CN115294805B - A video image-based aircraft conflict warning system and method for airport scenes - Google Patents

Abstract

The invention relates to an airport scene aircraft conflict warning system and method based on video images, which belong to the technical field of airport scene monitoring and airport management. The artificial intelligence target recognition subsystem is used to identify the airport scene target, and based on the coordinate transformation subsystem, the Two types of safety zones are set up for aircraft to realize real-time and accurate early warning of all types of aircraft conflicts on the airport scene; based on the complex taxiway blockage early warning subsystem, timely early warning of compound taxiway blockage. The recognition speed is fast, the hardware requirements are low, and it can be easily transplanted to small embedded platforms such as Raspberry Pi and FPGA, which has established the possibility for large-scale use in various airports and reduces the cost of use and maintenance. The early warning information is finally presented in the controller's airport automated scene monitoring terminal equipment, and different types of aircraft collision risk safety zones are identified based on different colors, which brings great convenience to the controller's real-time collision risk assessment of the aircraft on the scene.

Description

A video image-based early warning system and method for aircraft conflicts at an airport scene

technical field

The invention belongs to the technical field of airport scene monitoring and airport management, and in particular relates to an airport scene aircraft conflict early warning system and method based on video images.

Background technique

Aircraft operation safety is an eternal theme in the aviation industry and the basis for the survival and development of civil aviation. Aircraft accidents in the world mainly occur in the four stages of takeoff, climb, approach and landing. According to the "Commercial Jet Aircraft Accident Statistics (1959-2020)" report released by Boeing, there were 39 fatal accidents in Boeing's nearly 10-year data. The accident rate of the first flight phase accounts for only 13% of all fatal accidents. Instead, more than half of fatal accidents occurred during takeoff and landing. In the process of aircraft takeoff and landing, the accidents in the airport scene stage accounted for 20%. Nowadays, in order to expand the capacity of airports and increase the utilization rate of runways, airlines continue to expand their fleets, airports continue to increase the number of runways, the structure of taxiways is increasingly complex, and the scale of facilities is expanded, which often leads to traffic congestion in airports and increases aviation traffic. The probability of an unsafe operation accident occurring.

The airport is the starting and ending point of the entire air transport activity, undertaking various tasks such as aircraft take-off, landing, and parking, and playing an important supporting role. In the actual operation process, the airport surface operation management should arrange the operation plan of the surface movement target (including aircraft, support vehicles, and personnel) in an orderly manner to achieve safe and efficient operation, and this depends on the controllers having sufficient professional knowledge and skills. And with the support of the air traffic control command automation system, the identification, tracking, and conflict management and control of surface activity targets are completed in real time. Affected by factors such as the control blind area of the airport, the complexity of run-slip configurations, low-visibility operating conditions, and crew control, the superimposition of high-density flight support requirements, relying solely on manual monitoring can no longer meet the safety and efficiency requirements of surface operation management. . With the continuous maturity of video surveillance methods, the multi-point layout can realize the panoramic fusion perception of the traffic situation on the scene, and with the help of advanced video or image processing technology, it can support the controller to carry out real-time target recognition, interval dynamic detection and conflict early warning .

In addition, technically, deep learning methods have developed rapidly in recent years, and target recognition technology has been widely used in various fields such as national defense, national economy, social life, and space technology. For example, radio frequency identification technology (RFID), which is considered to be one of the most promising information technologies in the 21st century, stores and exchanges information through radio waves to achieve non-contact two-way communication. This technology has been widely used in material management, Production line automation management, vehicle control and access control systems; face recognition technology, which is highly recognized by the society in life, is a kind of biometric recognition technology, which mainly uses the characteristics of the face itself (the size of each organ, location, etc.) A comparison is made to identify the identity of each face. In addition, biometric technology includes fingerprint recognition, iris recognition, speech recognition, retina recognition, palmprint recognition, etc., which are used in e-commerce, social welfare, government, military, security and defense, etc.; optical character recognition technology (OCR) optically detects the dark and light patterns on the printing paper to determine the shape, and then uses the character recognition method to realize the conversion between the shape and the computer text, and complete the electronic printing, which brings great convenience to people's study and work; image recognition technology Also known as visual recognition technology, it uses computer technology to process and analyze captured images, identify objects, and identify object categories, which is conducive to the system to make correct judgments. It has great potential in many fields such as navigation, environmental monitoring, and weather forecasting. application value.

For now, major airports have been equipped with multilateration positioning (MLAT), surface surveillance radar (SMR), secondary radar (SSR), Automatic Dependent Surveillance (ADS-B), etc. However, such systems often have the characteristics of high cost, complicated maintenance, and high cost. In addition, whether it is through radar or ADS-B to display the position of the target aircraft is often not intuitive enough, it cannot accurately display the various parts of the fuselage, and cannot clearly display the aircraft. The collision safety zone will also produce a large error in judging the distance between the two aircraft. At present, there are no relatively mature solutions at home and abroad for the related technologies of aircraft collision warning at airports. The previous scene surveillance system has certain deficiencies in the comprehensive surveillance and judgment of the airport scene, the dynamic control of aircraft, and the assessment of collision risk. The specific performance is as follows: (1) There are many studies on the independent conflict types of aircraft on the scene, and there are few comprehensive and practical global early warning schemes on the scene; (2) The current cutting-edge aircraft collision early warning systems mostly use aircraft motion models to calculate the probability of collision risk , but the probability value is of little significance for the real-time early warning of the airport, and airport controllers need a qualitative value rather than a quantitative risk value as a judgment standard; (3) There is currently no technology for judging the collision risk through the aircraft safety zone. The present invention is based on the designation of the aircraft collision safety zone on the scene to facilitate the controller to carry out the collision risk assessment for each aircraft on the scene in real time; (4) in the research on the aircraft risk warning of the airport scene with the photoelectric camera at present, there is no camera pixel coordinate system and The actual coordinate system of the airport is transformed. In actual operation, aircraft identification and positioning will inevitably produce errors, but these photoelectric sensor-based surface surveillance and warning systems have not been designed for this.

Therefore, at this stage, it is necessary to design an airport scene aircraft conflict warning system and method based on video images to solve the above problems.

Contents of the invention

The purpose of the present invention is to provide an airport scene aircraft conflict warning system and method based on video images, which are used to solve the technical problems in the above-mentioned prior art, aiming at the comprehensive monitoring and research and judgment of the airport scene and the dynamic control of aircraft in the current airport scene monitoring system As well as the deficiencies in the collision risk assessment and other aspects, a qualitative early warning judgment method is given through the airport surface operation rules and the establishment of aircraft safety threshold models and aircraft safety zone delineation schemes. The invention proposes a coordinate conversion algorithm to obtain high-precision aircraft operation data and construct a full-scenario, multi-level early warning system, which is of great significance for ensuring the normal and safe operation of the airport.

For realizing the above object, technical scheme of the present invention is:

A method for early warning of aircraft conflicts at an airport scene based on video images, comprising the following steps:

Step 1: Introduce and collect relevant information on the airport surface situation that supports the operation of each system, and provide the required raw data for the early warning of aircraft conflicts on the airport surface;

Step 2: The original data obtained in step 1 and the current scene visibility data are transmitted to the image defogging and noise reduction subsystem of the edge computing subsystem; the image defogging and noise reduction subsystem generates a clear real-time monitoring video image of the airport scene;

Step 3: The real-time monitoring video image of the airport scene obtained in step 2 is transmitted to the deep learning subsystem of the edge computing subsystem; the deep learning subsystem will identify and track common targets on the scene;

Step 4: The pixel coordinates obtained in step 3 and identified by the aircraft on the scene surveillance video image are transmitted to the coordinate transformation subsystem of the edge computing subsystem; the coordinate transformation subsystem will obtain the current position of all aircraft on the scene in the airport Location;

Step 5: Upload the current position and current time of all aircraft on the scene obtained in step 4 to the scene aircraft conflict warning subsystem of the cloud computing subsystem; Combined with the actual operating conditions of aircraft at the airport, and based on the aircraft safety threshold model, a dynamic safety area is set for the aircraft; through two types of safety areas, common conflicts between aircraft on the scene are warned; through the compound taxiway congestion model, compound taxi Early warning of road blockage events;

Step 6: Transmit the two types of safety zones and early warning information set in step 5 to the terminal scene monitoring screen.

Further, in step 2, the defogging and noise reduction processing of the real-time monitoring video image of the airport scene is adapted according to the visibility value of the airport scene.

Further, in step 3, the deep learning subsystem needs to first carry out image acquisition and training on the target object in the actual operation of the airport; and after the system is put into use, continue to collect the image data enhancement training of the target object in the airport to improve the System accuracy and stability.

Further, in step 5, the movement status of the aircraft will be judged based on the previous stage position and the current position data of each aircraft, and the velocity and direction of the aircraft will be calculated; for the moving aircraft, the safety zone will be defined according to the design plan.

Further, in step 5, the early warning is divided into three levels:

1) Warning, mark the solid safety zone for the moving aircraft according to the aircraft operating parameters and airport operating regulations in red, and set the dynamic safety zone by calculating the relative orientation of two aircraft, aircraft operating parameters, airport operating regulations, and aircraft safety thresholds Marked in blue, it is convenient for controllers to control the operation status and risk assessment of aircraft in the area;

2) Early warning, detailed warning information will be issued when the designated solid safety zone and dynamic safety zone early warning conditions are reached; for compound taxiways, early warning will be issued according to the blocking early warning algorithm; at this moment, the controller is reminded to pay attention to the designated aircraft in time, for the control Instructions issued by the staff for reference;

3) Alarm. For detailed alarm information when the designated solid-state safety zone and dynamic security zone alarm conditions are reached, the system will issue warnings by flashing warning lights and beeping.

Further, in step 6, detailed warning information will be displayed by the display module in the airport scene monitoring terminal equipment.

A video image-based early warning system for aircraft conflicts at airports, including:

The edge computing subsystem includes the distributed field surveillance camera group for airport scene layout; the distributed field surveillance camera group includes cameras and embedded platform devices, which are divided into scene surveillance camera module, image defogging and noise reduction subsystem, and deep learning subsystem and a coordinate conversion subsystem, the embedded platform will perform defog and noise reduction processing on the scene surveillance video images collected by this set of distributed field surveillance cameras in real time, recognize the airport scene targets based on deep learning, and obtain the scene aircraft through the coordinate transformation algorithm. location at the airport scene;

The cloud computing subsystem includes a surface aircraft conflict early warning subsystem, a surface aircraft solid safety zone designation module, and a surface aircraft dynamic security zone designation module; the surface aircraft conflict early warning subsystem includes a conflict early warning module between active aircraft, active aircraft and The conflict early warning module between static aircraft and the airport composite taxiway block early warning module; the conflict early warning module between active aircraft is used for conflict early warning between two moving aircraft on the airport scene; between active aircraft and stationary aircraft The conflict warning module is used for the conflict warning between the moving aircraft and the stationary or waiting aircraft on the airport surface; the airport compound taxiway block warning module is used for two intersections connected by a taxiway, and the intersection is connected in the middle The length of the taxiway is greater than the detection area of the dynamic safety zone, that is, it is impossible to use the dynamic safety zone designation scheme for early warning, and it is impossible to use the solid safety zone for early warning of congestion warning on the taxiway topology;

Communication network subsystem: used to ensure the stable connection of each subsystem and subsystem, and complete the data transmission in the system;

Time service sub-system: a time service system based on IP data network;

Data storage subsystem: including surface aircraft position and status data storage module, surface aircraft solid safety area storage module, surface aircraft static safety area storage module, surface taxiway structure information storage module, and airport surface meteorological data storage module; Use various types of data for classified storage;

Airport scene monitoring terminal: including a display module and an alarm module; the display module is used to display airport scene map information, airport scene aircraft position information, airport scene aircraft safety area information and airport scene aircraft early warning information; the early warning module uses sound warning equipment, when the airport When the scene monitoring terminal receives the early warning information sent by the cloud computing subsystem, it sends out a warning;

Airport data acquisition system: connected to the deployed airport weather station to receive the visibility data of the airport; also connected to the deployed airport operation control center to receive the aircraft take-off and landing time and aircraft priority data at the airport; also connected to To the deployed airport management center to receive coordinate data of taxiway intersections on the airport scene.

Compared with prior art, the beneficial effect that the present invention has is:

One of the beneficial effects of this scheme is that the airport scene aircraft conflict warning system and method of the present invention use the artificial intelligence target recognition subsystem to identify the airport scene target, and based on the coordinate conversion subsystem, set up two types of safety zones for the aircraft to achieve Real-time and accurate early warning of all types of aircraft conflicts on the airport scene; based on the complex taxiway blockage early warning subsystem, timely early warning of compound taxiway blockage. The invention has fast recognition speed, low requirements on hardware conditions, and can be easily transplanted to small embedded platforms such as Raspberry Pi and FPGA, thereby laying the foundation for large-scale use in various airports and the possibility of reducing use and maintenance costs.

The present invention is closely related to the actual operating conditions of the airport, so that it has extremely high practicability. The data comes from airport video, which is easy to obtain and process, and has strong user-friendliness. The early warning information is finally presented in the controller's airport automated scene monitoring terminal equipment, and different types of aircraft collision risk safety zones are identified based on different colors, which brings great convenience to the controller's real-time collision risk assessment of the aircraft on the scene.

Compared with the existing airport scene monitoring system, the present invention is specially designed for the application requirements of civil aviation airport scene operation safety in terms of system architecture, working mode and deployment scheme, so it can be very well adapted to civil aviation airport scene monitoring and Alarm requirements have huge application prospects and commercial value.

Description of drawings

Fig. 1 is a schematic diagram of an airport scene aircraft conflict warning system based on a video image according to an embodiment of the present application.

FIG. 2 is a schematic diagram of an edge computing subsystem according to an embodiment of the present application.

FIG. 3 is a comparison diagram of an example scene surveillance video processed by an image defogging and noise reduction subsystem according to an embodiment of the present application.

FIG. 4 is a schematic diagram of target recognition output by the deep learning subsystem of the embodiment of the present application.

FIG. 5 is a schematic diagram of a camera calibration algorithm of a coordinate transformation subsystem according to an embodiment of the present application.

FIG. 6 is a schematic diagram of a cloud computing subsystem in an embodiment of the present application.

Fig. 7 is a schematic diagram of the algorithm of the aircraft solid state safety area demarcation module in the embodiment of the present application.

FIG. 8 is a schematic diagram of an algorithm of a dynamic safety zone designation module for an aircraft on the surface according to an embodiment of the present application.

FIG. 9 is a schematic diagram of an early warning situation of a conflict early warning module between active aircraft according to an embodiment of the present application.

FIG. 10 is a schematic diagram of the algorithm of the conflict warning module between active aircraft according to the embodiment of the present application.

FIG. 11 is a schematic diagram of an early warning module for a conflict early warning between an active aircraft and a stationary aircraft according to an embodiment of the present application.

Fig. 12 is a schematic flow diagram of the algorithm flow of the conflict warning module between the active aircraft and the stationary aircraft according to the embodiment of the present application.

Fig. 13 is a schematic diagram of 12 common composite taxiway blockages in the embodiment of the present application.

Fig. 14 is a schematic diagram of an early warning module of an airport compound taxiway congestion early warning module according to an embodiment of the present application.

Fig. 15 is a schematic flow diagram of the algorithm flow of the airport complex taxiway congestion early warning module according to the embodiment of the present application.

Fig. 16 is a schematic diagram of an airport scene monitoring terminal according to an embodiment of the present application.

FIG. 17 is a schematic diagram of a video of an airport scene intersection conflict monitoring terminal according to an embodiment of the present application.

Fig. 18 is a video schematic diagram of a composite taxiway congestion early warning monitoring terminal for an airport scene according to an embodiment of the present application.

Detailed ways

In order to make the object, technical solution and advantages of the present invention more clear, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, and are not intended to limit the present invention, that is, the described embodiments are only some of the embodiments of the present invention, but not all of the embodiments. The components of the embodiments of the invention generally described and illustrated in the figures herein may be arranged and designed in a variety of different configurations.

Accordingly, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts belong to the protection scope of the present invention. It should be noted that relative terms such as the terms "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any relationship between these entities or operations. There is no such actual relationship or order between them.

Furthermore, the term "comprises", "comprises" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article or apparatus comprising a set of elements includes not only those elements, but also includes elements not expressly listed other elements of, or also include elements inherent in, such a process, method, article or apparatus. Without further limitations, an element defined by the phrase "comprising a ..." does not preclude the presence of additional identical elements in the process, method, article or apparatus comprising said element.

The characteristics and performance of the present invention will be described in further detail below in conjunction with the examples.

Example 1:

A method for early warning of aircraft conflicts at airport scenes based on video images is proposed, including the following steps:

Step 1: Lead and collect relevant information on the airport surface situation that can support the operation of each system, and provide the required raw data for the early warning of aircraft conflicts on the airport surface. These include: local meteorological data from the data acquisition system, flight schedule and priority data from the operation control center, airport run-slip configuration data, and scene real-time surveillance video image data from the distributed edge computing system.

Step 2: The airport scene monitoring video data and current scene visibility data obtained in step 1 are transmitted to the image defogging and noise reduction subsystem of the edge computing subsystem; the image defogging and noise reduction subsystem will obtain a clear airport scene in real time Monitor video images.

Step 3: Transmit the clear airport scene real-time surveillance video image obtained in step 2 to the deep learning subsystem of the edge computing subsystem; the deep learning subsystem will identify and track common targets on the scene, including: aircraft, operations Vehicles, people, birds and covered bridges.

Step 4: The pixel coordinates obtained in step 3 that the airport scene aircraft is identified on the scene surveillance video image are transmitted to the coordinate transformation subsystem of the edge computing subsystem; the coordinate transformation subsystem will obtain all aircraft currently in the airport on the scene s position

Step 5: Upload the current position and current time of all aircraft on the scene obtained in step 4 to the scene aircraft conflict warning subsystem of the cloud computing subsystem; Solid safety zone; combined with the actual operation of aircraft at the airport, and based on the aircraft safety threshold model, a dynamic safety zone is set for the aircraft; two types of safety zones are used to warn common conflicts between aircraft on the scene; Early warning of taxiway blockage events.

Step 6: Transmit the two types of safety zones and early warning information set in step 5 to the terminal scene monitoring screen.

