Resource distribution path planning method and device, equipment and storage medium

Abstract

Embodiments of the present application provide a resource distribution path planning method and device, equipment and a storage medium, belonging to the field of artificial intelligence. The method comprises: acquiring monitoring image data, social public opinion data and power monitoring data of a target disaster area; performing casualty prediction on the target disaster area according to the monitoring image data, the social public opinion data and the power monitoring data to obtain casualty prediction data; constructing a disaster view according to the monitoring image data, the casualty prediction data and the power monitoring data to obtain a current disaster view; performing resource demand prediction according to the casualty prediction data to obtain resource demand prediction data; and performing resource distribution path planning according to the current disaster view, the resource demand prediction data and preset path planning reference data to obtain a resource distribution planning path. The present application can be applied to financial technology and health care and other business systems that require a large amount of data, can save manpower during distribution path planning, and improve the efficiency of relief resource distribution.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and is applied to the fields of fintech and healthcare, particularly to a resource distribution route planning method, apparatus, device, and storage medium. Background Technology

[0002] Disaster response systems can be used to collect disaster information during disasters and to distribute relief resources based on that information. In fintech scenarios, for example, disaster information stored in a disaster response system can be used to develop insurance payout plans, thus affecting the amount of compensation received by policyholders after a disaster. In healthcare scenarios, a disaster response system can provide medical rescue personnel with crucial rescue reference information and medical supply demand information, helping them to reach disaster areas promptly and provide assistance.

[0003] In related technologies, disaster response systems primarily rely on manual planning of relief resource delivery routes based on on-site disaster information. This requires a significant amount of manpower, resulting in low efficiency and severe delays in relief resource delivery. Therefore, how to reduce manpower spent on route planning and improve the efficiency of relief resource delivery has become an urgent technical problem to be solved. Summary of the Invention

[0004] The main objective of this application is to propose a resource distribution route planning method, apparatus, equipment, and storage medium, which aims to save manpower in distribution route planning and improve the efficiency of relief resource distribution.

[0005] To achieve the above objectives, a first aspect of this application proposes a resource distribution route planning method, the method comprising: Acquire disaster monitoring data; wherein, the disaster monitoring data includes: monitoring image data of the target disaster area, social media data, and power monitoring data; Based on the monitoring image data, the social media data, and the power monitoring data, casualty prediction data is obtained for the target disaster-stricken area. A disaster view is constructed based on the monitoring image data, the casualty prediction data, and the power monitoring data to obtain the current disaster view; wherein, the current disaster view represents the distribution of the affected population, road conditions, and power conditions within the target disaster area; Resource demand forecasting is performed based on the casualty prediction data to obtain resource demand forecasting data. Based on the current disaster situation view, the resource demand forecast data, and the preset route planning reference data, a resource delivery route is planned to obtain the resource delivery planning route.

[0006] In some embodiments, the step of predicting casualties in the target disaster area based on the monitoring image data, the social media data, and the power monitoring data to obtain casualty prediction data includes: Based on the aforementioned social media opinion data, intentions are categorized to obtain opinion categories; Based on the aforementioned public opinion categories and preset medical demand categories, keywords are extracted from the social public opinion data to obtain medical demand keywords; Feature extraction is performed on the monitoring image data to obtain monitoring image features; Feature extraction is performed on the power monitoring data to obtain power monitoring features; Casualty prediction is performed based on the medical demand keywords, the monitoring image features, and the power monitoring features to obtain the casualty prediction data.

[0007] In some embodiments, the step of constructing a disaster view based on the monitoring image data, the casualty prediction data, and the power monitoring data to obtain the current disaster view includes: Based on the monitoring image data, damage to fixed facilities is identified to obtain fixed facility damage identification data. Based on the power monitoring data, power fault areas are identified to obtain power fault area identification data. The grid edge parameters are determined based on the fixed facility damage identification data, the grid thermal value is determined based on the casualty prediction data, and the grid gray value is determined based on the power failure area identification data, so as to construct the current disaster situation view.

[0008] In some embodiments, the fixed facility damage identification data is a fixed facility damage category, and the step of identifying fixed facility damage based on the monitoring image data to obtain fixed facility damage identification data includes: Features are extracted from the fixed facilities in the monitoring image data to obtain fixed facility features; wherein, the fixed facility features include: facility texture features, facility shape features and facility structure features; Based on the facility texture features, facility shape features, and facility structural features, the probability of damage to the fixed facility is predicted to obtain the damage prediction probability; The damage category of the fixed facility is determined based on the predicted damage probability.

[0009] In some embodiments, the step of identifying power fault areas based on the power monitoring data to obtain power fault area identification data includes: The power monitoring data is subjected to anomaly data extraction to obtain power anomaly data; wherein, the power anomaly data includes: power anomaly duration and power anomaly area information; If the duration of the power anomaly exceeds a preset duration threshold, the power anomaly area information is identified as a fault to obtain power fault area identification data.

[0010] In some embodiments, the casualty prediction data includes: the number of casualties and information on the casualty targets; the step of predicting resource demand based on the casualty prediction data to obtain resource demand prediction data includes: Based on the number of casualties and the information on the casualties, the disaster situation is classified to obtain the disaster category; Based on the disaster category, the preset historical relief resource consumption data is filtered to obtain the target relief resource consumption data; Based on the target rescue resource consumption data, a target resource demand prediction model is selected from the preset candidate resource demand prediction models; Resource demand forecasting data is obtained by using the target resource demand forecasting model, the number of casualties, and the information on the casualties.

[0011] In some embodiments, the route planning reference data includes rescue risk information, rescue point information, and rescue vehicle information. The step of planning resource delivery routes based on the current disaster situation view, the resource demand forecast data, and the preset route planning reference data to obtain a resource delivery planned route includes: Resource allocation information is obtained by performing resource allocation processing based on the resource demand forecast data, the rescue point information, and the rescue vehicle information. Based on the rescue risk information and the current disaster situation view, a preliminary planned route is obtained; The resource distribution planning path is obtained by optimizing the preliminary planned path based on the resource allocation information.

