A real-time identification method for traffic operation risks based on low-altitude drones

Abstract

The present invention discloses a real-time identification method for traffic operation risk based on low-altitude UAVs, which relates to the technical field of UAV traffic monitoring, including the dispatch center assigning numbers and working areas to low-altitude UAVs, collecting UAV working environment data, analyzing the working hours of low-altitude UAVs and arranging the working schedule of low-altitude UAVs; low-altitude UAVs collect traffic data through cameras, generate multi-mode traffic data OD matrix, and analyze traffic behavior based on the multi-mode traffic data OD matrix to detect traffic accidents; UAVs implement treatment measures according to traffic accident detection results; low-altitude UAVs send traffic data to the dispatch center in real time for display, storage and real-time upload to the cloud. The present invention effectively improves the safety of UAV operation, improves the working efficiency of UAVs, and improves the accuracy of traffic accident analysis and the ability to respond to accidents in real time.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) traffic monitoring, and in particular to a real-time identification method for traffic operation risks based on low-altitude UAVs. Background Art

[0002] In recent years, low-altitude UAV technology has made significant progress in the field of traffic management. With the maturity of UAV technology and the expansion of application scenarios, low-altitude UAVs have gradually been used in real-time traffic monitoring, traffic accident detection and emergency response. UAVs have the advantages of rapid response and efficient patrol, and can provide timely traffic information in areas that are difficult to cover by traditional traffic monitoring systems. The progress of these technologies is due to the improvement of UAV flight control, sensor technology, communication technology and data processing capabilities. However, the existing low-altitude UAV traffic monitoring system still has shortcomings. It is not intelligent and flexible enough in the real-time management of UAVs, and the data utilization rate is insufficient, resulting in low UAV operation efficiency and low accuracy of traffic risk monitoring results. Summary of the invention

[0003] In view of the problems existing in the above-mentioned existing real-time identification method of traffic operation risks based on low-altitude drones, the present invention is proposed.

[0004] Therefore, the problem to be solved by the present invention is that the real-time management of drones is not intelligent and flexible enough, and the data utilization rate is insufficient, resulting in low operating efficiency of drones and low accuracy of traffic risk monitoring results.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a real-time identification method for traffic operation risks based on low-altitude UAVs, which includes: a dispatch center assigning numbers and working areas to low-altitude UAVs, collecting UAV working environment data, analyzing the working hours of low-altitude UAVs and arranging a working schedule for low-altitude UAVs;

[0006] Low-altitude drones collect traffic data through cameras, generate multi-modal traffic data OD matrix, analyze traffic behavior based on the multi-modal traffic data OD matrix, and detect traffic accidents;

[0007] The drone implements treatment measures based on the traffic accident detection results;

[0008] Low-altitude drones send traffic data to the dispatch center in real time for display, storage, and real-time upload to the cloud.

[0009] As a preferred solution of the real-time identification method of traffic operation risks based on low-altitude drones described in the present invention, the dispatch center assigns numbers and working areas to low-altitude drones, collects drone working environment data, analyzes the working hours of low-altitude drones, and arranges a working schedule for low-altitude drones, including:

[0010] The dispatch center assigns a unique FAA registration number to each low-altitude drone, encodes the number into a binary sequence in ASCII format, divides the work area according to the geographic information system, specifies the work area for each low-altitude drone, and binds the low-altitude drone FAA registration number to the work area;

[0011] When the low-altitude UAV flies in the working area, it collects and pre-processes the working area environmental data through sensors, and the working area environmental data includes temperature data, humidity data, wind speed data and air pressure data;

[0012] Construct a support vector machine (SVM) model, define the model input format as the working area environment data, and the model output as the predicted UAV flight time. Use the training set to train the model, define the cross loss function and Adam optimizer to iteratively optimize the model parameters to obtain a trained support vector machine (SVM) model.

