified in an image identification mode, the power element, the power analog quantity and the switching value information are compared with a threshold value of the measuring point signal, if the identification value of the polling signal exceeds a set signal threshold value, the polling robot judges that the polling signal is abnormal, and records the abnormal polling signal.

4. The automatic inspection robot of the power dispatching system according to claim 3, wherein: the power graphic picture of the power dispatching system comprises wiring diagrams and signal diagrams of all substations in the dispatching system, and the inspection client of the workstation is used for inspecting the power graphic picture of an inspection signal: inquiring which power pattern picture the polling signal belongs to according to the id name of the polling signal, generating a cutting skip track entering the power pattern picture where the signal to be polled is located from the current power pattern picture, according to the known power graphic picture file, the patrol client of the workstation inquires the key _ id attribute of the metadata tag in the svg format file and inquires the key id attributes of the GzpGzp, DText, Disconnectitor, GroudDisconnectitor and CBreaker tags in the G format file to find the svg format file and the G format file which are the same as the key _ id in the signal parameter to be patrolled, therefore, the power graphic picture to which the signal to be patrolled is belonged is determined, and the patrolling client of the workstation inquires the xlink of the a label in the svg format file: href attribute and poke attribute of a label in the query G format file, acquiring the jump relation between the power graphic pictures, and generating a cutting skip track entering the power graphic picture where the signal to be inspected is located from the current power graphic picture according to the principle of the shortest skip path of the power graphic picture.

5. The automatic inspection robot of the power dispatching system according to claim 4, wherein: the path planning method of the map cutting jumping track comprises the following steps:

6. The automatic point-aligning inspection robot of the power dispatching system according to claim 2, wherein: the warning window inspection method comprises the following steps:

7. The automatic point-aligning inspection robot of the power dispatching system according to claim 2, wherein: the preposed multi-channel interface patrol method comprises the following steps:

step 5, checking the signal information which has the same point number as the signal to be aligned in the pre-multi-channel interface image character recognition result, and whether other information in the image result is the same as the signal to be aligned: if the identification result is the same information as the point signal to be aligned, the signal aligns correctly; if the identification result is different from the signal to be aligned, the signal is aligned to the point in a failure mode; the table analysis algorithm is: and identifying a table transverse line in the preprocessed pre-multichannel image by adopting opencv, namely, a transverse line of a segmentation record, identifying a table vertical line in the image, namely, a segmentation line of each column, and finding a cell where data in the image is located.

8. And then identifying the text information of the cells in the image by utilizing pyteressact, wherein the text information comprises a signal name, a channel/point number and a signal value.

9. The automatic point-aligning inspection robot of the power dispatching system according to claim 2, wherein: the automatic graphical picture inspection method for the power dispatching system comprises the following steps:

10. The automatic point-to-point inspection robot of the power dispatching system according to claim 8, wherein: the image recognition method for the remote signaling signal primitive is used for recognizing the color on the middle point, the vertical center line or the specific area of the primitive to judge the state of the remote signaling signal primitive, and the specific recognition steps are as follows: