An intelligent offshore first-aid equipment automatic positioning method and system

Abstract

The application provides an intelligent offshore first-aid equipment automatic positioning method and system. Multiple distress signals are received and fused to generate optimal geographic position information and a timestamp. A distress area is determined on a preset sea area electronic map, and the distress position is corrected. The probability distribution of each potential distress area is evaluated to generate a primary search target with the highest probability. A thermal imaging and optical sensing system on a rescue drone is started to scan and generate a high-accuracy potential survivor position. The search and rescue path is dynamically adjusted and optimized, the search strategy is adjusted, the best flight route is updated in real time, and an economic and efficient search and rescue path planning is generated. The accurate coordinates of the potential survivor position are fed back to the nearest rescue coordination center in a timely manner, and the first-aid equipment is guided to go to the scene along the optimal path to implement rescue. The technical scheme provided by the application improves the search and rescue efficiency and accuracy, and ensures rapid and accurate positioning of the potential survivor position.

Description

TECHNICAL FIELD

[0001] The embodiment of the present application relates to the technical field of intelligent marine first-aid equipment automatic positioning, and particularly relates to an intelligent marine first-aid equipment automatic positioning method and system. BACKGROUND

[0002] Marine search and rescue tasks have complexity and urgency, especially in the vast ocean environment, quickly and accurately positioning the distressed personnel and ships is crucial. The traditional search and rescue method relies on a single signal source, which often fails to provide sufficient accuracy and reliability in the changing marine environment. In addition, factors such as adverse weather conditions, complex sea conditions, and unmanned aerial vehicle endurance also increase the difficulty of search and rescue. Therefore, an intelligent automatic positioning method is needed, which can integrate multiple distress signals to generate optimal geographic location information, and combine advanced machine learning algorithms to improve non-fixed object recognition ability and path planning efficiency, to ensure fast and accurate positioning of potential survivor locations.

[0003] Currently, marine search and rescue mainly relies on the following traditional technical means to receive distress signals through satellites and determine the distress location. However, relying solely on satellite signals is susceptible to weather and terrain, resulting in decreased positioning accuracy. Utilizing radio waves for communication and positioning, but its coverage is limited and ineffective in open sea areas. Relying on eyewitness reports or radio communication provided by nearby ships, but this method is slow in response and limited by ship distribution and communication distance.

[0004] However, the existing scheme has the following main defects: a single signal source is susceptible to interference and cannot provide high-precision distress locations, especially in complex and changing marine environments. Traditional methods rely on manual dispatch and manual path planning, making it difficult to achieve fast response and delaying valuable rescue time. Lack of adaptability to different weather conditions and complex sea conditions affects the effectiveness and reliability of search equipment. For example, traditional image recognition technology performs poorly in adverse weather conditions, and the endurance of unmanned aerial vehicles also limits long-term operation.

[0005] In order to overcome the above problems, an intelligent marine first-aid equipment automatic positioning method is proposed, which improves search and rescue efficiency and accuracy, and enhances the robustness and adaptability of the system in complex and changing marine environments, providing strong technical support for marine search and rescue. SUMMARY

[0006] The embodiment of the present application provides an intelligent marine first-aid equipment automatic positioning method and system to solve the problem of low search and rescue efficiency and accuracy in the prior art.

[0007] In a first aspect, the embodiment of the present application provides an intelligent marine first-aid equipment automatic positioning method, comprising:

[0008] receiving and fusing multiple distress signals from satellites, radio relay stations and nearby ships to generate optimal geographic position information and time stamp;

[0009] determining a distress area on a preset sea area electronic map based on the optimal geographic position information, and calculating signal propagation delay according to the time stamp to correct the distress position, while applying a Bayesian network prediction model to evaluate the probability distribution of each potential distress area to generate the primary search target with the highest probability;

[0010] According to the primary search target with the highest probability, starting the thermal imaging and optical sensing system on the rescue drone to scan, combining the convolutional neural network to enhance the non-fixed object recognition ability, and using the generative adversarial network to simulate the image features under different weather conditions, to generate the location of potential survivors with high accuracy;

[0011] Using sea current prediction algorithm and wind direction data to dynamically adjust and optimize the search path to ensure rapid coverage of the distress range where the potential survivor is located, and through reinforcement learning algorithm to adaptively adjust the search strategy, real-time update the best flight route, and consider the remaining power and endurance of the drone, to generate an economic and efficient search path planning;

[0012] Based on the search path planning, the accurate coordinates of the potential survivor's location are fed back to the nearest rescue coordination center in time, and the emergency equipment is guided along the optimal path to implement rescue.

[0013] Optionally, the determination of the distress area on the preset sea area electronic map based on the optimal geographic position information, and the calculation of the signal propagation delay according to the time stamp to correct the distress position, while applying a Bayesian network prediction model to evaluate the probability distribution of each potential distress area to generate the primary search target with the highest probability, includes:

[0014] Using the optimal geographic position information, the initial distress position is matched and processed on the preset sea area electronic map to determine the distress area;

[0015] According to the time stamp, the signal propagation delay is calculated to correct the initial distress position to obtain the corrected distress position;

[0016] Based on the corrected distress position, a Bayesian network prediction model is applied to evaluate the probability distribution of each distress area to generate the occurrence probability value of each distress area;

[0017] According to the occurrence probability value of each distress area, the distress area with the highest probability is selected to generate the primary search target with the highest probability, ensuring that resources are preferentially invested in the distress area with the highest probability.

[0018] Optionally, based on the corrected distress location, a Bayesian network prediction model is applied to evaluate and process the probability distribution of each distress area, and an occurrence probability value of each distress area is generated, including:

[0019] The corrected distress location and surrounding environment parameters are input into the pre-trained Bayesian network prediction model as input variables to obtain basic data;

[0020] According to the conditional probability distribution learned from the historical distress event data set, the possibility of each preset distress area becoming an actual distress location is evaluated and processed, and a preliminary occurrence probability value is output;

[0021] Based on the preliminary occurrence probability value, the occurrence probability value is further adjusted and optimized to ensure accuracy, and an optimized occurrence probability value is obtained;

[0022] According to the optimized occurrence probability value, an occurrence probability value of each distress area is generated to ensure that resources are preferentially invested in the distress area with the highest probability.

[0023] Optionally, according to the primary search target with the highest probability, the thermal imaging and optical sensing system on the rescue drone is started to scan, the convolutional neural network is used to enhance the non-fixed form object recognition capability, and the generative adversarial network is used to simulate image features under different weather conditions to generate a high-accuracy potential survivor location, including:

[0024] Based on the primary search target with the highest probability, the thermal imaging and optical sensing system on the rescue drone is started to scan, and preliminary image data is obtained;

[0025] Using the preliminary image data, the convolutional neural network is used to identify and process non-fixed form objects in the preliminary image data to obtain a preliminary object recognition result;

[0026] According to the preliminary object recognition result, the generative adversarial network is applied to simulate image features under different weather conditions to generate optimized image features;

[0027] Based on the optimized image features, the accurate position of the potential survivor is further analyzed and confirmed, and a high-accuracy potential survivor location is generated.

[0028] Optionally, the preliminary image data is used in combination with the convolutional neural network to identify and process non-fixed form objects in the preliminary image data to obtain a preliminary object recognition result, including:

[0029] The preliminary image data is fed into a pre-trained convolutional neural network as input to perform feature extraction processing on the preliminary image data, generating detailed feature information;

[0030] Based on the detailed feature information, the feature information is further refined through multi-layer convolution and pooling operations of the convolutional neural network, and an optimized feature representation is obtained according to visual characteristics of shape, texture and color;

[0031] The optimized feature representation is classified by a classification layer of the convolutional neural network, and accurate classification is performed on irregular or floating objects to generate a classification result;

[0032] According to the classification result, a preliminary object recognition result is generated, which contains preliminary position information of potential survivors.

[0033] Optionally, the search and rescue path is dynamically adjusted and optimized using the sea current prediction algorithm and wind direction data to ensure rapid coverage of the distress area where the potential survivor is located. The search strategy is adaptively adjusted by a reinforcement learning algorithm to update the best flight route in real time, and the remaining power and endurance of the unmanned aerial vehicle are considered to generate an economic and efficient search and rescue path plan, including:

[0034] The obtained high-accuracy potential survivor location is combined with real-time acquired sea current prediction algorithm and wind direction data to preliminarily plan the search and rescue path, obtaining a preliminary search and rescue path;

[0035] Based on the preliminary search and rescue path, the search and rescue path is dynamically adjusted and optimized according to the sea current prediction algorithm and wind direction data to adapt to the changing marine environment, ensuring the effectiveness and timeliness of the path, and generating an optimized search and rescue path;

[0036] The optimized search and rescue path is adaptively adjusted by a reinforcement learning algorithm to update the search strategy in real time according to the situation encountered during the actual search and rescue process, generating a real-time updated search and rescue path;

[0037] The real-time updated search and rescue path is further processed considering the remaining power and endurance of the unmanned aerial vehicle to ensure that the unmanned aerial vehicle can safely return after completing the task, while avoiding unnecessary energy waste, generating a final flight route;

[0038] Based on the high-accuracy potential survivor location, real-time sea current prediction algorithm and wind direction data, dynamically adjusted and optimized search and rescue path, adaptively adjusted search strategy, and remaining power and endurance of the unmanned aerial vehicle, the final flight route is evaluated and optimized to generate an economic and efficient search and rescue path plan.

[0039] Optionally, the optimized search and rescue path is adaptively adjusted by the reinforcement learning algorithm, and a real-time updated search and rescue path is generated by updating the search strategy in real time according to the situation encountered in the actual search and rescue process, comprising:

[0040] The optimized search and rescue path is used to provide a training basis for the reinforcement learning algorithm by combining the actual search and rescue environment data obtained in real time, and initial training data is obtained;

[0041] Based on the initial training data, the state space in the reinforcement learning algorithm represents the state of the current search and rescue environment, and the position and state of the unmanned aerial vehicle are taken as part of the environment, so as to ensure that the reinforcement learning algorithm can fully understand the current situation and generate an environment state representation;

[0042] According to the environment state representation, the adjustment operation defined by the action space in the reinforcement learning algorithm is generated, which includes changing the flight direction, adjusting the flight height, accelerating or decelerating, generating an adjustment scheme;

[0043] Using the adjustment scheme, the effect of each adjustment operation is evaluated by the reward mechanism in the reinforcement learning algorithm, which considers the factors of search and rescue efficiency, safety and resource consumption, and generates an evaluation result;

[0044] Based on the evaluation result, the reinforcement learning algorithm is applied for iterative learning, and the search strategy is continuously optimized according to the result of each adjustment operation, so that the reinforcement learning algorithm can gradually find the optimal path adjustment scheme in the actual search and rescue process, and generate an optimized search strategy;

[0045] According to the optimized search strategy, the search and rescue path is updated in real time to generate a real-time updated search and rescue path, so as to ensure that the unmanned aerial vehicle can dynamically adjust the flight route according to the latest environmental information and improve the search and rescue efficiency.

[0046] In a second aspect, the embodiments of the present application provide an intelligent offshore emergency equipment automatic positioning system, comprising:

[0047] The acceptance fusion module is used for receiving and fusing multiple distress signals from satellites, radio relay stations and nearby ships, estimating optimal geographic location information and time stamps;

[0048] The determination and evaluation module is used for determining the distress area on the preset sea area electronic map based on the optimal geographic location information, calculating the signal propagation delay according to the time stamp to correct the distress position, and applying a Bayesian network prediction model to evaluate the probability distribution of each potential distress area to generate the primary search target with the highest probability;

[0049] The scanning generation module is configured to start a thermal imaging and optical sensing system on the rescue drone to scan according to the primary search target with the highest probability, to combine a convolutional neural network to enhance the ability of non-fixed object recognition, and to use a generative adversarial network to simulate image features under different weather conditions to generate a high-accuracy potential survivor location.

[0050] The optimization updating module is configured to use a sea current prediction algorithm and wind direction data to dynamically adjust and optimize the search and rescue path, to ensure that the search and rescue path quickly covers a distress range in which the potential survivor location is located, to use a reinforcement learning algorithm to adaptively adjust a search strategy, to update a best flight route in real time, and to consider the remaining power and endurance of the drone to generate an economic and efficient search and rescue path plan.

[0051] The feedback guiding module is configured to feed back accurate coordinates of the potential survivor location to a nearest rescue coordination center in a timely manner based on the search and rescue path plan, and to guide emergency equipment to go to the potential survivor location along an optimal path to implement rescue.

[0052] In a third aspect, an embodiment of the present application provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement the automatic positioning method of the intelligent maritime emergency equipment according to any one of the first aspect.

[0053] In a fourth aspect, an embodiment of the present application provides a computer storage medium, which stores a computer program; when the computer program is executed by a computer, the automatic positioning method of the intelligent maritime emergency equipment according to any one of the first aspect is implemented.

[0054] In the embodiments of the present application, multiple distress signals from satellites, radio relay stations and nearby ships are received and fused to generate optimal geographic position information and a timestamp; based on the optimal geographic position information, a distress area is determined on a preset sea area electronic map, and signal propagation delay is calculated according to the timestamp to correct the distress position, while a Bayesian network prediction model is applied to evaluate the probability distribution of each potential distress area to generate a primary search target with the highest probability; according to the primary search target with the highest probability, a thermal imaging and optical sensing system on a rescue drone is started to scan, a convolutional neural network is used to enhance the non-fixed object recognition capability, and a generative adversarial network is used to simulate image features under different weather conditions to generate a high-accuracy potential survivor position; a sea current prediction algorithm and wind direction data are used to dynamically adjust and optimize the search path to ensure rapid coverage of the distress range where the potential survivor position is located, an adaptive search strategy is adjusted through a reinforcement learning algorithm to update the best flight route in real time, and the remaining power and endurance of the drone are considered to generate an economic and efficient search path planning; based on the search path planning, the accurate coordinates of the potential survivor position are fed back to the nearest rescue coordination center in time, and emergency equipment is guided to go to the scene along the optimal path to implement rescue.

[0055] The technical scheme of the present application has the following beneficial effects:

[0056] The application generates optimal geographic position information and time stamps by receiving and fusing multiple distress signals from satellites, radio relay stations and nearby ships, ensuring high-precision correction of the distress location. Based on the optimal geographic position information, the distress area is determined on the preset sea area electronic map, and the signal propagation delay is calculated combined with the time stamp, further improving the accuracy of the distress location. The Bayesian network prediction model is applied to evaluate the probability distribution of each potential distress area, and the primary search target with the highest probability is generated, so as to concentrate resources for efficient search and avoid invalid coverage. The thermal imaging and optical sensing system on the rescue drone is started to scan, combined with the convolutional neural network to enhance the recognition ability of non-fixed form objects, ensuring accurate identification of potential survivors in complex environments. The generative adversarial network is used to simulate image features under different weather conditions, generating high-accuracy potential survivor locations, adapting to various harsh environments and improving the robustness of identification. The search and rescue path is dynamically adjusted and optimized using the sea current prediction algorithm and wind direction data, ensuring rapid coverage of the distress area where the potential survivor is located. The search strategy is adaptively adjusted through reinforcement learning algorithm, and the best flight route is updated in real time, considering the remaining power and endurance of the drone, to generate an economical and efficient search and rescue path planning, maximizing the operation time of the drone and improving the success rate of search and rescue. The accurate coordinates of the potential survivor location are fed back to the nearest rescue coordination center in time, ensuring the timeliness and accuracy of information transmission. The first-aid equipment is guided along the optimal path to implement rescue, ensuring that the rescue action is quickly launched, shortening the response time and improving the survival probability of the survivors. The whole process fully considers the remaining power and endurance of the drone, ensuring effective use of resources, avoiding unnecessary energy waste, and improving the economy and sustainability of the overall search and rescue operation.

[0057] Further, the embodiments of the application further describe in detail the process of determining the distress area based on the optimal geographic position information, correcting the distress location and generating the primary search target. Specifically, it includes: using the optimal geographic position information to perform matching processing on the initial distress location on the preset sea area electronic map to determine the distress area; calculating the signal propagation delay according to the time stamp to perform correction processing on the initial distress location to obtain the corrected distress location; based on the corrected distress location, applying the Bayesian network prediction model to evaluate the probability distribution of each potential distress area to generate the occurrence probability value of each distress area; selecting the distress area with the highest probability as the primary search target to ensure that resources are preferentially invested in this area. In addition, according to the primary search target with the highest probability, the thermal imaging and optical sensing system on the rescue drone is started to scan, combined with the convolutional neural network to enhance the recognition ability of non-fixed form objects, and the generative adversarial network is used to simulate image features under different weather conditions, finally generating high-accuracy potential survivor locations.

[0058] Through the above method, the present application significantly improves the efficiency and accuracy of maritime search and rescue tasks. First, through multi-source signal fusion and timestamp correction, the high-precision determination of the distress location is ensured, and the positioning error is reduced. Second, the application of the Bayesian network prediction model enables the system to intelligently evaluate the probability distribution of each potential distress area, select the most likely search target, and concentrate resources for efficient search. Finally, combined with thermal imaging, optical sensing system, convolutional neural network and generative adversarial network, the recognition ability of non-fixed form objects and the ability to adapt to different weather conditions are enhanced, and high-accuracy potential survivor locations are generated. Overall, this method not only improves the success rate of search and rescue, but also enhances the robustness and adaptability of the system in complex and variable marine environments, providing strong technical support for fast and accurate rescue.

[0059] These aspects or other aspects of the present application will be more apparent in the following description of the embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0060] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0061] Figure 1 A flow chart of an intelligent maritime first aid equipment automatic positioning method provided by an embodiment of the present application;

[0062] Figure 2 A structural schematic diagram of an intelligent maritime first aid equipment automatic positioning system provided by an embodiment of the present application;

[0063] Figure 3 A structural schematic diagram of a computing device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0064] In order to enable personnel in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application.

[0065] In some of the flowcharts described in the specification and claims of the present application and in the above description of the drawings, a plurality of operations are included in the order in which they occur, but it should be clearly understood that these operations can be executed or performed in parallel or in the order in which they occur in this text, and the serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these flowcharts can include more or fewer operations, and the operations can be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this text are used to distinguish different messages, devices, modules, etc., and do not represent the order of precedence, nor do "first" and "second" represent different types.

[0066] The technical solutions in the embodiments of the present application will be described clearly and completely in the present application with reference to 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, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0067] Figure 1 A flowchart of an intelligent offshore first-aid equipment automatic positioning method is provided for the embodiments of the present application, as shown in Figure 1 The method comprises:

[0068] 101, receiving and fusing multiple distress signals from satellites, radio relay stations and nearby ships to generate optimal geographic location information and time stamps;

[0069] It is related to multi-source signal fusion technology, aiming to generate optimal geographic location information and time stamps by receiving and integrating distress signals from multiple different sources (such as satellites, radio relay stations and nearby ships). Specifically, satellite signals provide high-precision global positioning data, radio relay stations provide regional communication support, and nearby ships provide immediate eyewitness reports or auxiliary communication. These data are used together to determine the distress location and ensure the time synchronization of all signals through time stamps, thereby improving the accuracy and reliability of positioning.

[0070] The system first receives multiple distress signals from satellites, radio relay stations and nearby ships. Then, using signal processing algorithms, the signals are preprocessed to remove noise and interference, ensuring the integrity and accuracy of the data. Next, the signals from different sources are time and space aligned to generate a unified timestamp, eliminating the time deviation caused by signal propagation delay. Finally, through multi-source data fusion algorithms, the weights and reliabilities of each signal source are considered to generate optimal geographic location information. This process not only improves the positioning accuracy, but also enhances the robustness of the system.

[0071] In a simulated search and rescue mission, a fishing boat sends out a distress signal when it encounters trouble in a certain sea area. The satellite receives the signal and provides preliminary geographic coordinates; a nearby radio relay station also captures the signal and supplements detailed information within the region; at the same time, a nearby merchant ship reports the sighting through VHF radio. The system fuses these three signals together to generate accurate geographic location information and a timestamp. Based on this information, the search and rescue center quickly initiates follow-up rescue operations.

[0072] 102、Based on the optimal geographic location information, determine the distress area on the preset sea area electronic map, and calculate the signal propagation delay according to the timestamp to correct the distress location, while applying a Bayesian network prediction model to evaluate the probability distribution of each potential distress area, generating the primary search target with the highest probability;

[0073] The core is to accurately locate the distress area on the preset sea area electronic map using the optimal geographic location information, and correct the initial location by calculating the signal propagation delay based on the timestamp. In addition, a Bayesian network prediction model is applied to evaluate the probability distribution of each potential distress area, and the most likely search target is selected. This step ensures effective allocation of resources, avoids blind search, and improves search efficiency.

[0074] Based on the generated optimal geographic location information, the system matches the specific distress area on the preset sea area electronic map. Then, according to the timestamp, the signal propagation delay is calculated to correct the initial distress location, obtaining a more accurate distress location. Next, a Bayesian network prediction model is applied, combined with a historical dataset of distress events, to evaluate the probability distribution of each potential distress area, and finally select the distress area with the highest probability as the primary search target. In this way, the system can concentrate resources for efficient search and reduce ineffective coverage.

[0075] In the above-mentioned simulated search and rescue mission, the system determines the possible distress area of the fishing boat on the preset sea area electronic map based on the generated optimal geographic location information. Considering the signal propagation delay, the system corrects the initial location and obtains a more accurate distress location. Then, a Bayesian network prediction model is applied to evaluate the probability distribution of multiple potential distress areas, and the result shows that a certain specific area has the highest probability. The search and rescue center accordingly prioritizes sending a drone to search in this area, greatly improving search and rescue efficiency.

[0076] 103、According to the primary search target with the highest probability, start the thermal imaging and optical sensing system on the rescue drone for scanning, combine convolutional neural networks to enhance non-fixed object recognition capabilities, and use generative adversarial networks to simulate image features under different weather conditions, generating high-accuracy potential survivor locations;

[0077] The thermal imaging and optical sensing systems on the rescue drone are focused on identifying potential survivors. Advanced machine learning algorithms, such as convolutional neural networks and generative adversarial networks, are used to enhance the recognition of non-rigid objects and generate highly accurate potential survivor locations. The thermal imaging system detects human heat in low visibility conditions, while the optical sensing system captures visual information. Convolutional neural networks extract features from images, and generative adversarial networks simulate image features under different weather conditions to ensure accurate identification of potential survivors in various environments.

[0078] Based on the generated primary search target, the system activates the thermal imaging and optical sensing systems on the rescue drone to scan and obtain preliminary image data. Convolutional neural networks are then used to extract features and classify the image data, identifying non-rigid objects such as floating bodies. To account for complex weather conditions, generative adversarial networks simulate image features under different weather conditions, further optimizing the image recognition results. Finally, the system generates highly accurate potential survivor locations, providing reliable information for subsequent rescue operations.

[0079] In the aforementioned search and rescue mission, the drone travels to the designated area based on the primary search target. The thermal imaging system successfully detects several suspected human heat points at night, while the optical sensing system captures visual information of some floating objects. Convolutional neural networks analyze the image data and identify one floating object as a potential survivor. Generative adversarial networks simulate image features under current weather conditions, further confirming the object's identity. The search and rescue team quickly launches a rescue operation based on these highly accurate location information.

[0080] 104、Utilizing sea current prediction algorithms and wind direction data, the search and rescue path is dynamically adjusted and optimized to ensure rapid coverage of the distress area where the potential survivor is located. Reinforcement learning algorithms are used to adaptively adjust the search strategy, update the best flight route in real time, and consider the remaining battery life and endurance of the drone to generate an economic and efficient search and rescue path plan.

[0081] This step involves dynamically adjusting and optimizing the search and rescue path using sea current prediction algorithms and wind direction data, combined with reinforcement learning algorithms. This ensures rapid coverage of the area where the potential survivor is located. Sea current prediction algorithms are used to predict the direction and speed of ocean currents, while wind direction data helps the drone adjust its flight attitude. Reinforcement learning algorithms continuously learn and optimize the search strategy, updating the best flight route in real time while considering the remaining battery life and endurance of the drone to generate an economic and efficient path plan. This step ensures the efficiency and sustainability of the search and rescue operation.

[0082] The system utilizes sea current prediction algorithms and wind direction data to predict the direction and speed of ocean currents, as well as the impact of wind on the drone's flight. Based on this, the system adaptively adjusts the search strategy through reinforcement learning algorithms, updating the optimal flight path in real time to ensure that the drone can quickly cover the area where potential survivors are located. At the same time, the system continuously monitors the remaining power and endurance of the drone, dynamically adjusting the path planning to ensure that the drone completes the task before running out of power and returns safely. This not only improves search and rescue efficiency, but also maximizes the operating time of the drone.

[0083] In the search and rescue mission, the drone begins to execute the search path planning based on the high-accuracy potential survivor location information. The system utilizes sea current prediction algorithms and wind direction data to predict the direction and speed of ocean currents, adjusting the flight path of the drone. The reinforcement learning algorithm continuously optimizes the search strategy, updating the optimal flight path in real time to ensure that the drone can quickly cover the distress area. At the same time, the system monitors the remaining power of the drone, timely adjusts the path, and ensures that the drone successfully completes the task and returns safely to the base. This dynamic adjustment makes the search and rescue operation more efficient and economical.

[0084] 105、Based on the search path planning, the accurate coordinates of the potential survivor location are timely fed back to the nearest rescue coordination center, and the emergency equipment is guided along the optimal path to implement rescue.

[0085] The core is to timely feed back the accurate coordinates of the potential survivor location to the nearest rescue coordination center, and guide the emergency equipment along the optimal path to implement rescue. This process ensures the timeliness and accuracy of information transmission, enabling the rescue team to respond quickly, shorten the rescue time, and improve the survival probability of survivors. By integrating communication systems and navigation technology, the system realizes seamless connection from information collection to rescue implementation.

[0086] The system feeds back the accurate coordinates of the potential survivor location to the nearest rescue coordination center in a timely manner based on the generated search path planning. The rescue coordination center transmits this information to the ground rescue team and other relevant units through the communication system. At the same time, the system guides the emergency equipment (such as rescue ships, helicopters, etc.) along the optimal path to implement rescue. Navigation technology ensures that the equipment can quickly and accurately reach the designated location, improving the overall efficiency of the rescue operation.

[0087] In the aforementioned search and rescue mission, the system timely feeds back the precise coordinates of the potential survivor's location discovered by the UAV to the nearest rescue coordination center. The rescue coordination center quickly conveys this information to the ground rescue team and nearby merchant ships. At the same time, the system guides the rescue ships and helicopters to the implementation of rescue along the optimal path. Navigation technology ensures that these devices can quickly and accurately reach the designated location, greatly shortening the rescue time. Finally, the rescue team successfully rescued the distressed fishermen, significantly improving their survival probability.

[0088] Through the implementation of steps 101 to 105, the method of the present application significantly improves the efficiency and accuracy of maritime search and rescue missions. First, multi-source signal fusion and time stamp correction ensure high-precision determination of the distress location, reducing positioning errors. Second, the application of the Bayesian network prediction model enables the system to intelligently evaluate the probability distribution of each potential distress area, select the most likely search target, and concentrate resources for efficient search. Third, combining thermal imaging, optical sensing systems, convolutional neural networks, and generative adversarial networks enhances the ability to identify non-fixed objects and adapt to different weather conditions, generating high-accuracy potential survivor locations. Fourth, using sea current prediction algorithms and wind direction data, combined with reinforcement learning algorithms, dynamically adjusts and optimizes the search and rescue path to ensure rapid coverage of the potential survivor's area, while considering the remaining power of the UAV, generating an economical and efficient path planning. Finally, through the integration of communication systems and navigation technology, seamless connection from information collection to rescue implementation is achieved, ensuring that the rescue team can respond quickly, shorten the rescue time, and improve the survival probability of the survivors. Overall, this method not only improves the success rate of search and rescue, but also enhances the robustness and adaptability of the system in complex and variable marine environments, providing strong technical support for fast and accurate rescue.

[0089] To solve the problem of signal propagation delay and the accuracy of distress location correction, in some embodiments, the distress area determined based on the optimal geographic location information in step 102 on the preset sea area electronic map, and the signal propagation delay is calculated according to the time stamp to correct the distress location, while applying the Bayesian network prediction model to evaluate the probability distribution of each potential distress area, generating the primary search target with the highest probability, including:

[0090] The optimal geographic position information is used for matching processing of the initial distress position on a preset sea area electronic map to determine a distress area; a signal propagation delay is calculated according to the timestamp to correct the initial distress position to obtain a corrected distress position; a Bayesian network prediction model is applied to evaluate the probability distribution of each distress area based on the corrected distress position to generate an occurrence probability value of each distress area; and the distress area with the highest probability is selected as a primary search target according to the occurrence probability value of each distress area to ensure that resources are preferentially invested in the distress area with the highest probability.

[0091] In this embodiment, the optimal geographic position information generated by multi-source signal fusion is matched with a preset sea area electronic map. The preset sea area electronic map contains detailed marine geographic information such as water depth, seabed topography, navigation marks, etc. Through this matching processing, the system can accurately locate the specific area where the distress event occurs. The optimal geographic position information includes multiple distress signal data from satellites, radio relay stations and nearby ships to ensure the high accuracy and reliability of the positioning. The timestamp is used to record the receiving time of each distress signal. Since the signal needs a certain time to propagate from the distress location to the receiving point, there is a propagation delay. Through the accurate timestamp, the system can calculate this delay and correct the initial distress position accordingly. The correction process takes into account factors such as signal propagation speed, path loss, etc., to ensure that the final determined distress position is more accurate. The Bayesian network prediction model is a method based on probabilistic reasoning, which learns the conditional probability distribution from a historical distress event dataset to evaluate the likelihood of each potential distress area becoming an actual distress location. This model takes into account various environmental parameters (such as sea currents, wind direction, weather conditions, etc.) and generates an occurrence probability value for each distress area. This step improves the accuracy and reliability of the prediction. The system selects the distress area with the highest probability as the primary search target according to the occurrence probability value generated by the Bayesian network prediction model. This way, resources can be concentrated in the most likely area for search, avoiding ineffective coverage and improving search efficiency. In addition, it also ensures the effective allocation of resources, making the rescue operation more efficient.

[0092] In the embodiment of the present application, the system first uses the optimal geographic position information to match the initial distress position on the preset sea area electronic map to determine the specific distress area. Then, the signal propagation delay is calculated according to the timestamp to correct the initial distress position to obtain a more accurate corrected distress position. Then, the Bayesian network prediction model is applied to evaluate the probability distribution of each distress area to generate an occurrence probability value of each distress area. Finally, according to these probability values, the distress area with the highest probability is selected as the primary search target to ensure that resources are preferentially invested in this area. The whole process not only improves the accuracy of the distress position, but also enhances the targeting and efficiency of the search and rescue operation.

[0093] Here is a specific example:

[0094] In a specific maritime search and rescue mission, when a fishing boat encountered distress in a certain sea area, the system received multiple distress signals and generated optimal geographic location information through multi-source signal fusion. Next, the system matched this information with the preset electronic map of the sea area to determine the possible distress area of the fishing boat. Considering the signal propagation delay, the system corrected the initial position and obtained a more accurate distress location. Then, the Bayesian network prediction model was applied to evaluate the probability distribution of multiple potential distress areas, and the results showed that a specific area had the highest probability. The search and rescue center accordingly prioritized sending a drone to the area for search, significantly improving the search and rescue efficiency. After the drone arrived at the designated area, it successfully found the distressed fishermen and guided the subsequent rescue forces to quickly launch rescue operations, ultimately successfully rescuing the distressed personnel.

[0095] In this way, the system not only improves the accuracy of the distress location, but also enhances the robustness and adaptability of the system in complex and variable marine environments, providing strong technical support for fast and accurate rescue.

[0096] To solve the problem of accuracy and reliability of distress area evaluation, in some embodiments, the step 102 includes applying a Bayesian network prediction model to evaluate the probability distribution of each distress area based on the corrected distress location, generating the occurrence probability value of each distress area, including:

[0097] The corrected distress location and surrounding environmental parameters are input as input variables into the pre-trained Bayesian network prediction model to obtain basic data. According to the basic data, the conditional probability distribution learned from the historical distress event data set is combined to evaluate the possibility of each preset distress area becoming an actual distress location, outputting a preliminary occurrence probability value. Based on the preliminary occurrence probability value, the occurrence probability value is further adjusted and optimized to ensure accuracy, obtaining an optimized occurrence probability value. According to the optimized occurrence probability value, the occurrence probability value of each distress area is generated to ensure that resources are preferentially invested in the distress area with the highest probability.

[0098] In this embodiment, the corrected distress location and surrounding environmental parameters (such as sea currents, wind direction, weather conditions, etc.) are input into a pre-trained Bayesian network prediction model. These environmental parameters provide additional contextual information, which helps to more accurately assess the likelihood of potential distress areas. The base data refers to the probability distribution calculated by the model under various conditions. The Bayesian network prediction model learns the conditional probability distribution of each distress area becoming an actual distress location under different conditions using a historical distress event dataset. By combining the base data with these conditional probability distributions, the system can assess the likelihood of each preset distress area becoming an actual distress location and output a preliminary probability value. This step improves the accuracy of the assessment and ensures the reliability of the results. Although the preliminary probability value has considered multiple factors, to further improve accuracy, the system adjusts and optimizes the preliminary probability value based on the latest environmental data and other real-time information. For example, if new distress signals or environmental changes are discovered during the assessment process, the system will recalculate and update the probability value to ensure the final result is as accurate as possible. Finally, the system generates the probability of occurrence for each distress area based on the optimized probability value. The distress area with the highest probability is chosen as the primary search target, ensuring that resources are prioritized in that area. This not only improves search efficiency but also minimizes ineffective coverage, focusing resources on the most likely areas for search.

[0099] In the embodiments of the present application, the system first inputs the corrected distress location and surrounding environmental parameters as input variables into a pre-trained Bayesian network prediction model to obtain base data. Then, based on the conditional probability distribution learned from the historical distress event dataset, the system evaluates the likelihood of each preset distress area becoming an actual distress location and outputs a preliminary probability value. Next, based on the preliminary probability value, the system further adjusts and optimizes these probability values to ensure their accuracy. Finally, the system generates the probability of occurrence for each distress area and selects the distress area with the highest probability as the primary search target, ensuring that resources are prioritized in that area. The entire process not only improves the accuracy of the assessment but also enhances the robustness and adaptability of the system.

[0100] Here is a specific example:

[0101] In a specific maritime search and rescue mission, when a fishing boat encountered distress in a certain sea area, the system received multiple distress signals and generated the optimal geographic location information through multi-source signal fusion. Considering the signal propagation delay, the system corrected the initial position and obtained a more accurate distress location. Then, the system input the corrected distress location and surrounding environmental parameters (such as sea currents, wind direction, weather conditions, etc.) into the pre-trained Bayesian network prediction model to obtain basic data. Then, combined with the conditional probability distribution learned from the historical distress event dataset, the system evaluated the likelihood of multiple potential distress areas becoming the actual distress location and output the preliminary probability value. To further improve accuracy, the system adjusted and optimized based on the latest environmental data, and finally generated the probability value of each distress area. The results showed that a certain specific area had the highest probability, and the search and rescue center accordingly prioritized sending a drone to that area for search, significantly improving search and rescue efficiency. After the drone arrived at the designated area, it successfully found the distressed fishermen and guided the subsequent rescue forces to quickly launch rescue operations, ultimately successfully rescuing the distressed personnel.

[0102] In this way, the system not only improves the accuracy of distress location evaluation, but also enhances the robustness and adaptability of the system in complex and variable marine environments, providing strong technical support for fast and accurate rescue.

[0103] To solve the problem of accuracy of non-fixed form object recognition in complex weather conditions, in some embodiments, step 103 includes starting the thermal imaging and optical sensing system on the rescue drone to scan based on the primary search target with the highest probability, combining convolutional neural networks to enhance non-fixed form object recognition capabilities, and using generative adversarial networks to simulate image features under different weather conditions to generate high-accuracy potential survivor locations, including:

[0104] Based on the primary search target with the highest probability, the thermal imaging and optical sensing system on the rescue drone is started to scan and obtain preliminary image data. Using the preliminary image data, combined with convolutional neural networks, non-fixed form objects in the preliminary image data are identified and processed to obtain preliminary object recognition results. According to the preliminary object recognition results, generative adversarial networks are applied to simulate image features under different weather conditions to generate optimized image features. Based on the optimized image features, the precise location of potential survivors is further analyzed and confirmed to generate high-accuracy potential survivor locations.

[0105] In this embodiment, the thermal imaging and optical sensing systems on the rescue drone are activated to scan the primary search target with the highest probability. The thermal imaging system can detect human heat in low visibility or nighttime conditions, while the optical sensing system is used to capture visual information. The preliminary image data collected by these sensors serves as the basis for further processing to analyze and confirm the presence of potential survivors. Convolutional neural networks are deep learning models that excel at extracting complex features from images. By inputting the preliminary image data into a pre-trained convolutional neural network, the system can identify non-fixed objects in the image, such as floating bodies or shipwreck debris. The preliminary object identification results provide a preliminary judgment about the presence of potential survivors, laying the foundation for further analysis. The generative adversarial network consists of a generator and a discriminator, which can simulate image features under different weather conditions. Based on the preliminary object identification results, the system uses the generative adversarial network to generate optimized image features, taking into account the impact of various adverse weather conditions (such as fog, rain, strong wind, etc.) on the image. This step improves the robustness of image recognition, ensuring accurate identification of potential survivors in complex and variable environments. Finally, the system further analyzes and confirms the precise location of potential survivors based on the optimized image features. By integrating multiple image features and environmental parameters, the system can generate high-accuracy potential survivor locations, providing reliable evidence for subsequent rescue operations. This step not only improves the accuracy of identification but also enhances the adaptability and reliability of the system.

[0106] In the embodiments of the present application, the system first activates the thermal imaging and optical sensing systems on the rescue drone to scan the primary search target with the highest probability and obtain preliminary image data. Then, the preliminary image data is input into a pre-trained convolutional neural network to identify non-fixed objects in the image and obtain preliminary object identification results. Next, the system applies a generative adversarial network to simulate image features under different weather conditions and generate optimized image features. Finally, based on the optimized image features, the system further analyzes and confirms the precise location of potential survivors, generating high-accuracy potential survivor locations. The entire process not only improves the accuracy of non-fixed object identification but also enhances the robustness and adaptability of the system under complex and variable weather conditions.

[0107] Here is a specific example:

[0108] In a specific maritime search and rescue mission, after the system determines the primary search target with the highest probability, it immediately activates the thermal imaging and optical sensing systems on the rescue drone to scan the area. The drone conducts a detailed scan in the designated area and collects preliminary image data. These image data include the heat distribution captured by the thermal imaging system and the visual information captured by the optical sensing system. Next, the system inputs these preliminary image data into the pre-trained convolutional neural network and identifies multiple suspected floating objects, obtaining preliminary object recognition results. To cope with the current complex weather conditions (such as heavy fog), the system applies the generative adversarial network to generate optimized image features, taking into account the impact of fog on images. Finally, based on the optimized image features, the system further analyzes and confirms the precise location of potential survivors, generating high-accuracy potential survivor locations. Based on these high-accuracy location information, the search and rescue team quickly launches a rescue operation and successfully rescues the stranded personnel.

[0109] In this way, the system not only improves the accuracy of non-fixed form object recognition, but also enhances the robustness and adaptability of the system under complex and variable weather conditions, providing strong technical support for rapid and accurate rescue.

[0110] To solve the problem of non-fixed form object recognition accuracy under complex weather conditions, in some embodiments, the step 103 of using the preliminary image data, combined with the convolutional neural network, to identify non-fixed form objects in the preliminary image data and obtain preliminary object recognition results includes:

[0111] Using the preliminary image data as input to feed into the pre-trained convolutional neural network, the preliminary image data is subjected to feature extraction processing to generate detailed feature information; based on the detailed feature information, the convolutional neural network is further refined through multi-layer convolution and pooling operations, and the feature information is further refined according to the visual characteristics of shape, texture and color to obtain optimized feature representation; through the classification layer of the convolutional neural network, the optimized feature representation is subjected to classification processing, and the irregular or floating objects are accurately classified to generate classification results; according to the classification results, the preliminary object recognition results are generated, which contain preliminary location information of potential survivors.

[0112] In this embodiment, the drone captures preliminary image data using thermal imaging and optical sensing systems. This data is fed into a pre-trained convolutional neural network (CNN). CNNs are deep learning models that excel at extracting complex features from images. The preliminary image data includes thermal distribution captured by thermal imaging and visual information captured by optical sensing. This data is processed to extract detailed features such as edges, textures, and colors. The CNN refines these features through multiple layers of convolution and pooling operations. Convolution operations extract local features like edges and textures, while pooling operations reduce the size of the feature maps, retaining only the most important features. Through these operations, the system generates more optimized feature representations based on visual characteristics such as shape, texture, and color. This process enhances the ability to recognize non-fixed objects, such as floating bodies or shipwrecks, improving the accuracy of classification. The classification layer, the final part of the CNN, maps the optimized feature representations to different classes. For maritime search and rescue tasks, the classification layer focuses on identifying irregular or floating objects such as human bodies, life jackets, and debris. By classifying the optimized feature representations, the system generates accurate classification results, distinguishing potential survivors from other objects. Finally, the system generates preliminary object recognition results based on the classification results, which include not only the identified object classes but also the location information of these objects in the image. In particular, for potential survivors, the system generates preliminary location information, providing reliable evidence for subsequent rescue operations. This step ensures the practicality and reliability of the recognition results, improving search and rescue efficiency.

[0113] In this embodiment, the system first feeds the preliminary image data into a pre-trained CNN as input, extracts detailed features, and generates detailed feature information. Then, through multiple layers of convolution and pooling operations in the CNN, the system further refines the feature information and obtains optimized feature representations based on visual characteristics such as shape, texture, and color. Next, the system classifies the optimized feature representations through the classification layer of the CNN, accurately classifies irregular or floating objects, and generates classification results. Finally, based on the classification results, the system generates preliminary object recognition results, which include preliminary location information of potential survivors. The entire process not only improves the accuracy of non-fixed object recognition but also enhances the robustness and adaptability of the system in complex and variable environments.

[0114] Here is a specific example:

[0115] In a specific maritime search and rescue mission, after the system determines the highest probability primary search target, it immediately activates the thermal imaging and optical sensing systems on the rescue drone to scan the area. The drone conducts a detailed scan in the designated area, collecting preliminary image data. These image data contain the heat distribution captured by the thermal imaging system and the visual information captured by the optical sensing system. Next, the system inputs these preliminary image data into the pre-trained convolutional neural network for feature extraction, generating detailed feature information. Then, through multiple layers of convolution and pooling operations, the system further refines the feature information, obtaining optimized feature representations based on visual characteristics such as shape, texture, and color. Subsequently, the system classifies the optimized feature representations through the classification layer of the convolutional neural network, identifies multiple suspected floating objects, and generates classification results. Finally, the system generates preliminary object recognition results based on the classification results, confirming the preliminary location information of potential survivors. The search and rescue team quickly launches a rescue operation based on these high-accuracy location information, successfully rescuing the distressed personnel.

[0116] In this way, the system not only improves the accuracy of non-fixed object recognition, but also enhances the robustness and adaptability of the system under complex and variable weather conditions, providing strong technical support for rapid and accurate rescue.

[0117] To solve the efficiency and reliability problem of search and rescue path planning in complex marine environment, in some embodiments, the search and rescue path is dynamically adjusted and optimized by using the sea current prediction algorithm and wind direction data in step 104, to ensure rapid coverage of the distress range where the potential survivor is located. Through reinforcement learning algorithm, the search strategy is adaptively adjusted, the best flight route is updated in real time, and the remaining power and endurance of the drone are considered to generate an economic and efficient search and rescue path planning, including:

[0118] With the high-accuracy potential survivor position obtained, a preliminary search and rescue path is planned by combining real-time acquired sea current prediction algorithm and wind direction data, and a preliminary search and rescue path is obtained. Based on the preliminary search and rescue path, the search and rescue path is dynamically adjusted and optimized according to the sea current prediction algorithm and the wind direction data to adapt to the changing marine environment, ensure the effectiveness and timeliness of the path, and generate an optimized search and rescue path. Through a reinforcement learning algorithm, the optimized search and rescue path is adaptively adjusted, the search strategy is updated in real time according to the situation encountered in the actual search and rescue process, and a real-time updated search and rescue path is generated. Considering the residual power and endurance of the unmanned aerial vehicle, the real-time updated search and rescue path is further processed to ensure that the unmanned aerial vehicle can safely return after completing the task, while avoiding unnecessary energy waste, and a final flight route is generated. Based on the high-accuracy potential survivor position, real-time sea current prediction algorithm and wind direction data, dynamically adjusted and optimized search and rescue path, adaptively adjusted search strategy, and residual power and endurance of the unmanned aerial vehicle, the final flight route is evaluated and optimized to generate an economic and efficient search and rescue path planning.

[0119] In this embodiment, the high-accuracy potential survivor location information is combined with real-time acquired sea current prediction algorithms and wind direction data to perform preliminary search and rescue path planning. The sea current prediction algorithms are used to predict the direction and speed of ocean flow, while the wind direction data helps adjust the flight attitude and path of the UAV. These data are used together to generate a preliminary search and rescue path, ensuring the basic accuracy of the path planning. Although the preliminary search and rescue path has considered various factors, in order to respond to the changing marine environment (such as sudden changes in sea currents or changes in wind direction), the system dynamically adjusts and optimizes the search and rescue path according to the latest sea current prediction algorithms and wind direction data. This process ensures the effectiveness and timeliness of the path, enabling the UAV to quickly and accurately cover the distress area where the potential survivor is located. The reinforcement learning algorithm is a machine learning method that can adaptively adjust behavior strategies in a changing environment. The system uses the reinforcement learning algorithm to update the search strategy in real time based on the actual search and rescue process (such as new distress signals, environmental changes, etc.), generating real-time updated search and rescue paths. This step improves the flexibility and response speed of the system, ensuring that the search and rescue action can quickly adapt to new situations. In order to ensure the safety and energy saving of the UAV, the system considers the remaining power and endurance of the UAV to further process the real-time updated search and rescue path. The system dynamically adjusts the path to ensure that the UAV can safely return to base after completing the task, while avoiding unnecessary energy waste. This step not only improves search and rescue efficiency, but also prolongs the operation time of the UAV. Finally, the system evaluates and optimizes the final flight route based on all the above information, generating an economic and efficient search and rescue path planning. This step ensures the efficiency and sustainability of the entire search and rescue operation, maximizing the success rate of rescue.

[0120] In the embodiments of the present application, the system first uses high-accuracy potential survivor location information and real-time acquired sea current prediction algorithms and wind direction data to perform preliminary search and rescue path planning. Then, according to the latest sea current prediction algorithms and wind direction data, the search and rescue path is dynamically adjusted and optimized to ensure the effectiveness and timeliness of the path. Next, through the reinforcement learning algorithm, the optimized search and rescue path is adaptively adjusted to update the search strategy in real time. In order to ensure the safety and energy saving of the UAV, the system considers the remaining power and endurance of the UAV to further process the real-time updated search and rescue path, generating the final flight route. Finally, the system evaluates and optimizes the final flight route to generate an economic and efficient search and rescue path planning. The entire process not only improves the efficiency and reliability of search and rescue path planning, but also enhances the robustness and adaptability of the system in complex and variable marine environments.

[0121] Here is a specific example:

[0122] In a specific maritime search and rescue mission, after the system determines the high-accuracy potential survivor location, the search and rescue path planning is immediately initiated. First, the system combines real-time acquired sea current prediction algorithms and wind direction data to conduct preliminary search and rescue path planning, generating a preliminary search and rescue path. Then, the system dynamically adjusts and optimizes the search and rescue path according to the latest sea current prediction algorithms and wind direction data, ensuring the effectiveness and timeliness of the path. During the search and rescue process, the system updates the search strategy in real time according to the actual situation (such as new distress signals, sudden changes in sea currents, etc.) through reinforcement learning algorithms, generating real-time updated search and rescue paths. In order to ensure the safety and energy saving of the unmanned aerial vehicle, the system comprehensively considers the remaining power and endurance of the unmanned aerial vehicle, and further processes the real-time updated search and rescue path to generate the final flight route. Finally, the system evaluates and optimizes the final flight route to generate an economic and efficient search and rescue path planning. The unmanned aerial vehicle quickly launches rescue operations according to this path and successfully finds and rescues the distressed personnel.

[0123] In this way, the system not only improves the efficiency and reliability of search and rescue path planning, but also enhances the robustness and adaptability of the system in complex and variable marine environments, providing strong technical support for rapid and accurate rescue.

[0124] In order to solve the flexibility and efficiency of search and rescue path dynamic adjustment in complex marine environment, in some embodiments, the step 104 of adaptively adjusting the optimized search and rescue path through reinforcement learning algorithm, updating the search strategy in real time according to the actual search and rescue process, generating real-time updated search and rescue path, includes:

[0125] With the optimized search and rescue path, combined with the actual search and rescue environment data obtained in real time, a training basis is provided for the reinforcement learning algorithm, and initial training data is obtained; based on the initial training data, the state space in the reinforcement learning algorithm represents the current search and rescue environment state, and the position and state of the unmanned aerial vehicle are taken as part of the environment, ensuring that the reinforcement learning algorithm can fully understand the current situation and generate an environment state representation; according to the environment state representation, the adjustment operations defined by the action space in the reinforcement learning algorithm are generated, including changing the flight direction, adjusting the flight height, accelerating or decelerating, generating an adjustment scheme; using the adjustment scheme, the effect of each adjustment operation is evaluated by the reward mechanism in the reinforcement learning algorithm, which considers the factors of search and rescue efficiency, safety and resource consumption, and generates an evaluation result; based on the evaluation result, the reinforcement learning algorithm is applied for iterative learning, and the search strategy is continuously optimized according to the result of each adjustment operation, so that the reinforcement learning algorithm can gradually find the optimal path adjustment scheme in the actual search and rescue process, and generate an optimized search strategy; according to the optimized search strategy, the search and rescue path is updated in real time, and a real-time updated search and rescue path is generated, ensuring that the unmanned aerial vehicle can dynamically adjust the flight route according to the latest environmental information, and improving the search and rescue efficiency.

[0126] In this embodiment, the optimized search and rescue path is combined with real-time acquired actual search and rescue environment data (such as sea current, wind direction, weather conditions, etc.) as the training basis of the reinforcement learning algorithm. The initial training data not only includes the position and state of the unmanned aerial vehicle, but also includes various parameters of the surrounding environment, ensuring that the algorithm can fully understand the current situation. The state space in the reinforcement learning algorithm is used to represent the current search and rescue environment state. The state space includes the position, height, speed and other state information of the unmanned aerial vehicle, as well as various parameters of the surrounding environment (such as sea current direction and speed, wind direction and intensity, etc.). In this way, the system can generate a detailed environment state representation, ensuring that the reinforcement learning algorithm can fully understand the current search and rescue environment. The action space defines all possible operations that the unmanned aerial vehicle can perform, such as changing the flight direction, adjusting the flight height, accelerating or decelerating, etc. According to the environment state representation, the system generates a series of possible adjustment schemes to cope with the current search and rescue environment. These adjustment schemes aim to optimize the search and rescue path, improve the search and rescue efficiency and safety. The reward mechanism is used to evaluate the effect of each adjustment operation. The reward mechanism considers multiple factors, such as search and rescue efficiency (quickly covering the distress area), safety (avoiding dangerous areas), and resource consumption (saving power and fuel). Through this evaluation, the system can determine which adjustment operation is most beneficial to achieve the goal and generate an evaluation result. The system applies the reinforcement learning algorithm for iterative learning according to the evaluation result. After each adjustment operation, the system will feedback according to the actual effect, gradually optimizing the search strategy. Through continuous iterative learning, the system can gradually find the optimal path adjustment scheme in the actual search and rescue process, further improving the search and rescue efficiency and reliability. Finally, the system updates the search and rescue path in real time according to the optimized search strategy, generating the latest flight route. This process ensures that the unmanned aerial vehicle can dynamically adjust the flight route according to the latest environmental information, quickly respond to changing search and rescue conditions, and improve search and rescue efficiency and success rate.

[0127] In the embodiments of the present application, the system first combines the optimized search and rescue path with the real-time acquired actual search and rescue environment data as the training basis of the reinforcement learning algorithm, generating the initial training data. Then, the system represents the current search and rescue environment state through the state space, and takes the position and state of the unmanned aerial vehicle as part of the environment, generating the environment state representation. Next, the system generates a series of possible adjustment schemes through the adjustment operations defined by the action space according to the environment state representation. Using these adjustment schemes, the system evaluates the effect of each adjustment operation through the reward mechanism, generating an evaluation result. Based on the evaluation result, the system applies the reinforcement learning algorithm for iterative learning, continuously optimizing the search strategy. Finally, the system updates the search and rescue path in real time according to the optimized search strategy, generating the latest flight route. The whole process not only improves the flexibility and efficiency of dynamic adjustment of the search and rescue path, but also enhances the robustness and adaptability of the system in complex and variable environments.

[0128] The following is a specific example:

[0129] In a specific maritime search and rescue mission, after the system determines the optimized search and rescue path, it immediately starts the reinforcement learning algorithm for adaptive adjustment. First, the system combines the optimized search and rescue path with the real-time acquired actual search and rescue environment data (such as sea currents, wind direction, weather conditions, etc.) as the training basis to generate initial training data. Then, the state space represents the current search and rescue environment state, and the position and state of the unmanned aerial vehicle are part of the environment to generate a detailed environment state representation. Next, the system generates a series of possible adjustment schemes according to the environment state representation through the adjustment operations defined by the action space (such as changing the flight direction, adjusting the flight height, accelerating or decelerating). Using these adjustment schemes, the system evaluates the effect of each adjustment operation through the reward mechanism, considering factors such as search efficiency, safety, and resource consumption, and generates evaluation results. Based on the evaluation results, the system applies the reinforcement learning algorithm for iterative learning to continuously optimize the search strategy. Finally, the system updates the search and rescue path in real time according to the optimized search strategy to generate the latest flight route. The unmanned aerial vehicle quickly carries out rescue operations according to this path and successfully finds and rescues the distressed personnel.

[0130] In this way, the system not only improves the flexibility and efficiency of dynamic adjustment of the search and rescue path, but also enhances the robustness and adaptability of the system in complex and variable marine environments, providing strong technical support for fast and accurate rescue.

