CN115597659B

CN115597659B - Intelligent safety management and control method for transformer substation

Info

Publication number: CN115597659B
Application number: CN202211154079.4A
Authority: CN
Inventors: 欧健; 孙磊; 刘道寰; 宋斌; 孙鹏; 乔剑; 王佳; 路宴鹏; 周倩; 栾薇
Original assignee: Beijing Zhongyi Xinyuan Technology Co ltd; Shandong Zhongyuan Information Technology Co ltd; Shandong Ruiyi Electric Power Engineering Co ltd
Current assignee: Beijing Zhongyi Xinyuan Technology Co ltd; Shandong Zhongyuan Information Technology Co ltd; Shandong Ruiyi Electric Power Engineering Co ltd
Priority date: 2022-09-21
Filing date: 2022-09-21
Publication date: 2023-04-14
Anticipated expiration: 2042-09-21
Also published as: CN115597659A

Abstract

The embodiment of the disclosure relates to a transformer substation intelligent safety management and control method, which comprises the following steps: detecting the physical environment of the transformer substation based on space mobile detection equipment and ground mapping equipment, and constructing a three-dimensional point cloud model with centimeter-level positioning accuracy; determining a dangerous area vector library in the three-dimensional point cloud model; acquiring motion information of each monitoring target according to monitoring data sent by the monitoring device and the auxiliary monitoring equipment on each monitoring target in a first time period; based on the comparison between the motion information of each monitored target and the data of the dangerous area vector library, the processing system calculates and determines whether the monitored target is in a dangerous area, and if not, whether the monitored target is in the dangerous area in a second time period is predicted; and if the mobile terminal is in the dangerous area or in the dangerous area within the second time period, sending out early warning information. The transformer substation applying the method can integrally improve the efficiency of the safety control of the field operation of the transformer substation, and realize intelligent, autonomous and integrated management.

Description

Intelligent safety management and control method for transformer substation

Technical Field

The application belongs to a transformer substation control technology, and particularly relates to an intelligent transformer substation safety control method.

Background

The prior patent application with publication number CN114494630A discloses a transformer substation infrastructure intelligent safety control method and system based on a precise positioning technology, and the prior art discloses a visualized three-dimensional model formed based on a three-dimensional basic model and an information database; dividing corresponding dangerous areas on the three-dimensional model according to the charged equipment in the information database, and generating a first electronic fence and a boundary early warning area around the dangerous areas; acquiring first position information of an operator and second position information of a construction instrument; mapping the first position information and the second position information to the three-dimensional model for displaying; and judging whether the first position information and the second position information are in the first electronic fence or the boundary early warning area, and if so, outputting corresponding warning information.

The prior art described above has the following technical drawbacks:

1) The existing three-dimensional model of the transformer substation is that a three-dimensional basic model of the transformer substation is built in three dimensions according to a transformer substation design drawing, and the probability that deviation occurs between the geographic space data of the three-dimensional basic model of the transformer substation and the geographic space data of an actual transformer substation exists, so that the positioning accuracy is influenced.

2) In the prior art, the first position information of an operator is acquired through a Beidou system and a 5G communication module on an operator safety helmet, the second position information of a construction instrument is acquired through a Beidou vehicle-mounted positioning terminal on the construction instrument, and the accuracy of the positioned position data cannot be used in a real transformer substation.

3) The operation and maintenance of present transformer substation overhauls the operation and has 2 kinds of modes: manual work and mechanical work. The prior patent only discloses a safety control method of a mechanical operation mode, the mentioned operator refers to an assistant person of the mechanical operation, and a more effective safety control method cannot be provided for a manual operation mode.

4) The method in the prior art cannot realize effective control without safety helmets or insulating protective clothing.

5) In the prior art, an analysis result is obtained for a vertical hoisting structure based on a visible light image recognition, analysis and operation technology, and the analysis result is influenced by hardware parameters such as pixels, light sensitivity and focal length of a camera device and software algorithm accuracy, so that the analysis result cannot reach a high-precision (centimeter-level) standard.

In view of this, the embodiment of the present invention provides an intelligent transformer substation safety management and control method that can implement high-precision monitoring in combination with the spatial position of an actual transformer substation.

Disclosure of Invention

Technical problem to be solved

In view of the above disadvantages and shortcomings in the prior art, the embodiment of the present invention provides an intelligent security control method for a substation.

(II) technical scheme

In order to achieve the purpose, the following technical scheme is adopted in the application:

in a first aspect, an embodiment of the present invention provides a transformer substation intelligent security management and control method, including:

s10, detecting the physical environment of the transformer substation based on space movable detection equipment and ground mapping equipment, and acquiring three-dimensional space laser point cloud data with centimeter-level positioning accuracy to construct a three-dimensional point cloud model with centimeter-level positioning accuracy;

s20, determining a dangerous area vector library in the three-dimensional point cloud model based on construction information, equipment information of a transformer substation, operation information and the three-dimensional point cloud model;

s30, acquiring motion information of each monitoring target, including attitude information, motion trail and spatial position data, according to monitoring data sent by the monitoring device and the auxiliary monitoring equipment on each monitoring target in a first time period;

the monitoring target includes: mobile construction equipment and operators; the monitoring device is fixed on a monitoring target body or a moving end of the monitoring target and uses an RTK positioning technology;

the auxiliary monitoring device includes: the system comprises monitoring equipment and an image acquisition device which are arranged on an unmanned aerial vehicle and use data of an RTK positioning technology;

s40, based on the comparison of the motion information of each monitored target and the data of the dangerous area vector library, a processing system of the computing equipment measures and calculates to determine whether the monitored target is in a dangerous area, and if the monitored target is not in the dangerous area, whether the monitored target is in the dangerous area in a second time period is predicted;

s50, if the mobile terminal is in a dangerous area or in a dangerous area within a second time period, sending out early warning information; the second time period is a future time period.

Optionally, the method further comprises:

s60, loading the motion information, the dangerous area vector library and the three-dimensional point cloud model on a Cesium three-dimensional map by combining a digital twin model to generate a visual dynamic three-dimensional scene; loading a dangerous area vector library by using a WebGL rendering technology, and realizing area boundary visualization;

alternatively, the first and second electrodes may be,

s60a, loading a dangerous area vector library and a three-dimensional point cloud model on a Cesum three-dimensional map by combining a digital twin model to generate a visual dynamic three-dimensional scene; loading a dangerous area vector library by using a WebGL rendering technology, and realizing area boundary visualization;

the contour/geometric model of the monitoring target is rendered on a dynamic three-dimensional scene.

