An optimization method, device and equipment for the inspection path of an unmanned aerial vehicle group in a chemical industrial park

Abstract

The present application provides an optimization method, device and equipment for the inspection path of an unmanned aerial vehicle (UAV) group in a chemical industrial park, which relates to the technical field of UAV inspection and includes: obtaining the environmental data of the target chemical industrial park, as well as the attributes and initial inspection paths of each UAV; for any UAV, obtaining the real-time monitoring values of each inspection position and the corresponding inspection position on the initial inspection path; for any inspection target responsible for the UAV, obtaining the weight when the UAV reaches the inspection target; inputting each inspection position and the corresponding real-time monitoring value, the target position and weight of the inspection target, the environmental data, and the initial inspection path into a pre-constructed multi-objective optimization model to calculate the score of the initial inspection path; determining the optimal inspection path of the UAV by using dynamic neighborhood search according to the weight, target position, environmental data, attributes, and the initial inspection paths and positions of each UAV; the present application can generate an optimal inspection path that takes into account efficiency, safety, high quality and high coverage rate.

Description

Technical Field

[0001] This application relates to the technical field of drone inspection, and more particularly, to an optimization method, device, and equipment for the inspection path of a drone swarm in a chemical industrial park. Background Art

[0002] As a concentrated area of hazardous chemicals, the safety management of chemical industrial parks has long been regarded as a key and difficult point in the field of industrial safety. The traditional manual safety inspection mode is difficult to fully meet the actual needs of modern chemical industrial park safety management due to problems such as low efficiency, limited coverage, and many safety hazards. In recent years, with the rapid development of drone technology, this technology has provided a new solution for the safety inspection of chemical industrial parks. Drones, with their highly flexible operating performance, wide coverage, and the ability to carry a variety of sensors, significantly enhance the safety of inspections while improving inspection efficiency.

[0003] However, the operating environment of chemical industrial parks is often very complex, facing various potential risk factors such as fire hazards and toxic gas leaks. Relying solely on a single drone to perform inspection tasks is difficult to achieve the goal of comprehensive and efficient safety monitoring. Therefore, the use of drone swarms for collaborative operations has gradually become a research hotspot in the academic and industrial fields. At the same time, how to design a reasonable inspection path for drone swarms to balance risk control and data collection quality while ensuring inspection efficiency has become an urgent and challenging issue. Summary of the Invention

[0004] The purpose of the embodiments of this application is to provide an optimization method, device, and equipment for the inspection path of a drone swarm in a chemical industrial park, which solves the above problems existing in the prior art, can generate an optimal inspection path that takes into account efficiency, safety, high quality, and high coverage, and controls the drone swarm to perform efficient and high-quality inspections.

[0005] In a first aspect, the present invention provides an optimization method for the inspection path of a drone swarm in a chemical industrial park, the method comprising:

[0006] Obtain the environmental data of the target chemical industrial park, as well as the attributes and initial inspection paths of each unmanned aerial vehicle (UAV) used to inspect the target chemical industrial park; wherein, the initial inspection path corresponding to any UAV includes at least one inspection target that the corresponding UAV is responsible for; for any UAV, control the UAV to perform its respective inspection tasks according to the corresponding initial inspection path, and obtain the real-time monitoring values of each inspection position on the initial inspection path and the corresponding inspection position; for any inspection target that the UAV is responsible for, obtain the weight when the UAV reaches the inspection target; input each inspection position and the corresponding real-time monitoring value, the target position of the inspection target and the weight, the environmental data, and the initial inspection path into a pre-constructed multi-objective optimization model to calculate the score of the initial inspection path; according to the weight, the target position, the environmental data, the attributes, and the initial inspection paths and positions of each UAV, use dynamic neighborhood search to determine the optimal inspection path of the UAV; if the score of the optimal inspection path is greater than the score of the initial inspection path, then use the optimal inspection path as the new initial inspection path, and return to execute the steps: for any UAV, control the UAV to conduct inspections according to the new initial inspection path until the UAV completes the inspection task.

[0007] In an alternative embodiment, the attributes include: flight speed, endurance time, communication range, safe flight range, maximum weight, minimum weight, maximum neighborhood range, and minimum neighborhood range; the environmental data includes: all inspection targets in the chemical industrial park and the dangerous characteristics of the corresponding inspection targets, as well as each risk area in the target chemical industrial park and the risk level of the corresponding risk area; the initial inspection path is generated based on each inspection target and the corresponding dangerous characteristics, each risk area and the corresponding risk level, the number of UAVs, and the flight speed, endurance time, communication range, and safe flight range of the corresponding UAV; when the UAV reaches any inspection target, the number of optimization times of the inspection path corresponding to the UAV is incremented by one.

[0008] In an alternative embodiment, the method for obtaining the weight includes: determining the weight of the UAV when it reaches the inspection target based on the configured total number of optimization times, the maximum weight, the minimum weight, and the number of optimization times of the inspection path corresponding to the UAV.

[0009] In an alternative embodiment, inputting each inspection position and the corresponding real-time monitoring value, the target position of each inspection target and the weight, the environmental data, and the initial inspection path into a pre-constructed multi-objective optimization model includes:

[0010] For any inspection target on the initial inspection path, the inspection position whose distance from the target position of the inspection target is less than the configured distance threshold is used as the target inspection position of the inspection target; based on the distance between the target inspection position and the corresponding inspection target, the accuracy of the real-time monitoring value corresponding to the target inspection position is determined; based on the accuracy of the real-time monitoring values corresponding to the UAV at each target inspection position, the data collection quality of the UAV on the initial inspection path is determined; according to the ratio of the area covered by the UAV on the initial inspection path to the total area covered by the initial inspection path, the inspection coverage rate of the UAV on the initial inspection path is obtained; according to the risk area covered by the initial inspection path and the risk level of the corresponding risk area, the risk level of the initial inspection path is calculated; according to the flight time of the UAV on the initial inspection path and the risk level of the initial inspection path, the risk degree of the UAV on the initial inspection path is calculated; the data collection quality, the inspection coverage rate, the risk degree and the path length of the initial inspection path are input into a pre-constructed multi-objective optimization model to obtain the score of the initial inspection path.

[0011] In an alternative embodiment, the multi-objective optimization model is constructed with the minimization of the path length, the minimization of the risk degree, the maximization of the inspection coverage rate and the maximization of the data collection quality as the objective functions, and the battery life, communication range, safe flight range of the UAV and the configured inspection task allocation rules as the constraints;

[0012] The inspection task allocation rule includes: the inspection paths corresponding to each UAV are different.

[0013] In an alternative embodiment, according to the weights, the target positions, the environmental data, the attributes, and the initial inspection paths and positions of each UAV, using dynamic neighborhood search to determine the optimal inspection path of the UAV includes:

[0014] According to the configured total number of optimizations, the maximum neighborhood range, the minimum neighborhood range, and the number of optimizations of the inspection path corresponding to the UAV, the neighborhood range of the UAV at the target position is determined; according to the environmental data, neighborhood search is performed within the neighborhood range of the target position to obtain multiple next alternative positions of the UAV; based on the weights, the target position, the multiple next alternative positions, and the initial inspection paths and positions of each UAV, the optimal inspection path is determined.

[0015] In an alternative embodiment, the weights include individual weights and group weights;

[0016] Determining the optimal inspection path based on the weights, the target location, multiple next alternative locations, and the initial inspection paths and locations of each unmanned aerial vehicle (UAV) includes: obtaining multiple alternative paths according to the multiple next alternative locations; calculating the scores of each alternative path using the multi-objective optimization model; taking the alternative location corresponding to the alternative path with the highest score as the optimal alternative location; determining the first location under the influence of the individual UAV according to the individual weight of the UAV, the target location, and the optimal alternative location; determining the second location under the influence of the UAV group according to the target location, the group weight, and the initial inspection paths and locations of each UAV; performing weighted fusion on the first location and the second location to obtain the optimal location of the UAV; and determining the optimal inspection path based on the target location and the optimal location.

[0017] In a second aspect, the present invention provides an optimization device for the inspection path of a UAV group in a chemical industrial park. The device includes:

[0018] An acquisition unit, configured to acquire the environmental data of the target chemical industrial park, and the attributes and initial inspection paths of each UAV for inspecting the target chemical industrial park; wherein, the initial inspection path corresponding to any UAV includes at least one inspection target responsible for the corresponding UAV.