[0012] To achieve the above objectives, a second aspect of this application provides a resource distribution route planning apparatus, the apparatus comprising: The data acquisition module is used to acquire disaster monitoring data; wherein, the disaster monitoring data includes: monitoring image data of the target disaster area, social media data, and power monitoring data; The casualty prediction module is used to predict casualties in the target disaster area based on the monitoring image data, the social media data, and the power monitoring data, and to obtain casualty prediction data. The disaster view construction module is used to construct a disaster view based on the monitoring image data, the casualty prediction data, and the power monitoring data to obtain the current disaster view; wherein, the current disaster view represents the distribution of the affected population, road conditions, and power conditions within the target disaster area; The demand forecasting module is used to forecast resource demand based on the casualty forecasting data to obtain resource demand forecasting data. The route planning module is used to plan resource delivery routes based on the current disaster situation view, the resource demand forecast data, and the preset route planning reference data, so as to obtain the resource delivery planned route.

[0013] To achieve the above objectives, a third aspect of the present application provides a computer device, the computer device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method described in the first aspect.

[0014] To achieve the above objectives, a fourth aspect of the present application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in the first aspect.

[0015] The resource distribution route planning method, apparatus, equipment, and storage medium proposed in this application collect monitoring image data, social media data, and power monitoring data of the target disaster area, and combine these three types of data to complete the casualty prediction of the target disaster area, obtaining accurate casualty prediction data. Simultaneously, based on the casualty prediction data, power monitoring data, and monitoring image data, a current disaster situation view is constructed that accurately represents the distribution of affected people, road conditions, and power status in the target disaster area. Then, based on the casualty prediction data, resource demand prediction data of the target disaster area is determined. Combining the current disaster situation view and resource demand prediction data, a distribution route planning is completed to obtain a resource distribution planning route, thus constructing an accurate resource distribution planning route. Therefore, in the resource distribution route planning process, route planning is completed automatically, saving manpower. Furthermore, the route planning requires the combination of accurate current disaster situation view and resource demand prediction data, enabling the planning of accurate resource distribution routes and improving the timeliness and safety of resource distribution. Attached Figure Description

[0016] Figure 1 This is a flowchart of the resource delivery route planning method provided in the embodiments of this application; Figure 2 yes Figure 1 The flowchart of step S102 in the document; Figure 3 yes Figure 1 The flowchart of step S103 in the process; Figure 4 This is a schematic diagram of the current disaster situation view provided in the embodiments of this application; Figure 5 yes Figure 3 The flowchart of step S301 in the process; Figure 6 yes Figure 3 The flowchart of step S302 in the document; Figure 7 yes Figure 1 The flowchart of step S104 in the process; Figure 8 This is a flowchart of a resource delivery route planning method provided in another embodiment of this application; Figure 9 yes Figure 1 The flowchart of step S105 in the process; Figure 10 This is a schematic diagram of the structure of the resource distribution route planning device provided in the embodiments of this application; Figure 11 This is a schematic diagram of the hardware structure of the computer device provided in the embodiments of this application. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0018] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, and the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0020] First, let's analyze some of the terms used in this application: Artificial intelligence (AI) is a new branch of computer science that studies, develops, and applies theories, methods, technologies, and systems to simulate, extend, and expand human intelligence. It aims to understand the essence of intelligence and produce intelligent machines that can react in a way similar to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing, and expert systems. AI can simulate the information processes of human consciousness and thought. Furthermore, AI utilizes digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceiving the environment, acquiring knowledge, and using that knowledge to achieve optimal results.

[0021] Multimodal data refers to data collections from different modalities (such as text, images, audio, and video). Each of these data types carries rich information, and by fusing these modalities, a more comprehensive analysis and understanding can be achieved. Multimodal learning utilizes technologies such as deep learning to overcome the limitations of single-modal data and enhance the model's perception and decision-making capabilities.

[0022] U-Net segmentation network: A deep learning-based convolutional neural network primarily used for image segmentation tasks, especially biomedical image segmentation. It consists of two parts: an encoder (downsampling path) and a decoder (upsampling path), and its shape is U-shaped, hence the name U-Net.

[0023] BERT model: A Transformer-based neural network used to understand and generate human-like language. BERT employs an encoder-only architecture. The original Transformer architecture has both encoder and decoder modules. The decision to use an encoder-only architecture in BERT indicates a primary emphasis on understanding the input sequence rather than generating the output sequence.

[0024] GIS grid: A GIS grid is an intelligent information platform that utilizes grid technology and spatial information infrastructure to create a virtual spatial information management and processing environment, providing users with integrated spatial information application services. The characteristics of a GIS grid include heterogeneity, dynamism, and resource sharing across multiple management domains.

[0025] YOLOv networks are a family of machine learning algorithms for real-time object detection. Object detection is a computer vision task that uses neural networks to locate and classify objects in images. This task has a wide range of applications, from medical imaging to autonomous vehicles. Various machine learning algorithms are used for object detection, one of which is the convolutional neural network (CNN).

[0026] A graded heatmap is a statistical chart that displays data by coloring blocks. When drawing the chart, rules for color mapping must be specified. For example, larger values ​​are represented by darker colors, and smaller values ​​by lighter colors; larger values ​​are represented by warmer colors, and smaller values ​​by cooler colors.

[0027] Spatiotemporal Graph Convolutional Network (ST-GCN) is a neural network architecture used for skeleton-based action recognition. The core idea of ​​ST-GCN is to extend traditional convolution operations to graph-structured data, thereby capturing the spatiotemporal features of skeletal data. The structure of ST-GCN mainly consists of two parts: spatial graph convolution and temporal graph convolution, with spatial graph convolution being its core component.

[0028] Disaster response systems are comprehensive solutions integrating technologies such as the Internet of Things (IoT), big data, cloud computing, and artificial intelligence. When a disaster occurs, the system collects disaster information and plans resource delivery routes based on this information, enabling rapid delivery of relief resources. For example, in a fintech scenario, the disaster information stored in the system can be used for insurance claims, providing a reference for developing claim plans. In a healthcare scenario, the system can provide medical rescue personnel with resource delivery route planning, assisting them in quickly and safely reaching rescue points to provide assistance.

[0029] In related technologies, the distribution routes of relief resources in disaster response systems mainly rely on manual planning. This requires manual collection of disaster information and planning of resource distribution routes based on that information. However, manually planned resource distribution routes consume a lot of manpower and are prone to errors due to complex road conditions and disaster situations, thus affecting the efficiency of resource distribution.