Optionally, when the monitoring target is a mobile construction equipment, the S30 includes:

s31, at least one first RTK positioning device arranged on the movable construction equipment main body acquires first spatial position data and a first moving speed of the movable construction equipment main body in real time;

s32, at least two second RTK positioning devices arranged on the moving end of the mobile construction equipment acquire second spatial position data and second moving speed of the moving end of the mobile construction equipment in real time;

s33, calculating motion information and a first safe distance based on the first spatial position data, the first moving speed, the second spatial position data and the second moving speed, wherein the motion information comprises: the position information of the main body of the mobile construction equipment, the position information of the mobile end of the mobile construction equipment, the motion track and the motion rate; the first safety distance is a safety distance with the nearest dangerous area;

when the monitoring target is the operator, S30 includes:

s34, monitoring the area where the operator is located by using an RTK unmanned aerial vehicle carrying visible light camera equipment, infrared thermal imaging equipment and laser ranging equipment, and acquiring a first monitoring image of visible light, infrared imaging data and laser ranging data;

the visible light camera device, the infrared thermal imaging device and the laser ranging device are centimeter-level precision data or images acquired based on an RTK positioning technology; the visible light camera shooting equipment, the infrared thermal imaging equipment and the laser ranging equipment are all used for acquiring data on the ground vertically based on a 90-degree posture with the unmanned aerial vehicle body as a reference;

s35, measuring and calculating the monitoring image I, the infrared imaging data and the laser ranging data by adopting an AI intelligent identification model and a spatial position algorithm, identifying position information, a motion track and a motion rate of an operator, and identifying whether the operator belongs to an illegal operator who does not wear a safety helmet and an insulating protective clothing;

and judging whether the operator deviates from the operation track or not based on the operation track defined in advance.

Optionally, the S35 includes:

acquiring basic parameters of an unmanned aerial vehicle and related equipment on the unmanned aerial vehicle; the basic parameters include: the current ground height of the unmanned aerial vehicle, the pixels of a video/photo, the width and the length of an imaging frame, the focal length of a camera lens and the CMOS (complementary metal oxide semiconductor) parameters of the camera lens;

according to the basic parameters, the processing system acquires a first distance between the positive center point of the unmanned aerial vehicle and the edge of the visual field, calculates a second distance between the positive center point of the unmanned aerial vehicle and the four corners of the edge of the visual field by using the first distance and the width/2 of the imaging picture, and calculates the longitude and latitude coordinates of the actual positions of the four corners by using the longitude and latitude of the center point and the second distance;

identifying a human body target in the image by using an AI human body target detection model with the aid of user auxiliary information, obtaining pixel coordinates of the human body target after identification, and converting the human body pixel coordinates into RTK coordinates by using a GIS geographic information coordinate system conversion method according to the position information of the known current position/camera imaging central point and the RTK coordinates of the four corners of the boundary, thereby obtaining the actual position of the operator; the AI human body target detection model is one of AI intelligent recognition models;

alternatively, S35 includes:

acquiring centimeter-level longitude and latitude coordinates of the current position/camera imaging central point of the unmanned aerial vehicle and basic parameters of a camera through an unmanned aerial vehicle-mounted RTK sensor;

according to the basic parameters, calculating to obtain a first distance between the positive center point of the unmanned aerial vehicle and the edge of the visual field through a triangle pythagorean theorem, calculating to obtain a second distance between the positive center point of the unmanned aerial vehicle and the four corners of the edge of the visual field through the first distance and the width/2 of the imaging frame, and calculating to obtain the longitude and latitude coordinates of the actual positions of the four corners through the longitude and latitude of the center point and the second distance;

recognizing a human body target in the photo/video by using an AI human body target detection model by means of the auxiliary information of the user, and obtaining the pixel coordinate of the human body target after the human body target is recognized;

specifically, given the position information of the current position/camera imaging center point of the unmanned aerial vehicle and the boundary four-corner RTK coordinates, the pixel coordinates of the human body are converted into the RTK coordinates by using a GIS geographic information coordinate system conversion method, so that the actual position of the operator is obtained.

Optionally, S35 further includes:

the unmanned aerial vehicle is used for carrying a laser range finder to measure the height to the ground, the height difference between the laser range finder and a camera lens and the data of the left-right placing distance are corrected through a target position correction model to obtain the actual position of the operator, and centimeter-level precision spatial position information of the operator is finally obtained.

Optionally, the S35 further includes:

judging the area with the operator, and extracting the head subregion of the operator;

and (4) adopting a binary classification method for the extracted head subarea to judge whether the operator wears the safety helmet or not.

And finishing judging whether the safety helmet is worn or not by using a binary classification method.

Optionally, the S40 includes:

calculating whether the single element is in a polygon range of a dangerous area vector library based on the gis geographic information engine;

firstly, calculating the distance between a point and a plane where a polygon is located, and if the distance is greater than 0, not locating on the plane; then, reducing the three-dimensional plane into a two-dimensional plane, reducing the three-dimensional points into two-dimensional points, and judging whether the points are in the three-dimensional polygon according to a method for judging whether the points are in the polygon in the two-dimensional plane;

and (3) establishing a three-dimensional space model, wherein each point of a dangerous area vector library has own X and Y coordinates, cesium can calculate the distance according to the coordinates when the distance between the monitored target and the polygon is calculated through the existing coordinate information, and if the distance is less than 0, the digital twin model and the electronic fence boundary generate intersection, namely collision, so that the monitored target is in a dangerous area.

Optionally, the S60 includes:

based on the profile information of the mobile construction equipment pre-established in a construction equipment profile model library, a mobile construction equipment twin model of 1 is made by using 3 DMAX; rendering the twin model of the mobile construction equipment on a visual dynamic three-dimensional scene;

in the construction process, the first space position data, the first moving speed, the second space position data and the second moving speed are mapped to the three-dimensional scene, so that the twin model of the mobile construction equipment is dynamically changed and is consistent with the motion of the mobile construction equipment in actual operation.

Optionally, the S50 includes:

if one monitoring target is in a dangerous area or in a dangerous area within a second time period, pushing alarm information to the monitoring target or a related terminal controlling the monitoring target;

the alarm information already in the danger area and the alarm information in the danger area in the second time period are different in level.

In a second aspect, an embodiment of the present invention provides a computing device, which includes a memory and a processor, where the memory stores a computer program, and the processor executes the computer program stored in the memory and performs the steps of the transformer substation intelligent security management and control method according to any one of the above first aspects.

In a third aspect, an embodiment of the present invention provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the substation intelligent security management and control method according to any one of the first aspect above are implemented.

(III) advantageous effects

The technical scheme provided by the application can comprise the following beneficial effects:

in the method, RTK satellite positioning devices are configured on the monitoring unmanned aerial vehicle through sensor equipment, and the positioning accuracy of positioning data acquired by the RTK satellite positioning devices or result data measured and calculated by the data can reach centimeter level.

In the method, an RTK unmanned aerial vehicle is used for carrying a laser radar to acquire data; collecting field remote sensing data by using an RTK monitoring unmanned aerial vehicle; and acquiring the position data of the construction equipment by using an RTK positioning device. Thus, the data collected is in centimeters, and the results are also in centimeters.

The method can be effectively popularized to each transformer substation, the cost is low, the transformer substations applying the method can integrally improve the efficiency of field operation safety control of the transformer substations, intelligent, autonomous and integrated management is realized, the depth perception level of the situation of the power grid is improved, the existing field operation mode and the safety protection capability of operators/mechanical operators are greatly improved, and the operation efficiency is improved.