[0013] Input the collected UAV working area environmental data into the trained support vector machine SVM model to obtain the predicted UAV flight time t;

[0014] Based on the working area environment data content, the Monte Carlo simulation method is used to generate uncertainty scenarios, and the uncertainty scenarios are brought into the trained support vector machine (SVM) model to obtain the corresponding predicted UAV flight time t k ;

[0015] Using uncertainty sets to predict drone flight time optimization, the uncertainty sets include polyhedron sets, box sets, and ellipse sets;

[0016] The polyhedron set is:

[0017]

[0018] Among them U p is a polyhedron set, t0 is the UAV flight time under the benchmark environment data, y s is the weight of the sth uncertainty scenario, is the predicted UAV flight time of the sth uncertainty scenario, Γ is the uncertainty parameter, and S is the number of uncertainty scenarios;

[0019] The box set is:

[0020]

[0021] The ellipse set is:

[0022]

[0023] The weights and uncertainty parameters are obtained based on the sensitivity analysis method, and the maximum flight time reduction is calculated based on the polyhedron set, box set and ellipse set. as well as And select the maximum value as the maximum flight time reduction t z ;

[0024] By predicting the UAV flight time t and the maximum flight time reduction t z Calculate the drone flight time T:

[0025] T=tt z ;

[0026] Calculate the number of drones required in the work area daily based on the drone flight time to develop a drone work schedule, and record the FAA registration numbers of all drones in each work area.

[0027] As a preferred solution of the real-time identification method of traffic operation risks based on low-altitude drones described in the present invention, the drone records the flight time in real time when flying in the working area, and marks the drone position as the starting point and the dispatch center position as the end point according to the geographic information system, and identifies the obstacle position, plans the return route of the drone using a path planning algorithm, and calculates the return time according to the return route and the flight speed;

[0028] When the drone detects that the remaining flight time is twice the return time, it automatically sends a return notification and return route to the dispatch center. The dispatch center dispatches a replacement drone to the work area according to the return route. After the replacement drone arrives at the work area, the drone automatically returns to the dispatch center according to the return route.

[0029] As a preferred solution of the real-time identification method of traffic operation risks based on low-altitude drones described in the present invention, wherein: the low-altitude drone collects traffic data through a camera, and generates a multi-mode traffic data OD matrix, which means that the low-altitude drone collects road traffic data through a camera when flying in the working area, uses the YOLO model to perform real-time detection on the road traffic data, identifies traffic objects, and uses a multi-target tracking algorithm to track the detected traffic objects to generate traffic trajectories;

[0030] Divide the working area into sub-areas by gridding, count the number of traffic objects in each sub-area, and calculate the traffic flow between each pair of sub-areas:

[0031]

[0032] OD ij is the traffic flow from sub-area i to sub-area j, N is the number of traffic objects, k is the traffic object, δ i (k) and δ j(k) are the occurrence probabilities of traffic object k in sub-area i to sub-area j, where the occurrence is 1 and the non-appearance is 0;

[0033] The traffic flows among all sub-regions are calculated and mapped into the sub-regions to form a comprehensive multi-modal traffic data OD matrix.

[0034] As a preferred solution of the real-time identification method of traffic operation risks based on low-altitude drones described in the present invention, wherein: the analysis of traffic behavior based on the multi-modal traffic data OD matrix and the detection of traffic accidents refer to extracting traffic flow data from the multi-modal traffic data OD matrix and extracting traffic features from the traffic flow data using feature engineering;

[0035] Use the ARIMA model to analyze traffic flow data and predict future traffic flow, and extract future traffic characteristics from future traffic flow;

[0036] The random forest algorithm is used to detect traffic flow anomalies based on traffic characteristics and future traffic characteristics, and the detected traffic flow anomalies are treated as traffic accidents. The drone automatically sends traffic flow anomalies and traffic accidents to the dispatch center for recording.