[0030] Specifically, during earthquakes, floods, and fires, demand and road conditions are constantly changing. Route planning is often based solely on static road networks and resource needs, without integrating real-time information such as road damage and traffic control. For example, during wildfire relief efforts, failure to avoid burning sections of road in time forces some supply vehicles to turn back, impacting the efficiency and timeliness of resource delivery.

[0031] Based on this, embodiments of this application provide a resource distribution route planning method, apparatus, device, and storage medium, aiming to collect monitoring image data, social media data, and power monitoring data of a target disaster-stricken area, and predict casualty data for the target disaster-stricken area based on the monitoring image data and social media data. A current disaster situation view is constructed based on the casualty prediction data, monitoring image data, and power monitoring data. This current disaster situation view determines the distribution of affected people, road conditions, and power status in the target disaster-stricken area, and then resource demand prediction data is determined based on the casualty prediction data. Finally, route planning is performed based on the resource demand prediction data, the current disaster situation view, and route planning reference data to plan a resource distribution route. Therefore, by automatically collecting data and constructing a current disaster situation view representing the distribution of affected people, road conditions, and power status, and then automatically planning a resource distribution route based on the current disaster situation view, resource demand, and route planning reference data, automatic route planning is achieved. This planning is based on real-time demand and road conditions, saving manpower, planning accurate routes, and achieving efficient resource distribution.

[0032] The resource distribution route planning method, apparatus, equipment, and storage medium provided in this application are specifically described through the following embodiments. First, the resource distribution route planning method in this application embodiment is described.

[0033] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0034] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0035] The resource delivery path planning method provided in this application relates to the field of artificial intelligence technology, and is specifically applied to fintech and healthcare. The resource delivery path planning method provided in this application can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, etc.; the server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application implementing the resource delivery path planning method, but is not limited to the above forms.

[0036] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer computer devices, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0037] It should be noted that in all specific embodiments of this application, when processing data related to user identity or characteristics, such as user information, user behavior data, user historical data, and user location information, user permission or consent is obtained first. Furthermore, the collection, use, and processing of this data comply with relevant laws, regulations, and standards. In addition, when embodiments of this application require access to sensitive personal information of users, separate permission or consent from the user is obtained through pop-ups or redirection to confirmation pages. Only after obtaining the user's separate permission or consent is the necessary user-related data required for the proper functioning of these embodiments acquired.

[0038] Figure 1 This is an optional flowchart of the resource delivery route planning method provided in the embodiments of this application. Figure 1The method may include, but is not limited to, steps S101 to S105.

[0039] Step S101: Obtain disaster monitoring data; wherein, disaster monitoring data includes: monitoring image data of the target disaster area, social media data, and power monitoring data; Step S102: Based on monitoring image data and social media opinion data, perform casualty prediction on the target disaster area to obtain casualty prediction data; Step S103: Construct a disaster view based on monitoring image data, casualty prediction data, and power monitoring data to obtain the current disaster view; wherein, the current disaster view represents the distribution of the affected population, road conditions, and power conditions within the target disaster area; Step S104: Based on the casualty prediction data, resource demand is predicted to obtain resource demand prediction data; Step S105: Based on the current disaster situation view, resource demand forecast data, and preset route planning reference data, resource distribution route planning is performed to obtain the resource distribution planning route.

[0040] Steps S101 to S105 of this embodiment involve collecting monitoring image data, social media data, and power monitoring data of the target disaster area. Based on the monitoring image data and social media data, casualty prediction data is obtained by predicting casualties in the target disaster area. Then, a current disaster situation view is constructed based on the casualty prediction data, monitoring image data, and power monitoring data. The current disaster situation view represents the distribution of affected people, road conditions, and power supply status within the target disaster area. Simultaneously, resource demand prediction data is determined based on the casualty prediction data. Finally, a resource distribution planning route is planned based on the current disaster situation view, resource demand prediction data, and route planning reference data. Therefore, by automatically collecting monitoring image data, social media data, and power monitoring data of the target disaster area, and combining these three aspects of data, a current disaster situation view representing the distribution of affected people, road conditions, and power supply status is constructed. Finally, the resource distribution planning route is automatically generated by combining the current disaster situation view, resource demand, and route planning parameters, saving manpower. Furthermore, the generated resource distribution planning route conforms to dynamically changing road conditions, resource demands, and power supply status, enabling accurate and efficient resource distribution.

[0041] In step S101 of some embodiments, the disaster monitoring data is multimodal data, and it is data collected and fused in real time. The disaster monitoring data includes: monitoring image data, social media data, and power monitoring data. Monitoring image data refers to image data captured within the target disaster area, which can be at least one of satellite image data, drone aerial photography data, monitoring equipment recording data, and photographic data. This embodiment does not specifically limit the source of the monitoring image data. Monitoring image data can clearly show the population distribution, casualty status, and road conditions within the target disaster area. Social media data is collected from the target disaster area using a web crawler engine; it is also called social media data, specifically including public opinion videos, audio, and text. This embodiment does not specifically limit the social media data. Specifically, social media data is collected by collecting social media data with geographic tags, where the geographic tags are the location information of the target disaster area. Power monitoring data characterizes the power status within the target disaster area, specifically by collecting the electricity consumption data of smart meters within the target disaster area. The electricity consumption data includes historical and current electricity consumption.

[0042] For example, if an earthquake occurs in a target disaster area, satellite imagery data from satellites and drone aerial photography data from drones are collected as monitoring imagery data. Simultaneously, social media data with geographic tags pointing to the target disaster area is collected from various social media platforms, along with electricity consumption data from every smart meter within the target disaster area. Therefore, by collecting data on regional imagery, public opinion, and electricity consumption, a clear picture of the disaster situation within the target disaster area can be obtained, facilitating the construction of a more accurate current disaster situation view.

[0043] In some embodiments, after collecting monitoring image data, social media opinion data, and power monitoring data, in order to more accurately construct the current disaster situation view, it is necessary to perform spatiotemporal alignment processing on the monitoring image data, social media opinion data, and power monitoring data. It should be noted that spatiotemporal alignment is performed according to time and space. Specifically, the monitoring image data, social media opinion data, and power monitoring data are mapped to a unified GIS network, and the generation timestamps of the monitoring image data, social media opinion data, and power monitoring data are determined respectively. Then, the monitoring image data, social media opinion data, and power monitoring data are aligned to the same sliding time window according to their generation timestamps. In this embodiment, the sliding time window is 15 minutes. Therefore, by performing spatiotemporal alignment on the three types of data, the consistency of different data in time and space can be ensured, facilitating subsequent fusion analysis.