According to the safety management and control system, by integrating the technologies in multiple fields such as unmanned aerial vehicle laser radar, remote sensing, RTK positioning, AI artificial intelligence recognition, virtual reality and Internet of things, a software and hardware system with higher intelligence and autonomy is constructed, the intelligent safety management and control on manual and mechanical operation and maintenance operation sites of the transformer substation are effectively realized, and the risk identification capability and the site safety management and control level of the production site operation of the transformer substation are comprehensively improved.

And by applying a virtual reality technology, multidimensional visual presentation of a substation operation scene, real-time monitoring of operation conditions, and interconnection and intercommunication of operation site real-time data and digital twin mode data are realized.

Through the RTK spatial position monitoring devices of the operation site, the airborne remote sensing equipment of the aerial unmanned aerial vehicle is combined, the aerial data on the ground are combined, and therefore accurate spatial position data (with the accuracy reaching centimeter level) are obtained.

Ground fixed point real time monitoring combines unmanned aerial vehicle to carry out the totality or single object in the air to the transformer substation and monitors, realizes all-round accurate control, ensures that each developments of transformer substation is mastered completely. Meanwhile, night operation monitoring can be achieved through an infrared monitoring technology and an RTK accurate monitoring technology.

Based on the dangerous area vector library and real-time spatial position monitoring data, the system automatically measures and calculates the position of an object and the distance between the object and the early warning line electronic fence by using an AI artificial intelligence technology. When the safety control system monitors that an operator does unsafe behaviors or the operator collides the warning line electronic fence, the safety control early warning system can automatically determine corresponding terminal equipment according to the positioning data of the target, and sends early warning information to a handheld terminal (mobile phone) of the operator/construction equipment operator, so that multiple safety guarantee measures combining passive warning and active sensing are realized.

Drawings

The application is described with the aid of the following figures:

FIG. 1 is a schematic diagram of real-time acquisition of precise spatial location data of a job site by augmented reality technology in accordance with the present invention;

FIG. 2 is a schematic diagram of a transformer substation space point cloud data acquisition implementation scenario of the present invention;

FIG. 3 is a schematic diagram of the operation of the laser radar surveying and mapping technique of the unmanned aerial vehicle of the present invention;

FIG. 4 is a diagram of a danger area vector library according to an embodiment of the present invention;

FIGS. 5A, 5B and 6 are schematic views of the present invention showing the positioning of spatial position data of an operator at a work site;

FIG. 7 is a schematic diagram of the safety management and control system of the present invention;

fig. 8 is a schematic flow chart of a transformer substation intelligent safety control method according to an embodiment of the present invention.

Detailed Description

For a better understanding of the present invention, reference will now be made in detail to the present embodiments of the invention, which are illustrated in the accompanying drawings. It is to be understood that the following specific examples are illustrative of the invention only and are not to be construed as limiting the invention. In addition, it should be noted that, in the case of no conflict, the embodiments and features in the embodiments in the present application may be combined with each other; for convenience of description, only portions related to the present invention are shown in the drawings.

In the scene, the high-precision (centimeter-level) spatial position data of the hoisted object can be obtained in real time by an effective technical means, and the moving track of the hoisted object can be measured in real time, so that accurate early warning judgment can be made.

The prior art does not provide a technical means and a method for acquiring centimeter-level positioning precision data, and whether the monitoring target high-precision (centimeter-level) positioning data can be acquired in a high-risk operation environment of a transformer substation is a key technical requirement for realizing effective safety control.

In the embodiment of the invention, an RTK (Real time kinematic) carrier phase differential positioning technology is a Real-time kinematic positioning technology based on a carrier phase observed value, is a combined differential positioning system formed by a satellite measurement technology and a data transmission technology, a base station sends carrier observed quantity and station coordinate information to a mobile terminal in Real time through a data chain, and the mobile terminal receives the carrier phase of a satellite and the carrier phase from the base station to form a phase differential observed value for Real-time processing, so that a centimeter-level positioning result can be given in Real time.

The RTK unmanned aerial vehicle in the embodiment of the invention can be understood as an unmanned aerial vehicle provided with a data acquisition device using an RTK technology, so the RTK unmanned aerial vehicle is called as the RTK unmanned aerial vehicle for short.

As shown in fig. 1 to 8, the present embodiment provides a method for intelligently managing and controlling a substation, where an execution subject of the method of the present embodiment may be a computing device of a background server of the substation, and the method may include the following steps:

s10, detecting the physical environment of the transformer substation based on the space movable detection equipment and the ground mapping equipment, and obtaining three-dimensional space laser point cloud data with centimeter-level positioning accuracy so as to construct a three-dimensional point cloud model with centimeter-level positioning accuracy.

In this embodiment, a plurality of ground lidar devices, a mobile backpack lidar device (as shown in fig. 2), and an unmanned lidar device (as shown in fig. 3) are arranged at intervals in the physical environment of the substation; the method is used for obtaining centimeter-level three-dimensional space laser point cloud data of the transformer substation in a full-coverage physical environment.

Namely, the RTK unmanned aerial vehicle is used for carrying the laser radar and collecting centimeter-level three-dimensional space laser point cloud data by combining a ground laser radar surveying and mapping means, and a three-dimensional point cloud model is synthesized through three-dimensional software.

In this step, the data for constructing the three-dimensional point cloud module may be data of a centimeter level, and then a three-dimensional point cloud model of a centimeter level is synthesized by using the existing three-dimensional modeling technology.

The three-dimensional point cloud model in this step can be a digital twin three-dimensional model or a holographic three-dimensional model.

It should be noted that fig. 1 may be a schematic diagram illustrating an architecture of the entire substation, and the computing device in fig. 1 executes the method of the first embodiment.

And S20, determining a dangerous area vector library in the three-dimensional point cloud model, such as a limited area shown in figure 4, based on the construction information, the equipment information of the transformer substation, the operation information and the three-dimensional point cloud model.

For example, the definition and division can be made according to the actual situation of the change station. Before each operation, according to actual requirements such as working content, power failure equipment, safety control operation, construction requirements and construction content in the transformer substation, a live equipment area, a power failure equipment area, a safety operation area, a warning line and other dangerous area vector libraries with different levels can be divided in the three-dimensional point cloud model.

And S30, acquiring motion information including attitude information, motion trail and spatial position data of each monitoring target according to monitoring data sent by the monitoring device and the auxiliary monitoring equipment on each monitoring target in a first time period.

The monitoring target of the present embodiment includes: mobile construction equipment and operators; the monitoring device is fixed on a monitoring target body or a moving end of the monitoring target and uses an RTK positioning technology;

the auxiliary monitoring device comprises: the data monitoring device and the image acquisition device are arranged on the unmanned aerial vehicle and use an RTK positioning technology.

In a specific application, real-time data of an RTK positioning device in different time periods can be mapped into a 1. The real-time data in this embodiment may be millisecond-level updated data.

