Multi-source data fusion disaster situation dynamic assessment method and system

Abstract

The invention provides a disaster situation dynamic assessment method and system based on multi-source data fusion, and relates to the technical field of disaster assessment, and the method comprises the steps: collecting disaster multi-source parameters according to a plurality of data sources, and obtaining a first disaster data set; performing time scale unification by adopting a time alignment algorithm to obtain a second disaster data set; mapping the second disaster data set to a grid space framework by adopting a spatial interpolation algorithm to obtain a third disaster data set; performing multi-source data correction fusion to obtain a disaster fusion data set; inputting the disaster fusion data set into a disaster situation dynamic evaluation model to obtain a disaster situation dynamic evaluation result; and constructing a disaster situation map. According to the disaster situation assessment method and device, the technical problem that the accuracy of disaster situation assessment is poor due to the fact that different types of data are different and multi-source heterogeneous data are difficult to integrate is solved, and the accuracy and timeliness of disaster situation assessment are improved by fusing the multi-source data and data correction.

Description

Technical Field

[0001] The present application relates to the field of disaster assessment technology, and in particular to a method and system for dynamic disaster situation assessment based on multi-source data fusion. Background Art

[0002] When a disaster strikes, data from systems such as towers, communication networks, population heat maps, traffic monitoring, online public opinion, and video analysis often occurs in varying formats, frequencies, and accuracies. This heterogeneity of data presents challenges for real-time integration. For example, tower power outage information may be updated at a frequency of minutes, while population heat maps are generated every half hour. This inconsistency in timescales can lead to biased analysis results. Furthermore, the spatial resolution of data from each system varies: tower data is based on base stations, traffic data is based on road sections, and video data is pixel-level. This makes it difficult to unify these data at different scales within a common spatial framework for comprehensive analysis. Furthermore, the reliability of various data sources is affected by various factors. For example, video analysis data can become unstable due to weather conditions or changes in lighting. This can affect the accuracy of video data during disaster assessment, thereby impacting the overall disaster assessment results.

[0003] In summary, the existing technology has the technical problem that it is difficult to integrate multi-source heterogeneous data due to the differences between different types of data, resulting in poor accuracy of disaster situation assessment. Summary of the Invention

[0004] The purpose of this application is to provide a method and system for dynamic disaster situation assessment based on multi-source data fusion, so as to solve the technical problem in the existing technology that due to the differences between different types of data, it is difficult to integrate multi-source heterogeneous data, resulting in poor accuracy of disaster situation assessment.

[0005] In view of the above problems, this application provides a method and system for dynamic disaster situation assessment based on multi-source data fusion.

[0006] In the first aspect, the present application provides a dynamic assessment method for disaster situation based on multi-source data fusion, which is implemented by a dynamic assessment system for disaster situation based on multi-source data fusion, wherein the dynamic assessment method for disaster situation based on multi-source data fusion includes: collecting multi-source parameters of disasters based on multiple data sources to obtain a first disaster data set; using a time alignment algorithm to unify the time scale of the first disaster data set to obtain a second disaster data set; using a spatial interpolation algorithm to map the second disaster data set to a grid space framework to obtain a third disaster data set; performing multi-source data correction and fusion based on the third disaster data set to obtain a disaster fusion data set; inputting the disaster fusion data set into a disaster situation dynamic assessment model to obtain a disaster situation dynamic assessment result; and constructing a disaster situation map based on the disaster situation dynamic assessment result.

[0007] Optionally, the multiple data sources include a tower system, a communication network system, a population thermal system, a traffic monitoring system, an online public opinion system, and a video large model analysis system.

[0008] Optionally, a unified time axis is constructed according to a predetermined time scale; multiple update frequencies corresponding to the multiple data sources are obtained; based on the multiple update frequencies, the first disaster data set is time-aligned according to the unified time axis to generate the second disaster data set.

[0009] Optionally, the grid space framework is constructed according to standard grid units; data mapping constraints are set; based on the data mapping constraints, the second disaster dataset is spatially interpolated and projected according to the grid space framework to generate the third disaster dataset.

[0010] Optionally, the third disaster data set is corrected for environmental interference to obtain a fourth disaster data set; weights are assigned to the multiple data sources to obtain weight values ​​for each data source; and multi-source data fusion is performed on the fourth disaster data set based on the weight values ​​for each data source to obtain the disaster fusion data set.

[0011] Optionally, the tower system is used as the benchmark data source, and a data quality assessment model is used to perform reliability scoring on the multiple data sources to obtain a multi-source reliability score; feature identification is performed on the fourth disaster data set to obtain disaster scene features; and based on the multi-source reliability score and the disaster scene features, weights are assigned to the multiple data sources to obtain weight values ​​for each data source.