[0131] The present application considers that in maritime search and rescue missions, accurately evaluating the likelihood of each pre-set distress area becoming the actual distress location is crucial. Traditional methods often rely on a single environmental parameter or historical data, making it difficult to fully consider the influence of multiple factors, resulting in inaccurate evaluation results. To improve the accuracy of the evaluation, the research and development team introduced a Bayesian network prediction model and combined the conditional probability distribution learned from the historical distress event dataset to propose an optimized occurrence probability calculation formula. This formula aims to consider the influence of preliminary probability values and prior probabilities, adjusting through weight factors and exponential factors to ensure the accuracy and reliability of the evaluation results, thus proposing a new optional scheme, which includes:

[0132] Based on the corrected distress location, a Bayesian network prediction model is applied to evaluate the probability distribution of each distress area to generate the occurrence probability value of each distress area, including:

[0133] The corrected distress location and surrounding environmental parameters are input as input variables X into the pre-trained Bayesian network prediction model to obtain basic data D, where X = [x1, x2, …, x n} represents the set of corrected distress location and surrounding environmental parameters, and D represents the basic data obtained after processing by the Bayesian network prediction model, including the model's preliminary analysis results of environmental parameters;

[0134] Based on the base data D and the conditional probability distribution P(Y|X) learned from the dataset of historical distress events, the probability of each preset distress area becoming the actual distress location is evaluated, and a preliminary occurrence probability value P is output. pre (Y i |D), where Y i Let Y represent the i-th preset distress area. P(Y|X) is a conditional probability distribution learned from the dataset of historical distress events, representing the probability that a distress will occur in a certain area Y given environmental parameters X.

[0135] Based on the initial probability value P pre (Y i |D), further adjust and optimize the occurrence probability value to ensure accuracy, and obtain the optimized occurrence probability value P. opt (Y i |D);

[0136] The optimized probability value P is calculated using the following formula. opt (Y i |D):

[0137]

[0138] Among them, P opt (Y i |D) is the optimized probability value, which improves the accuracy of probability estimation through adjustment and optimization; α is a weighting factor (0 < α < 1), used to balance the influence of the initial probability value and the prior probability; β and γ are exponential factors, used to adjust the importance of the initial probability value and the prior probability, respectively; P prior (Y i ) is the prior probability, representing the probability of a region Y in the absence of a specific environmental parameter X. i The probability of encountering danger; j is an index variable that iterates through all preset danger zones, ranging from 1 to m; m is the number of preset danger zones;

[0139] Based on the optimized probability value P opt (Y i |D), generate the probability value of each distress zone, and ensure that resources are prioritized for the distress zone with the highest probability.

[0140] The following is a detailed explanation of each parameter:

[0141] X: is the input variable, representing the corrected distress location and surrounding environmental parameters. Through multi-source signal fusion and timestamp correction, the optimal geographic position information is obtained, and combined with real-time acquired environmental parameters such as sea current, wind direction, weather conditions, etc. These data together constitute the input variable X.

[0142] D: is the basic data, which is obtained after processing by the Bayesian network prediction model, containing the preliminary analysis results of the model on environmental parameters. The input variable X is input into the pre-trained Bayesian network prediction model, and the basic data D output by the model contains the preliminary analysis results of the environmental parameters.

[0143] P(Y|X): is the conditional probability distribution, which is learned from the historical distress event data set, representing the probability of distress occurring in a certain region Y given the environmental parameters X. Through machine learning methods, a large number of historical distress event data sets are used for training to obtain the conditional probability distribution P(Y|X).

[0144] P pre (Y i |D): is the preliminary occurrence probability value, which is evaluated and processed according to the basic data D and the conditional probability distribution P(Y|X) to evaluate the possibility of each preset distress region becoming the actual distress location, and the output preliminary occurrence probability value. Based on the basic data D and the conditional probability distribution P(Y|X), the preliminary occurrence probability value of each preset distress region Y i is calculated.

[0145] P prior (Y i ) : is the prior probability, representing the probability of a certain region Y i occurring distress without specific environmental parameters X. The prior probability of each region is calculated by statistical frequency distribution in the historical distress event data set.

[0146] α: is the weight factor, used to balance the influence of preliminary probability value and prior probability, with the value range of 0<α<1. It is usually set by experiment or expert experience, and can also be optimized by cross-validation method, etc.

[0147] β and γ: are exponential factors, respectively used to adjust the importance of preliminary probability value and prior probability. They are set by experiment or expert experience, and can also be optimized by cross-validation method, etc.

[0148] j: is the index variable. Traverse all preset distress regions, the range is 1 to m. It is automatically determined according to the number of preset distress regions m.

[0149] m: is the number of preset distress regions, representing the total number of distress regions preset by the system. It is determined according to the specific task requirements and geographic information.

[0150] The reasons for each sub-design are introduced as follows:

[0151] α·P pre (Y i ∣D) β This part reflects the impact of the preliminary probability value P pre (Y i ∣D) on the final optimized occurrence probability value. The weight factor α controls its importance, while the exponential factor β is used to further adjust the influence of the preliminary probability value.

[0152] (1-α)·P prior (Y i ) γ This part reflects the impact of the prior probability P prior (Y i ) on the final optimized occurrence probability value. The weight factor 1-α controls its importance, while the exponential factor γ is used to further adjust the influence of the prior probability.

[0153] This part is a weighted sum of the preliminary probability values of all preset distress areas, ensuring the normalization of the final optimized occurrence probability value.

[0154] This part is a weighted sum of the prior probabilities of all preset distress areas, also ensuring the normalization of the final optimized occurrence probability value.

[0155] Both the numerator and the denominator use the addition method to consider the influence of the preliminary probability value and the prior probability. By adding, the influence of both can be reasonably reflected, while maintaining the normalized nature of the final result.

[0156] The overall design of this formula aims to improve the accuracy and reliability of distress area assessment. By introducing the Bayesian network prediction model and learning from the historical distress event dataset, the system can more comprehensively consider the influence of multiple factors. Specifically, through the weight factors α and 1-α, the formula can find a reasonable balance point between the preliminary probability value and the prior probability, avoiding excessive dependence on a certain factor. Through the exponential factors β and γ, the formula can flexibly adjust the relative importance of the preliminary probability value and the prior probability according to the actual situation, thereby improving the flexibility and adaptability of the assessment. Through the weighted sum of the denominator part, the formula ensures the normalization of the final optimized occurrence probability value, making the result comparable and interpretable.

[0157] In summary, this formula not only improves the accuracy of distress area assessment in complex marine environments, but also enhances the robustness and adaptability of the system, providing strong technical support for fast and accurate rescue.

[0158] Here is a specific example:

[0159] In a specific maritime search and rescue mission, the system receives multiple distress signals and generates optimal geographic location information through multi-source signal fusion. Considering signal propagation delay, the system corrects the initial location and obtains a more accurate distress location. Next, the system needs to evaluate the likelihood of multiple potential distress areas becoming the actual distress site to determine the primary search target. To achieve this, the system applies the above-optimized occurrence probability calculation formula.

[0160] Assume that the system has input the corrected distress location and surrounding environmental parameters (such as sea currents, wind direction, weather conditions, etc.) as input variables X into the pre-trained Bayesian network prediction model to obtain basic data D. These data include:

[0161] Corrected distress location and surrounding environmental parameter set: X = {x1, x2, …, x n}

[0162] Basic data D, containing the model's preliminary analysis results of environmental parameters

[0163] Prescribed number of distress areas m = 5

[0164] Here are the calculation steps:

[0165] According to the basic data D, combined with the conditional probability distribution P(Y|X) learned from the historical distress event data set, the likelihood of each prescribed distress area becoming the actual distress site is evaluated and processed, outputting the preliminary occurrence probability value P(Y|D). pre (Y i ∣D)。

[0166] Assume the following preliminary occurrence probability values are obtained:

[0167] P pre (Y1∣D)=0.85;

[0168] P pre (Y2∣D)=0.70;

[0169] P pre (Y3∣D)=0.60;

[0170] P pre (Y4∣D)=0.45;

[0171] P pre (Y5∣D)=0.30;

[0172] Get the prior probability:

[0173] Prior probability P prior (Y i ) represents the probability of a certain region Y i occurring distress without specific environmental parameters X. Assume that the following prior probabilities are obtained from historical data:

[0174] P prior (Y1) = 0.20;

[0175] P prior (Y2) = 0.15;

[0176] P prior (Y3) = 0.10;

[0177] P prior (Y4) = 0.08;

[0178] P prior (Y5) = 0.07;

[0179] Set parameters:

[0180] Set weight factor a = 0.7, exponential factor β = 2, γ = 1.5.

[0181] Use the formula to calculate the optimized occurrence probability value P opt (Y i | D) of each preset distress region:

[0182]

[0183] Substitute the numerical values to calculate:

[0184] For Y1:

[0185]

[0186] P opt (Y1| D) ≈ 0.402

[0187] Similarly, the optimized occurrence probability values of other regions can be calculated:

[0188] P opt (Y2| D) ≈ 0.308;

[0189] P opt (Y3| D) ≈ 0.212;

[0190] P opt (Y4| D) ≈ 0.065;

[0191] P opt (Y5| D) ≈ 0.013;

[0192] Through the above calculation, the optimized occurrence probability value of each preset distress area is obtained. According to these probability values, the system can preferentially invest resources into the distress area Y1 with the highest probability, whose optimized occurrence probability value is 0.402. This not only improves the search and rescue efficiency, but also maximizes the reduction of invalid coverage, concentrates on the most likely area for search, thereby significantly improving the success rate of rescue.

[0193] By applying the optimized occurrence probability calculation formula, the system significantly improves the accuracy and reliability of distress area evaluation in complex marine environments. Specifically, the formula takes into account the influence of preliminary probability values and prior probability, and adjusts through weight factors and exponential factors to ensure high precision of evaluation results. Finally, the system can generate occurrence probability values for each distress area according to the optimized occurrence probability values, ensuring that resources are preferentially invested in the distress area with the highest probability, improving search and rescue efficiency and success rate. Overall, this method not only improves the flexibility and efficiency of search and rescue, but also enhances the robustness and adaptability of the system in complex and variable marine environments, providing strong technical support for fast and accurate rescue.