[0037] As a preferred solution of the real-time identification method of traffic operation risks based on low-altitude drones described in the present invention, wherein: the drone implements processing measures according to the traffic accident detection results, which means that the drone automatically reports to the dispatch center when a traffic accident is detected, flies to the air above the traffic accident location and hovers, and automatically identifies the traffic objects in the traffic accident, sends voice inquiries to the traffic objects and collects the voice data of the traffic objects. If the voice data of the traffic objects is collected, the semantics of the traffic objects are identified through natural language technology to assist in traffic accident processing; if the voice data of the traffic objects is not collected, the emergency number is automatically dialed to notify the emergency personnel of the location data of the traffic accident area and the traffic object information, and simultaneously sends voice and flash alarms to remind other traffic objects to avoid the traffic accident area.

[0038] As a preferred solution of the real-time identification method of traffic operation risks based on low-altitude drones described in the present invention, wherein: the low-altitude drone sends traffic data to the dispatching center for display in real time, which means that the low-altitude drone sends the traffic data and analysis results to the dispatching center in real time after collecting traffic data and analyzing traffic objects, and the dispatching center displays the traffic data and analysis results in real time according to the working area where the low-altitude drone is located.

[0039] As a preferred solution of the real-time identification method of traffic operation risks based on low-altitude drones described in the present invention, the storage and real-time uploading to the cloud refers to the drone storing the collected traffic data through the database and uploading it to the cloud synchronously. When the drone returns to the dispatch center, the drone will compare the stored traffic data with the traffic data received by the dispatch center for consistency, and generate a detection report after the comparison for storage.

[0040] A computer device comprises: a memory and a processor; the memory stores a computer program, and the processor implements the steps of the above-mentioned method for real-time identification of traffic operation risks based on low-altitude unmanned aerial vehicles when executing the computer program.

[0041] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned method for real-time identification of traffic operation risks based on low-altitude unmanned aerial vehicles.

[0042] The beneficial effects of the present invention are as follows: the present invention effectively improves the safety of drone operation and the work efficiency of drones by allocating working areas to drones and collecting environmental data in real time to analyze the flight time of drones, planning the return routes of drones and performing drone rotation. In addition, by generating a multi-mode traffic data OD matrix and analyzing traffic accidents using machine learning and data mining technology, the accuracy of traffic accident analysis and the real-time response capability of accidents are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0044] Figure 1 The figure is a flowchart of the real-time identification method of traffic operation risks based on low-altitude UAVs. DETAILED DESCRIPTION

[0045] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0046] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0047] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

[0048] Example 1

[0049] Reference Figure 1 , which is the first embodiment of the present invention, and which provides a real-time identification method for traffic operation risks based on low-altitude UAVs. The real-time identification method for traffic operation risks based on low-altitude UAVs includes:

[0050] S1. The dispatch center assigns numbers and work areas to low-altitude drones, collects drone working environment data, analyzes the working hours of low-altitude drones and arranges the working schedule of low-altitude drones;

[0051] Specifically, the dispatch center assigns numbers and work areas to low-altitude drones, collects drone working environment data, analyzes the working hours of low-altitude drones, and arranges the working schedule of low-altitude drones, including:

[0052] The dispatch center assigns a unique FAA registration number to each low-altitude drone, encodes the number into a binary sequence in ASCII format, divides the work area according to the geographic information system, specifies the work area for each low-altitude drone, and binds the low-altitude drone FAA registration number to the work area;

[0053] When the low-altitude UAV flies in the working area, it collects and pre-processes the working area environmental data through sensors, and the working area environmental data includes temperature data, humidity data, wind speed data and air pressure data;

[0054] Construct a support vector machine (SVM) model, define the model input format as the working area environment data, and the model output as the predicted UAV flight time. Use the training set to train the model, define the cross loss function and Adam optimizer to iteratively optimize the model parameters to obtain a trained support vector machine (SVM) model.