[0012] Optionally, the video data in the third disaster data set is extracted; the video data is preprocessed using an environmental factor correction model to obtain corrected video data; the third disaster data set is updated according to the corrected video data to generate the fourth disaster data set.

[0013] In the second aspect, the present application also provides a disaster situation dynamic assessment system with multi-source data fusion, which is used to execute the disaster situation dynamic assessment method with multi-source data fusion as described in the first aspect, wherein the disaster situation dynamic assessment system with multi-source data fusion includes: a data acquisition module, which is used to collect multi-source parameters of disasters based on multiple data sources to obtain a first disaster data set; a time alignment module, which is used to unify the time scale of the first disaster data set using a time alignment algorithm to obtain a second disaster data set; a spatial mapping module, which is used to map the second disaster data set to a grid space framework using a spatial interpolation algorithm to obtain a third disaster data set; a data correction module, which is used to perform multi-source data correction and fusion based on the third disaster data set to obtain a disaster fusion data set; a dynamic assessment module, which is used to input the disaster fusion data set into a disaster situation dynamic assessment model to obtain a disaster situation dynamic assessment result; and a situation map construction module, which is used to construct a disaster situation map based on the disaster situation dynamic assessment result.

[0014] One or more technical solutions provided in this application have at least the following beneficial effects:

[0015] The method collects multi-source disaster parameters from multiple data sources to obtain a first disaster dataset; uses a time alignment algorithm to unify the time scale of the first disaster dataset to obtain a second disaster dataset; uses a spatial interpolation algorithm to map the second disaster dataset to a grid spatial framework to obtain a third disaster dataset; performs multi-source data correction and fusion based on the third disaster dataset to obtain a disaster fusion dataset; inputs the disaster fusion dataset into a dynamic disaster situation assessment model to obtain a dynamic disaster situation assessment result; and constructs a disaster situation map based on the dynamic disaster situation assessment result. In other words, by collecting disaster-related data from various data sources, unifying the data scale using time alignment and spatial interpolation algorithms, processing the data using an environmental factor correction model, and analyzing the fused dataset using the dynamic disaster situation assessment model, an accurate disaster situation map is generated. This achieves effective integration and analysis of multi-source heterogeneous data, significantly improving the accuracy and timeliness of disaster situation assessment.

[0016] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, which can be implemented in accordance with the contents of the description, and to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically listed below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in this application or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely exemplary, and a person of ordinary skill in the art can obtain other drawings based on the provided drawings without creative work.

[0018] Figure 1 This is a flowchart of the multi-source data fusion method for dynamic disaster situation assessment in this application;

[0019] Figure 2 This is a structural diagram of the disaster situation dynamic assessment system based on multi-source data fusion in this application.

[0020] Description of the accompanying drawings: data acquisition module 11, time alignment module 12, space mapping module 13, data correction module 14, dynamic evaluation module 15, situation map construction module 16. DETAILED DESCRIPTION

[0021] This application provides a method and system for dynamic disaster situation assessment based on multi-source data fusion, addressing the existing technical issues of poor accuracy in disaster situation assessment due to the difficulty in integrating multi-source heterogeneous data due to differences in different types of data. By collecting disaster-related data from various data sources, unifying the data scale using time alignment and spatial interpolation algorithms, processing the data using an environmental factor correction model, and analyzing the fused data set using a dynamic disaster situation assessment model, an accurate disaster situation map is generated. This achieves effective integration and analysis of multi-source heterogeneous data, significantly improving the accuracy and timeliness of disaster situation assessment.

[0022] Below, the technical solutions in this application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of this application, rather than all the embodiments of this application. It should be understood that this application is not limited to the example embodiments described herein. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application. It should also be noted that, for the convenience of description, only the parts related to this application, rather than all of them, are shown in the accompanying drawings.

[0023] For example, see the attached Figure 1 The present application provides a method for dynamic assessment of disaster situation based on multi-source data fusion, wherein the method is executed by a dynamic assessment system for disaster situation based on multi-source data fusion, and the method specifically includes the following steps:

[0024] S100: Collect disaster multi-source parameters based on multiple data sources to obtain a first disaster data set.

[0025] Furthermore, the present application S100 includes:

[0026] The multiple data sources include a tower system, a communication network system, a population thermal system, a traffic monitoring system, an online public opinion system, and a video large model analysis system.