[0194] The present application takes into account that in complex weather conditions, maritime search and rescue tasks face many challenges, such as low visibility, strong winds, waves and other environmental factors, making it extremely difficult to identify non-fixed objects such as floating human bodies or shipwreck debris. Traditional image processing methods are difficult to effectively deal with these challenges, resulting in low recognition accuracy. To solve this problem, the research and development team introduced a convolutional neural network and combined specific formulas to extract, optimize and classify preliminary image data to improve the recognition accuracy of non-fixed objects. The formula scheme aims to gradually refine feature information through multiple convolution and pooling operations, and generate high-confidence object recognition results through the classification layer, so a new optional scheme is proposed, which includes:

[0195] Using the preliminary image data, combined with the convolutional neural network, the non-fixed objects in the preliminary image data are identified and processed to obtain preliminary object recognition results, including:

[0196] Using the preliminary image data I as input, the pre-trained convolutional neural network is fed to extract features from the preliminary image data and generate detailed feature information F (l) ;

[0197] Through the following formula, detailed feature information F (l) is extracted:

[0198] F (l) = σ(W (l) *I+b (l) )

[0199] where I is the preliminary image data; W (l) and b (l) are the weight matrix and bias term of the l-th layer, respectively; * denotes the convolution operation; σ is the activation function; F (l) represents the detailed feature information;

[0200] Based on the detailed feature information F (l) , the feature information is further refined through multi-layer convolution and pooling operations of the convolutional neural network, and the optimized feature representation is obtained according to the visual characteristics of shape, texture, and color

[0201] The optimized feature representation is obtained by the following formula

[0202]

[0203] where represents the optimized feature representation of the l-th layer, which is a more refined feature obtained after multi-layer convolution and pooling operations; γ l is the exponential factor of the l-th layer, used to adjust the importance of the output of each layer; Pooling represents the pooling operation; σ is the activation function; F (l-1) represents the detailed feature information of the previous layer;

[0204] The optimized feature representation F is classified by the classification layer of the convolutional neural network, and the classification result C i is generated for accurate classification of irregular or floating objects;

[0205] The classification result C i is calculated by the following formula:

[0206]

[0207] where C i represents the classification result, which is a probability distribution calculated by the classification layer and the softmax function; W c and b c are the weight matrix and bias term of the classification layer, respectively; N is the number of categories; β' is the exponential factor, used to adjust the sharpness of the classification result; i represents the index of the category, used to traverse all possible categories; j is the index variable, ranging from 1 to N, used to traverse all possible categories;

[0208] According to the classification result C i , a preliminary object recognition result R is generated, which contains the preliminary position information of potential survivors;

[0209] The preliminary object recognition result R is calculated using the following formula:

[0210]

[0211] Where R represents the preliminary object recognition result; arg max(C i) The function returns the classification result C. i The category index with the highest probability; θ is the confidence threshold used to filter out results with low confidence; max(C i ) represents the classification result C i The maximum probability value in the range; the if conditional statement is used to determine whether the classification result is reliable. If the maximum probability value is max(C i If the probability is greater than the set threshold θ, the category with the highest probability is selected as the final recognition result; otherwise, "not recognized" is output, indicating that there is not enough confidence to determine any category.

[0212] The following is a detailed explanation of each parameter:

[0213] I: This is the initial image data, the raw image input to the convolutional neural network, typically a two-dimensional matrix where each element represents a pixel value. It is acquired through cameras, drones, or other image acquisition devices.

[0214] W (l) and b (l) These are the convolutional kernel weight matrix and bias term, respectively. They are parameters used in the convolutional layer to extract features. The weight matrix is ​​used to detect different types of features (such as edges and textures), while the bias term allows the model to better fit the data. They are automatically adjusted through backpropagation during training to minimize the loss function.

[0215] σ: is the activation function, which introduces nonlinear factors, enabling the model to learn more complex patterns.

[0216] F (l) This represents detailed feature information; the output of each convolutional operation represents the degree of presence of certain patterns or features in the image. It is generated by the combined action of convolutional operations and activation functions.

[0217] This is an optimized feature representation, obtained after multiple convolutional and pooling operations, resulting in a more abstract and discriminative feature representation. Pooling is applied after each convolutional layer, and an exponential factor may be applied to enhance or diminish the importance of that layer.

[0218] γ l: is an exponential factor that adjusts the importance of each layer's output. It can be seen as a hyperparameter that influences the composition of the final feature representation. It is adjusted based on specific task requirements and experimental results to achieve optimal performance.

[0219] W c b c : are the classification layer weight matrix and bias term, respectively. The classification layer is used to map the last layer's features to the class space. These parameters are also automatically adjusted during the training process.

[0220] N: is the number of classes, which is the number of different objects the system tries to identify. It is determined based on the application scenario, for example, in a search and rescue scenario, it could be human bodies, life jackets, other objects, etc.

[0221] β': is an exponential factor that adjusts the sharpness of the classification results. A higher value will make the probability distribution more concentrated on a certain class. It is adjusted based on actual needs to obtain more accurate classification results.

[0222] θ: is the confidence threshold, which determines the standard for accepting classification results. Results above this value are considered reliable. It is set to ensure high-precision recognition results and avoid false positives.

[0223] C i : is the classification result, which is a probability distribution calculated using the softmax function. It represents the likelihood of each class appearing. It provides a probabilistic classification result, making it easier to choose the most likely class.

[0224] R: is the preliminary object recognition result, which contains the location information of potential targets. Based on the classification result Ci and the confidence threshold θ, it determines whether there is a target and its location.

[0225] Here is a specific example:

[0226] In a specific maritime search and rescue task, the system receives multiple distress signals and generates optimal geographic location information through multi-source signal fusion. Considering signal propagation delay, the system corrects the initial location to obtain a more accurate distress location. Next, the system needs to process the preliminary image data I taken by the unmanned aerial vehicle to identify the location of potential survivors. To achieve this goal, the system applies the detailed feature extraction and classification formulas described above.

[0227] Assume that the system has obtained the preliminary image data I and processed it through a pre-trained convolutional neural network.

[0228] The specific parameters are as follows:

[0229] Preliminary image data I: a grayscale image with a size of 256x256 pixels.

[0230] Convolution kernel weight matrix W (l) and bias term b (l) : the convolution kernel weight matrix and bias term of the l-th layer, respectively, which are assumed to have been pre-trained.

[0231] Activation function σ: ReLU activation function is adopted.

[0232] Exponential factor γ l : used to adjust the importance of the output of each layer, set to 1.5.

[0233] Classification layer weight matrix W c and bias term b c : the weight matrix and bias term of the classification layer, respectively, which are assumed to have been pre-trained.

[0234] Number of categories N: set to 3 categories (human body, clothing, and others).

[0235] Exponential factor β': used to adjust the sharpness of the classification result, set to 0.5.

[0236] Confidence threshold θ: set to 0.7.

[0237] The following are the calculation steps:

[0238] Using the preliminary image data I as input, feed it into the pre-trained convolutional neural network to perform feature extraction processing on the preliminary image data, and generate detailed feature information F (l) .

[0239] F (l) = σ(W (l) *I + b (l) )

[0240] Assuming that the detailed feature information obtained after the first layer convolution operation is F (1) .

[0241] Based on the detailed feature information F (l) , further refine the feature information through multi-layer convolution and pooling operations of the convolutional neural network, and obtain the optimized feature representation according to the visual characteristics of shape, texture, and color

[0242]

[0243] Assuming that the optimized feature representation obtained after the second layer convolution and pooling operation is

[0244] Through the classification layer of the convolutional neural network, the optimized feature representation Classification processing is performed, and irregular or floating objects are accurately classified to generate a classification result C i ;

[0245]

[0246] The calculated classification result is:

[0247] C1≈0.85 (human body);

[0248] C2≈0.10 (clothes);

[0249] C3≈0.05 (other);

[0250] A preliminary object recognition result R is generated.

[0251] According to the classification result C i , a preliminary object recognition result R is generated.

[0252]

[0253] In this example, the maximum probability value max(C i ) = 0.85 is greater than the set threshold θ = 0.7, so the class with the highest probability is selected as the final recognition result.

[0254] R = arg max(C i ) = 1 (human body)

[0255] Through the above calculation, the preliminary object recognition result R is obtained, confirming that the object in the image is "human body" with a corresponding probability of 0.85. This not only improves the accuracy of recognition, but also enhances the robustness and adaptability of the system, ensuring accurate identification of potential survivors in complex and variable environments. Finally, the system can quickly launch rescue operations based on these high-confidence recognition results, significantly improving the success rate of rescue.

[0256] By applying detailed feature extraction and classification formulas, the system significantly improves the accuracy of non-fixed form object recognition in complex weather conditions. Specifically, this formula scheme improves the recognition ability of non-fixed form objects (such as floating human bodies or shipwreck debris) by gradually refining feature information in the following ways. Through the softmax function to calculate the probability distribution, ensure the high confidence of the classification result. By setting a threshold θ, filter low-confidence results to avoid misjudgment.

[0257] Overall, this method not only improves search and rescue efficiency, but also enhances the robustness and adaptability of the system in complex and variable marine environments, providing strong technical support for fast and accurate rescue.

[0258] Figure 2 This application provides a schematic diagram of the structure of an automatic positioning system for intelligent maritime emergency rescue equipment, as shown in the embodiment of the present application. Figure 2 As shown, the system includes:

[0259] The receiving fusion module 21 is used to receive and fuse multiple distress signals from satellites, radio relay stations and nearby ships, and estimate the optimal geographical location information and timestamp;

[0260] The assessment module 22 is used to determine the distress area on a preset electronic map of the sea area based on the optimal geographical location information, calculate the signal propagation delay according to the timestamp to correct the distress location, and apply a Bayesian network prediction model to evaluate the probability distribution of each potential distress area and generate the primary search target with the highest probability.

[0261] The scan generation module 23 is used to activate the thermal imaging and optical sensing system on the rescue drone to scan based on the primary search target with the highest probability, combine convolutional neural networks to enhance the recognition ability of non-fixed-shape objects, and use generative adversarial networks to simulate image features under different weather conditions to generate highly accurate potential survivor locations.

[0262] The optimization and update module 24 is used to dynamically adjust and optimize the search and rescue path using ocean current prediction algorithms and wind direction data, to ensure rapid coverage of the distress area where the potential survivors are located, to adaptively adjust the search strategy through reinforcement learning algorithms, to update the best flight route in real time, and to generate an economical and efficient search and rescue path plan considering the remaining battery power and endurance of the UAV.

[0263] The feedback guidance module 25 is used to promptly feed back the precise coordinates of the potential survivor's location to the nearest rescue coordination center based on the search and rescue route planning, and guide emergency rescue equipment to carry out rescue along the optimal route.

[0264] Figure 2 The aforementioned intelligent maritime emergency rescue equipment automatic positioning system can perform... Figure 1 The implementation principle and technical effects of the automatic positioning method for intelligent maritime emergency rescue equipment described in the illustrated embodiment will not be repeated here. The specific operation methods of each module and unit in the automatic positioning system for intelligent maritime emergency rescue equipment in the above embodiments have been described in detail in the embodiments related to this method, and will not be elaborated upon here.

[0265] In one possible design, Figure 2 The intelligent maritime emergency rescue equipment automatic positioning system of the illustrated embodiment can be implemented as a computing device, such as... Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0266] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32.

[0267] The processing component 32 is configured to receive and fuse multiple distress signals from satellites, radio relay stations and nearby ships, generate optimal geographic position information and time stamps; based on the optimal geographic position information, determine the distress area on the preset sea area electronic map, and calculate the signal propagation delay according to the time stamp to correct the distress position, while applying the Bayesian network prediction model to evaluate the probability distribution of each potential distress area, and generate the primary search target with the highest probability; according to the primary search target with the highest probability, start the scanning of the thermal imaging and optical sensing system on the rescue drone, combine the convolutional neural network to enhance the non-fixed object recognition ability, and use the generative adversarial network to simulate the image features under different weather conditions, to generate the position of the potential survivor with high accuracy; use the sea current prediction algorithm and wind direction data to dynamically adjust and optimize the search path, to ensure rapid coverage of the distress range where the potential survivor is located, use the reinforcement learning algorithm to adaptively adjust the search strategy, to update the best flight route in real time, and consider the remaining power and endurance of the drone, to generate an economic and efficient search path planning; based on the search path planning, timely feedback the accurate coordinates of the potential survivor to the nearest rescue coordination center, and guide the first-aid equipment to the optimal path to implement rescue.