[0019] A control unit, configured to, for any UAV, control the UAV to perform its respective inspection tasks according to the corresponding initial inspection path, and acquire the real-time monitoring values of each inspection location and the corresponding inspection location on the initial inspection path.

[0020] A calculation unit, configured to, for any inspection target responsible for the UAV, acquire the weight when the UAV reaches the inspection target; input each inspection location and the corresponding real-time monitoring value, the target location of each inspection target and the weight, the environmental data, and the initial inspection path into a pre-constructed multi-objective optimization model, and calculate the score of the initial inspection path.

[0021] A determination unit, configured to determine the optimal inspection path of the UAV by using dynamic neighborhood search according to the weight, the target location, the environmental data, the attributes, and the initial inspection paths and locations of each UAV.

[0022] An optimization unit, configured to, if the score of the optimal inspection path is greater than the score of the initial inspection path, take the optimal inspection path as the new initial inspection path, and return to execute the steps: for any UAV, control the UAV to perform inspections according to the new initial inspection path until the UAV completes the inspection task.

[0023] In a third aspect, the present invention provides an electronic device, which includes a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete communication with each other through the communication bus;

[0024] The memory is used to store computer programs;

[0025] The processor is configured to implement the method according to any one of the foregoing embodiments when executing the program stored on the memory.

[0026] In a fourth aspect, the present invention provides a computer-readable storage medium, in which a computer program is stored, and when the computer program is executed by a processor, the method according to any one of the foregoing embodiments is implemented.

[0027] This application effectively solves the problem of collaborative patrol of unmanned aerial vehicle (UAV) swarms in complex environments by constructing a multi-objective optimization model including path length, risk level, patrol coverage rate, and data collection quality, and designing an improved multi-objective jellyfish search algorithm. Experimental results show that the method proposed in this application is superior to traditional algorithms in terms of patrol efficiency, risk control, and data collection quality, providing an innovative intelligent solution for safety patrol in chemical industrial parks.

[0028] The multi-objective optimization model of this application includes four objectives. The first objective is to minimize the total path length to reduce patrol time and energy consumption; the second objective is to minimize the exposure time in high-risk areas to reduce the risks of UAVs and operators; the third objective is to maximize the patrol coverage rate to ensure that key areas are fully monitored; the fourth objective is to maximize the data collection quality, including the accuracy of flame recognition and the precision of gas detection; at the same time, the multi-objective optimization model also includes various constraint conditions, such as UAV endurance time, communication range, obstacle avoidance requirements, etc.; by introducing a penalty function mechanism, the degree of constraint violation is incorporated into the objective function, and the collaborative constraints between UAVs, such as collision avoidance and task allocation, are integrated, so that the multi-objective optimization model can comprehensively reflect the requirements of actual patrol tasks and comprehensively assist in the optimization of UAV patrol paths. Description of the Drawings

[0029] In order to more clearly illustrate the technical solutions of the embodiments of this application, the following will briefly introduce the drawings required to be used in the embodiments of this application. It should be understood that the following drawings only show certain embodiments of this application, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0030] Figure 1 It is a flowchart of an optimization method for the patrol path of a UAV swarm in a chemical industrial park provided by an embodiment of this application;

[0031] Figure 2 This is a schematic structural diagram of an optimization device for the inspection path of an unmanned aerial vehicle (UAV) group in a chemical industrial park provided by an embodiment of the present application;

[0032] Figure 3 This is a schematic structural diagram of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0033] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0034] The optimization method for the inspection path of the UAV group in the chemical industrial park provided by the embodiment of the present application can be applied to the optimization system for the inspection path of the UAV group in the chemical industrial park. The system includes: multiple UAVs equipped with flame recognition devices and hazardous gas detection devices, a ground control station, and communication devices; the flame recognition device can detect and locate the flame hazards in the park in real time, featuring high accuracy and fast response; the hazardous gas detection device can detect various toxic and harmful gases, such as hydrogen sulfide, ammonia, etc., and transmit the concentration data in real time. The UAV group maintains communication with the ground control station through a wireless network to achieve real-time data transmission and instruction reception.

[0035] In the embodiment of the present application, the UAV can adopt DJI Matrice 300 / custom four-axis (equipped with Pixhawk flight controller) or other UAVs that support SDK (such as DJI MSDK / PX4 MAVSDK) and sensor expansion interfaces; the flame recognition device includes a camera component; the camera component can select Sony IMX477 (supporting 4K); the hazardous gas detection device adopts Figaro TGS2600 (VOC detection) or other multi-sensors with unified power supply and I2C / SPI interfaces; the UAV is provided with NVIDIA Jetson Orin Nano (embedded GPU) for performing preprocessing such as image compression and data fusion to reduce the transmission bandwidth requirement.

[0036] In the embodiments of the present application, data transmission can adopt Holybro SiK Radio (915MHz, 10km) or cellular network: Quectel 5G module; adopt a dual-link redundancy design (data transmission + 4G) for automatic switching; the communication protocols include: MAVLink / ROS; path planning adopts PX4 Avoidance / Open Motion Planning; MAVLink defines ODOMETRY and POSITION_TARGET messages, and ROS2 is used for task orchestration.

[0037] In the embodiments of the present application, the ground control station can be a cloud server (AWS EC2 / GCP instance, equipped with NVIDIA T4 GPU); it runs the ROS / GAZEBO simulation environment to process the multi-aircraft data stream in real time;

[0038] The ground control station can be a server or a terminal with strong computing power. The server can be a physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, Content Delivery Network (CDN), and big data and artificial intelligence platforms. The terminal can be a user equipment (UE) such as a mobile phone, smart phone, laptop, digital broadcast receiver, personal digital assistant (PDA), tablet computer (PAD), handheld device, vehicle-mounted device, wearable device, computing device or other processing devices connected to a wireless modem, mobile station (MS), mobile terminal, etc. The terminal and the server can be directly or indirectly connected through wired or wireless communication methods, and the present application does not limit this here.

[0039] The preferred embodiments of the present application will be described below with reference to the accompanying drawings of the specification. It should be understood that the preferred embodiments described herein are only for explaining and illustrating the present application, and are not used to limit the present application. And without conflict, the embodiments in the present application and the features in the embodiments can be combined with each other.

[0040] Figure 1 It is a schematic flow chart of an optimization method for the inspection path of an unmanned aircraft group in a chemical industrial park provided by the embodiments of the present application. As Figure 1 shown, the method may include:

[0041] Step S110: Obtain the environmental data of the target chemical industrial park, as well as the attributes and initial inspection paths of each unmanned aerial vehicle (UAV) for inspecting the target chemical industrial park; for any UAV, control the UAV to perform its respective inspection tasks according to the corresponding initial inspection path, and obtain the real-time monitoring values of each inspection position on the initial inspection path and the corresponding inspection positions.

[0042] In the embodiment of the present application, the environmental data of the target chemical industrial park includes: all inspection targets in the chemical industrial park and the dangerous characteristics of the corresponding inspection targets, as well as each risk area in the target chemical industrial park and the risk level of the corresponding risk area.

[0043] In the embodiment of the present application, the initial inspection paths of each UAV are different to complete the inspection tasks of the entire park through division of labor.

[0044] In the embodiment of the present application, the risk areas in the target chemical industrial park and the risk levels of the corresponding risk areas are obtained through the following method:

[0045] First, obtain the basic information of each building in the target chemical industrial park, the basic information of the pipelines and tank farms set in the target chemical industrial park, the hazardous chemical storage, production and logistics map, and other data; among them, the basic information of each building includes: location, use, structure, etc.; the basic information of the pipelines and tank farms includes: direction, location or area, and the types of hazardous chemicals stored, etc.; the hazardous chemical storage, production and logistics map includes: the storage, transportation routes and production areas of hazardous chemicals; other data includes: the population distribution in the park, emergency facilities and the surrounding environment, etc. Secondly, perform risk identification on the target chemical industrial park according to the above-obtained data, including: hazardous chemical identification, risk source identification and potential accident type identification; hazardous chemical identification includes: identifying all hazardous chemicals in the target chemical industrial park and the dangerous characteristics of the corresponding hazardous chemicals (such as toxicity, flammability, etc.); risk source identification includes: identifying possible risk sources, such as storage tanks, pipelines, production devices, etc.; potential accident type identification includes: identifying possible accident types, such as leakage, fire, explosion, etc.; furthermore, based on the risk identification results, perform accident consequence analysis, risk level division and vulnerability analysis to obtain the risk assessment results; among them, accident consequence analysis: evaluate the influence range and severity of different accidents; risk level division: divide the risk level according to the accident possibility and consequences (such as high, medium, low); vulnerability analysis: evaluate the vulnerability of personnel, equipment and environment in the park to accidents; finally, generate a risk distribution map including the risk areas in the target chemical industrial park and the risk levels of the corresponding risk areas; the risk distribution map uses the architectural map or pipeline and tank farm layout map of the target chemical industrial park as the base map; mark the risk sources on the base map, and use different colors or symbols to represent the risk levels; mark the possible influence range of accidents, such as the leakage diffusion range or the explosion shock wave range; mark the emergency facilities such as fire stations and emergency material storage points.