[0027] Specifically, multi-source disaster parameters are collected from multiple data sources, such as tower systems, communication network systems, population thermal systems, traffic monitoring systems, online public opinion systems, and video large-scale model analysis systems. That is, during the occurrence or evolution of disasters, observation data parameters from different fields, such as power status, network connectivity, crowd density, traffic flow, etc., are collected to obtain the first disaster data set. The tower system collects information on power outage status (whether power is supplied) and offline status (station loss alarms). The communication network system collects information on the number of online users and loss rates at base stations. The population and thermal system collects information on the number of active users in each area. The transportation system collects information on speed and volume for each road section. The online public opinion system uses multimodal analysis (text, image, and video) to automatically identify and locate disaster events. This includes extracting disaster keywords and geolocating them (e.g., "a certain street has lost power"), assessing the impact area, tracing rumors, and constructing a semantic map of emergencies. The video large model analysis system uses a multimodal large model to identify video surveillance, adaptively correcting for environmental interference (such as rain, snow, fog, and low light), and identifying crowd density, unusual behavior events, and key targets (such as collapsed buildings and trapped people) in disaster scenarios. All data is assigned a unified sampling timestamp. By collecting multi-source disaster parameters from multiple data sources, more comprehensive and accurate disaster information is obtained.

[0028] S200: Using a time alignment algorithm to unify the time scale of the first disaster dataset to obtain a second disaster dataset.

[0029] Furthermore, the present application S200 includes:

[0030] According to a predetermined time scale, a unified time axis is constructed; a plurality of update frequencies corresponding to the plurality of data sources are obtained; based on the plurality of update frequencies, the first disaster data set is time-aligned according to the unified time axis to generate the second disaster data set.

[0031] Specifically, different data sources have different update frequencies. For example, tower data may be updated every minute, while population heatmaps are updated every half hour. This can lead to data temporal inconsistencies. Therefore, when data is collected asynchronously from multiple sources, a time alignment algorithm is used to uniformly map the data to a common time scale. First, a unified timeline is constructed, which is usually at the minute level. The update frequency of each data source is identified, and the data in the first disaster dataset is aligned to this unified timeline. For example, for tower data, there will be a data point every minute, while for population heatmaps, there will be a data point every half hour. For high-frequency data (such as video analysis), sliding window aggregation (such as maximum / average) is used for minute-level aggregation. For low-frequency data (such as heatmaps), time padding is performed using the zero-value hold (ZOH) or linear interpolation. For medium-frequency data, adjacent time points are directly matched or interpolated to the target time step.

[0032] By using time alignment algorithms, data at different time scales are unified onto a common time base, which helps eliminate data analysis bias caused by inconsistent time scales.

[0033] S300: Mapping the second disaster dataset to a grid space framework using a spatial interpolation algorithm to obtain a third disaster dataset.

[0034] Furthermore, the present application S300 includes:

[0035] According to the standard grid units, the grid space framework is constructed; data mapping constraints are set; based on the data mapping constraints, the second disaster dataset is spatially interpolated and projected according to the grid space framework to generate the third disaster dataset.

[0036] Specifically, due to the different spatial resolutions of different data sources, for example, tower data is based on base stations, traffic data is based on road sections, and video data is at the pixel level, which will lead to spatial inconsistency of the data. Using a spatial interpolation algorithm, we first construct a standard grid space framework, such as in kilometers. Spatial interpolation algorithms are a method of estimating the values ​​of other unknown locations based on data points at known spatial locations, including nearest neighbor interpolation (NNI), inverse distance weighting (IDW), Kriging interpolation (Kriging), bilinear interpolation, etc. The grid space framework divides the spatial area into regular grids of a uniform scale (such as 1km×1km), which is used to unify the mapping and analysis platform of spatial data from different sources and different granularities.

[0037] Data mapping constraints are set to limit or correct interpolation accuracy using geographic, logical, or semantic conditions during spatial interpolation, such as boundary restrictions, coverage limits, and data validity weights. Based on these constraints, the data in the secondary disaster dataset is mapped onto this grid space framework. For example, for tower data, interpolate from the base station location to the nearest grid cell; for traffic data, distribute the data evenly from road sections to the grid cells it passes through; and for video data, aggregate the data from the pixel level to the corresponding grid cell.

[0038] Through spatial mapping and projection, all data were unified onto a kilometer-scale grid, resulting in the third disaster dataset. Using spatial interpolation algorithms, data of varying spatial resolutions were unified onto a standard grid, helping to eliminate data analysis biases caused by inconsistent spatial resolution. This dataset, with its unified spatial resolution, provides a more accurate and consistent spatial data foundation for subsequent disaster situation assessments.