The motion information includes: spatial position data of the mobile construction equipment, which is data of centimeter level acquired by means of RTK technology.

S40, based on the comparison of the motion information of each monitored target and the data of the dangerous area vector library, a processing system of the computing equipment measures and calculates to determine whether the monitored target is in a dangerous area, and if the monitored target is not in the dangerous area, whether the monitored target is in the dangerous area in a second time period is predicted.

S50, if the mobile terminal is in a dangerous area or in a dangerous area within a second time period, sending early warning information; the second time period is a future time period.

In practical applications, in order to better visually display the content and facilitate visual management and real-time control, the motion information needs to be presented in a three-dimensional modeling manner, and therefore, the method of this embodiment further includes the following step S60 or step S60a.

S60, loading the motion information, the dangerous area vector library and the three-dimensional point cloud model on a Cesum three-dimensional map by combining a digital twin model to generate a visual dynamic three-dimensional scene; loading a dangerous area vector library by using a WebGL rendering technology, and realizing area boundary visualization;

or S60a, loading the dangerous area vector library and the three-dimensional point cloud model on a Cesum three-dimensional map by combining a digital twin model to generate a visual dynamic three-dimensional scene; and loading a dangerous area vector library by using a WebGL rendering technology, and realizing area boundary visualization.

In the step, technologies such as a digital twin model, detection or measurement data, a space collision algorithm and the like can be combined together to realize visualization and dynamic display of a digital twin three-dimensional scene, and real-time monitoring can be realized for each mobile construction device and operator. In this embodiment, the contour model/geometric model of the monitoring target is rendered on the dynamic three-dimensional scene.

In addition, centimeter-level positioning data are adopted in the embodiment of the invention, so that the power failure accident caused by collision with substation equipment/facilities can be effectively avoided when mechanical operation is carried out in a narrow space.

In a specific implementation process, when the monitoring target is a mobile construction device, the S30 includes:

and S32, at least two second RTK positioning devices arranged on the moving end of the mobile construction equipment acquire second spatial position data and second moving speed of the moving end of the mobile construction equipment in real time.

That is, each construction equipment is provided with a plurality of RTK positioning devices, and each RTK positioning device is installed in a different area, so that a plurality of spatial position data can be acquired, and spatial position data and moving rate of a body, a moving end, and the like of the mobile construction equipment can be further acquired according to profile information (information such as length, width, height, thickness, shape, and the like) of the construction equipment.

S33, calculating motion information and a first safe distance based on the first spatial position data, the first moving speed, the second spatial position data and the second moving speed, wherein the motion information comprises: the position information of the main body of the mobile construction equipment, the position information of the mobile end of the mobile construction equipment, the motion track and the motion rate; the first safety distance is the safety distance to the nearest hazardous area.

For example, determining the safe distance to the nearest polygon's danger zone may be accomplished according to a distance algorithm.

In practical applications, for example, it may be calculated based on the GIS geographic information engine whether a single element (pixel coordinates of the current position of each construction equipment) is within a polygon. Specifically, the distance between the point and the plane where the polygon is located may be calculated first, and if the distance is greater than 0, the point is not on the plane and is less likely to be inside the polygon; then, the three-dimensional plane can be reduced to a two-dimensional plane (a component is directly deleted), the three-dimensional point can be reduced to a two-dimensional point, and then whether the point is in the three-dimensional polygon or not can be judged according to the method for judging whether the point is in the polygon or not in the two-dimensional plane. The reason why the three-dimensional plane is reduced to the two-dimensional plane in this embodiment is that when a plane is projected to a certain coordinate plane, if the point before projection is within the polygon, the point after projection is still within the polygon. Namely, the distance calculation mode is realized on the premise that the plane A where the polygon is located cannot be perpendicular to the projected coordinate plane B; if a certain component in the normal vector is 0, the plane A where the polygon is located is perpendicular to a certain coordinate plane B, and if the polygon is projected to the coordinate plane B, the projection result is a line, it cannot be determined whether the point is in the polygon.

In this embodiment, a three-dimensional space model is established, after spatial reference, each point in the dangerous area vector library has its own X and Y coordinates, and when distance between an object and a polygon is calculated through existing coordinate information, the distance can be calculated according to the coordinates by using the cesum technique, and if the distance is less than 0, it indicates that an intersection (collision) is generated between the construction equipment and the boundary of the electronic fence.

When the monitoring target is the operator, S30 includes:

the visible light camera device, the infrared thermal imaging device and the laser ranging device are centimeter-level precision data or images acquired based on an RTK positioning technology; the visible light camera equipment, the infrared thermal imaging equipment and the laser ranging equipment are all used for acquiring data vertically to the ground based on a 90-degree posture with the unmanned aerial vehicle body as a reference;

s35, measuring and calculating the monitoring image I, the infrared imaging data and the laser ranging data by adopting an AI intelligent identification model and a spatial position algorithm, identifying position information, a motion track and a motion rate of an operator, and identifying whether the operator belongs to an illegal operator who does not wear a safety helmet or an insulating protective clothing;

For example, in one implementation, the step S35 can be implemented by:

according to the basic parameters, the processing system acquires a first distance between the center point of the unmanned aerial vehicle and the edge of the visual field, calculates a second distance between the center point of the unmanned aerial vehicle and the four corners of the edge of the visual field by using the first distance and the width/2 of the imaging frame, and calculates actual position longitude and latitude coordinates of the four corners by using the longitude and latitude of the center point and the second distance;

identifying a human body target in the image by using the AI human body target detection model with the help of the auxiliary information of the user, and obtaining the pixel coordinate of the human body target after identification; the AI human body target detection model is one of AI intelligent recognition models. In addition, with the help of the position information of the current position of the known unmanned aerial vehicle/camera imaging center point and the boundary four-corner RTK coordinates, the pixel coordinates of the human body are converted into the RTK coordinates by using a GIS geographic information coordinate system conversion method, so that the actual position of the operator is obtained, as shown in FIG. 6.

In another achievable manner, for better understanding, step S35 may also be embodied as follows:

the method comprises the following steps that visible light camera shooting equipment carried by an unmanned aerial vehicle vertically shoots the ground in a 90-degree posture, and therefore centimeter-level longitude and latitude coordinates of the current position of the unmanned aerial vehicle/a camera imaging central point and basic parameters of a camera (the current ground height of the unmanned aerial vehicle, pixels of a video/photo, the width and length of an imaging frame, the focal length of a camera lens and CMOS (complementary metal oxide semiconductor) parameters of the camera lens are obtained through an airborne RTK (real time kinematic) sensor of the unmanned aerial vehicle;

according to the basic parameters, calculating to obtain a first distance between the positive center point of the unmanned aerial vehicle and the edge of the visual field through a triangle pythagorean theorem, calculating to obtain a second distance between the positive center point of the unmanned aerial vehicle and the four corners of the edge of the visual field through the first distance and the width/2 of the imaging frame, and calculating to obtain the longitude and latitude coordinates of the actual positions of the four corners through the longitude and latitude of the center point and the second distance; as shown in fig. 5A.

specifically, given the position information of the current position/camera imaging center point of the unmanned aerial vehicle and the boundary four-corner RTK coordinates, the pixel coordinates of the human body are converted into the RTK coordinates by using a GIS geographic information coordinate system conversion method, as shown in fig. 5B, so as to obtain the actual position of the operator.