[0046] In the embodiments of the present application, the risk distribution map has a legend for explaining the meanings of colors and symbols in the risk distribution map; the risk levels and corresponding countermeasures of each risk source are detailed in the risk distribution map or in the attached schedule; the risk distribution map is updated regularly to reflect the changes in the park; specifically, the risk distribution map can be obtained through the following tools or software: GIS software: such as ArcGIS, QGIS, for spatial data management and risk map drawing; CAD software: such as AutoCAD, for processing architectural drawings and pipeline layout drawings; risk assessment software: such as PHAST, ALOHA, for accident consequence simulation.

[0047] In the embodiments of the present application, the attributes of the drone include: flight speed, endurance time, communication range, safe flight range, maximum weight value, minimum weight value, maximum neighborhood range, and minimum neighborhood range.

[0048] In the embodiments of the present application, the inspection targets include: industrial pipelines, factories, tanks, warehouses, and / or other facilities for storing or transporting hazardous chemicals in the target chemical industrial park; the inspection targets can also be other buildings such as machine rooms that are prone to safety accidents; a risk level is set for each inspection target in advance according to the risk identification results and risk assessment results; the number of drones used for inspecting the target chemical industrial park is greater than or equal to 2; any one drone is responsible for at least one inspection target; the inspection targets responsible for by different drones can be the same or different.

[0049] In the embodiments of the present application, the initial inspection path corresponding to any one drone includes at least one inspection target that the corresponding drone is responsible for; the inspection task of the drone consists of inspection speed, inspection target, inspection path, and total inspection time; before all drones execute their respective inspection tasks, inspection tasks are generated for all drones in advance according to the environmental data of the target chemical industrial park and the attributes of each drone, and the initial inspection paths are assigned; the initial inspection path corresponding to any one drone is generated based on each inspection target and its corresponding hazard characteristics, each risk area and its corresponding risk level, the number of drones, and the flight speed, endurance time, communication range, and safe flight range of the corresponding drone.

[0050] For example, before the inspection task of the target chemical industrial park starts, multiple drones will form a "population". At the same time, the initial parameters of this inspection task are determined, such as setting the maximum flight speed of the drone to 15 m / s, the maximum endurance time to 30 minutes, the communication range to 500 meters, etc. In addition, according to the actual situation of the park, different risk areas are divided, the locations of dangerous facilities such as toxic gas pipelines and storage tanks are marked, and a 3D model of the park is constructed and imported into the simulation platform to prepare for subsequent path planning.

[0051] In the embodiment of the present application, the inspection path of the drone consists of multiple inspection positions, and the inspection path of the drone covers at least one inspection target that the drone is responsible for; the target position of the inspection target is also an inspection position of the drone. For example, there are 2 inspection targets and 3 drones in a chemical industrial park, and the area of the chemical industrial park is 150 mu; drones A and B are responsible for inspecting target p1, and drone C is responsible for inspecting target p2; the inspection paths of the 3 drones are all different, and there is an inspection position every 10 meters on each inspection path. One inspection path contains 50 inspection positions, and p1 coincides with inspection position d5; drones A and B pass by inspection target p1 at the same time, but the remaining 49 inspection positions are all different.

[0052] In the embodiment of the present application, when the drone passes by each inspection position or at each position on the inspection path, it needs to use the monitoring equipment carried by itself for real-time monitoring; specifically, the monitoring equipment includes a flame recognition device and a hazardous gas detection device; the flame recognition device consists of an image acquisition component and an image analysis component, which is used to acquire the image at the inspection position and output the flame recognition result through image analysis; the hazardous gas detection device includes a hazardous gas collection component and a hazardous gas analysis component, which is used to collect the air at the inspection position and output the types of hazardous gases contained in the air (such as hydrogen sulfide, ammonia, etc.) and the concentration of the corresponding hazardous gas through gas analysis, and compare it with the configured concentration thresholds of each hazardous gas. When it exceeds the configured concentration threshold, it means that the hazardous gas at this inspection position exceeds the standard.

[0053] In the embodiment of the present application, the real-time monitoring values include: flame recognition results and hazardous gas recognition results; the image analysis component is equipped with a YOLO flame recognition model, which can quickly identify whether the image at the inspected position contains a flame. If it contains a flame, it will output the detection frame of the flame and the confidence of the corresponding detection frame.

[0054] Step S120: For any inspection target that the drone is responsible for, obtain the weight when the drone reaches the inspection target; input each inspection position and the corresponding real-time monitoring value, the target position and weight of the inspection target, environmental data, and the initial inspection path into the pre-constructed multi-objective optimization model, and calculate the score of the initial inspection path.

[0055] In the embodiment of the present application, when any drone reaches an inspection target, the path is automatically optimized once, and the number of optimization times of the inspection path corresponding to the drone is incremented by one.

[0056] In another embodiment of the present application, it can also be set that when any drone reaches an inspection position, the path is automatically optimized once, and the number of optimization times of the inspection path corresponding to the drone is incremented by one.

[0057] In the embodiments of the present application, the weights include individual weights and group weights; specifically, the individual weights are obtained based on the configured total number of optimizations, the maximum weight, the minimum weight, and the number of optimizations of the inspection path corresponding to the drone; the specific formula is as follows: ; where is the individual weight when the drone reaches any inspection target, and are the configured minimum weight and maximum weight, obtained from the drone attributes; preferably, = 0.1, = 0.9; t is the current number of optimizations, and T is the total number of optimizations.

[0058] In the embodiments of the present application, the group weight = 1 - individual weight.

[0059] In the embodiments of the present application, the weight of each drone is not fixed, but changes with "iteration" (which can be understood as an optimization of the inspection path). Based on the current iteration number (i.e., the number of optimizations) and the total iteration number (i.e., the total number of optimizations), the individual weight (the adjustment of the flight path of each drone) and the group weight (the cooperative flight strategy among multiple drones) are dynamically adjusted.

[0060] In the embodiments of the present application, adaptively adjusting the weight of the drone according to the number of optimizations can give the drone a large autonomous exploration space at the beginning of the inspection path optimization, allowing the drone to search for possible inspection paths within a large range; as the optimization approaches the end, gradually increase the weight of group cooperation to enable better cooperation among the drones and jointly optimize the inspection path.

[0061] In the embodiments of the present application, inputting each inspection position and the corresponding real-time monitoring value, the target position and weight of each inspection target, environmental data, and the initial inspection path into a pre-constructed multi-objective optimization model includes:

[0062] For any inspection target on the initial inspection path, the inspection position whose distance from the target position of the inspection target is less than the configured distance threshold is used as the target inspection position of the inspection target; based on the distance between the target inspection position and the corresponding inspection target, the accuracy of the real-time monitoring value corresponding to the target inspection position is determined; based on the accuracy of the real-time monitoring values corresponding to the drone at each target inspection position, the data acquisition quality of the drone on the initial inspection path is determined; according to the ratio of the area covered by the drone on the initial inspection path to the total area covered by the initial inspection path, the inspection coverage rate of the drone on the initial inspection path is obtained; according to the risk area covered by the initial inspection path and the risk level of the corresponding risk area, the risk level of the initial inspection path is calculated; according to the flight time of the drone on the initial inspection path and the risk level of the initial inspection path, the risk degree of the drone on the initial inspection path is calculated; according to the path length of the initial inspection path and the maximum configured path length, the path length score is calculated; according to the risk degree of the initial inspection path and the maximum configured risk degree, the risk degree score is calculated; the data acquisition quality, inspection coverage rate, risk degree, and path length score are weighted and summed to obtain the score of the initial inspection path.