[0055] Input the collected UAV working area environmental data into the trained support vector machine SVM model to obtain the predicted UAV flight time t;

[0056] Based on the working area environment data content, the Monte Carlo simulation method is used to generate uncertainty scenarios, and the uncertainty scenarios are brought into the trained support vector machine (SVM) model to obtain the corresponding predicted UAV flight time t k ;

[0057] Using uncertainty sets to predict drone flight time optimization, the uncertainty sets include polyhedron sets, box sets, and ellipse sets;

[0058] The polyhedron set is:

[0059]

[0060] Among them U p is a polyhedron set, t0 is the UAV flight time under the benchmark environment data, y s is the weight of the sth uncertainty scenario, is the predicted UAV flight time of the sth uncertainty scenario, Γ is the uncertainty parameter, and S is the number of uncertainty scenarios;

[0061] The box set is:

[0062]

[0063]

[0064]

[0065]

[0066] The ellipse set is:

[0067] The weights and uncertainty parameters are obtained based on the sensitivity analysis method, and the maximum flight time reduction is calculated based on the polyhedron set, box set and ellipse set. as well as And select the maximum value as the maximum flight time reduction t z ;

[0068] By predicting the UAV flight time t and the maximum flight time reduction t z Calculate the drone flight time T:

[0069] T=tt z ;

[0070] Calculate the number of drones required in the work area daily based on the drone flight time to develop a drone work schedule, and record the FAA registration numbers of all drones in each work area.

[0071] By assigning a unique FAA registration number and a clear working area, the operation of each drone can be effectively managed and monitored, avoiding conflicts between drones in the air and ensuring that the area for mission execution is clear. Environmental data is collected in real time through sensors and preprocessed to ensure the accuracy and consistency of the data. These environmental data are important inputs for predicting drone flight time, which improves the reliability of the prediction. The use of support vector machine models can effectively process multidimensional environmental data and predict the flight time of drones. By defining a cross loss function and using the Adam optimizer to optimize model parameters, the accuracy and robustness of the model can be improved. The use of Monte Carlo simulation methods to generate uncertainty scenarios can comprehensively evaluate the changes in the flight time of drones under different environmental conditions. Through optimization of uncertainty sets, the maximum possible reduction in flight time can be calculated to ensure the robustness of the prediction results. According to the predicted flight time and the actual needs of the working area, a scientific and reasonable drone work schedule is formulated to ensure that each working area has enough drones to perform tasks and improve the efficiency and accuracy of task completion. Formulating a drone work schedule can optimize the allocation of drone resources, avoid resource waste, and ensure the smooth progress of drone missions in various working areas.

[0072] Furthermore, the UAV records the flight time in real time when flying in the work area, and marks the UAV location as the starting point and the dispatch center location as the end point according to the geographic information system, identifies the location of obstacles, plans the return route of the UAV using a path planning algorithm, and calculates the return time based on the return route and flight speed;

[0073] When the drone detects that the remaining flight time is twice the return time, it automatically sends a return notification and return route to the dispatch center. The dispatch center dispatches a replacement drone to the work area according to the return route. After the replacement drone arrives at the work area, the drone automatically returns to the dispatch center according to the return route.

[0074] When the drone flies in the work area, it records the flight time in real time, and marks the drone position and the dispatch center position according to the GIS system to ensure accurate tracking of the flight path. Through the path planning algorithm, the starting point, end point and obstacle positions are considered to plan the best route for the drone to return, ensuring that the drone can return to the dispatch center safely and efficiently. When the drone detects that the remaining flight time is twice the return time, it automatically sends a return notification and return route to the dispatch center to ensure that the dispatch center can arrange a replacement drone in time to avoid mission interruption. The dispatch center dispatches a replacement drone to the work area according to the return route to ensure the continuity of the mission and the efficient use of the drone. After the replacement drone arrives at the work area, the original drone automatically returns to the dispatch center according to the planned return route to ensure the automation and efficiency of the return process. The automatic return of the drone can reduce human intervention, improve the automation level of drone operation, ensure that the drone returns safely before the battery is exhausted, and improve the drone's endurance and mission completion rate.