Of course, the determination process after the identification is as follows:

the unmanned aerial vehicle is used for carrying the laser range finder to measure the height to the ground, the height difference between the laser range finder and the camera lens and the data of the left placing distance and the right placing distance are corrected to the actual position of the operator through the target position correction model, and the centimeter-level precision spatial position information of the operator is finally obtained.

The purpose of the correction is mainly that a certain error exists in the laser range finder, the unmanned aerial vehicle and the camera under different environments, flight heights and monitoring angles, for example, the geometric position of an image pixel generated in the imaging process of the unmanned aerial vehicle is extruded, stretched, deviated, distorted and the like relative to a reference system (ground actual position), so that the geometric position, size, shape and orientation of a ground object in an image are changed, and the difference of data needs to be corrected and calculated through a geometric correction algorithm and a spatial positioning algorithm to obtain relatively accurate spatial position data.

The area where the operator exists is determined, and the head sub-area of the operator is extracted. For example, the face is marked, and then a head subregion is extracted; the YOLOv5 algorithm can be adopted to realize face detection and marking; and (4) adopting a binary classification method for the extracted head subarea to judge whether the operator wears the safety helmet or not. That is, the judgment of whether the helmet is worn is completed by a binary classification method.

In practical application, before the YOLOv5 algorithm is used, the YOLOv5 algorithm can be trained by means of a training data set labeled manually, so that the trained YOLOv5 algorithm is used for identifying the pixel coordinates of a human body target in a video stream or a picture, the position information of the current position/imaging center point of the unmanned aerial vehicle and the RTK coordinates of four corners of a boundary are known, and the pixel coordinates of the human body can be converted into the RTK coordinates by using a GIS geographic information coordinate system conversion method to obtain the position information of an operator.

Generally, a laser range finder, an unmanned aerial vehicle and a camera have certain errors in different environments, flight heights and monitoring angles, for example, the geometric position of an image pixel generated in the imaging process of the unmanned aerial vehicle is extruded, extended, deviated, distorted and the like relative to a reference system (ground actual position), so that the geometric position, size, shape and orientation of a ground object in an image are changed, and the difference of data needs to be corrected and calculated through a geometric correction and spatial positioning algorithm to obtain relatively accurate spatial position data. If the unmanned aerial vehicle is used for carrying the laser range finder to measure the height to the ground, the actual position of the operator is corrected through the target position correction model according to the height difference between the laser range finder and the camera lens, the left-right placement distance and other data, and finally centimeter-level precision space position information of the operator is obtained.

In the embodiment, a precise (centimeter-level) three-dimensional point cloud model of the transformer substation is established by using a high-precision RTK unmanned aerial vehicle laser radar modeling technology, and the three-dimensional effect dynamic display of the actual working environment of the transformer substation is realized by matching a digital twin model. According to the requirements of work content, power failure equipment, safety control operation and the like in the station, different levels of dangerous area vector libraries such as a charged equipment area, a power failure equipment area, a safety operation area, a warning line and the like are divided in the three-dimensional laser point cloud model.

By using an RTK positioning technology, high-precision RTK positioning devices are arranged on the main body of the construction equipment and other parts (such as a lifting hook, the front end of an extension arm and the maximum outer edge around a hoisted object) and are used for acquiring accurate spatial position data and moving rate of a monitored target in real time and transmitting the data back to the safety management and control system in real time.

The RTK unmanned aerial vehicle is monitored through an in-station operation site, carries on visible light camera equipment, infrared thermal imaging equipment and laser ranging equipment to monitor operators and construction equipment entering a safe operation area, and remote sensing data such as visible light, infrared and laser ranging acquired by monitoring of the RTK unmanned aerial vehicle are transmitted back to a safety control system of a background server in real time.

The method comprises the steps of automatically calculating the position, moving track, moving speed, safety distance and other data of an operator and a construction apparatus according to the spatial position data of the construction apparatus returned by an RTK spatial position monitoring device on an operation site and the monitoring remote sensing data of the operator returned by an RTK unmanned aerial vehicle on the operation site, mapping the data onto a three-dimensional point cloud model, calculating based on technologies such as a dangerous area vector library, a spatial collision detection algorithm and the like, and autonomously judging whether a target deviates from a predefined specified route, enters a charged interval by mistake and invades a non-safety area of charged equipment, namely a dangerous area.

When the target enters a set dangerous area by mistake or exits a preset operation area by mistake, the computing system automatically sends out alarm signals and shouting of different levels in stages according to the distance that the target approaches the warning line electronic fence. Meanwhile, the remote sensing data of visible light transmitted back by the RTK unmanned aerial vehicle is monitored on the operation site, the AI intelligent recognition model is used for timely finding out the illegal operation condition that an operator does not wear safety helmets and insulating protective clothing, and warning information and shouting are sent to the target in real time.

The safety distances mentioned above can be understood as: on the three-dimensional map, the position of the target is a distance from a pre-defined fence (fence). Sending an alarm when approaching an electronic fence (a warning line) in a safety area; an alarm is raised when entering or approaching an unsecured area.

In addition, the invention also provides a handheld terminal App which provides real-time data such as three-dimensional model data of the transformer substation, spatial position data, an activity track, a moving speed, a safety distance and the like for field operation personnel and construction equipment operators, and displays the data in a visual digital twin three-dimensional scene (if field data in different time periods are mapped onto a digital twin model, a three-dimensional and dynamic scene can be realized), and the data is consistent with a real operation field.

When the safety control cloud platform monitors that the working personnel do unsafe behaviors or detect collision warning, the corresponding terminal equipment can be determined according to the positioning data, and warning information is automatically pushed to the terminal equipment. The technology is used as an effective technical support for the line operator to sense the hidden danger in advance, and multiple safety guarantee measures combining passive warning and active sensing are achieved.

Example two

The safety management and control system of the embodiment calculates and calculates based on technologies such as a dangerous area vector library, a space collision detection algorithm and the like, autonomously judges whether a target (an operator and a construction apparatus) deviates from a specified route, mistakenly enters a charged interval and invades into an unsafe area of charged equipment, and sends out warning signals and shouting of different levels in stages according to the distance of the monitored target approaching a warning line electronic fence when the target (the operator and the construction apparatus) mistakenly enters a set dangerous area or mistakenly exits a preset operation area.

Furthermore, visible light remote sensing data returned by the RTK unmanned aerial vehicle is monitored through the operation site, the AI intelligent recognition model is used for timely finding out the illegal operation condition that an operator does not wear safety helmets and insulating protective clothing, and warning information and shouting are sent to the target in real time.