[0063] The score of the initial inspection path is the initial fitness value of the drone. This fitness value comprehensively considers multiple objective functions, such as path length, exposure time in high-risk areas, inspection coverage rate, and data acquisition quality. For example, if the initial path planned by a certain drone is short, it spends less time in high-risk areas, can cover more key areas, and has high data acquisition quality, then its initial fitness value is relatively high. Specifically, the maximum risk degree can be obtained from the risk levels of all risk areas in the chemical industrial park and the maximum configured residence time of the drone in each risk area.

[0064] In the embodiment of the present application, according to the distance between the target inspection position and the corresponding inspection target, the target accuracy corresponding to the corresponding distance is matched from the configured comparison table of different distances and different real-time monitoring value accuracies; it can also be determined according to the accuracy of the gas detection device or the flame recognition device. For example, it can be set that when the distance between the target inspection position and the corresponding inspection target is less than 5 meters, the accuracy is 70%, and when it is less than 1 meter, it is 90%.

[0065] In another embodiment of the present application, the attributes of the drone further include: the attitude of the drone. According to the attitude of the drone, the risk level and risk characteristics corresponding to the inspection target, the data acquisition quality of the drone at the target inspection position corresponding to the inspection target is determined.

[0066] In the embodiment of the present application, the accuracy of the real-time monitoring value at the corresponding inspection position can be determined according to the distance between the inspection position and the corresponding inspection target, or the accuracy of the real-time monitoring value at each inspection position can be estimated by using a pre-trained monitoring accuracy estimation model.

[0067] In another embodiment of the present application, when flames and / or hazardous gases are detected at any inspection position, the corresponding inspection position and detection results (including detection frames, confidence levels, concentration values, etc.) can be input into a data quality assessment model pre-trained using historical data to obtain the data acquisition quality at that inspection position, that is, the data acquisition quality of the real-time monitoring values.

[0068] In the embodiments of the present application, the data quality assessment model is trained using historical data. The method for collecting historical data includes:

[0069] Obtain the historical monitoring values collected by the UAV at the corresponding historical inspection positions within the historical time period in the target chemical industrial park, the accuracy rate of the corresponding historical flame recognition results, and the precision of the historical gas detection results; specifically, the accuracy rate of the historical flame recognition results is calculated through the detection results of the YOLO flame recognition model. The specific method is as follows:

[0070] Flame recognition accuracy Calculated by the following formula: ;

[0071] Among them: TP (True Positive) represents the number of correctly recognized flames; FP (False Positive) represents the number of false alarms as flames; FN (False Negative) represents the number of undetected flames; the above data is obtained by manually annotating and counting the images on the UAV inspection path after the UAV completes an inspection task; for example, if the YOLO flame recognition model detects 10 flames, 8 of which are correct (TP = 8), 2 are false alarms (FP = 2), and 1 flame is not detected (FN = 1), then: ;

[0072] The method for determining the precision of the historical gas detection results includes:

[0073] Gas detection precision Calculated by the following formula: ;

[0074] Among them: represents the detected gas concentration value; represents the standard gas concentration value; det represents the number of detections; for example, if 5 groups of gas concentration values are detected, and the deviations from the standard values are: 0.1, 0.2, 0.05, 0.15, 0.1, and the sum of the standard values is 10, then After calculation, it is equal to 0.94;

[0075] The corresponding historical data collection quality is obtained by weighted summation of the accuracy rate of historical flame recognition results and the precision of historical gas detection results. ; where: represents the weight of the flame recognition accuracy rate (usually set to 0.5); represents the weight of the gas detection precision (usually set to 0.5); For example, if the flame recognition accuracy rate is 72.7% and the gas detection precision is 94%, then Q = 0.8335;

[0076] The historical monitoring value, the corresponding historical data collection quality, and the target location or historical inspection location of the historical inspection target for collecting the historical monitoring value are used as a piece of historical data of the UAV. The trained data quality assessment model is obtained by training the data quality assessment model with multiple pieces of historical data of the UAV.

[0077] In the embodiment of the present application, taking the data collection quality as the objective function can guide the algorithm to optimize in the direction of being conducive to obtaining high-quality data during the path planning stage. If this objective is not considered and the path is planned only from aspects such as path length, risk level, and inspection coverage rate, it may cause the UAV to fly frequently in some areas with poor data collection conditions, and the obtained data has low reliability and cannot effectively meet the needs of safety inspections. By setting the data collection quality as the objective function, the algorithm will comprehensively weigh various factors, and while ensuring the coverage of key areas and reducing risks, it will try to select paths that are conducive to improving the data collection quality, making the finally planned inspection path more scientific and practical.

[0078] In practical applications, after the risk distribution assessment of the target chemical industrial park is performed, the characteristics, potential risks, and environmental conditions of different regions are known or can be estimated. For example, in areas close to production devices, storage tank areas, etc., the possibility of dangerous gas leakage and fire is higher. Correspondingly, when the UAV collects data in these areas, the difficulty and importance of its flame recognition and gas detection are also different. Based on information such as past chemical industrial park accident statistics data and facility layout characteristics, a prediction model can be established to theoretically predict the quality of the collected data in different regions. For example, in high-risk areas, due to many potential interference factors and great prediction difficulty, the data quality may be low; while in relatively open and less interfered areas, the data quality may be high. This prediction method can provide basic data support for incorporating the data collection quality into the objective function.

[0079] Meanwhile, the UAV real-time monitoring device has specific performance parameters, such as the accuracy rate of flame recognition, the precision range of gas detection, the detection limit, etc. According to these known sensor performance indicators, the data acquisition quality under different conditions can be inferred. Combining risk distribution assessment and environmental assessment, in an environment with little signal interference and obvious characteristics of the detection object, the sensor is more likely to obtain high-quality data; on the contrary, in a situation with complex signals and the concentration of the detection object close to the detection limit, the data quality may be affected. The speculation of data acquisition quality based on sensor performance does not depend on the actually collected data, but starts from the capabilities of the sensor itself, evaluates in advance the data quality that may be obtained for different inspection paths, and provides a reference for optimizing the inspection paths.

[0080] In the embodiment of the present application, according to the risk areas covered by the initial inspection path and the risk levels of the corresponding risk areas, the risk level of the initial inspection path is calculated, including: determining the risk value according to the proportion of the risk area in the area covered by the initial inspection path; determining the risk weight according to the risk level corresponding to the risk area; and determining the risk level of the initial inspection path based on the risk value and the risk weight.

[0081] In the embodiment of the present application, the multi-objective optimization model is constructed with the minimization of path length, the minimization of risk degree, the maximization of inspection coverage rate, and the maximization of data acquisition quality as the objective functions, and the endurance time, communication range, safe flight range of the UAV, and the configured inspection task allocation rules as the constraints; the inspection task allocation rules include: the inspection paths corresponding to each UAV are different.

[0082] Specifically, the objective functions of the multi-objective optimization model formula include:

[0083] ;

[0084] ;

[0085] ;

[0086]

[0087] Among them, represents the distance that UAV i moves from a certain point on path j to the next point (the line connecting these two points corresponds to a path segment), is a binary variable indicating whether UAV i passes through path j; represents the risk level of path j, is the residence time of UAV i on path j; represents the area covered by UAV i on path j, is the total coverage area of path j; Indicates the data quality collected by the UAV i on the path j.

[0088] The constraint conditions of the multi-objective optimization model formula include:

[0089] 1. Flight time constraint: ; where is the maximum flight time of the UAV i.

[0090] 2. Communication range constraint: ; where represents the distance between the UAV i and the UAV λ , and R is the communication range;

[0091] 3. Obstacle avoidance constraint: ; where represents the minimum distance between the UAV i and the obstacle, and S is the safety distance;

[0092] 4. Task assignment constraint: ; The task assignment constraint means that each path j can only be responsible for one UAV i.

[0093] In the embodiment of the present application, the multi-objective optimization model further includes: a scoring calculation formula for the inspection path; the scoring calculation formula for the inspection path is as follows:

[0094] ;

[0095] where L represents the path length of the inspection path; L max represents the maximum length (or the preset theoretical maximum path length) among all possible inspection paths (i.e., each path in the path set); Ris represents the risk degree of the inspection path, which can be obtained by weighted summation of the exposure time in the high-risk area or the risk level; Ris max represents the maximum risk value among all possible inspection paths (or the preset theoretical maximum risk degree); Cov represents the inspection coverage rate of the inspection path, which is the ratio of the covered area to the total area; Cov max represents the theoretical maximum coverage rate (usually 1, i.e., full coverage); Q represents the data collection quality of the inspection path, which can be obtained by weighted summation of the flame recognition accuracy and the gas detection accuracy; Q max represents the theoretical maximum data collection quality (usually 1, i.e., 100% accurate); , , and represent the weights of the path length, risk degree, inspection coverage rate, and data collection quality, ; , , and The value can be dynamically adjusted according to actual requirements; if safety is prioritized, increase (risk level) and (data acquisition quality); if efficiency is prioritized, increase (path length) and (inspection coverage rate); if there are no special requirements, take the default weights: = 0.25, = 0.3, = 0.2, = 0.25; this default weight is the comprehensive result of experimental data, multi-objective optimization theory, and the safety-prioritized characteristics of chemical industrial parks, ensuring an optimal balance among path length, risk, coverage rate, and data quality. In practical applications, it can be further calibrated through simulation.