[0075] S2, low-altitude drones collect traffic data through cameras, generate multi-modal traffic data OD matrix, and analyze traffic behavior based on the multi-modal traffic data OD matrix to detect traffic accidents;

[0076] Specifically, low-altitude drones collect traffic data through cameras and generate multi-modal traffic data OD matrix refers to low-altitude drones collecting road traffic data through cameras when flying in the working area, using the YOLO model to detect road traffic data in real time, identify traffic objects, and use multi-target tracking algorithms to track detected traffic objects to generate traffic trajectories;

[0077] Divide the working area into sub-areas by gridding, count the number of traffic objects in each sub-area, and calculate the traffic flow between each pair of sub-areas:

[0078]

[0079] OD ij is the traffic flow from sub-area i to sub-area j, N is the number of traffic objects, k is the traffic object, δ i (k) and δ j (k) are the occurrence probabilities of traffic object k in sub-area i to sub-area j, where the occurrence is 1 and the non-appearance is 0;

[0080] The traffic flows among all sub-regions are calculated and mapped into the sub-regions to form a comprehensive multi-modal traffic data OD matrix.

[0081] The YOLO model can complete target detection in a single forward propagation process, with extremely high real-time performance and detection accuracy. It can identify a variety of traffic objects, such as vehicles, pedestrians, bicycles, etc. The multi-target tracking algorithm can track the detected traffic objects in real time and generate the motion trajectory of each object. These trajectory data are an important basis for analyzing traffic behavior and patterns. Gridding the working area can refine the statistics and analysis of traffic data, making the traffic flow statistics in each sub-area more accurate. Gridding makes the statistics of traffic flow and the generation of OD matrix more refined, which helps to deeply analyze the traffic flow and behavior patterns between different areas. By calculating the traffic flow between each sub-area, an OD matrix reflecting the traffic behavior between different areas can be generated. The OD matrix can fully display the flow of multi-modal traffic data. The OD matrix is ​​an important tool for traffic behavior analysis and prediction. By analyzing the OD matrix, traffic congestion points can be identified, traffic management strategies can be optimized, and the overall efficiency of the transportation system can be improved.

[0082] Furthermore, analyzing traffic behaviors based on the OD matrix of multimodal traffic data and detecting traffic accidents refers to extracting traffic flow data from the OD matrix of multimodal traffic data and extracting traffic features from the traffic flow data using feature engineering;

[0083] Use the ARIMA model to analyze traffic flow data and predict future traffic flow, and extract future traffic characteristics from future traffic flow;

[0084] The random forest algorithm is used to detect traffic flow anomalies based on traffic characteristics and future traffic characteristics, and the detected traffic flow anomalies are treated as traffic accidents. The drone automatically sends traffic flow anomalies and traffic accidents to the dispatch center for recording.

[0085] Extracting traffic flow data from the multimodal traffic data OD matrix and extracting meaningful traffic features through feature engineering can help to more accurately analyze traffic behavior and detect anomalies. Traffic features include flow mean, variance, peak hours, etc., which can provide strong support for subsequent traffic flow prediction and anomaly detection. The ARIMA model can perform time series analysis on traffic flow, capture the trend and seasonal changes of traffic flow, and predict future traffic flow. By predicting future traffic flow, potential traffic problems can be identified in advance, traffic management and scheduling strategies can be optimized, and the operating efficiency of the traffic system can be improved. The random forest algorithm can detect anomalies in traffic flow based on the extracted traffic features and future traffic features, and identify these anomalies as traffic accidents. The detected traffic flow anomalies can be used as early warning signals for traffic accidents, helping the dispatch center to take timely measures to reduce the occurrence of traffic accidents and ensure road traffic safety. The drone can automatically send the detected traffic flow anomalies and traffic accident information to the dispatch center to ensure the timely transmission and recording of information. Through automated information transmission and recording, the response speed of the dispatch center can be improved, ensuring that traffic accidents are handled in a timely manner and reducing the impact of traffic accidents on traffic flow.