Example 2:

As shown in Figure 1, the airport scene aircraft conflict early warning system based on video images proposed by the present invention is characterized in that: it includes a data acquisition subsystem 1 connecting the airport weather station and the airport operation control center, a data storage subsystem 2, and a time distribution system. System 3, edge technology subsystem 4, cloud computing subsystem 5, airport scene monitoring terminal 6, communication network subsystem 7. Each part is distributed in various positions of the airport use environment. The present invention firstly needs to lead and collect information related to the situation of the airport scene that can support the operation of each system, and provide the required original data for the early warning of aircraft conflicts on the airport scene. These include: the local meteorological data from the data collection subsystem 1, the flight schedule and priority data of the operation control center, the airport run-slip configuration data, and the scene real-time surveillance video images from the distributed edge computing subsystem 4 data. Secondly, the obtained airport scene monitoring video data and current scene visibility data are transmitted to the image defogging and noise reduction subsystem 41 of the edge computing subsystem 4 to obtain clear real-time monitoring video images of the airport scene. Again, the obtained clear airport scene real-time surveillance video images are transmitted to the deep learning subsystem 42 of the edge computing subsystem 4, and the deep learning subsystem 42 will identify and track common targets on the scene. Then, the obtained pixel coordinates of the airport scene aircraft identified on the scene surveillance video image are transmitted to the coordinate conversion subsystem 43 of the edge computing subsystem 4 to obtain the current positions of all aircraft on the scene in the airport. Upload the obtained current position and current time of all aircraft on the scene to the scene aircraft conflict warning subsystem 51 of the cloud computing subsystem 5 . The aircraft solid safety area designation module 501 on the scene designates a solid safety area for the aircraft according to the operation rules of the maneuvering area; area; the conflict early warning module 511 between the active aircraft and the conflict early warning module 512 between the active aircraft and the static aircraft carry out early warning to common conflicts between the aircraft on the scene through two types of safety zones; through the airport compound taxiway blocking module 513, the Early warning of complex taxiway blockage events. Finally, the two types of safety zones and the early warning information are transmitted to the display module 61 of the airport scene monitoring terminal 6, and the early warning module 62 emits an early warning sound. The data transmission all passes through the communication network subsystem 7 , and the data for each transmission is stored in the data storage subsystem 2 . The whole system performs time service through the time service sub-system 3. The function of each subsystem of the present invention will be described in detail below.

A data acquisition subsystem 1

In order to realize that the aircraft conflict warning system on the airport scene accurately reflects the real-time dynamics of the airport where the system is deployed, it is necessary to determine the relevant information and data connected from the data acquisition subsystem 1 according to the design of the related systems of the edge computing subsystem 4 and the cloud computing subsystem 5 . When setting the type of data to be linked, factors such as data quality, availability, and acquisition time are mainly considered. Data comes mainly from three sources:

a. Visibility detection of the local weather station;

b. The flight schedule and priority of the airport operation control center;

c. Data related to the running and sliding structure of the field management center.

The data acquisition subsystem 1 is connected to the airport weather station deployed by the system, and is used to receive the visibility data of the airport, and transmit the visibility data to the data storage subsystem 2 through the communication network subsystem 7;

The data acquisition subsystem 1 is connected to the airport operation control center deployed by this system, and is used to receive the aircraft take-off and landing time and aircraft priority data on the airport scene, and transmit the taxiway related data on the scene through the communication network subsystem 7 to data storage subsystem 2;

The data acquisition subsystem 1 is connected to the airport management center deployed by the system, and is used to receive the coordinate data of taxiway intersections on the airport scene, and transmit them to the data storage subsystem 2 through the communication network subsystem 7 .

2 Data storage subsystem 2

The data storage subsystem 2 realizes the storage and fusion processing of the connection data, and sends the data to the specified subsystem to form a unified and dynamically updated airport scene situation data center.

Specifically, the data storage subsystem 2 needs to obtain the visibility data of the local airport, the take-off and landing time of the aircraft at the local airport, the priority data of the aircraft, and the coordinate data of the taxiway intersection on the scene from the data acquisition subsystem 1 through the communication network subsystem 7; Obtain the local time from the time service subsystem 3; obtain the position coordinate data of the aircraft on the surface from the edge computing subsystem 3; obtain the solid safety area and dynamic safety area data of the aircraft on the surface, the early warning information data, and activities between active aircraft from the cloud computing subsystem 5 Early warning information data between aircraft and stationary aircraft and composite taxiway early warning information data. Through data fusion processing, and storage, real-time update of the field scene situation database.

The data storage subsystem 2 transmits the visibility data to the image defogging and noise reduction subsystem 41 inside the edge computing subsystem 4 through the communication network subsystem 7; the flight time and its priority and the scene taxiway related data are communicated The network subsystem 7 transmits to the scene aircraft conflict warning subsystem 51 inside the cloud computing subsystem 5; transmits the current position coordinates and historical position coordinate data of the scene aircraft to the scene aircraft conflict inside the cloud computing subsystem 5 through the communication network subsystem 7 The early warning subsystem 51; transmits the two types of collision risk safety zone data of the aircraft on the scene, the early warning information data between active aircraft, the early warning information data between the active aircraft and the stationary aircraft, and the composite taxiway early warning information data through the communication network subsystem 7 to the Airport scene monitoring terminal 6;

Three time transfer system 3

The time service sub-system obtains time information through the Internet or time service satellites, and internally calibrates and converts it to the current local time, and at the same time releases it to each component in the system to ensure that the system time of each device is consistent.

Quad Edge Computing Subsystem 4

As shown in Figure 2, in some embodiments, there are several distributed edge computing subsystems 4, and the edge computing subsystems 4 include a scene monitoring camera module 401, an image defogging and noise reduction subsystem 41, and a deep learning subsystem 42 , Coordinate transformation subsystem 43.

The scene monitoring camera module 401 is used to capture video information of the airport scene in real time, and directly transmit the video to the image defogging and noise reduction subsystem 41 of the sub-system.

The image defogging and noise reduction subsystem 41 is an important part of the present invention, and its purpose is to reduce or eliminate the interference of fog in the image and improve the image quality. In the outdoor imaging system, there is a great dependence on the weather, especially under severe weather conditions, there are often a large amount of dust and particles suspended in the atmosphere, which is amplified by the absorption and scattering of light, affecting the light transmittance, resulting in the obtained Images are severely degraded. The detailed features in the picture are covered, the clarity and contrast are reduced, the dynamic range is reduced, and the picture may even be distorted due to the reduction of image color saturation, which directly affects the recognition of objects.

Design of the image defogging and noise reduction subsystem 41: the scene monitoring video image transmitted from the scene monitoring camera module 401, the data storage subsystem 2 transmits the field visibility data through the communication network subsystem 7. The system determines whether defog and noise reduction processing is required, and if defog and noise reduction processing is not required, the scene monitoring video image is directly transmitted to the deep learning subsystem 42 of this sub-system; if defog and noise reduction processing is required, the After the noise reduction algorithm is processed, it is transmitted to the deep learning subsystem 42 of this subsystem. In the process of implementation, the key lies in the choice of image processing method. Although there are many methods in the field of image processing, each has advantages and disadvantages. The image defogging and noise reduction subsystem 41 of the present invention can not only use one method for image defogging and noise reduction, but also can combine two or more methods to achieve complementary advantages and improve the processing effect.