[0096] In another embodiment of the present application, input each inspection location and the corresponding real-time monitoring value, the target location and weight of each inspection target, environmental data, and the initial inspection path into a pre-constructed multi-objective optimization model, including:

[0097] According to the target location and weight of each inspection target and environmental data, determine the path set of the drone; for any inspection target on the initial inspection path, take the inspection location whose distance from the target location of the inspection target is less than the configured distance threshold as the target inspection location of the inspection target; based on the distance between the target inspection location and the corresponding inspection target, determine the accuracy of the real-time monitoring value corresponding to the target inspection location; based on the accuracy of the real-time monitoring values of the drone at each target inspection location, determine the data acquisition quality of the drone on the initial inspection path; according to the ratio of the area covered by the drone on the initial inspection path to the total area covered by the initial inspection path, obtain the inspection coverage rate of the drone on the initial inspection path; according to the risk area covered by the initial inspection path and the risk level of the corresponding risk area, calculate the risk level of the initial inspection path; according to the path length of the initial inspection path and the maximum configured path length, calculate the path length score; according to the flight time of the drone on the initial inspection path and the risk level of the initial inspection path, calculate the risk level of the drone on the initial inspection path; calculate the risk levels of all paths in the path set; take the risk level of the path with the highest risk level as the maximum risk level; take the ratio of the risk level of the initial inspection path to the maximum risk level as the risk level score; weight and sum the data acquisition quality, inspection coverage rate, risk level score, and path length score to obtain the score of the initial inspection path.

[0098] In the embodiment of the present application, the path set contains multiple paths; the path set can be composed of all possible inspection paths within a chemical industrial park, or can be composed of all possible inspection paths of the drone. If it is composed of all possible inspection paths of the drone, each path in the path set has the same starting and ending points as the initial inspection path, and any path covers at least one inspection target responsible for by the drone.

[0099] Step S130: According to the weights, target positions, environmental data, attributes, and the initial inspection paths and positions of each drone, use dynamic neighborhood search to determine the optimal inspection path of the drone.

[0100] In the embodiment of the present application, the determination of the optimal inspection path of the drone adopts a multi-objective jellyfish search algorithm improved based on adaptive weights and dynamic neighborhood search; in each inspection path optimization process, each drone adjusts its inspection path according to the adaptive weight and dynamic neighborhood search strategy; at the same time, an external archive mechanism is introduced to store the non-dominated solutions (i.e., solutions that cannot be surpassed by other solutions in all objectives, such as a certain path is not worse than other paths in terms of path length, risk exposure time, inspection coverage rate, and data collection quality, etc.) generated in each iteration into the external archive; then, through non-dominated sorting and crowding degree calculation, the solutions in the archive are screened and optimized to maintain the diversity of the solution set and ensure that multiple different but relatively optimal inspection path schemes can be found.

[0101] Specifically, it includes: determining the neighborhood range of the drone at the target position according to the configured total number of optimizations, the maximum neighborhood range, the minimum neighborhood range, and the number of optimizations of the inspection path corresponding to the drone; performing neighborhood search within the neighborhood range of the target position according to the environmental data to obtain multiple next alternative positions of the drone; obtaining multiple alternative paths according to the multiple next alternative positions; calculating the scores of each alternative path using a multi-objective optimization model; taking the alternative position corresponding to the alternative path with the highest score as the optimal alternative position; determining the first position affected by the individual drone according to the individual weight of the drone, the target position, and the optimal alternative position; determining the second position affected by the drone group according to the target position, the group weight, and the initial inspection paths and positions of each drone; performing weighted fusion on the first position and the second position to obtain the optimal position of the drone; and determining the optimal inspection path based on the target position and the optimal position.

[0102] In the embodiment of the present application, the formula for determining the neighborhood range of any drone at any target position is as follows:

[0103] ;

[0104] Wherein, represents the neighborhood range of any drone at any target position, and represent the maximum and minimum values of the neighborhood range; it can be set , .

[0105] In the embodiment of the present application, a dynamic neighborhood search strategy is adopted. At the beginning stage, a relatively large "neighborhood range" is set for each UAV. Here, the neighborhood can be understood as the range within which each UAV refers to the paths of other surrounding UAVs when looking for a better path. For example, when initially planning a path, a certain UAV can refer to the path information of other UAVs that are relatively far away from it, which can expand the search range, enhance the global search ability, and find more potential high-quality inspection paths. As the iteration progresses, the neighborhood range is gradually reduced according to the formula of the neighborhood range, so that in the later stage, the UAV can focus more on the local area near itself, conduct a more refined search near the explored better paths, improve the local search accuracy, and further optimize its own inspection path.

[0106] In the embodiment of the present application, methods such as random sampling, grid division, or proximity search can be used to perform neighborhood search within the neighborhood range of the target position to obtain multiple next alternative positions of the UAV.

[0107] In the embodiment of the present application, after obtaining multiple alternative positions, a smooth interpolation method can be used to generate a smooth path between the current position and each alternative position, thereby obtaining alternative paths; or multiple alternative positions can be used to replace the next inspection position in the initial inspection path to obtain multiple alternative paths.

[0108] In the embodiment of the present application, the determination of the optimal path and the optimal alternative position is based on multiple iterations. Instead of directly outputting an optimal alternative position or an optimal path, it is necessary to judge whether it meets the termination condition. The termination condition can be that the score of the corresponding alternative path has not been significantly improved for several consecutive times. When the termination condition is met, the optimal inspection path will be output.

[0109] In the embodiment of the present application, the optimal inspection path is the path with the optimal comprehensive performance, which can take into account multiple objectives such as path length, risk level, inspection coverage rate, and data collection quality, and can meet the actual needs of chemical industrial park safety inspection to the greatest extent, realizing efficient, comprehensive, and safe inspection tasks.

[0110] In the embodiment of the present application, non-dominated sorting and crowding degree calculation are used to maintain the diversity and convergence of the solution set:

[0111] ;

[0112] Among them, represents the crowding degree of the individual , is the individual ( represents the value of a solution that covers the overall task allocation and path planning scheme of the UAV swarm on the k-th objective function, and are the maximum and minimum values of the k-th objective function.

[0113] Step S140: If the score of the optimal inspection path is greater than the score of the initial inspection path, then use the optimal inspection path as the new initial inspection path, and return to execute the step: for any UAV, control the UAV to perform inspections according to the new initial inspection path until the UAV completes the inspection task.

[0114] In the embodiment of the present application, after obtaining the optimal inspection path for the optimization of the current inspection path, compare it with the initial inspection path that the UAV is currently executing. If its score is better than the initial inspection path that the UAV is currently executing, it means that the optimal inspection path is better. Use the optimal inspection path to replace the currently executing inspection path, and continue to perform the inspection task according to the optimal inspection path until the next inspection target is reached, and then re-optimize the inspection path; if the score of the optimal inspection path is less than the currently executing inspection path, it means that the currently executing inspection path is better. At this time, control the UAV to continue to execute the currently executing inspection path.

[0115] In the embodiment of the present application, after the UAV completes the inspection task, it will uniformly summarize and analyze the real-time monitoring values collected by the UAV during the inspection process for the management of the chemical industrial park.

[0116] In another embodiment of the present application, when the UAV detects that the concentration of flames or dangerous gases exceeds the preset concentration threshold during the inspection process, it will immediately start the alarm program and send the current time and location to the remote end; at the same time, the UAV will first expand the detection range, collect images and air containing more objects for secondary precise detection; at the same time, the UAV will descend in height and perform image collection and air detection again. Combining the results of the three detections, generate a dangerous detection result for the inspection location; the dangerous detection result includes: dangerous location, dangerous type, dangerous level, and dangerous range; among them, the dangerous type is flames or dangerous gases; the dangerous level is determined according to the flame area or combustion degree or the type and concentration of dangerous gases, and the dangerous range is determined according to the scope involved in the danger.