[0086] S3. The drone implements treatment measures based on the traffic accident detection results;

[0087] Specifically, the drone implements handling measures based on the traffic accident detection results, which means that the drone automatically reports the traffic accident to the dispatch center upon detection, flies to the location of the traffic accident and hovers, automatically identifies the traffic objects in the traffic accident, sends voice inquiries to the traffic objects and collects the voice data of the traffic objects. If the voice data of the traffic objects is collected, the traffic object semantics are identified through natural language technology to assist in traffic accident handling; if the voice data of the traffic objects is not collected, the emergency number is automatically dialed to notify the emergency personnel of the location data and traffic object information of the traffic accident area, and voice and flash alarms are simultaneously issued to remind other traffic objects to avoid the traffic accident area.

[0088] After detecting a traffic accident, the drone can immediately report the accident to the dispatch center to ensure the timely transmission and recording of accident information. The automatic reporting function can improve the response speed of traffic accident handling, ensure that the dispatch center can understand the accident situation in a timely manner, and arrange subsequent handling measures. The drone hovers above the location of the traffic accident, which can provide a stable perspective and data collection platform, which helps to fully understand the situation at the scene of the accident. Through the hovering of the drone, the dispatch center can view the video data of the accident scene in real time, providing an intuitive basis for decision-making. The drone can automatically identify traffic objects in the accident, including vehicles, pedestrians, etc., to improve the accuracy of information collection at the accident scene. The automatic identification function can help the drone quickly lock the accident scene. The UAV can identify key objects of traffic, avoid missing important information, and provide comprehensive data support for accident handling. By sending voice inquiries to traffic objects, the UAV can obtain real-time feedback from traffic objects and understand their injuries and needs. The voice inquiry function can improve the efficiency of information collection at the accident scene and help the dispatch center make emergency decisions quickly. The natural language processing technology is used to identify the voice data of traffic objects, and the content expressed can be accurately understood to assist the UAV in accident handling. When the UAV fails to obtain the voice data of the traffic object, it automatically calls the emergency number to ensure the timeliness and accuracy of emergency rescue. By issuing voice and flash alarms, the UAV can remind other traffic objects to avoid the accident area and reduce the occurrence of secondary accidents.

[0089] S4. Low-altitude drones send traffic data to the dispatch center in real time for display, storage, and real-time upload to the cloud.

[0090] Specifically, the low-altitude UAV sends the traffic data to the dispatch center for display in real time, which means that the low-altitude UAV sends the traffic data and analysis results to the dispatch center in real time after collecting traffic data and analyzing traffic objects. The dispatch center displays the traffic data and analysis results in real time according to the working area where the low-altitude UAV is located.

[0091] After collecting traffic data and analyzing traffic objects, low-altitude drones can send data and analysis results to the dispatch center in real time. This real-time nature ensures that the traffic management department can grasp the current traffic conditions in a timely manner. The real-time transmission function can significantly improve the efficiency of traffic monitoring, ensuring that the dispatch center can respond quickly to traffic changes and optimize traffic management decisions. The dispatch center displays traffic data and analysis results in real time according to the working area of ​​the low-altitude drone to ensure the accuracy and regionality of the information display. The regional display function can help traffic management personnel intuitively understand the traffic conditions of each working area, facilitate the adoption of corresponding management measures for different areas, and improve the accuracy of overall traffic management.

[0092] Furthermore, storage and real-time uploading to the cloud means that the drone stores the collected traffic data through the database and uploads it to the cloud simultaneously. When the drone returns to the dispatch center, the drone will compare the stored traffic data with the traffic data received by the dispatch center for consistency, and generate a detection report for storage after the comparison.