[0044] Furthermore, the collection of monitoring image data needs to meet preset conditions, namely, that the monitoring image data includes all buildings in the target disaster area, so as to construct a current disaster situation view corresponding to the target disaster area based on the monitoring image data.

[0045] Please see Figure 2 In some embodiments, step S102 may include, but is not limited to, steps S201 to S205: Step S201: Classify the intentions based on social media opinion data to obtain opinion categories; Step S202: Extract keywords from social media opinion data based on opinion categories and preset medical need categories to obtain medical need keywords; Step S203: Extract features from the monitoring image data to obtain monitoring image features; Step S204: Extract features from the power monitoring data to obtain power monitoring features; Step S205: Casualty prediction is performed based on medical demand keywords, monitoring image features, and power monitoring features to obtain casualty prediction data.

[0046] In step S201 of some embodiments, intent classification identifies the intent or purpose behind social media opinion data. This embodiment uses the BERT model to classify the intent of social media opinion data to determine the opinion category of each piece of social media opinion data. It should be noted that on social media platforms, opinion categories can include entertainment, food, film and television, clothing, and medical needs, etc.

[0047] In step S202 of some embodiments, social media data categorized as medical needs is selected as public opinion data, and this selected data is public opinion data related to medical needs within the target disaster area. Keywords are extracted from the selected public opinion data as medical need keywords. It should be noted that medical need keywords can be {"fracture", "bleeding", "dehydration", "difficulty breathing", etc.}, and the extent of injury or death can be determined using these keywords.

[0048] In step S203 of some embodiments, monitoring image features are extracted from the monitoring image data, mainly the personnel features of each grid in the monitoring image data. The casualty situation of each grid point can be determined through the personnel features.

[0049] In step S204 of some embodiments, power monitoring features are extracted from power monitoring data. Power monitoring features can be used to determine the power situation of the target disaster area, i.e., the power consumption characteristics, and power monitoring features can be used to assist in subsequent casualty prediction.

[0050] In step S205 of some embodiments, monitoring image features, power monitoring features, and medical demand keywords are fused to obtain multimodal data. This multimodal data is then input into a casualty prediction model, which predicts the casualty data based on the multimodal data. It should be noted that the casualty prediction model is a YOLOv5 network, and a temporal convolutional layer is added to the YOLOv5 network to process continuous frame data and capture time series. Specifically, the YOLOv5 network uses a temporal convolutional layer to process the continuous frame multimodal data, outputting the predicted number of minor injuries, serious injuries, and deaths for each grid, thus obtaining the casualty prediction data. The YOLOv5 network performs multi-layer convolutional operations on the input multimodal data to extract features related to the distribution of the affected population. These features are used to classify casualties and determine the predicted number of minor injuries, serious injuries, and deaths for each grid, thus obtaining the casualty prediction data.

[0051] In steps S201 to S205 of this embodiment, medical demand keywords with the category of medical demand are collected from social media opinion data, and then monitoring image features and power monitoring features are extracted from monitoring image data. By combining medical demand keywords, monitoring image features and power monitoring features, casualty prediction data can be predicted. This can clearly show the casualty information in each grid, which is convenient for accurate prediction of resource demand in the future.

[0052] Please see Figure 3 In some embodiments, step S103 may include, but is not limited to, steps S301 to S303: Step S301: Based on the monitoring image data, identify the damage to fixed facilities to obtain fixed facility damage identification data; Step S302: Identify power fault areas based on power monitoring data to obtain power fault area identification data; Step S303: Determine grid edge parameters based on fixed facility damage identification data, determine grid thermal values ​​based on casualty prediction data, and determine grid grayscale values ​​based on power failure area identification data to construct the current disaster situation view.

[0053] In step S301 of some embodiments, fixed facilities are determined based on the disaster type of the target disaster area. If the disaster type is earthquake or urban flooding, the fixed facilities are buildings, fixed equipment, and fixed structures; if the disaster type is forest fire, the fixed facilities are trees and fixed structures. Fixed facility loss identification data is the fixed facility damage category for each fixed facility. The fixed facility damage category is obtained by identifying the damage category of fixed facilities in the monitoring image data. The road passability probability is determined based on the fixed facility damage category to make more accurate road planning.

[0054] In step S302 of some embodiments, as disclosed above, the power monitoring data is the power consumption data of the smart meter. Specifically, the power fault area identification data is determined based on the decline rate of the power consumption data, which can accurately assess the power status of each area.

[0055] In step S303 of some embodiments, the current disaster view is a hierarchical heat map with grids. The hierarchical heat map can determine the casualty situation, road conditions, and power status in each area. The grid edge parameters in the hierarchical heat map are determined by fixed facility damage identification data, and then the grid heat value is determined based on the predicted number of minor injuries, serious injuries, and deaths in the casualty prediction data. Finally, the grid gray value is determined based on the power failure area identification data.

[0056] Specifically, the grid edge parameter is the bandwidth coefficient. Road traffic capacity is determined based on fixed facility damage identification data and traffic maps of the target disaster area, and the bandwidth coefficient is determined based on the road traffic capacity. The bandwidth coefficient is calculated as: Base value 0.5 + (1 - Road traffic capacity) * 0.3. It should be noted that the road traffic capacity reflects the current road accessibility, and the bandwidth coefficient increases when the road traffic capacity is low, thus expanding the coverage of the tiered heat map. The grid heat value is the sum of the product of a preset heat value formula and a Gaussian decay function, specifically: Grid heat value = ∑Number of injured persons within the grid × Gaussian decay function - Displacement of the center point. The Gaussian decay function is used to simulate the change in injured person density with distance. In the current disaster view, different areas are colored differently based on the grid heat value. If the grid heat value is greater than the first threshold, it is set to red; if the grid heat value is less than the first threshold but greater than the second threshold, it is set to yellow; if the grid heat value is less than the second threshold, it is set to green. It should be noted that the first threshold is greater than the second threshold. In addition, casualty prediction data can also be represented by person icons on the current disaster view.