[0117] When any drone detects a danger at any inspection location, the drone will downgrade the priority of the inspection task it is currently performing; the priority of the task of exploring the danger will be regarded as the top - priority task to be executed, that is, the drone will stay at the dangerous location instead of moving forward; the drone will determine whether there is an inspection target at the location where the danger is detected based on the images collected by the imaging device. If there is, it will identify the type of the inspection target and obtain the corresponding danger characteristics and danger level of the inspection target; at the same time, the drone will conduct a re - inspection based on the environment around the inspection target and the location or scope involved of the inspection target to determine whether the danger has spread, and timely feedback the inspection result to the remote end.

[0118] When the remote end receives the danger detection result feedback from any drone, it will call other drones that have completed the inspection task or drones in the standby state to conduct an investigation at the corresponding location to avoid false detection at a single location, etc.; if two drones both feedback the danger detection result, a danger control plan will be generated according to the danger location, danger type, danger level, and danger scope, and a drone that can execute the danger control plan will be dispatched to the dangerous location to execute the corresponding plan; for example, when a fire is detected, a drone that can carry fire extinguishing agent will be controlled to go to the dangerous location to extinguish the fire.

[0119] In the embodiment of the present application, when any drone detects a danger, the weights or priorities of all inspection locations within the danger range, or other inspection targets associated with the inspection targets within the danger range (such as connected pipelines), or other inspection locations within a certain range from the danger range will increase, so that other drones that are currently performing inspection tasks will give priority to inspecting the above - mentioned locations to avoid the spread of danger.

[0120] In the embodiment of the present application, when an individual drone hovers for confirmation due to identifying data fluctuations or suspected fire leakage, or when an individual drone fails to complete the original path inspection according to the original planned time due to uncontrollable reasons such as insufficient battery power, environmental disturbances, or malfunctions, or a temporary recall, etc., the dynamic adjustment of the path of the drone swarm of the present application can be triggered, so that the cluster can flexibly respond to emergencies and cooperate to complete the inspection task of the park.

[0121] In the embodiment of the present application, when any unmanned aerial vehicle (UAV) detects a danger at any inspection position, it will send its inspection target, inspection path, and current position to the remote end; when the remote end receives the danger detection result feedback by any UAV and the inspection target, inspection path, and current position of the corresponding UAV, it determines whether the UAV can continue to perform the inspection task according to the danger control plan; if not, it calls a standby UAV to perform the inspection task of this UAV; or, retrieves the attributes, current positions, and inspection paths of all UAVs, and inputs the inspection targets not inspected by this UAV and the attributes, current positions, and inspection paths of the UAVs that can normally perform the inspection task into a multi-objective optimization model to re-optimize the paths of the UAVs, so that other UAVs that can normally perform the inspection task complete the inspection task of the UAV that cannot normally perform the inspection task.

[0122] In an embodiment of the present application, if the UAV cannot execute the danger control plan, the UAV can continue to perform the inspection task; if the UAV can execute the danger control plan, it determines the danger control time required to execute the danger control plan, the first estimated inspection completion time of the inspection task it performs, and the second estimated inspection completion time of other UAVs that are performing the inspection task; if the difference between the sum of the danger control time and the first estimated inspection completion time and the second estimated inspection completion time is greater than the configured time threshold, the UAV cannot continue to perform the inspection task; wherein, the estimated inspection completion time is obtained according to the speed of the UAV, the inspection path, and the acquisition time required at each inspection position.

[0123] The technical solution of the present application will be further described below in combination with specific experiments.

[0124] To verify the effectiveness of the proposed method, the present application selects a typical chemical industrial park as the experimental scenario. The park contains various facilities such as toxic gas pipelines, storage tanks, office buildings, production plants, and living areas, with high complexity and representativeness. The present application sets up UAV groups of different scales (5, 10, 15) for comparative experiments, and compares IMOJSA with traditional NSGA-II and MOPSO algorithms.

[0125] 1.1 Experimental Setup

[0126] 1.1.1 Experimental Environment

[0127] The experiment selects a typical chemical industrial park as the test scenario, which contains various facilities such as toxic gas pipelines, storage tanks, office buildings, production plants, and living areas. The experimental scenario is constructed by 3D modeling software and imported into the simulation platform for testing. The simulation platform is built based on Gazebo and ROS (Robot Operating System) and can simulate the flight behavior of UAVs and sensor data acquisition.

[0128] 1.1.2 UAV Parameter Settings

[0129] In the experiment, multiple quadrotor UAVs were used, and their parameter settings are as follows: maximum flight speed: 15 m / s; maximum endurance time: 30 minutes; communication range: 500 meters; sensors: YOLO flame recognition model, chemical gas detection equipment (capable of detecting hydrogen sulfide, ammonia, etc.).

[0130] 1.1.3 Comparative Algorithms

[0131] To verify the performance of IMOJSA, two classic multi-objective optimization algorithms were selected for comparison: NSGA-II: Non-dominated Sorting Genetic Algorithm and MOPSO: Multi-Objective Particle Swarm Optimization Algorithm;

[0132] 1.1.4 Evaluation Metrics

[0133] The following evaluation metrics were used in the experiment to evaluate the performance of the algorithms: Hypervolume (HV): Measures the coverage and quality of the Pareto front solution set. Patrol coverage rate: The proportion of key areas covered by the UAV swarm. Risk exposure time: The total residence time of the UAVs in high-risk areas. Data acquisition quality: A comprehensive score of flame recognition accuracy and gas detection precision.

[0134] 1.2 Experimental Process

[0135] 1.2.1 Algorithm Initialization

[0136] First, initialize the parameters of the IMOJSA, NSGA-II, and MOPSO algorithms: population size: 50; maximum number of iterations: 100; external archive size: 100; adaptive weight range: , ; dynamic neighborhood range: =1, =10;

[0137] 1.2.2 Path Planning and Simulation

[0138] Path planning: Use IMOJSA, NSGA-II, and MOPSO to plan patrol paths for the UAV swarm respectively. After each planning, record the Pareto front solution set. Simulation run: Run the planned paths in the Gazebo simulation platform to simulate the flight behavior of the UAVs and the acquisition of sensor data. Data recording: Record the patrol coverage rate, risk exposure time, and data acquisition quality of each simulation.

[0139] 1.2.3 Experimental Scene Setup

[0140] The experiment set up three different scales of UAV swarms: 5 UAVs: suitable for small chemical industrial parks. 10 UAVs: suitable for medium-sized chemical industrial parks. 15 UAVs: suitable for large chemical industrial parks.

[0141] 1.3 Experimental Results

[0142] 1.3.1 Hypervolume Index Comparison

[0143] The hypervolume index (HV) is used to measure the coverage range and quality of the Pareto front solution set. The larger the HV value, the higher the quality of the solution set. The experimental results are shown in Table 1:

[0144] Table 1

[0145]

[0146] As can be seen from Table 1, the HV values of IMOJSA in all scenarios are higher than those of NSGA-II and MOPSO, indicating that IMOJSA can find a Pareto front solution set with higher quality.

[0147] 1.3.2 Patrol Coverage Rate Comparison

[0148] The patrol coverage rate is an important indicator to measure the proportion of key areas covered by the UAV swarm. The results are shown in Table 2:

[0149] Table 2

[0150]

[0151] As can be seen from Table 2, the patrol coverage rates of IMOJSA in all scenarios are higher than those of NSGA-II and MOPSO, indicating that IMOJSA can cover key areas more effectively.

[0152] 1.3.3 Risk Exposure Time Comparison

[0153] The risk exposure time is an indicator to measure the total residence time of the UAV in high-risk areas. The experimental results are shown in Table 3:

[0154] Table 3

[0155]

[0156] As can be seen from Table 3, the risk exposure times of IMOJSA in all scenarios are lower than those of NSGA-II and MOPSO, indicating that IMOJSA can more effectively reduce the residence time of the UAV in high-risk areas.

[0157] 1.3.4 Data Acquisition Quality Comparison

[0158] The data acquisition quality is a comprehensive score that measures the accuracy of flame recognition and the precision of gas detection. The experimental results are shown in Table 4 as follows:

[0159] Table 4

[0160]

[0161] As can be seen from Table 4, the data acquisition quality of IMOJSA is higher than that of NSGA-II and MOPSO in all scenarios, indicating that IMOJSA can more effectively improve the data acquisition quality.