In order to better clarify the defogging and noise reduction technology proposed in the present invention, the improvement of the recognition rate of aircraft and other targets under low visibility by deep neural network is used as an index to evaluate the pros and cons of the defogging and noise reduction technology. Select the scene surveillance video image as the research object, as shown in Figure 3, for example, use a frame of the morning surveillance video of an airport as an example image to demonstrate the processing effect of the image defogging and noise reduction subsystem 41:

Judging from the change in the recognition percentage of the indicator "airplane", after dehazing and noise reduction processing and then using the deep learning neural network for recognition, the target recognition rate is significantly improved, and the effect is remarkable.

The deep learning subsystem 41 is based on the deep neural network model to classify the scene targets (including: aircraft, work vehicles, people, birds and bridges), and obtain the target coordinates (pixel coordinates), especially the scene aircraft coordinates. As shown in Figure 4, the sample airport scene surveillance video is output image through deep learning subsystem 41

Figure 4 shows that after using the airport scene target recognition model based on the output of the deep learning subsystem 42, whether it is a moving aircraft or a waiting aircraft, the six aircraft in the figure have been recognized with high accuracy. Since the scene is a dynamic process, the recognition rate of each aircraft shown in the red box in the figure is constantly changing, and this effect is far superior to other currently available target-based recognition models.

In the previous step, defogging and noise reduction processing and target recognition have been performed on the surveillance video image of the airport scene under low visibility, and the coordinates of the aircraft on the scene are sent to the coordinate conversion subsystem 43 . However, the above steps only solve the recognition, classification and acquisition of pixel coordinates of airport targets under low visibility. The biggest difference between the present invention and other aircraft risk warning schemes based on photoelectric cameras is that this paper uses the actual airport coordinates of the aircraft instead of pixel coordinates for high-precision risk warning.

The coordinate transformation algorithm of the coordinate transformation subsystem 43 starts from the two two-dimensional planes of reality and the camera, and has the characteristics of few external parameters required, simple algorithm structure, high precision, and no cumbersome chessboard for calibration. Figure 5 below shows a comprehensive coordinate map based on airport vision and camera vision. Through this system, the mutual conversion of the two coordinate systems can be realized.

In Fig. 5, the scene coordinate system O ₁ and the airport coordinate system O ₂ are established by measuring the known aircraft pixel position as (C _x , _Cy ), and the pixel size is 1280×720, AB=a, BC ₁ =b , BC ₄ =c, C ₁₁ C ₁₂ =g, C ₄₁ C ₄₂ =k, and then have:

e＝d cosθ，

因为

令

(360为720像素的一半)，

C₁C₃＝arctan(α+θ+γ)-b，C₁₂在O₁坐标系下的坐标为

C₄₂在点O₁坐标系下的坐标为

得到C₃C₄＝c-b-C₁C₃，P₁P₂在O₁坐标系下的直线方程表示为

将y＝C₁C₃-C₁C₂带入得C₃₂C₃₄，因为

(640为1280像素的一半)，再将O₁坐标系转换为O₂坐标系，可得(C₃₁C_C3-C₃₁C₃₂+h+g，C₁C₃-C₁C₂+v)，机位飞机在机场O₂坐标系中的坐标。

e = d cos θ,

because

make

(360 is half of 720 pixels),

C ₁ C ₃ =arctan(α+θ+γ)-b, the coordinates of C ₁₂ in the O ₁ coordinate system are

The coordinates of C ₄₂ in the point O ₁ coordinate system are

Obtain C ₃ C ₄ =cbC ₁ C ₃ , the linear equation of P ₁ P ₂ in the O ₁ coordinate system is expressed as

Put y=C ₁ C ₃ -C ₁ C ₂ into C ₃₂ C ₃₄ , because

(640 is half of 1280 pixels), and then convert the O ₁ coordinate system to the O ₂ coordinate system, which can be obtained (C ₃₁ C _C3 -C ₃₁ C ₃₂ +h+g, C ₁ C ₃ -C ₁ C ₂ +v ), the coordinates of the plane in the airport O ₂ coordinate system.

By using the coordinate conversion algorithm, the conversion between the actual coordinates of the airport and the coordinates of the camera view can be realized, and then the high-precision position coordinate data of each aircraft scene can be calculated, and transmitted to the data storage subsystem through the communication network subsystem.

Five Cloud Computing Subsystems 5

Usually an important factor for aircraft collisions at airports is that the controller fails to detect the conflict in time and stops the aircraft. However, due to the increase in the number of aircraft sorties, the controllers have limited energy, and the radar screen cannot intuitively display the fine target volume and the collision safety zone, which brings great pressure to the ground controllers. Therefore, based on target recognition, research on the designation of a single aircraft collision safety zone, visual display on the screen, building a refined early warning model, and providing qualitative risk indicators can reduce the workload of controllers, reduce airport unsafe incidents, and improve The efficiency of airport operations is of great significance.

As shown in FIG. 6 , in some implementations, the cloud computing subsystem 4 includes a solid-state aircraft safety zone designation module 501 , a dynamic aircraft safety zone designation module 502 , and an early-warning subsystem 51 for aircraft conflicts on the surface. The surface aircraft conflict early warning subsystem 51 includes a conflict early warning module 511 between active aircraft, a conflict early warning module 512 between active aircraft and stationary aircraft, and an airport compound taxiway block early warning module 513 .

Since the location of the airport studied is located in the jurisdiction of China, the main parameters of the following airport operations are specified in accordance with the Ministry of Transport of the People's Republic of China No. 30 Regulations on Civil Aviation Air Traffic Management (CCAR-93TM-R5) Limitations, the main limitations are shown in Table 1 below. In the follow-up research of this paper, these actual data will be used as the basis for research.

Table 1 Actual operating restrictions of the airport

The on-the-scene aircraft solid-state safety zone designation module 501 constructs aircraft collision safety zones based on parameters such as my country's actual airport operating restrictions and on-the-scene surveillance videos, so that controllers can discover potential aircraft risks on the scene in a timely manner and control the operation status of the airport. Since parameters such as the size, shape, and direction of the safety zone are relatively fixed to the target aircraft body, the present invention refers to it as a solid-state safety zone for the aircraft on the ground.