[0057] For example, if the first threshold is 100 and the second threshold is 30, when the grid heat value is greater than or equal to 100, the corresponding area is set to red; when the grid heat value is greater than 30 and less than 100, the corresponding area is set to yellow; and when the grid heat value is less than or equal to 30, the corresponding area is set to green.

[0058] If the power failure area identification data is "power paralysis area", the corresponding grid is set to grayscale; if the power failure area identification data is "power normal area", the corresponding grid is not set to grayscale. Therefore, the power status of each area can be determined by the grayscale areas in the current disaster view.

[0059] For example, please refer to Figure 4 The current disaster situation view of the target disaster area shown is generated by... Figure 4It can be seen that the traffic status of the road and the number of character icons at different grid points determine the casualties in different areas. At the same time, the power status of each grid point can be determined by the grayscale of the grid points.

[0060] In steps S301 to S303 of this embodiment, fixed facility damage identification data is identified from monitoring image data, and power failure area identification data is identified from power monitoring data. The fixed facility damage identification data is used to determine the grid edge parameters of the graded heat map. Casualty prediction data is used as the grid heat value, and power failure area identification data is used as the grid gray value to construct the current disaster view. The current disaster view is used to determine the distribution of casualties, road conditions, and power status in the target disaster area to assist in the accurate planning of delivery routes.

[0061] Please see Figure 5 In some embodiments, step S301 may include, but is not limited to, steps S501 to S503: Step S501: Extract features from the fixed facilities in the monitoring image data to obtain fixed facility features; wherein, the fixed facility features include: facility texture features, facility shape features and facility structure features; Step S502: Based on the facility texture features, facility shape features and facility structural features, the damage probability of the fixed facility is predicted to obtain the damage prediction probability. Step S503: Determine the damage category of fixed facilities based on the damage prediction probability.

[0062] In step S501 of some embodiments, in this embodiment, the identification of the damage category of fixed facilities in the monitoring image data is completed by a damage identification model, and the damage identification model is a U-Net segmentation network, which includes an encoder, a residual convolutional layer, and a decoder. The encoder encodes the monitoring image data to obtain coded monitoring image data. The residual convolutional layer extracts the fixed facility features from the coded monitoring image data. The residual convolutional layer learns the texture, shape, and structural features in the monitoring image features to accurately extract the facility texture features, facility shape features, and facility structural features.

[0063] In step S502 of some embodiments, facility texture features, facility shape features, and facility structural features are combined into a multi-layer feature map, and then a decoder predicts the damage prediction probability based on the multi-layer feature map. It should be noted that the damage prediction probability can determine the degree of damage to each fixed facility.

[0064] Specifically, the damage prediction probability is calculated as follows: Damage prediction probability = (Number of collapsed pixels * 10 + Number of half-damaged pixels * 0.5) / Total number of pixels.

[0065] In step S503 of some embodiments, the decoder outputs a fixed facility damage category from three classifications: "collapse category," "partially damaged category," and "completed category," based on the damage prediction probability. Furthermore, the monitoring image data is updated according to a preset time period, and the fixed facility damage category is also updated according to a preset time period. For example, if the preset time period is 10 minutes, the fixed facility damage category will be updated every 10 minutes, ensuring that the current disaster view matches the current road conditions and improving the accuracy of route planning.

[0066] In steps S501 to S503 of this embodiment, the texture features, shape features, and structural features of fixed facilities in the monitoring image data are extracted. These features are then combined to predict the damage rate of each fixed facility, resulting in a damage prediction probability. Finally, the damage category of the fixed facility is determined based on the damage prediction rate, achieving accurate damage classification and constructing a current disaster situation view that better reflects actual road conditions.

[0067] Please see Figure 6 In some embodiments, step S302 may include, but is not limited to, steps S601 to S602: Step S601: Extract abnormal data from the power monitoring data to obtain abnormal power data; wherein, the abnormal power data includes: power abnormal duration and power abnormal area information; Step S602: If the duration of the power anomaly exceeds a preset duration threshold, the power anomaly area information is marked with a fault to obtain power fault area identification data.

[0068] In step S601 of some embodiments, as disclosed above, the power monitoring data is power consumption data. By identifying the power decrease value in the power consumption and extracting the time period when the power decrease value exceeds a preset decrease threshold, the power abnormality duration is determined, and the area where the smart meter corresponding to the power consumption data is located is extracted as power abnormality area information.

[0069] In step S602 of some embodiments, when a disaster occurs, the power status may experience a brief interruption and restoration. To accurately determine the power status in each area, the duration of the power anomaly is compared with a preset duration threshold. If the duration of the power anomaly exceeds the preset duration threshold, the area containing the power anomaly information is marked as faulty to determine power fault area identification data. Conversely, if the duration of the power anomaly does not exceed the preset duration threshold, the area containing the power anomaly information is not marked. Therefore, by comparing the duration of the power anomaly with the preset duration threshold, the power fault area identification data for each area can be determined more accurately.

[0070] Specifically, the power consumption decrease within a preset time period is calculated based on electricity consumption data. The duration for which the power consumption decrease exceeds a preset threshold is identified as the power anomaly duration, and areas where the power anomaly duration exceeds the preset threshold are marked as power fault area identification data. If the power consumption decrease within the preset time period exceeds the preset threshold, the area is identified as a "power paralysis area." For example, if the preset time period is 2 hours, and the preset decrease threshold is 80% for cities and 60% for rural areas, if the power consumption in the urban area of ​​the target disaster area decreases by 80% within 2 hours, it is marked as a "power paralysis area"; if the power consumption in the rural area of ​​the target disaster area decreases by 60% within 2 hours, it is marked as a "power paralysis area."

[0071] In steps S601 to S602 of this embodiment, the duration of power anomalies and the corresponding power anomaly area information in the power monitoring data are identified. If the duration of power anomalies exceeds a preset duration threshold, the power anomaly area information is marked as an "infrastructure paralysis area" to obtain power fault area identification data. This data is then used to improve the current disaster situation view and make a more accurate resource distribution planning path based on the current disaster situation view.