[0268] The processing component 32 can include one or more processors to execute the computer instructions to complete all or part of the steps in the above method. Of course, the processing component can also be one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic elements, for executing the above method.

[0269] The storage component 31 is configured to store various types of data to support the operation of the terminal. The storage component can be realized by any type of volatile or non-volatile storage device or their combination, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0270] Of course, the computing device can also include other components, such as input / output interfaces, display components, communication components, etc.

[0271] The input / output interface provides an interface between the processing component and peripheral interface modules, which can be output devices, input devices, and the like.

[0272] The communication component is configured to facilitate wired or wireless communication between the computing device and other devices, and the like.

[0273] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform, and the processing component, the storage component, and the like can be basic server resources rented or purchased from the cloud computing platform.

[0274] The embodiments of the present application also provide a computer storage medium storing a computer program, and the computer program can implement the above-mentioned Figure 1 The intelligent offshore first-aid device automatic positioning method of the embodiments shown.

[0275] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-mentioned system, device and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0276] The device embodiments described above are only schematic, and the units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiments. Those skilled in the art can understand and implement without creative labor.

[0277] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software and necessary general hardware platforms, and of course, can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of software products, and the computer software products can be stored in a computer readable storage medium, such as ROM / RAM, magnetic disk, optical disk, and the like, and include a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0278] Finally, it should be noted that the above examples are only used to illustrate the technical solutions of the present application, and are not intended to limit the same; although the present application has been described in detail with reference to the foregoing examples, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing examples, or make equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An automatic positioning method for intelligent maritime emergency rescue equipment, characterized in that, include: It receives and merges multiple distress signals from satellites, radio relay stations, and nearby ships to generate optimal geographic location information and timestamps; Based on the optimal geographical location information, the distress area is determined on the preset electronic map of the sea area, and the signal propagation delay is calculated according to the timestamp to correct the distress location. At the same time, a Bayesian network prediction model is applied to evaluate the probability distribution of each potential distress area and generate the primary search target with the highest probability. Based on the primary search target with the highest probability, the thermal imaging and optical sensing system on the rescue drone is activated to scan, combined with a convolutional neural network to enhance the recognition of non-fixed-shape objects, and a generative adversarial network is used to simulate image features under different weather conditions to generate highly accurate locations of potential survivors. By using ocean current prediction algorithms and wind direction data, the search and rescue route is dynamically adjusted and optimized to ensure rapid coverage of the distress area where the potential survivors are located. The search strategy is adaptively adjusted through reinforcement learning algorithms, the best flight route is updated in real time, and the remaining battery power and endurance of the UAV are taken into account to generate an economical and efficient search and rescue route plan. Based on the search and rescue route planning, the precise coordinates of potential survivors' locations are promptly fed back to the nearest rescue coordination center, and emergency medical equipment is guided to the optimal route to carry out the rescue.

2. The method according to claim 1, characterized in that, Based on the optimal geographical location information, the distress area is determined on a preset electronic map of the sea area. The signal propagation delay is calculated based on the timestamp to correct the distress location. Simultaneously, a Bayesian network prediction model is applied to evaluate the probability distribution of each potential distress area, generating the primary search target with the highest probability, including: Using the optimal geographical location information, the initial distress location is matched on a preset electronic map of the sea area to determine the distress area; The signal propagation delay is calculated based on the timestamp, and the initial distress position is corrected to obtain the corrected distress position. Based on the corrected distress location, a Bayesian network prediction model is applied to evaluate the probability distribution of each distress area and generate the occurrence probability value of each distress area. Based on the probability value of each distress area, the distress area with the highest probability is selected, and the primary search target with the highest probability is generated, ensuring that resources are prioritized for the distress area with the highest probability.

3. The method according to claim 2, characterized in that, Based on the corrected distress location, a Bayesian network prediction model is applied to evaluate the probability distribution of each distress area, generating a probability value for each distress area, including: The corrected distress location and surrounding environmental parameters are used as input variables and fed into a pre-trained Bayesian network prediction model to obtain basic data. Based on the aforementioned basic data and the conditional probability distribution learned from the dataset of historical distress events, the probability of each preset distress area becoming an actual distress location is evaluated, and a preliminary probability value of occurrence is output. Based on the initial probability value of occurrence, the probability value of occurrence is further adjusted and optimized to ensure accuracy, resulting in an optimized probability value of occurrence. Based on the optimized occurrence probability values, the occurrence probability value of each distress area is generated to ensure that resources are prioritized for the distress areas with the highest probability.

4. The method according to claim 1, characterized in that, Based on the primary search target with the highest probability, the thermal imaging and optical sensing system on the rescue drone is activated to scan, combined with a convolutional neural network to enhance the recognition capability of non-fixed-shape objects, and a generative adversarial network is used to simulate image features under different weather conditions to generate highly accurate locations of potential survivors, including: Based on the primary search target with the highest probability, the thermal imaging and optical sensing system on the rescue drone is activated to scan and obtain preliminary image data. Using the preliminary image data, combined with a convolutional neural network, non-fixed-shape objects in the preliminary image data are identified to obtain preliminary object recognition results. Based on the preliminary object recognition results, a generative adversarial network is applied to simulate image features under different weather conditions to generate optimized image features; Based on the optimized image features, the precise location of potential survivors is further analyzed and confirmed, generating highly accurate potential survivor locations.

5. The method according to claim 4, characterized in that, The process of using the preliminary image data, combined with a convolutional neural network, to identify non-fixed-shape objects in the preliminary image data, and obtaining preliminary object recognition results, includes: Using the preliminary image data as input, a pre-trained convolutional neural network is fed into the preliminary image data to perform feature extraction processing and generate detailed feature information. Based on the detailed feature information, the feature information is further refined through multi-layer convolution and pooling operations of the convolutional neural network, and an optimized feature representation is obtained according to the visual characteristics of shape, texture and color. The optimized feature representation is classified through the classification layer of a convolutional neural network to accurately classify irregular or floating objects and generate classification results. Based on the classification results, preliminary object recognition results are generated, which include preliminary location information of potential survivors.

6. The method according to claim 1, characterized in that, The method utilizes ocean current prediction algorithms and wind direction data to dynamically adjust and optimize search and rescue routes, ensuring rapid coverage of the distress area where the potential survivors are located. It adaptively adjusts the search strategy using reinforcement learning algorithms, updates the optimal flight route in real time, and considers the drone's remaining battery power and endurance to generate an economical and efficient search and rescue route plan. This includes: Using the highly accurate locations of potential survivors obtained, combined with real-time ocean current prediction algorithms and wind direction data, a preliminary search and rescue route is planned to obtain a preliminary search and rescue route. Based on the initial search and rescue route, the search and rescue route is dynamically adjusted and optimized according to ocean current prediction algorithms and wind direction data to adapt to the ever-changing marine environment, ensure the effectiveness and timeliness of the route, and generate an optimized search and rescue route. The optimized search and rescue path is adaptively adjusted using a reinforcement learning algorithm. The search strategy is updated in real time based on the situation encountered during the actual search and rescue process, and a real-time updated search and rescue path is generated. Taking into account the drone's remaining battery power and endurance, the real-time updated search and rescue path is further processed to ensure that the drone can return safely after completing the mission, while avoiding unnecessary energy waste, and to generate the final flight path. Based on the highly accurate location of potential survivors, real-time ocean current prediction algorithms and wind direction data, dynamically adjusted and optimized search and rescue routes, adaptive search strategies, and the remaining battery power and endurance of the UAV, the final flight route is evaluated and optimized to generate an economical and efficient search and rescue route plan.

7. The method according to claim 6, characterized in that, The process of adaptively adjusting the optimized search and rescue path using a reinforcement learning algorithm, updating the search strategy in real time based on the actual search and rescue situation, and generating a real-time updated search and rescue path includes: Using the optimized search and rescue path and combined with real-time acquired actual search and rescue environment data, a training basis is provided for the reinforcement learning algorithm, and initial training data is obtained. Based on the initial training data, the state of the current search and rescue environment is represented by the state space in the reinforcement learning algorithm, and the position and state of the UAV are taken as part of the environment to ensure that the reinforcement learning algorithm can fully understand the current situation and generate an environmental state representation. Based on the environmental state representation, an adjustment scheme is generated through adjustment operations defined in the action space of the reinforcement learning algorithm, including changing flight direction, adjusting flight altitude, accelerating or decelerating. Using the aforementioned adjustment scheme, the effect of each adjustment operation is evaluated through the reward mechanism in the reinforcement learning algorithm, which takes into account factors such as search and rescue efficiency, safety, and resource consumption, to generate evaluation results; Based on the evaluation results, the reinforcement learning algorithm is applied for iterative learning. The search strategy is continuously optimized according to the results of each adjustment operation, so that the reinforcement learning algorithm can gradually find the optimal path adjustment scheme in the actual search and rescue process and generate an optimized search strategy. Based on the optimized search strategy, the search and rescue path is updated in real time, and a real-time updated search and rescue path is generated to ensure that the UAV can dynamically adjust its flight route according to the latest environmental information, thereby improving search and rescue efficiency.

8. An automatic positioning system for intelligent maritime emergency rescue equipment, characterized in that, include: The receiving fusion module is used to receive and fuse multiple distress signals from satellites, radio relay stations and nearby ships, and estimate the optimal geographical location information and timestamp; The assessment module is used to determine the distress area on a preset electronic map of the sea area based on the optimal geographical location information, calculate the signal propagation delay according to the timestamp to correct the distress location, and apply a Bayesian network prediction model to evaluate the probability distribution of each potential distress area and generate the primary search target with the highest probability. The scanning generation module is used to activate the thermal imaging and optical sensing system on the rescue drone to scan based on the primary search target with the highest probability. It combines convolutional neural networks to enhance the recognition of non-fixed-shape objects and uses generative adversarial networks to simulate image features under different weather conditions to generate highly accurate locations of potential survivors. The optimization and update module is used to dynamically adjust and optimize the search and rescue path using ocean current prediction algorithms and wind direction data, ensuring rapid coverage of the distress area where the potential survivors are located. It adaptively adjusts the search strategy through reinforcement learning algorithms, updates the best flight route in real time, and takes into account the remaining battery power and endurance of the UAV to generate an economical and efficient search and rescue path plan. The feedback guidance module is used to promptly feed back the precise coordinates of potential survivors' locations to the nearest rescue coordination center based on the search and rescue route planning, and guide emergency medical equipment to carry out rescue operations along the optimal route.

9. A computing device, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement the automatic positioning method for an intelligent maritime emergency rescue device as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that, The device contains a computer program that, when executed by a computer, implements an automatic positioning method for an intelligent maritime emergency rescue device as described in any one of claims 1 to 7.

Images