[0039] S400: performing multi-source data correction and fusion according to the third disaster data set to obtain a disaster fusion data set.

[0040] Furthermore, the present application S400 includes:

[0041] The third disaster data set is corrected for environmental interference to obtain a fourth disaster data set; weights are assigned to the multiple data sources to obtain weight values ​​of each data source; and multi-source data fusion is performed on the fourth disaster data set based on the weight values ​​of each data source to obtain the disaster fusion data set.

[0042] Taking the tower system as the benchmark data source, a data quality assessment model is used to perform reliability scoring on the multiple data sources to obtain a multi-source reliability score; feature identification is performed on the fourth disaster data set to obtain disaster scene features; based on the multi-source reliability score and the disaster scene features, weights are assigned to the multiple data sources to obtain weight values ​​for each data source.

[0043] Furthermore, the present application further comprises the following steps:

[0044] Extract the video data in the third disaster data set; use the environmental factor correction model to preprocess the video data to obtain corrected video data; update the third disaster data set according to the corrected video data to generate the fourth disaster data set.

[0045] Specifically, video data is affected by environmental factors such as weather and lighting and contains noise or distortion, which will affect the accuracy of video analysis data. In order to improve data quality, video data is extracted from the third disaster data set, that is, the original video image sequence under each grid (or monitoring point). The data format is a grayscale image / color video frame sequence, and the timestamps are unified. The environmental factor correction model is used to preprocess the video data to obtain corrected video data. The environmental factor correction model is an algorithm model used to compensate for or remove the impact of environmental interference (such as lighting changes, rain and snow occlusion, and haze blur) on visual data. Common methods include brightness normalization, image enhancement, GAN dehazing, image restoration network, etc.

[0046] Image enhancement is performed on the video data, such as histogram equalization and adaptive contrast enhancement (CLAHE) to improve low-light conditions. Next, dehazing and deraining are performed on the video data using AOD-Net, DehazeNet, and RainNet to address haze and rain streaks. The Retinex algorithm is used to separate illumination and reduce the effects of glare and shadows. After model processing, video data with higher quality and more stable visual features is obtained, known as rectified video data. The original fields are replaced with the rectified video data to generate the fourth disaster dataset.

[0047] Tower data was selected as the baseline data source due to its high frequency of power outage / offline updates (once per minute), minimal impact from external interference, and high reliability. A data quality assessment model was developed, inputting metrics such as temporal integrity, spatial coverage, anomaly rate, volatility, and latency for each data source and outputting a quality score. The data quality assessment model is a mechanism for quantifying the quality and reliability of multi-source data in disaster scenarios. Evaluation metrics may include data integrity, timeliness, robustness, accuracy, and anti-interference capabilities.

[0048] The data quality assessment model is used to assess the reliability of other data sources (such as communication network systems, population and thermal systems, traffic monitoring systems, online public opinion systems, and video large-scale model analysis systems). This multi-source reliability score is calculated by taking into account factors such as data completeness, consistency, accuracy, and timeliness. For example, if video analysis data is less reliable under certain weather conditions, its score will be lowered accordingly.

[0049] Based on the fourth disaster dataset, feature recognition is performed to extract disaster scenario characteristics, namely key contextual attributes of the current environment (such as severe weather, nighttime scenes, high crowd density, and network outages). These include the disaster type, impact area, and duration. These characteristics are used to assist in determining the dynamic changes in the reliability of different data sources. Based on the multi-source reliability scores and disaster scenario characteristics, a dynamic weighting algorithm is used to assign weight coefficients to each data source, resulting in weight values ​​for each data source. If the importance of video analysis data in the disaster scenario is high, the weight of video analysis is increased accordingly. For example, suppose that during a hurricane, the reliability score of the tower system is 0.95, while the reliability score of the video model analysis system drops to 0.80 due to weather conditions. If the disaster scenario characteristics indicate that video analysis data is critical for assessing the hurricane's impact on buildings, the dynamic weighting algorithm may assign a relatively high weight, such as 0.6, to the video model analysis system, while the tower system may receive a weight of 0.4.

[0050] Multi-source data, which have been time-aligned, spatially unified, and environmentally corrected, are fused using a multi-source data fusion algorithm, combined with the weights assigned to each data source. This involves weighting each data source in the fourth disaster dataset with its corresponding weighted value to produce a fused disaster dataset. Multi-source data fusion integrates information from diverse data sources, improving the accuracy and comprehensiveness of disaster situation assessments. Multi-source data fusion algorithms combine data from diverse sources, formats, and precision levels within a unified spatial-temporal framework using weighted methods. Common strategies include weighted averaging.