[0072] In some embodiments, after the current disaster view is constructed, and a certain grid building in the current disaster view is intact, but there is a surge in SOS messages in the social media data of that grid area, the DS evidence theory will be triggered to verify the monitoring image data, social media data, and power monitoring data. Specifically, if the buildings in the target area of ​​the current disaster view are intact, but there is a surge in SOS messages in the social media data of the target area, or a surge in power failure messages in the social media data, it is necessary to calculate the data source credibility of the monitoring image data, social media data, and power monitoring data separately, and optimize the current disaster view based on the credibility of each data source to construct a more accurate current disaster view and achieve accurate resource distribution path planning.

[0073] Specifically, data credibility is determined by the source information of each data point and a preset candidate credibility level. The credibility of the current disaster situation view is calculated as follows: Image data credibility × Q1 + Public opinion data credibility × Q2 + Power data credibility × Q3. Here, Q1 is the weight value of the monitoring image data, Q2 is the weight value of the social media public opinion data, and Q3 is the weight value of the power monitoring data, with Q1 > Q2 > Q3. It should be noted that the weight value of each data point is dynamically updated and adjusted based on the reliability and real-time nature of the data source. For example, if the update delay of the monitoring image data is significant, Q1 will decrease.

[0074] If the calculated reliability of the current disaster situation view is less than the preset reliability threshold, it is necessary to initiate on-site drone inspection to update the current disaster situation view based on the current image data fed back by the on-site drone inspection, thereby improving the accuracy of resource distribution route planning.

[0075] In some embodiments, the casualty prediction data includes: the number of casualties and the information of the casualties. The number of casualties includes the predicted number of minor injuries, the predicted number of serious injuries, and the predicted number of deaths. The information of the casualties includes the age, gender, height, weight, medical history, and drug allergy information of the minorly injured, seriously injured, and deceased individuals. This embodiment does not impose specific limitations on the content of the information of the casualties.

[0076] Please see Figure 7 In some embodiments, step S104 includes, but is not limited to, steps S701 to S704: Step S701: Classify the disaster situation based on the number of casualties and the information on the casualties to obtain the disaster category; Step S702: Filter the preset historical relief resource consumption data according to the disaster category to obtain the target relief resource consumption data; Step S703: Select the target resource demand prediction model from the preset candidate resource demand prediction models based on the target rescue resource consumption data; Step S704: Resource demand is predicted using the target resource demand prediction model, the number of casualties, and information on casualties to obtain resource demand prediction data.

[0077] In step S701 of some embodiments, information on the casualty group and the casualty object is input into the disaster classification model, and the disaster classification model is used to determine the disaster category of the target disaster area. The disaster category is divided into level 1 disaster, level 2 disaster, level 3 disaster, etc.

[0078] In step S702 of some embodiments, the historical relief resource consumption data is the resource consumption data of the same disaster type. The target relief resource consumption data is selected from the historical relief resource consumption data by disaster category. The resource consumption prediction of the current disaster-stricken area can be improved by using the historical relief resource consumption data.

[0079] In step S703 of some embodiments, resource demand in the future can be predicted by using the resource demand prediction model that matches the disaster category of the target disaster-stricken area and the target relief resource consumption data.

[0080] Specifically, the target relief resource consumption data is cleaned to remove duplicate, erroneous, or incomplete data. The cleaned data is then standardized to ensure consistency in format across different sources. Target disaster data is selected from historical disaster data using the target relief resource consumption data, and disaster characteristics are extracted, including disaster intensity, affected area, number of affected people, and geographical location. Simultaneously, resource consumption characteristics are extracted from the target relief resource consumption data, including medical resource consumption characteristics and daily necessities consumption characteristics. Medical resource consumption characteristics include antibiotic usage, plasma demand, and medical resource allocation. These disaster and resource consumption characteristics are used to train a pre-defined candidate resource demand prediction model to construct a target resource demand prediction model that can accurately predict resource needs.

[0081] In step S704 of some embodiments, the number of casualties and information on the casualties are input into the target resource demand prediction model to predict resource demand, thereby obtaining resource demand prediction data. It should be noted that the resource demand prediction data represents resource demand for a future preset period, specifically including medical resource demand data and daily necessities demand data.

[0082] Specifically, the formula for calculating medical resource demand data is as follows: Antibiotic demand = predicted number of seriously injured persons × 0.8 doses / person + predicted number of minorly injured persons × 0.3 doses / person; Plasma requirement = predicted number of seriously injured people × 400ml / person × trauma type correction factor.

[0083] In steps S701 to S704 of this embodiment, the disaster category is determined by the number of casualties and information on the casualties. Target relief resource consumption data matching the disaster category is then selected from historical relief resource consumption data. Next, the candidate resource demand prediction model is trained based on the target relief resource consumption data to obtain a target resource demand prediction model that conforms to the disaster category. This target resource demand prediction model predicts resource demand based on the current number of casualties and information on the casualties, resulting in accurate resource demand prediction. This facilitates more accurate resource allocation planning and improves relief efficiency.

[0084] Please refer to Figure 8 In some embodiments, after step S704, the resource delivery route planning method may also include, but is not limited to, steps S801 to S802: Step S801: Collect actual resource consumption data; Step S802: Adjust the parameters of the target resource demand forecasting model based on actual resource consumption data and resource demand forecasting data.

[0085] In step S801 of some embodiments, the actual resource consumption data is the actual amount of resources consumed in each area within the target disaster area after the route planning is completed and the resources are delivered. This can be the actual consumption of medical resources and the actual consumption of daily necessities.

[0086] In step S802 of some embodiments, resource loss data is obtained by calculating the actual resource consumption data and resource demand prediction data, and the parameters of the target resource demand prediction model are adjusted according to the resource loss data to improve the accuracy of the target resource demand prediction model.

[0087] In steps S801 to S802 of this embodiment, after resource distribution, actual resource consumption data and resource demand forecast data are collected to adjust the target resource demand forecasting model in order to construct a more accurate target resource demand forecasting model.

[0088] In some embodiments, route planning reference data includes rescue risk information, rescue point information, and rescue vehicle information. Rescue risk information is characterized as a rescue risk value for the target disaster-stricken area, and this risk value is determined based on road damage type and disaster recurrence rate. Rescue point information includes the location, size, and number of rescue personnel at rescue points within the target disaster-stricken area. Rescue vehicle information includes the types of vehicles capable of performing rescue operations in the target disaster-stricken area, vehicle load capacity, maximum load capacity, and current vehicle location.