Considering the influence of the aircraft wake, a 30° arc of 50 meters behind the tail is taken as the area affected by the wake, and a 200m linear protection zone is set in front of the nose. The aircraft solid-state security zone delineation algorithm of the aircraft solid-state security zone demarcation module 501 on the scene is shown in Figure 7 below:

有As shown in Figure 7, the point S (x, y) is known, the velocity angle θ can be obtained by measurement, the size of the aircraft and the size of the safety zone are known, set

have

Point I: (x+ksinθ, y-kcosθ),

Point V: (x-ksinθ, y+kcosθ),

Point T: (x+lcosθ, y+lsinθ),

R point: (x-lcosθ, y-lsinθ),

Point H: (x+ksinθ-lcosθ, y-kcosθ-lsinθ),

Point J: (x+ksinθ+lcosθ, y-kcosθ+lsinθ),

K point: (x+ksinθ+lcosθ+200cosθ, y+kcosθ+lsinθ+200sinθ),

N points: (x-ksinθ-lcosθ, y+kcosθ-lsinθ),

M point: (x-ksinθ+lcosθ, y+kcosθ+lsinθ),

Point L: (x-ksinθ+lcosθ+200cosθ, y+kcosθ+lsinθ+200sinθ),

Q points:

(x+ksinθ-lcosθ-50cos(θ+15), y-kcosθ-lsinθ-50sin(θ+15)),

point p:

(x-ksinθ-lcosθ-50cos(θ-15), y+kcosθ-lsinθ-50sin(θ-15)).

It should be noted that the present invention only sets a safety zone for aircraft in motion, because according to the provisions of the "Civil Aviation Air Traffic Management Regulations", no matter whether it is a safety interval of 200m or a wake interval of 50m, it is only for aircraft in motion. .

Since the solid safety area always points directly in front of the aircraft taxiing, for the aircraft waiting at the taxiway intersection, the solid safety area designation scheme proposed above cannot effectively warn. Therefore, it is necessary to establish a safety early warning method for the active aircraft and the aircraft waiting at the intersection—the safety area based on the safety threshold calculation model. Since the size parameters, activation conditions and direction of the safety area are controlled by the relative state of the two aircraft, Therefore the present invention claims it as the scene aircraft dynamic safety zone. When the aircraft is traveling on the taxiway, it will pass through many taxiway intersections, and it is very easy to collide with the stationary aircraft in the waiting position outside the taxiway. The critical safety distance between the two machines is called the safety threshold.

The aircraft dynamic safety zone designation module 502 on the scene draws a dynamic safety zone according to the threshold value obtained by the safety threshold calculation model, and sets the distance between the center points of the two aircrafts to be smaller than the safety threshold. When a moving aircraft appears a dynamic safety zone, it acts as a warning. When it is less than the safety threshold two, the target aircraft is immediately detected by the dynamic safety zone of the moving aircraft, and an early warning is issued. Since there is no anti-skid braking and the redundancy of the safety threshold of wet road surface is the largest, it is set as the safety threshold 1. Anti-skid braking and the minimum redundancy of the safety threshold of arterial surface, let it be the safety threshold 2. As shown in FIG. 8 , the length of S ₁ W is the threshold 2, and XV is the diagonal length of the aircraft target frame obtained based on the deep learning subsystem 42 . The specific scene aircraft dynamic safety zone designation module 502 algorithm is as follows:

Its calculation formula is as follows: Suppose S ₁ (x ₁ , y ₁ ), S ₂ (x ₂ , y ₂ ), then there is

Point V coordinates:

Point X coordinates:

The two types of safety zone information are obtained by the aircraft solid safety area designation module 501 and the aircraft solid safety area designation module 502 on the surface, and are transmitted to the conflict early warning module 511 between active aircraft, active aircraft and stationary aircraft inside the aircraft conflict early warning system 51 on the surface. The conflict early warning module 512 and the airport complex taxiway block early warning module 513.

The conflict early warning module 511 between described active aircraft aims at several specific conflict situations (follow-up, head-to-head, crossing) that often occur between airport taxiway active targets. The aircraft in the overlapping area of the dynamic safety zone (blue zone) will provide timely early warning and alarm, which will help reduce the workload of controllers and pilots, and reduce the incidence of aviation unsafe incidents.

Figure 9 shows several common conflicts between aircraft moving on the airport taxiway and the early warning situations of the solid safety zone (red zone) and the dynamic security zone (blue zone). As shown in (a), it is a situation where two moving aircraft follow up and conflict. When the speed of the following aircraft is greater than the speed of the preceding aircraft, the safe distance between the two is less than the minimum separation stipulated in the "Civil Aviation Air Traffic Management Regulations". When the solid safety zone (red zone) established by the two moving aircraft overlaps, a solid safety zone warning is issued. If the two aircraft continue to move forward, when the distance between the two aircraft is less than the safety threshold 1, the aircraft dynamic safety zone will appear, and when the distance between the two aircraft is less than the safety threshold 2, the aircraft dynamic safety zone will issue an early warning. If the dynamic and solid safety areas are simultaneously warned at this time, a collision warning message will be issued; (b) is when the two aircraft collide with each other, and the distance is less than the specified distance, an early warning will be issued; similarly, in (c), (d), ( In the scenario shown in e), whether it is a vertical intersection or an oblique intersection conflict, the solid safety areas (red areas) established by the two aircraft will overlap, and an early warning will be issued. The flowchart of the early warning algorithm of the conflict early warning module 511 between specific active aircraft is shown in Figure 10 below:

The conflict early warning module 512 between the active aircraft and the stationary aircraft is shown in FIG. 11 , which shows a schematic diagram of the early warning between the active aircraft and the stationary aircraft. As shown in (a) stage 1, it is assumed that there are four aircraft (one is driving on the taxiway, and the other three are waiting in their respective positions), and the driving aircraft is aircraft 1, which passes through the vertical intersection in order Aircraft 2, Aircraft 3 waiting at the diagonal intersection, and Aircraft 4 in an emergency stop directly ahead of the taxiway. It can be seen that aircraft 1 has a solid safety zone (red zone), but no other waiting aircraft is detected and the safety zone does not overlap, no warning is issued, and aircraft 1 continues taxiing. Then, as shown in stage 2 of (b), the distance between aircraft 1 and aircraft 2 is less than the safety threshold 1. At this time, the dynamic safety zone (blue area) of aircraft 1 appears, but aircraft 2 is not captured by the two aircraft 1. A similar safety zone is detected, and no warning occurs. To the state shown in (c) stage 3, aircraft 2 is detected by the dynamic safety zone (blue area) of aircraft 1, and an early warning is issued to remind the controller to pay attention to the status of the aircraft at the designated location; there is an intersection in front of the pilot of aircraft 1 Do not turn right while waiting for the aircraft; at the same time, remind the pilot of aircraft 2 that an aircraft will pass through the fork in front of the road, and do not enter the intersection to prevent further escalation of the conflict. When the state in (d) stage 4 is reached, aircraft 1 has passed the center line of the vertical intersection, and it is judged that it has no turning ability, and the dynamic safety zone (blue area) of aircraft 1 pointing to aircraft 2 disappears, so the warning disappears. (e) The state shown in the figure of stage 5 is that the distance between aircraft 1 and aircraft 3 is less than the safety threshold 1, and the dynamic safety zone (blue area) of aircraft 1 pointing to aircraft 3 appears, but aircraft 3 is not detected, and no notification is issued. early warning. After (f) stage 6, aircraft 1 points to aircraft 3 in the dynamic safety zone (blue area) and detects aircraft 3, and the distance to aircraft 4 is less than the safety threshold 1, and the dynamic safety area (blue area) pointing to aircraft 4 appears But the aircraft 4 is not detected, and the aircraft 1 points to the dynamic safety zone (blue area) of the aircraft 3 to issue an early warning, reminding the pilot of the aircraft 1 that there is a waiting aircraft on the taxiway on the left side of the intersection ahead. At the same time, the solid-state safety zone (red zone) of aircraft 1 detects aircraft 4 and sends an early warning to the controller to remind the pilot of aircraft 1 to brake in time and there is an aircraft ahead. In order to better explain the linkage early warning of the safety zone, we assume that the pilot of aircraft 1 continues to drive until the state shown in stage 7 of figure (g) is reached. The dynamic safety area (blue area) pointing to aircraft 4 detects aircraft 4, and at the same time, the solid safety area (red area) of aircraft 1 points to aircraft 4 detects aircraft 4. At this moment, an alarm is issued instead of an early warning, notifying the controller and aircraft 1, The pilot of Airplane 4, Airplane 1 and Airplane 4 are about to collide! If aircraft 1 continues to drive, as shown in (h) stage 8, aircraft 1 points to the dynamic safety zone (blue area) of aircraft 3 and continuously issues warnings, and aircraft 1 points to the safety zone of aircraft 3 to continuously issue warnings, and the two aircraft collide . The algorithm flow chart of the conflict early warning module 512 between specific active aircraft and stationary aircraft is shown in Figure 12 below:

In order to build a multi-level and three-dimensional aircraft early warning system for airport scenes, the present invention considers follow-up on taxiways, confrontation conflicts, and cross conflicts on cross taxiways. For this reason, two types of safety early warning areas are constructed. However, in the actual airport taxiway operation system, there are still complex taxiway layouts composed of simple intersections, which will cause temporary traffic jams although they will not cause taxiway conflicts. First, the compound taxiway is defined as: two intersections connected by a taxiway, and the length of the taxiway connecting the intersection in the middle is greater than the safety threshold 1, that is, the dynamic safety area designation scheme cannot be used for early warning and the solid safety area cannot be used. Taxiway topology for early warning. Combined with common airport layouts, the following 12 common compound taxiways are listed, as shown in Figure 13 below.

The types of traffic congestion caused by the 12 compound taxiways in Figure 13 are consistent, so one of the taxiway congestion will be described in detail. As shown in (a), two aircrafts drive from two parallel taxiways to the middle taxiway respectively. If the two aircraft continue to enter the middle taxiway, although the solid safety zone can be used to give timely warnings, the two aircrafts will not be killed. A collision event, but at this time the intermediate taxiway connecting the two intersections is blocked. Since aircraft generally do not have the function of reversing, they must wait for the tractor to drag them out of the blocked road. This process takes at least 20 minutes. During this period, the surrounding taxiways connected to the blocked taxiway cannot be used, causing the congestion to spread and seriously affecting the airport. of normal operation. In addition, before the complex taxiway blockage event occurred, neither of the two types of safety areas could provide timely warnings for such situations. Therefore, in view of this situation, the early warning model needs to be further improved. The schematic diagram of the early warning module 513 of the airport compound taxiway blockage early warning is shown in Figure 14 .

It is calculated as follows:

H = distance between two intersection points + fuselage length + 50 meters,

d(t) = the distance from the midpoint of the aircraft to its intersection point at this moment,

v(t)=the speed of aircraft 1 at this moment,

estimated time

The aircraft is identified through the deep learning subsystem 42 established above, and the coordinates, speed, direction and instructions of the aircraft can be obtained in real time through the communication network subsystem 7. When the solid safety zone (red zone) of the aircraft detects that there is an intersection ahead , it will automatically monitor whether there are other intersections detected by other aircraft at this moment, and detect whether the two intersections are on the same taxiway. If there is no intersection between the two intersections, a composite taxiway structure can be formed. At this time, it is determined whether the driving intentions of the two aircraft conflict. The low aircraft will be warned and the estimated waiting time will be given. The specific algorithm flow of the airport compound taxiway block early warning module 513 is shown in Figure 15 below.

Through the cloud computing subsystem 5, the visualization of aircraft collision risk assessment on the airport scene can be realized, and the output of the solid safety area and dynamic safety area information of the aircraft on the scene, the early warning information of the aircraft on the scene and the composite taxiway block early warning information are transmitted through the communication network subsystem 7 To data storage subsystem 2.

Six Airport Scene Surveillance Terminals 6

As shown in Figure 16, the airport scene monitoring terminal 6 includes a display module 61 and an early warning module 62, and the said display module 61 receives the airport scene map information and the airport scene aircraft position information from the data storage subsystem 2 through the communication network subsystem 7 , Airport scene aircraft safety area information and early warning prompt information. Said early warning module 62 sends out " early warning " sound based on airport surface aircraft early warning information to remind controller to pay attention to the scene designated area.

As shown in FIG. 17 , the video screenshot of a local monitoring terminal of an airport scene shown in the display module 61 shows that two aircrafts are heading towards the same taxiway intersection at the same time in the figure, forming a cross conflict. As the two planes gradually approached the intersection, the solid safety zone (red area) of the two planes approached and overlapped near the intersection, and the system issued an early warning. If they continue to approach each other, they are detected by the other party's dynamic safety zone (blue zone), and the system will send out an alarm "warning!", indicating that the aircraft is about to collide.

As shown in Figure 18 video screenshot of the composite taxiway congestion early warning monitoring terminal, the two planes are heading towards the middle transverse taxiway from different directions. The aircraft in the lower corner also detects the intersection. At this time, the system assigns priority according to "first come, first served". In the scene monitoring window in the lower right corner, the aircraft in the lower right corner issues an instruction of "aircraft congestion warning, please brake immediately! Wait at least 35 seconds". Among them, 35 seconds is the approximate waiting time calculated by the composite taxiway block pre-warning module 513 of the present invention.