The embodiment provides a processing system of which the calculation processing process in the safety management and control system can correspond to the computing equipment, the safety management and control system constructs an intelligent safety management and control system of an unmanned aerial vehicle laser radar transformer substation through a three-layer framework of a front-end acquisition layer (sensing layer), a data transmission layer (network layer) and an intelligent decision analysis layer (application layer), and intelligent monitoring and safety management and control on the operation site of the transformer substation are realized, as shown in fig. 7.

Front end acquisition layer

The front-end acquisition (perception layer) is used for acquiring the point cloud data of the laser radar of the transformer substation and the accurate spatial position data of the operation site by the augmented reality technology, and the front-end data acquisition comprises the following steps: (1) collecting point cloud data of a laser radar of a transformer substation; (2) collecting spatial position data of an operator; (3) and collecting the spatial position data of the working machine.

For example, three geographical mapping modes such as unmanned aerial vehicle laser radar equipment, ground laser radar equipment and backpack laser radar equipment are combined to obtain holographic space point cloud data (the positioning data precision reaches centimeter level) of transformer substation full coverage, no dead angle and no loss.

The unmanned aerial vehicle airborne laser radar technology is a novel measurement technology which is generated by combining a laser ranging technology, a computer technology, an unmanned aerial vehicle RTK positioning technology and the like. The laser ranging technology is based on the fact that a laser emits an electromagnetic wave signal to a measuring area, and therefore relevant measuring data information such as angles and distances of the measuring area is obtained. The airborne laser radar system uses the unmanned aerial vehicle as a working platform, utilizes the advantages of the laser in measuring angle and height, obtains the vertical height data of the laser by combining the RTK technology, and simultaneously obtains the laser emission azimuth data based on the INS, thereby rapidly and accurately calculating the three-dimensional coordinates of each light spot of a target area.

The RTK unmanned aerial vehicle is monitored through the operation site in the station to the operating personnel who get into the safe operation area, and remote sensing data such as visible light, infrared and laser ranging obtained by monitoring of the RTK unmanned aerial vehicle are transmitted back to the safe management and control system in real time (the positioning data precision reaches centimeter level).

The on-site monitoring RTK unmanned aerial vehicle's of operation in station airborne equipment includes: 1) The visible light camera equipment is used for monitoring construction operation in the daytime; 2) The infrared thermal imaging equipment is used for monitoring construction operation at night; 3) The laser ranging equipment is used for carrying out laser ranging on the target and improving the precision of an AI target detection algorithm; 4) The RTK positioning module enables data collected by equipment carried by the unmanned aerial vehicle to reach centimeter-level precision; 5) And the megaphone is used for shouting and alarming for operators or operators of construction equipment on the operation site.

Operation requirements of an operation site monitoring RTK unmanned aerial vehicle are as follows: 1) The unmanned aerial vehicle carries visible light camera equipment, an infrared thermal imager and laser ranging equipment; 2) And (4) setting parameters of a holder, so that the airborne equipment vertically acquires data to the ground in a 90-degree posture.

By applying an RTK positioning technology, high-precision RTK spatial position monitoring devices are arranged on a main body of a construction apparatus and other positions (a lifting hook, the front end of an extension arm, each bending joint point of a movable arm, the maximum outer edge of the periphery of a hoisted object and the like) for acquiring accurate spatial position data and moving rate of a monitored target in real time and transmitting the accurate spatial position data and moving rate data back to a safety management and control system (the positioning data precision reaches centimeter level) in real time. The high-precision positioning mode of the embodiment can be used for safety monitoring of construction operation at night.

Data transmission layer

The unmanned aerial vehicle ground station transmits the load data of the unmanned aerial vehicle to a nearby base station, the base station transmits the data to a switch network where a server is located, and the server decodes and distributes the data to each client after receiving the load data of the unmanned aerial vehicle to finish data transmission.

Intelligent decision analysis layer, application layer

The front-end acquisition (sensing layer) acquires accurate spatial position data of an operation field in real time through a virtual reality technology, and the data is identified, compared and analyzed, and an intelligent analysis result sends alarm information to the operation field through data transmission (network layer). The intelligent analysis (application layer) includes: data center, artificial intelligence analysis, service center and safety early warning system.

1) Point cloud data model preprocessing

In the embodiment, the holographic spatial point cloud data of the transformer substation acquired by unmanned aerial vehicle laser radar equipment in combination with ground laser radar and backpack laser radar equipment is synthesized into a high-precision point cloud data three-dimensional model (rough model), and the positioning precision reaches the centimeter level.

2) Point cloud data model post-processing

And post-processing the point cloud data three-dimensional model (rough model) through professional software, cutting redundant data, thinning, removing noise points and the like, and finely trimming and synthesizing a high-precision holographic three-dimensional model (fine model), wherein the positioning precision reaches the centimeter level.

3) Constructing digital twin models

The digital twin model is utilized to simulate and simulate the construction site of the transformer substation in a map of a system, data collected by the sensor can be displayed in real time through the model, and the positions, the movement tracks and the working conditions of operating personnel and operating machinery in the working process are visualized.

4) Classifying dangerous areas of different grades

Based on a high-precision three-dimensional laser point cloud precision model, according to the requirements of work content, power failure equipment and the like in a station, different-level dangerous area vector libraries such as a charged equipment area, a power failure equipment area, a safe operation area, a warning line and the like are divided in a three-dimensional laser point cloud model, and the electronic fence and the area boundary visualization of a three-dimensional map are realized by utilizing a WebGL rendering technology and are used for being used as a real-time collision detection comparison library.

5) Three-dimensional model rendering

Rendering a three-dimensional holographic model of a transformer substation, a digital twin model, a contour model of a working machine, a geometric model of a hoisted object, an AI sample model library and the like on a Cesium three-dimensional map; mapping the dangerous area vector library to a Cesium three-dimensional map;

and mapping real-time dynamic space position data of the operation machine returned from the operation site onto a three-dimensional map, so that the twin model of the operation machine dynamically adjusts the action of the operation machine which is the same as that of the operation machine which actually operates in real time.

6) AI measurement and calculation of spatial position data of operator

The AI target detection model automatically identifies a human body, and automatically identifies the aerial video by combining a spatial position detection algorithm and calculates spatial position data of an operator. The real-time dynamic spatial position data of the operator obtained by AI measurement and calculation is mapped onto a three-dimensional map, so that the dynamic adjustment of the twin model of the operator keeps the same action as the real-time adjustment of the field operator, as shown in fig. 5A to 6.

For example, (1) the RTK coordinates and the current ground height of the unmanned aerial vehicle are acquired through the RTK sensor, and the visible light camera is carried to view and acquire pixels, the width and the length of an imaging frame, the focal length of a camera lens and the CMOS parameters of the camera lens of a returned video/photo. By using the acquired parameters, RTK positioning of four edges of the edge in the visual field of the visible light camera of the unmanned aerial vehicle can be calculated;

(2) and recognizing the human body target in the picture/video by using the AI human body target detection model, and obtaining the pixel coordinate of the human body target after recognition.

(3) When the RTK coordinates of the imaging center point and the four corners of the boundary are known, the pixel coordinates of the human body can be converted into the RTK coordinates by using a coordinate system conversion method, so that the actual position of the operator can be obtained.