[0089] Please see Figure 9 In some embodiments, step S105 may include, but is not limited to, steps S901 to S903: Step S901: Based on the resource demand forecast data, rescue point information, and rescue vehicle information, resource allocation is processed to obtain resource allocation information; Step S902: Based on the rescue risk information and the current disaster situation view, a preliminary planned route is obtained; Step S903: Optimize the preliminary planned path based on the resource allocation information to obtain the resource distribution planned path.

[0090] In step S901 of some embodiments, the resource demand forecast data characterizes the resource demand of each grid point within the target disaster-stricken area, including the demand for medical resources and the demand for daily necessities. Resource allocation is performed using aid point information, aid vehicle information, and the resource demand forecast data to determine the resource allocation information for each grid point.

[0091] In step S902 of some embodiments, candidate planning paths can be determined through the current disaster situation view. Risk assessment data is obtained by performing a risk assessment on the candidate planning paths based on the rescue risk information. Candidate planning paths with risk assessment data lower than a preset risk threshold are used as preliminary planning paths.

[0092] In step S903 of some embodiments, the preliminary planned path is optimized based on resource allocation information, mainly using a multi-level cost function. Specifically, the path cost of each preliminary planned path is calculated using the multi-level cost function, and the preliminary planned path with the lowest path cost is selected as the resource distribution planned path. The path cost includes travel time cost, risk cost, and load cost. It should be noted that the travel time from the starting point to the destination of the preliminary planned path is predicted based on the preliminary planned path. The risk cost is mainly determined based on the road damage category and disaster recurrence rate, and the road damage category is the fixed facility damage category; the risk cost is road damage category × 0.7 + disaster recurrence rate × 0.3. The load cost is determined based on the current load, maximum load, and road slope of each vehicle, specifically: load cost = (current load / maximum load)^2 × road slope coefficient. Finally, the travel time cost, risk cost, and load cost are weighted and summed to obtain the path cost.

[0093] In steps S901 to S903 of this embodiment, resource allocation information is first determined, then preliminary planning paths are planned based on rescue risk information and the current disaster situation view, and then the path cost of each preliminary planning path is calculated according to the resource allocation information. The preliminary planning path with the lowest path cost is taken as the resource delivery planning path, so as to achieve accurate path and save resource consumption.

[0094] In some embodiments, after the resource distribution route is planned, resources are distributed according to the planned route to improve the efficiency of disaster relief.

[0095] It should be noted that resource delivery route planning is mainly implemented through the Internet of Things (IoT), which consists of multiple vehicle-mounted terminals, with one terminal installed on each resource delivery vehicle. If the backbone network is interrupted during resource delivery route planning, a LoRaMesh self-organizing network needs to be activated to extend the communication distance between vehicle-mounted terminals to 10km, and a TDMA time slot allocation protocol is adopted to reduce data transmission latency.

[0096] This application uses an earthquake as an example. When an earthquake occurs, satellite imagery data and drone aerial photography data are used as monitoring imagery data for the earthquake zone. An improved U-Net network is used to identify monitoring imagery features in the monitoring imagery data to output three categories of fixed facility damage: collapsed, partially damaged, and intact. Simultaneously, a web crawler engine is used to capture geographic tag information in real time to obtain social media opinion data for the earthquake zone. A BERT model is used to classify the social media opinion data by intent to obtain opinion categories. Medical demand keywords are extracted for opinion categories related to medical needs, specifically "fracture," "bleeding," "injury," and "difficulty breathing," etc. Casualty opinion data is extracted from the social media opinion data based on medical demand keywords. The casualty opinion data, power monitoring data, and monitoring imagery data are spatiotemporally aligned, and the aligned three types of data are merged into multimodal data. A casualty prediction model is then used to predict casualties from the multimodal data to obtain casualty prediction data. The casualty prediction data includes the predicted number of minor injuries, the predicted number of serious injuries, and the predicted number of deaths. Simultaneously, electricity consumption data from each smart meter is collected. If the electricity consumption in the target area decreases by more than a preset threshold for two consecutive hours, the target area is marked as an "infrastructure paralysis zone," thus obtaining power failure area identification data. Then, a current disaster situation view is constructed based on the predicted number of casualties, power failure area identification data, fixed facility damage categories, and building and road features from monitoring images.

[0097] When planning routes begins, resource demand forecasts are determined based on casualty prediction data. Candidate routes are identified using the current disaster situation view, and a risk value is calculated for each candidate route based on rescue risk information. Routes with risk values ​​exceeding a preset risk threshold are filtered to obtain preliminary planned routes. The travel time cost, risk cost, and load cost of each preliminary planned route are calculated, and then these costs are weighted and summed to obtain the route cost. The preliminary planned route with the lowest cost is selected as the resource distribution route, achieving accurate and efficient resource distribution route planning. Furthermore, resource distribution based on the planned route improves the accuracy and timeliness of resource distribution, enhances rescue efficiency, and reduces the risks associated with the resource distribution process.

[0098] Please see Figure 10 This application also provides a resource distribution route planning device that can implement the above-described resource distribution route planning method. The device includes: The data acquisition module 1001 is used to acquire disaster monitoring data, which includes: monitoring image data of the target disaster area, social media data, and power monitoring data. The casualty prediction module 1002 is used to predict casualties in the target disaster area based on monitoring image data, social media data and power monitoring data, and to obtain casualty prediction data. The disaster view construction module 1003 is used to construct a disaster view based on monitoring image data, casualty prediction data and power monitoring data to obtain the current disaster view; wherein, the current disaster view represents the distribution of the disaster-stricken population, road conditions and power conditions in the target disaster area; Demand forecasting module 1004 is used to forecast resource demand based on casualty forecasting data to obtain resource demand forecasting data. The route planning module 1005 is used to plan resource distribution routes based on the current disaster situation view, resource demand forecast data, and preset route planning reference data, and obtain the resource distribution planning route.

[0099] The specific implementation of this resource distribution route planning device is basically the same as the specific implementation of the resource distribution route planning method described above, and will not be repeated here.

[0100] This application also provides a computer device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described resource delivery path planning method. This computer device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0101] Please see Figure 11 , Figure 11 The hardware structure of a computer device according to another embodiment is illustrated. The computer device includes: The processor 1101 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application. The memory 1102 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 1102 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 1102 and is called and executed by the processor 1101 using the resource delivery path planning method of the embodiments of this application. Input / output interface 1103 is used to implement information input and output; The communication interface 1104 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 1105 transmits information between various components of the device (e.g., processor 1101, memory 1102, input / output interface 1103, and communication interface 1104); The processor 1101, memory 1102, input / output interface 1103 and communication interface 1104 are connected to each other within the device via bus 1105.