[0093] During the flight, the drone will store the collected traffic data in the local database to ensure the real-time and integrity of the data. Local database storage can prevent data loss during transmission and ensure high availability and reliability of the data. The drone will upload the collected traffic data to the cloud in real time to ensure the timeliness and accessibility of the data. Through cloud storage, traffic data can be accessed and analyzed by multiple users and systems in real time, improving the utilization and value of the data. When the drone returns to the dispatch center, the locally stored data will be compared with the data received by the dispatch center to ensure the integrity and consistency of the data. After the consistency comparison is passed, the drone will generate a test report and store it to provide a basis for subsequent analysis and auditing of the data.

[0094] Example 2

[0095] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0096] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0097] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0098] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used to implement: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0099] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A real-time identification method for traffic operation risks based on low-altitude drones, characterized by: include, The dispatch center assigns numbers and work areas to low-altitude drones, collects drone working environment data, analyzes the working hours of low-altitude drones and arranges a work schedule for low-altitude drones; Low-altitude drones collect traffic data through cameras, generate multi-modal traffic data OD matrix, analyze traffic behavior based on the multi-modal traffic data OD matrix, and detect traffic accidents; The drone implements treatment measures based on the traffic accident detection results; Low-altitude drones send traffic data to the dispatch center in real time for display, storage, and real-time upload to the cloud; The dispatch center assigns numbers and work areas to low-altitude drones, collects drone working environment data, analyzes the working hours of low-altitude drones, and arranges a low-altitude drone work schedule, including: The dispatch center assigns a unique FAA registration number to each low-altitude drone, encodes the number into a binary sequence in ASCII format, divides the work area according to the geographic information system, specifies the work area for each low-altitude drone, and binds the low-altitude drone FAA registration number to the work area; When the low-altitude UAV flies in the working area, it collects and pre-processes the working area environmental data through sensors, and the working area environmental data includes temperature data, humidity data, wind speed data and air pressure data; Construct a support vector machine (SVM) model, define the model input format as the working area environment data, and the model output as the predicted UAV flight time. Use the training set to train the model, define the cross loss function and Adam optimizer to iteratively optimize the model parameters to obtain a trained support vector machine (SVM) model. Input the collected UAV working area environmental data into the trained support vector machine SVM model to obtain the predicted UAV flight time t; Based on the working area environment data content, the Monte Carlo simulation method is used to generate uncertainty scenarios, and the uncertainty scenarios are brought into the trained support vector machine (SVM) model to obtain the corresponding predicted UAV flight time t k ; Using uncertainty sets to predict drone flight time optimization, the uncertainty sets include polyhedron sets, box sets, and ellipse sets; The polyhedron set is: Among them U p is a polyhedron set, t0 is the UAV flight time under the benchmark environment data, y s is the weight of the sth uncertainty scenario, is the predicted UAV flight time of the sth uncertainty scenario, Γ is the uncertainty parameter, and S is the number of uncertainty scenarios; The box set is: The ellipse set is: The weights and uncertainty parameters are obtained based on the sensitivity analysis method, and the maximum flight time reduction is calculated based on the polyhedron set, box set and ellipse set. as well as And select the maximum value as the maximum flight time reduction t z ; By predicting the UAV flight time t and the maximum flight time reduction t z Calculate the drone flight time T: T=t-t z ; Calculate the number of drones required in the work area daily based on the drone flight time to develop a drone work schedule, and record the FAA registration numbers of all drones in each work area.

2. The method for real-time identification of traffic operation risks based on low-altitude unmanned aerial vehicles according to claim 1 is characterized in that: The UAV records the flight time in real time when flying in the working area, and marks the UAV position as the starting point and the dispatch center position as the end point according to the geographic information system, identifies the obstacle position, plans the UAV return route using a path planning algorithm, and calculates the return time based on the return route and flight speed; When the drone detects that the remaining flight time is twice the return time, it automatically sends a return notification and return route to the dispatch center. The dispatch center dispatches a replacement drone to the work area according to the return route. After the replacement drone arrives at the work area, the drone automatically returns to the dispatch center according to the return route.