(4) The spatial position information measured and calculated based on the visible light data has errors, the error source is mainly the current ground height of the unmanned aerial vehicle, therefore, the unmanned aerial vehicle is required to carry a laser range finder to measure the ground height, the height difference and the left and right placement distances between the laser range finder and a camera lens are considered, and the high-precision spatial position information is finally obtained.

(5) Certain errors can exist in the laser range finder, the unmanned aerial vehicle and the camera under different environments, flight heights and monitoring angles, for example, the geometric position of an image pixel generated in the imaging process of the unmanned aerial vehicle is extruded, extended, deviated, distorted and the like relative to a reference system (ground actual position), so that the geometric position, size, shape and orientation of a ground object in an image are changed, and the difference of data needs to be corrected and calculated through a geometric correction algorithm and a spatial positioning algorithm to obtain relatively accurate spatial position data.

7) Method for acquiring spatial position data of construction equipment

(1) The construction equipment comprises different types and models, for example, a transport vehicle, a crane, a bulldozer, an excavator, a pile driver and the like, the overall length, width, height and other contour dimensions of the construction equipment of different types and models are obtained through technical parameters of a construction equipment supplier and actual measurement, and a construction equipment contour model base is established in a safety early warning system in advance according to the dimension data.

(2) Before construction operation, a construction equipment number is recorded in a safety early warning system, the system automatically finds the corresponding contour dimension of the construction equipment from a construction equipment contour model library, and a construction equipment twin model of 1. And rendering the twin model of the construction equipment on a Cesium three-dimensional map. During construction, the system maps real-time dynamic spatial position data of the construction equipment returned by an operation site to a three-dimensional map, so that the twin model dynamic adjustment of the construction equipment keeps the same action as that of the construction equipment in actual operation in real time.

8) Helmet identification

And automatically comparing and identifying the sample model library of the AI identification module to judge whether the safety helmet is worn by the human body of the field operator.

The method comprises the following steps: judging the area where the operator exists, and marking the face by using a face detection module; by adopting the YOLOv5 algorithm, the target detection framework can realize the detection of various target objects.

Step two: extracting a head subregion of an operator; uniformly extracting the upper part of the middle part of each region, fixing the extracted region as a square, and inputting the separated head region into a subsequent network for subsequent analysis.

Step three: and (4) adopting a binary classification method for the extracted images to judge whether the worker wears the safety helmet or not. And finishing judging whether the safety helmet is worn or not by using a binary classification method.

9) Spatial collision detection

9.1 Safety work detection for workers

And designing a space collision detection algorithm based on the dangerous area vector library and real-time space position monitoring data, automatically measuring and calculating the position of the human body and the distance between the human body and the early warning line electronic fence, sending out early warning information when the human body is predicted to approach the early warning line electronic fence, and sending out warning information when the early warning line electronic fence is detected to collide.

9.2 Safety operation detection for working machine

And designing a space collision detection algorithm based on the dangerous area vector library and real-time space position monitoring data, automatically measuring and calculating the position of the machine and the distance between the machine and the early warning line electronic fence, sending early warning information when predicting that the mechanical part is close to the early warning line electronic fence, and sending warning information when detecting that the mechanical part collides with the early warning line electronic fence.

9.3 ) spatial collision calculation method

The gis-based geographic information engine calculates whether the single element is within the polygon. Firstly, calculating the distance between a point and a plane where a polygon is located, if the distance is greater than 0, the point is not on the plane and is more unlikely to be in the polygon; then, the three-dimensional plane is reduced to a two-dimensional plane (one component is directly deleted), the three-dimensional point is reduced to a two-dimensional point, and then whether the point is in the three-dimensional polygon is judged according to the method for judging whether the point is in the polygon in the two-dimensional plane. The reason that a three-dimensional plane can be reduced to a two-dimensional plane is that when a plane is projected to a certain coordinate plane, if a point before projection is within a polygon, the point after projection is still within the polygon; of course, the above conclusion holds if the plane a of the polygon is not perpendicular to the projected coordinate plane B. If a certain component in the normal vector is 0, the plane A where the polygon is located is perpendicular to a certain coordinate plane B, and if the polygon is projected to the coordinate plane B, the projection result is a line, it cannot be determined whether the point is in the polygon.

And (3) establishing a three-dimensional space model, wherein after spatial reference is provided, each point of the vector file has own X and Y coordinates, cesium can further calculate the distance according to the coordinates when the distance between the object and the polygon is calculated according to the existing coordinate information, and if the distance is less than 0, the intersection (collision) is generated between the digital twin model and the boundary of the electronic fence.

The ground station mentioned in this embodiment communicates with a safety control system, and the operation site RTK monitoring unmanned aerial vehicle and the ground station are interconnected and intercommunicated in real time. The safety control system can provide real-time data such as three-dimensional model data, spatial positions, moving tracks, moving speed, safety distances and the like of the transformer substation for terminals of field operators and construction equipment operators, and the data is displayed in a visual digital twin three-dimensional scene and is consistent with a real operation field, so that front-line operators can know the safety condition of the environment where the front-line operators are located in real time and adjust operation routes in time; on the other hand, when the safety control system monitors that the operator does unsafe behaviors or the collision with the warning line electronic fence is detected, the safety control early warning can automatically determine corresponding terminal equipment according to the positioning data and send early warning information to a handheld terminal (mobile phone) of the operator/mechanical operator, so that multiple safety guarantee measures combining passive warning and active sensing are realized.

Before the safety management and control system realizes the safety management and control method of the first embodiment, initialization may be performed, for example, a vector library of different levels of dangerous areas such as a charged equipment area, a power failure equipment area, a safety operation area, a warning line and the like may be pre-partitioned in the three-dimensional laser point cloud model/according to the work content in a station, the power failure equipment, the safety management and control operation and the like before construction. And inputting a construction equipment number into the safety management and control system, wherein the safety management and control system can automatically find the contour dimension of the corresponding construction equipment from the construction equipment contour model library based on the construction equipment number and generate a construction equipment digital twin model of 1. And rendering the twin model of the construction equipment on a Cesium three-dimensional map.

The method comprises the steps that geometric data of a hoisted object are input into a safety control system, the safety control system automatically finds a corresponding geometric model of the hoisted object from a geometric model library of the hoisted object, and a digital twin model of the hoisted object is generated by 1. And rendering the twin model of the hoisted object on a Cesium three-dimensional map.

The safety management and control system is pre-loaded with a construction instrument outline model library, a hoisting object geometric model, a pre-trained AI intelligent calculation model/AI intelligent identification model and the like.

The mobile terminal or the cloud server of the embodiment may include: the order processing method comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein when the computer program is executed by the processor, the steps of the order processing method of the cloud application are implemented as in any one of the above embodiments.