[0162] 1.5 Result Analysis

[0163] From the above experimental data and charts, it can be seen that IMOJSA shows significant advantages in multi-objective optimization performance. The specific analysis is as follows: Hypervolume indicator: The HV value of IMOJSA is higher than that of NSGA-II and MOPSO in all scenarios, indicating that IMOJSA can find a Pareto front solution set with higher quality. Patrol coverage rate: The patrol coverage rate of IMOJSA is higher than that of NSGA-II and MOPSO in all scenarios, indicating that IMOJSA can more effectively cover key areas. Risk exposure time: The risk exposure time of IMOJSA is lower than that of NSGA-II and MOPSO in all scenarios, indicating that IMOJSA can more effectively reduce the stay time of drones in high-risk areas. Data acquisition quality: The data acquisition quality of IMOJSA is higher than that of NSGA-II and MOPSO in all scenarios, indicating that IMOJSA can more effectively improve the data acquisition quality. In summary, IMOJSA shows significant advantages in the problem of optimizing the safety patrol path of unmanned aerial vehicle swarms in chemical industrial parks, and can effectively improve the patrol efficiency, reduce the risk exposure time, and improve the data acquisition quality.

[0164] The experimental results show that IMOJSA shows significant advantages in the problem of optimizing the safety patrol path of unmanned aerial vehicle swarms in chemical industrial parks. By introducing an adaptive weight mechanism and a dynamic neighborhood search strategy, IMOJSA can effectively balance the exploration and exploitation capabilities of the algorithm and find a Pareto front solution set with higher quality. In addition, IMOJSA is superior to the traditional NSGA-II and MOPSO algorithms in terms of patrol coverage rate, risk exposure time, and data acquisition quality.

[0165] Experimental results show that IMOJSA exhibits significant advantages in terms of multi-objective optimization performance. In the scenario of 5 drones, the Pareto front solution set found by IMOJSA has increased by 12% and 8% respectively compared to NSGA-II and MOPSO in terms of the hypervolume metric. As the number of drones increases, the advantages of IMOJSA become more obvious. In the scenario of 15 drones, IMOJSA has increased the patrol coverage rate metric by more than 15% compared to the comparison algorithms, while maintaining a low risk exposure time and high data collection quality.

[0166] By analyzing the algorithm convergence curve and solution set distribution, this application finds that IMOJSA has a faster convergence speed and better solution set diversity. This indicates that IMOJSA can effectively explore the solution space and find a good balance among multiple objectives. In addition, this application also conducts actual flight tests to verify the feasibility and effectiveness of the proposed method in practical applications.

[0167] Corresponding to the above method, the embodiment of this application also provides an optimization device for the patrol path of a chemical industrial park drone swarm, as Figure 2 shown. The optimization device for the patrol path of the chemical industrial park drone swarm includes:

[0168] An acquisition unit 210, configured to acquire the environmental data of the target chemical industrial park, as well as the attributes and initial patrol paths of each drone for patrolling the target chemical industrial park; wherein, the initial patrol path corresponding to any one drone includes at least one patrol target responsible for the corresponding drone;

[0169] A control unit 220, configured to, for any one drone, control the drone to perform its respective patrol tasks according to the corresponding initial patrol path, and acquire the real-time monitoring values of each patrol position and the corresponding patrol position on the initial patrol path;

[0170] A calculation unit 230, configured to, for any one patrol target responsible for the drone, acquire the weight when the drone reaches the patrol target; input each patrol position and the corresponding real-time monitoring value, the target position and weight of each patrol target, the environmental data, and the initial patrol path into a pre-constructed multi-objective optimization model, and calculate the score of the initial patrol path;

[0171] A determination unit 240, configured to determine the optimal patrol path of the drone by using dynamic neighborhood search according to the weight, target position, environmental data, attributes, and the initial patrol paths and positions of each drone;

[0172] An optimization unit 250, configured to, if the score of the optimal patrol path is greater than the score of the initial patrol path, use the optimal patrol path as the new initial patrol path, and return to execute the step: for any one drone, control the drone to patrol according to the new initial patrol path until the drone completes the patrol task.

[0173] The functions of the functional units of the device for optimizing the inspection path of the UAV swarm in the chemical industrial park provided in the above embodiments of the present application can be realized by the above method steps. Therefore, the specific working processes and beneficial effects of each unit in the device for optimizing the inspection path of the UAV swarm in the chemical industrial park provided in the embodiments of the present application will not be repeated here.

[0174] The embodiments of the present application also provide an electronic device, as Figure 3 shown, including a processor 310, a communication interface 320, a memory 330, and a communication bus 340. Among them, the processor 310, the communication interface 320, and the memory 330 complete mutual communication through the communication bus 340.

[0175] The memory 330 is used to store a computer program; when the processor 310 executes the program stored on the memory 330, the following steps are implemented: obtaining environmental data of a target chemical industrial park, as well as the attributes and initial inspection paths of each UAV used to inspect the target chemical industrial park; where the initial inspection path corresponding to any UAV includes at least one inspection target responsible for the corresponding UAV; for any UAV, controlling the UAV to perform its respective inspection tasks according to the corresponding initial inspection path, and obtaining real-time monitoring values of each inspection position and the corresponding inspection position on the initial inspection path; for any inspection target responsible for the UAV, obtaining the weight when the UAV reaches the inspection target; inputting each inspection position and the corresponding real-time monitoring value, the target position and weight of the inspection target, the environmental data, and the initial inspection path into a pre-constructed multi-objective optimization model to calculate the score of the initial inspection path; determining the optimal inspection path of the UAV using dynamic neighborhood search according to the weight, target position, environmental data, attributes, and the initial inspection paths and positions of each UAV; if the score of the optimal inspection path is greater than the score of the initial inspection path, then taking the optimal inspection path as the new initial inspection path, and returning to execute the step: for any UAV, controlling the UAV to perform inspections according to the new initial inspection path until the UAV completes the inspection task.

[0176] The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience of representation, only a thick line is used in the figure, but it does not mean that there is only one bus or one type of bus. The communication interface is used for communication between the above-mentioned electronic device and other devices. The memory can include a Random Access Memory (RAM), and can also include a Non-Volatile Memory (NVM), such as at least one disk memory. Optionally, the memory can also be at least one storage device located far from the aforementioned processor. The above-mentioned processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. Since the implementation manners and beneficial effects of the various devices of the electronic device in the above embodiments can be seen from Figure 1 the steps in the embodiments shown, therefore, the specific working process and beneficial effects of the electronic device provided in the embodiments of the present application will not be elaborated herein.

[0177] In another embodiment provided by the present application, there is also provided a computer-readable storage medium, in which instructions are stored, and when it runs on a computer, it causes the computer to execute the optimization method for the inspection path of the chemical industrial park UAV fleet in any one of the above embodiments.

[0178] In another embodiment provided by the present application, there is also provided a computer program product containing instructions, and when it runs on a computer, it causes the computer to execute the optimization method for the inspection path of the chemical industrial park UAV fleet in any one of the above embodiments.

[0179] Those skilled in the art should understand that the embodiments in the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the embodiments in the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the embodiments in the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0180] Embodiments of the present application are described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate a means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or a means for implementing the functions specified in multiple blocks.

[0181] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction means that implements the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or a means for implementing the functions specified in multiple blocks.

[0182] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or a means for implementing the functions specified in multiple blocks.

[0183] Although the preferred embodiments in the embodiments of the present application have been described, those skilled in the art can make additional changes and modifications once they know the basic creative concept. Therefore, the appended claims are intended to be construed as including the preferred embodiments and all changes and modifications falling within the scope of the embodiments of the present application.

[0184] Obviously, those skilled in the art can make various changes and variations to the embodiments in the embodiments of the present application without departing from the spirit and scope of the embodiments in the embodiments of the present application. Thus, if these modifications and variations of the embodiments in the embodiments of the present application fall within the scope of the claims of the embodiments of the present application and their equivalent technologies, the embodiments in the embodiments of the present application are also intended to include these changes and variations.