3. The method for real-time identification of traffic operation risks based on low-altitude drones as claimed in claim 2 is characterized in that: The low-altitude UAV collects traffic data through a camera and generates a multi-mode traffic data OD matrix. The low-altitude UAV collects road traffic data through a camera when flying in the working area, uses the YOLO model to perform real-time detection on the road traffic data, identifies traffic objects, and uses a multi-target tracking algorithm to track the detected traffic objects to generate traffic trajectories; Divide the working area into sub-areas by gridding, count the number of traffic objects in each sub-area, and calculate the traffic flow between each pair of sub-areas: OD ij is the traffic flow from sub-area i to sub-area j, N is the number of traffic objects, k is the traffic object, δ i (k) and δ j (k) are the occurrence probabilities of traffic object k in sub-area i to sub-area j, where the occurrence is 1 and the non-appearance is 0; The traffic flows among all sub-regions are calculated and mapped into the sub-regions to form a comprehensive multi-modal traffic data OD matrix.

4. The method for real-time identification of traffic operation risks based on low-altitude unmanned aerial vehicles as claimed in claim 3 is characterized by: Analyzing traffic behavior based on the multi-modal traffic data OD matrix and detecting traffic accidents refers to extracting traffic flow data from the multi-modal traffic data OD matrix and extracting traffic features from the traffic flow data using feature engineering; Use the ARIMA model to analyze traffic flow data and predict future traffic flow, and extract future traffic characteristics from future traffic flow; The random forest algorithm is used to detect traffic flow anomalies based on traffic characteristics and future traffic characteristics, and the detected traffic flow anomalies are treated as traffic accidents. The drone automatically sends traffic flow anomalies and traffic accidents to the dispatch center for recording.

5. The method for real-time identification of traffic operation risks based on low-altitude unmanned aerial vehicles according to claim 4 is characterized in that: The drone implements handling measures according to the traffic accident detection results, which means that the drone automatically reports to the dispatch center when a traffic accident is detected, flies to the location of the traffic accident and hovers, automatically identifies the traffic objects in the traffic accident, sends voice inquiries to the traffic objects and collects the voice data of the traffic objects. If the voice data of the traffic objects is collected, the semantics of the traffic objects are identified through natural language technology to assist in traffic accident handling; if the voice data of the traffic objects is not collected, the emergency number is automatically dialed to inform the emergency personnel of the location data of the traffic accident area and the information of the traffic objects, and simultaneously sends out voice and flash alarms to remind other traffic objects to avoid the traffic accident area.

6. The method for real-time identification of traffic operation risks based on low-altitude unmanned aerial vehicles according to claim 5, characterized in that: The low-altitude UAV sends the traffic data to the dispatch center for display in real time, which means that the low-altitude UAV sends the traffic data and analysis results to the dispatch center in real time after collecting traffic data and analyzing traffic objects. The dispatch center displays the traffic data and analysis results in real time according to the working area of ​​the low-altitude UAV.

7. The method for real-time identification of traffic operation risks based on low-altitude unmanned aerial vehicles according to claim 6 is characterized in that: The storage and real-time uploading to the cloud refers to the drone storing the collected traffic data through the database and uploading it to the cloud simultaneously. When the drone returns to the dispatch center, the drone will compare the stored traffic data with the traffic data received by the dispatch center for consistency, and generate a detection report for storage after the comparison.

8. A computer device comprising: Memory and processor; The memory stores a computer program, which is characterized in that when the processor executes the computer program, the steps of the real-time identification method of traffic operation risks based on low-altitude unmanned aerial vehicles described in any one of claims 1 to 7 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the real-time identification method of traffic operation risks based on low-altitude unmanned aerial vehicles described in any one of claims 1 to 7 are implemented.

Images