The method disclosed by the embodiment of the invention can be applied to a processor or realized by the processor. The processor may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware in a processor or by instructions in the form of software. The processor described above may be a general purpose processor, a digital signal processor, an application specific integrated circuit, an off-the-shelf programmable gate array or other programmable logic device, discrete gate or transistor logic, discrete hardware components. The various methods, steps and logic blocks disclosed in the embodiments of the present invention may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present invention may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software elements in the decoding processor. The software elements may be located in ram, flash, rom, prom, or eprom, registers, among other storage media that are well known in the art. The storage medium is located in a memory, and a processor reads information in the memory and combines hardware thereof to complete the steps of the method.

In addition, an embodiment of the present invention may provide a computer-readable storage medium, where a computer program is stored, and when the computer program is executed by a processor, the steps of any one of the above method embodiments are implemented.

It should be noted that in the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. In the description of the present specification, the description of "one embodiment", "some embodiments", "an embodiment", "an example", "a specific example" or "some examples", etc., means that a specific feature, structure, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention. While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, the claims should be construed to include preferred embodiments and all changes and modifications that fall within the scope of the invention.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention should also include such modifications and variations.

Claims

1. An intelligent safety control method for a transformer substation is characterized by comprising the following steps:

s30, acquiring motion information including attitude information, motion trail and spatial position data of each monitoring target according to monitoring data sent by the monitoring device and the auxiliary monitoring equipment on each monitoring target in a first time period;

the auxiliary monitoring device comprises: the system comprises monitoring equipment and an image acquisition device which are arranged on an unmanned aerial vehicle and use data of an RTK positioning technology;

when the monitoring target is a mobile construction equipment, the S30 includes:

s33, calculating motion information and a first safety distance based on the first spatial position data, the first moving speed, the second spatial position data and the second moving speed, wherein the motion information comprises: the position information of the main body of the mobile construction equipment, the position information of the mobile end of the mobile construction equipment, the motion track and the motion rate; the first safety distance is a safety distance with the nearest dangerous area;

when the monitoring target is an operator, S30 includes:

s35, measuring and calculating the monitoring image I, the infrared imaging data and the laser ranging data by adopting an AI intelligent identification model and a spatial position algorithm, identifying position information, a motion track and a motion rate of an operator, and identifying whether the operator belongs to an illegal operator who does not wear a safety helmet or an insulating protective clothing; judging whether the operator deviates from the operation track or not based on the operation track defined in advance;

the S35 comprises: acquiring basic parameters of an unmanned aerial vehicle and related equipment on the unmanned aerial vehicle; the basic parameters include: the current ground height of the unmanned aerial vehicle, the pixels of a video/photo, the width and the length of an imaging frame, the focal length of a camera lens and the CMOS parameters of the camera lens;

according to the basic parameters, the processing system obtains a first distance between the positive center point of the unmanned aerial vehicle and the edge of the visual field, calculates a second distance between the positive center point of the unmanned aerial vehicle and the four corners of the edge of the visual field by using the first distance and the width/2 of the imaging picture, and calculates the longitude and latitude coordinates of the actual positions of the four corners by using the longitude and latitude of the center point and the second distance;

identifying a human body target in the image by using the AI human body target detection model with the help of the auxiliary information of the user, and obtaining the pixel coordinate of the human body target after identification;

converting the pixel coordinates of the human body into RTK coordinates by using the position information of the known current position/camera imaging central point of the unmanned aerial vehicle and RTK coordinates of four corners of a boundary by using a GIS geographic information coordinate system conversion method, thereby acquiring the actual position of an operator;

carrying a laser range finder by using an unmanned aerial vehicle to measure the height to the ground, correcting the actual position of the obtained operator by using a target position correction model according to the height difference between the laser range finder and a camera lens and the data of the left and right placement distances, and finally obtaining centimeter-level precision spatial position information of the operator;

alternatively, S35 includes: acquiring centimeter-level longitude and latitude coordinates of the current position/camera imaging central point of the unmanned aerial vehicle and basic parameters of a camera through an unmanned aerial vehicle-mounted RTK sensor;

recognizing the human body target in the picture/video by using an AI human body target detection model by means of the auxiliary information of the user, and obtaining the pixel coordinate of the human body target after recognition;

specifically, knowing the position information of the current position of the unmanned aerial vehicle/the imaging center point of the camera and RTK coordinates of four corners of a boundary, converting human body pixel coordinates into the RTK coordinates by using a GIS (geographic information system) coordinate system conversion method so as to obtain the actual position of an operator, measuring the ground height by using the unmanned aerial vehicle carried laser range finder, correcting the actual position of the operator by using a target position correction model according to the height difference between the laser range finder and a camera lens and the data of the left-right placement distance, and finally obtaining centimeter-level precision spatial position information of the operator;

the S40 includes: calculating whether the single element is in a polygon range of a dangerous area vector library based on the gis geographic information engine;

firstly, calculating the distance between a point and a plane where a polygon is located, and if the distance is greater than 0, not locating on the plane; then, the three-dimensional plane is reduced to a two-dimensional plane, the three-dimensional points are reduced to two-dimensional points, and then whether the points are in the three-dimensional polygon or not is judged according to the method that whether the points are in the polygon or not is judged in the two-dimensional plane;

establishing a three-dimensional space model, wherein each point of a dangerous area vector library has own X and Y coordinates, cesium calculates the distance according to the coordinates when the distance between the monitored target and the polygon is calculated according to the existing coordinate information, and if the distance is less than 0, the digital twin model and the electronic fence boundary generate intersection, namely collision, and the monitored target is in a dangerous area;

s50, if the mobile terminal is in a dangerous area or in a dangerous area within a second time period, sending out early warning information; the second time period is a future time period;

wherein, discerning whether the staff belongs to not wearing the safety helmet in S35, include:

adopting a binary classification method for the extracted head subarea to judge whether the worker wears a safety helmet or not;

and finishing the judgment on whether the safety helmet is worn or not by using a binary classification method.

2. The substation intelligent safety management and control method of claim 1, further comprising:

s60, loading the motion information, the dangerous area vector library and the three-dimensional point cloud model on a Cesium three-dimensional map by combining a digital twin model to generate a visual dynamic three-dimensional scene; and loading a dangerous area vector library by using a WebGL rendering technology, and realizing area boundary visualization.

3. The substation intelligent safety management and control method of claim 1, further comprising:

4. A substation intelligent safety control method according to claim 2,

the S60 includes: based on the profile information of the mobile construction equipment pre-established in a construction equipment profile model library, a mobile construction equipment twin model of 1 is manufactured by using 3 DMAX; rendering the twin model of the mobile construction equipment on a visual dynamic three-dimensional scene;

in the construction process, the first space position data, the first movement rate, the second space position data and the second movement rate are mapped to a three-dimensional scene, so that the twin model of the mobile construction equipment is dynamically changed and is consistent with the motion of the mobile construction equipment in actual operation.

5. The substation intelligent safety control method according to claim 1, wherein the S50 comprises:

6. A computing device comprising a memory storing a computer program and a processor executing the computer program stored in the memory and performing the steps of a substation intelligent security management method as claimed in any one of claims 1 to 5.