[0102] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described resource delivery path planning method.

[0103] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0104] The resource distribution route planning method, apparatus, equipment, and storage medium provided in this application collect monitoring image data, social media data, and power monitoring data of the target disaster area, and combine these three types of data to complete the casualty prediction of the target disaster area, thereby obtaining accurate casualty prediction data. Simultaneously, based on the casualty prediction data, power monitoring data, and monitoring image data, a current disaster situation view is constructed that accurately represents the distribution of affected people, road conditions, and power status in the target disaster area. Resource demand prediction data for the target disaster area is determined based on the casualty prediction data, and then combined with the current disaster situation view and resource demand prediction data to complete the distribution route planning, thereby constructing an accurate resource distribution planning route and improving the timeliness and safety of resource distribution.

[0105] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0106] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0107] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0108] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0109] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0110] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0111] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0112] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0113] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0114] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0115] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A resource distribution route planning method, characterized in that, The method includes: Acquire disaster monitoring data; wherein, the disaster monitoring data includes: monitoring image data of the target disaster area, social media data, and power monitoring data; Based on the monitoring image data, the social media data, and the power monitoring data, casualty prediction data is obtained for the target disaster-stricken area. A disaster view is constructed based on the monitoring image data, the casualty prediction data, and the power monitoring data to obtain the current disaster view; wherein, the current disaster view represents the distribution of the affected population, road conditions, and power conditions within the target disaster area; Resource demand forecasting data is obtained by forecasting resource demand based on the casualty prediction data. Based on the current disaster situation view, the resource demand forecast data, and the preset route planning reference data, a resource delivery route is planned to obtain the resource delivery planning route.

2. The method according to claim 1, characterized in that, The step of predicting casualties in the target disaster area based on the monitoring image data, the social media data, and the power monitoring data, to obtain casualty prediction data, includes: Based on the aforementioned social media opinion data, intentions are categorized to obtain opinion categories; Based on the aforementioned public opinion categories and preset medical demand categories, keywords are extracted from the social public opinion data to obtain medical demand keywords; Feature extraction is performed on the monitoring image data to obtain monitoring image features; Feature extraction is performed on the power monitoring data to obtain power monitoring features; Casualty prediction is performed based on the medical demand keywords, the monitoring image features, and the power monitoring features to obtain the casualty prediction data.

3. The method according to claim 1, characterized in that, The process of constructing a disaster view based on the monitoring image data, the casualty prediction data, and the power monitoring data to obtain the current disaster view includes: Based on the monitoring image data, damage to fixed facilities is identified to obtain fixed facility damage identification data. Based on the power monitoring data, power fault areas are identified to obtain power fault area identification data. The grid edge parameters are determined based on the fixed facility damage identification data, the grid thermal value is determined based on the casualty prediction data, and the grid gray value is determined based on the power failure area identification data, so as to construct the current disaster situation view.

4. The method according to claim 3, characterized in that, The fixed facility damage identification data is the fixed facility damage category. The fixed facility damage identification based on the monitoring image data includes: Features are extracted from the fixed facilities in the monitoring image data to obtain fixed facility features; wherein, the fixed facility features include: facility texture features, facility shape features and facility structure features; Based on the facility texture features, facility shape features, and facility structural features, the probability of damage to the fixed facility is predicted to obtain the damage prediction probability; The damage category of the fixed facility is determined based on the predicted damage probability.

5. The method according to claim 3, characterized in that, The step of identifying power fault areas based on the power monitoring data to obtain power fault area identification data includes: The power monitoring data is subjected to anomaly data extraction to obtain power anomaly data; wherein, the power anomaly data includes: power anomaly duration and power anomaly area information; If the duration of the power anomaly exceeds a preset duration threshold, the power anomaly area information is identified as a fault to obtain power fault area identification data.

6. The method according to any one of claims 1 to 5, characterized in that, The casualty prediction data includes: the number of casualties and information on the casualties; the step of predicting resource demand based on the casualty prediction data to obtain resource demand prediction data includes: Based on the number of casualties and the information on the casualties, the disaster situation is classified to obtain the disaster category; Based on the disaster category, the preset historical relief resource consumption data is filtered to obtain the target relief resource consumption data; Based on the target rescue resource consumption data, a target resource demand prediction model is selected from the preset candidate resource demand prediction models; Resource demand forecasting data is obtained by using the target resource demand forecasting model, the number of casualties, and the information on the casualties.

7. The method according to any one of claims 1 to 5, characterized in that, The route planning reference data includes rescue risk information, rescue point information, and rescue vehicle information. The resource delivery route planning, based on the current disaster situation view, the resource demand forecast data, and the preset route planning reference data, yields the resource delivery planned route, including: Resource allocation information is obtained by performing resource allocation processing based on the resource demand forecast data, the rescue point information, and the rescue vehicle information. Based on the rescue risk information and the current disaster situation view, a preliminary planned route is obtained; The resource distribution planning path is obtained by optimizing the preliminary planned path based on the resource allocation information.

8. A resource distribution route planning device, characterized in that, The device includes: The data acquisition module is used to acquire disaster monitoring data; wherein, the disaster monitoring data includes: monitoring image data of the target disaster area, social media data, and power monitoring data; The casualty prediction module is used to predict casualties in the target disaster area based on the monitoring image data, the social media data, and the power monitoring data, and to obtain casualty prediction data. The disaster view construction module is used to construct a disaster view based on the monitoring image data, the casualty prediction data, and the power monitoring data to obtain the current disaster view; wherein, the current disaster view represents the distribution of the affected population, road conditions, and power conditions within the target disaster area; The demand forecasting module is used to forecast resource demand based on the casualty forecasting data to obtain resource demand forecasting data. The route planning module is used to plan resource delivery routes based on the current disaster situation view, the resource demand forecast data, and the preset route planning reference data, so as to obtain the resource delivery planned route.

9. A computer device, characterized in that, The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the resource distribution path planning method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the resource distribution path planning method as described in any one of claims 1 to 7.

Images