7 Communication network subsystem 7

The communication network subsystem 7 is stably connected with each subsystem and subsystem to complete the data transmission in the system.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical scheme of the present invention, rather than limiting it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: It is still possible to modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the technical solutions of the various embodiments of the present invention. All of them should be covered by the scope of the claims and description of the present invention.

The above are the preferred embodiments of the present invention, and all changes made according to the technical solution of the present invention, when the functional effect produced does not exceed the scope of the technical solution of the present invention, all belong to the protection scope of the present invention.

Claims (7)

1. A video image-based airport scene aircraft conflict early warning method, is characterized in that, comprises the following steps: Step 1: Introduce and collect relevant information on the airport surface situation that supports the operation of each system, and provide the required raw data for the early warning of aircraft conflicts on the airport surface; Step 2: The original data obtained in step 1 and the current scene visibility data are transmitted to the image defogging and noise reduction subsystem of the edge computing subsystem; the image defogging and noise reduction subsystem generates a clear real-time monitoring video image of the airport scene; Step 3: The real-time monitoring video image of the airport scene obtained in step 2 is transmitted to the deep learning subsystem of the edge computing subsystem; the deep learning subsystem will identify and track common targets on the scene; Step 4: The pixel coordinates obtained in step 3 and identified by the aircraft on the scene surveillance video image are transmitted to the coordinate transformation subsystem of the edge computing subsystem; the coordinate transformation subsystem will obtain the current position of all aircraft on the scene in the airport Location; Step 5: Upload the current position and current time of all aircraft on the scene obtained in step 4 to the scene aircraft conflict warning subsystem of the cloud computing subsystem; Combined with the actual operating conditions of aircraft at the airport, and based on the aircraft safety threshold model, a dynamic safety area is set for the aircraft; through two types of safety areas, common conflicts between aircraft on the scene are warned; through the compound taxiway congestion model, compound taxi Early warning of road blockage events; Step 6: Transmit the two types of safety zones and early warning information set in step 5 to the terminal scene monitoring screen. 2. a kind of airport scene aircraft conflict early warning method based on video image as claimed in claim 1, is characterized in that, in step 2, according to airport scene visibility value adaptation airport scene real-time monitoring video image defog noise reduction processing. 3. a kind of airport scene aircraft conflict early warning method based on video image as claimed in claim 2, it is characterized in that, in step 3, described deep learning subsystem needs to carry out image acquisition and the target object in the actual operation of this airport earlier Training; and after the deep learning subsystem is put into use, continue to collect the image data of the target object at the airport for enhanced training to improve the accuracy and stability of the system. 4. a kind of airport scene aircraft conflict early-warning method based on video image as claimed in claim 3, is characterized in that, in step 5, will judge the state of motion of aircraft by the last stage position of each aircraft and current position data, calculate The magnitude and direction of the speed of the aircraft; for moving aircraft, the safety zone shall be established according to the design plan. 5. a kind of airport scene aircraft conflict early warning method based on video images as claimed in claim 4, is characterized in that, in step 5, early warning is divided into three levels: 1) Warning, mark the solid safety zone for the moving aircraft according to the aircraft operating parameters and airport operating regulations in red, and set the dynamic safety zone by calculating the relative orientation of two aircraft, aircraft operating parameters, airport operating regulations, and aircraft safety thresholds Marked in blue, it is convenient for the controller to control the aircraft operating status and risk assessment in the dynamic safety zone; 2) Early warning, detailed warning information will be issued when the designated solid safety zone and dynamic safety zone early warning conditions are reached; for compound taxiways, early warning will be issued according to the blocking early warning algorithm; at this moment, the controller is reminded to pay attention to the designated aircraft in time, for the control Instructions issued by the staff for reference; 3) Alarm. For detailed alarm information when the designated solid-state safety zone and dynamic security zone alarm conditions are reached, the system will issue a warning by flashing the warning light and beeping. 6. a kind of airport scene aircraft conflict early warning method based on video image as claimed in claim 5 is characterized in that, in step 6, will display detailed early warning information by display module in airport scene monitoring terminal equipment. 7. A video image-based airport scene aircraft conflict early warning system is characterized in that it comprises: The edge computing subsystem includes the distributed field surveillance camera group for airport scene layout; the distributed field surveillance camera group includes cameras and embedded platform devices, which are divided into scene surveillance camera module, image defogging and noise reduction subsystem, and deep learning subsystem and a coordinate transformation subsystem, the embedded platform device will perform defog and noise reduction processing on the scene surveillance video images collected by this set of distributed field surveillance cameras in real time, recognize the airport scene target based on deep learning, and obtain the scene through the coordinate transformation algorithm the position of the aircraft on the airport scene; The cloud computing subsystem includes a surface aircraft conflict early warning subsystem, a surface aircraft solid safety zone designation module, and a surface aircraft dynamic security zone designation module; the surface aircraft conflict early warning subsystem includes a conflict early warning module between active aircraft, active aircraft and The conflict early warning module between static aircraft and the airport composite taxiway block early warning module; the conflict early warning module between active aircraft is used for conflict early warning between two moving aircraft on the airport scene; between active aircraft and stationary aircraft The conflict warning module is used for the conflict warning between the moving aircraft and the stationary or waiting aircraft on the airport surface; the airport compound taxiway block warning module is used for two intersections connected by a taxiway, and the intersection is connected in the middle The length of the taxiway is greater than the detection area of the dynamic safety zone, and it is impossible to use the dynamic safety zone designation scheme for early warning or the solid safety zone for early warning of congestion warning on the taxiway topology; Communication network subsystem: used to ensure the stable connection of each subsystem and subsystem, and complete the data transmission in the system; Time service sub-system: a time service system based on IP data network; Data storage subsystem: including surface aircraft position and status data storage module, surface aircraft solid safety area storage module, surface aircraft static safety area storage module, surface taxiway structure information storage module, and airport surface meteorological data storage module; Use various types of data for classified storage; Airport scene monitoring terminal: including a display module and an alarm module; the display module is used to display airport scene map information, airport scene aircraft position information, airport scene aircraft safety zone information and airport scene aircraft early warning information; the early warning module uses sound warning equipment, when the airport When the scene monitoring terminal receives the early warning information sent by the cloud computing subsystem, it sends out a warning; Airport data acquisition system: connected to the deployed airport weather station to receive the visibility data of the airport; also connected to the deployed airport operation control center to receive the aircraft take-off and landing time and aircraft priority data on the airport scene; also connected to To the deployed airport management center to receive coordinate data of taxiway intersections on the airport scene.

Images