Claims

1. An optimization method for the inspection path of an unmanned aerial vehicle (UAV) swarm in a chemical industrial park, characterized in that, The method includes: Obtaining the environmental data of the target chemical industrial park, as well as the attributes and initial inspection paths of each unmanned aerial vehicle (UAV) for patrolling the target chemical industrial park; wherein, the initial inspection path corresponding to any UAV includes at least one inspection target responsible for the corresponding UAV; the attributes include: the maximum neighborhood range and the minimum neighborhood range; the environmental data includes: each risk area of the target chemical industrial park and the risk level of the corresponding risk area; For any UAV, controlling the UAV to perform its respective inspection tasks according to the corresponding initial inspection path, and obtaining the real-time monitoring values of each inspection position on the initial inspection path and the corresponding inspection position; For any inspection target responsible for the UAV, obtaining the weight when the UAV reaches the inspection target; when the UAV reaches any inspection target, the optimization times of the inspection path corresponding to the UAV is incremented by one; Inputting each inspection position and the corresponding real-time monitoring value, the target position of the inspection target and the weight, the environmental data, and the initial inspection path into a pre-constructed multi-objective optimization model, and calculating the score of the initial inspection path; According to the configured total number of optimizations, the maximum neighborhood range, the minimum neighborhood range, and the optimization times of the inspection path corresponding to the UAV, determining the neighborhood range of the UAV at the target position; According to the environmental data, performing a neighborhood search within the neighborhood range of the target position to obtain multiple next alternative positions of the UAV; Based on the weight, the target position, multiple next alternative positions, and the initial inspection paths and positions of each UAV, determining the optimal inspection path; If the score of the optimal inspection path is greater than the score of the initial inspection path, then taking the optimal inspection path as the new initial inspection path, and returning to execute the step: for any UAV, controlling the UAV to conduct inspections according to the new initial inspection path until the UAV completes the inspection task; The inputting each inspection position and the corresponding real-time monitoring value, the target position of each inspection target and the weight, the environmental data, and the initial inspection path into a pre-constructed multi-objective optimization model includes: For any inspection target on the initial inspection path, taking the inspection position whose distance from the target position of the inspection target is less than the configured distance threshold as the target inspection position of the inspection target; Based on the distance between the target inspection position and the corresponding inspection target, determining the accuracy of the real-time monitoring value corresponding to the target inspection position; Based on the accuracy of the real-time monitoring values corresponding to the UAV at each target inspection position, determining the data collection quality of the UAV on the initial inspection path; According to the ratio of the area covered by the UAV on the initial inspection path to the total area covered by the initial inspection path, obtaining the inspection coverage rate of the UAV on the initial inspection path; According to the risk areas covered by the initial inspection path and the risk levels of the corresponding risk areas, calculating the risk level of the initial inspection path; Calculate the risk level of the UAV on the initial inspection path according to the flight time of the UAV on the initial inspection path and the risk level of the initial inspection path; Input the data acquisition quality, the inspection coverage rate, the risk level, and the path length of the initial inspection path into a pre-constructed multi-objective optimization model to obtain the score of the initial inspection path.

2. The method according to claim 1, characterized in that The attributes include: flight speed, endurance time, communication range, safe flight range, maximum weight value, and minimum weight value; The environmental data further includes: all inspection targets in the chemical industrial park and the dangerous characteristics of the corresponding inspection targets; The initial inspection path is generated based on each inspection target and the corresponding dangerous characteristics, each risk area and the corresponding risk level, the number of UAVs, and the flight speed, endurance time, communication range, and safe flight range of the corresponding UAVs.

3. The method according to claim 2, wherein The method for obtaining the weight includes: Determine the weight of the UAV when inspecting the inspection target based on the configured total number of optimization times, the maximum weight value, the minimum weight value, and the number of optimization times of the inspection path corresponding to the UAV.

4. The method according to claim 2, wherein The multi-objective optimization model is constructed with the minimization of the path length, the minimization of the risk level, the maximization of the inspection coverage rate, and the maximization of the data acquisition quality as the objective functions, and the endurance time, communication range, safe flight range of the UAV, and the configured inspection task allocation rules as the constraints; The inspection task allocation rule includes: the inspection paths corresponding to each UAV are different.

5. The method according to claim 2, wherein The weight includes an individual weight and a group weight; The determining of the optimal inspection path based on the weight, the target position, a plurality of next alternative positions, and the initial inspection paths and positions of each UAV includes: Obtain multiple alternative paths according to the plurality of next alternative positions; calculate the scores of each alternative path by using the multi-objective optimization model; Take the alternative position corresponding to the alternative path with the highest score as the optimal alternative position; Determine the first position under the influence of the UAV individual according to the individual weight of the UAV, the target position, and the optimal alternative position; Determine the second position under the influence of the UAV group according to the target position, the group weight, and the initial inspection paths and positions of each UAV; Perform weighted fusion on the first position and the second position to obtain the optimal position of the UAV; Determine the optimal inspection path based on the target position and the optimal position.

6. An optimization device for the inspection path of an unmanned aerial vehicle group in a chemical industrial park, characterized in that, The device includes: An acquisition unit, configured to acquire the environmental data of the target chemical industrial park, and the attributes and initial inspection paths of each UAV for inspecting the target chemical industrial park; wherein, the initial inspection path corresponding to any UAV includes at least one inspection target responsible for the corresponding UAV; the attributes include: the maximum neighborhood range and the minimum neighborhood range; the environmental data includes: each risk area in the target chemical industrial park and the risk level of the corresponding risk area; A control unit, for any unmanned aerial vehicle, to control the unmanned aerial vehicle to perform its respective inspection tasks according to the corresponding initial inspection path, and obtain the real-time monitoring values of each inspection position on the initial inspection path and the corresponding inspection position; A calculation unit, for any inspection target responsible for by the unmanned aerial vehicle, to obtain the weight when the unmanned aerial vehicle reaches the inspection target; when the unmanned aerial vehicle reaches any inspection target, the number of optimization times of the inspection path corresponding to the unmanned aerial vehicle is incremented by one; input the real-time monitoring values of each inspection position and the corresponding inspection position, the target positions of each inspection target and the weight, the environmental data, and the initial inspection path into a pre-constructed multi-objective optimization model, and calculate the score of the initial inspection path; The inputting the real-time monitoring values of each inspection position and the corresponding inspection position, the target positions of each inspection target and the weight, the environmental data, and the initial inspection path into a pre-constructed multi-objective optimization model includes: For any inspection target on the initial inspection path, take the inspection position whose distance from the target position of the inspection target is less than the configured distance threshold as the target inspection position of the inspection target; Based on the distance between the target inspection position and the corresponding inspection target, determine the accuracy of the real-time monitoring value corresponding to the target inspection position; Based on the accuracy of the real-time monitoring values corresponding to the unmanned aerial vehicle at each target inspection position, determine the data collection quality of the unmanned aerial vehicle on the initial inspection path; According to the ratio of the area covered by the unmanned aerial vehicle on the initial inspection path to the total area covered by the initial inspection path, obtain the inspection coverage rate of the unmanned aerial vehicle on the initial inspection path; According to the risk area covered by the initial inspection path and the risk level of the corresponding risk area, calculate the risk level of the initial inspection path; According to the flight time of the unmanned aerial vehicle on the initial inspection path and the risk level of the initial inspection path, calculate the risk degree of the unmanned aerial vehicle on the initial inspection path; Input the data collection quality, the inspection coverage rate, the risk degree, and the path length of the initial inspection path into a pre-constructed multi-objective optimization model to obtain the score of the initial inspection path; A determination unit, for determining the neighborhood range of the unmanned aerial vehicle at the target position according to the configured total number of optimizations, the maximum neighborhood range, the minimum neighborhood range, and the number of optimization times of the inspection path corresponding to the unmanned aerial vehicle; performing neighborhood search within the neighborhood range of the target position according to the environmental data to obtain multiple next alternative positions of the unmanned aerial vehicle; determining the optimal inspection path based on the weight, the target position, the multiple next alternative positions, and the initial inspection paths and positions of each unmanned aerial vehicle; An optimization unit, for if the score of the optimal inspection path is greater than the score of the initial inspection path, then take the optimal inspection path as the new initial inspection path, and return to execute the steps: for any unmanned aerial vehicle, control the unmanned aerial vehicle to perform inspections according to the new initial inspection path until the unmanned aerial vehicle completes the inspection task.

7. An electronic device, characterized in that, The electronic device includes a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete communication with each other through the communication bus; The memory is used for storing computer programs; The processor is configured to implement the method according to any one of claims 1-5 when executing the programs stored on the memory.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer programs, and when the computer programs are executed by the processor, the method according to any one of claims 1-5 is implemented.

Images