[0051] Through time alignment, spatial unification, environmental correction and multi-source data fusion, reliable integration of data at a unified scale is achieved, avoiding error amplification caused by a single data source, significantly improving the robustness and stability of disaster perception information, supporting accurate situation assessment, and improving the practical integration capability of multi-source heterogeneous data. It is particularly suitable for complex disaster scenarios.

[0052] S500: Inputting the disaster fusion data set into a disaster situation dynamic assessment model to obtain a disaster situation dynamic assessment result.

[0053] Specifically, a dynamic disaster situation assessment model is constructed based on real-world scenarios. It analyzes and assesses the current state of disasters and their evolving trends, identifies the affected areas and severity, and predicts their development. This model supports real-time updates and is dynamically adaptable. The model input is a fused disaster dataset, and its output includes the current risk level for each region, the impact spread trends within each region, identification of hotspots, and predictions of the dynamic evolution of disasters.

[0054] Historical disaster data, including tower outage information, communication network status, population heat maps, traffic monitoring data, and video analysis data, is collected and labeled to determine actual disaster situation results, such as confirmed affected areas, casualty figures, and traffic disruptions. Data is cleaned to remove noise and outliers and address missing data. Time alignment and spatial interpolation are performed to ensure data consistency and comparability. Meaningful features are extracted from the data, such as the area of ​​the power outage, the duration of the communication network outage, and the degree of traffic congestion. New features are created, such as the season and time of day when the disaster occurred, which may influence the severity of the disaster. An appropriate model, such as a GNN, is selected based on the nature of the problem. First, a two- or multi-layer GNN is designed. The first layer projects the input features into a latent space and aggregates neighbor information to extract local spatiotemporal relationships. The middle layer further expands the receptive field to capture wider spatial correlations. The output layer maps the latent vector to a final risk prediction score or classification probability. Based on the node representation output by the last GNN layer, a multi-layer perceptron (MLP) readout module is added to classify each node's risk level or perform risk score regression. If changes across multiple moments need to be captured, historical fusion values ​​from the previous N moments can be concatenated into each node's features. Alternatively, a time series model (such as an LSTM) can be added to the GNN to process the node features, and the time series output can then be fed into the GNN.

[0055] All cases are randomly divided into training, validation, and test sets (common ratios are 70%, 15%, and 15%) by time or space to ensure uniform distribution across disaster types and regions. The training set is used to train the model, and model parameters are adjusted to optimize performance. During training, graphs are processed in batches: each batch contains graphs from several moments or sub-regions to alleviate memory pressure and accelerate convergence. Within a batch, forward propagation, loss calculation, and backpropagation are performed on each graph in sequence. Cross-entropy loss is calculated for the classification results of each node, and the overall loss is the sum or average of all nodes. Hyperparameters are set, including learning rate, batch size, hidden dimension size, regularization coefficient, and number of training rounds. To account for the dimensional differences of different physical quantities, the mean and standard deviation of each dimension of the feature are first calculated on the training set, and then all samples are normalized to zero mean and unit variance to ensure that the features have comparable scales during model training.

[0056] Use the test set to evaluate model performance, such as accuracy, recall, and F1 score. Adjust model parameters or try different model architectures based on the evaluation results. Repeat the training and evaluation process until satisfactory performance is achieved. Deploy the trained model in a real-world environment for real-time or periodic disaster situation assessments. Ensure that the model can receive real-time data and output evaluation results.

[0057] The fused disaster dataset is fed into the Disaster Situation Dynamic Assessment Model, which uses built-in algorithms and parameters to analyze the impact of disasters. This quantitative analysis of the impacts of disasters helps us understand their severity and urgency.

[0058] S600: Constructing a disaster situation map according to the dynamic assessment result of the disaster situation.

[0059] Specifically, the results of the dynamic disaster situation assessment are presented in an intuitive manner, employing visualization technology to construct a disaster situation map. Using Geographic Information System (GIS) software, the severity of the disaster is represented by color-coded layers, the impact area is indicated by shaded or marked areas, and the development trend is depicted by animations or trend lines. This allows for rapid identification of the most severely affected areas, assessment of the overall impact, and prediction of future disaster trends. Based on the model's predicted spread direction, arrows or streamlines are added to each or key grid cell to visually demonstrate the disaster's spread path. Important nodes, such as points of greatest risk, locations of rescue resources, and evacuation routes, are also annotated. By constructing a disaster situation map, complex disaster assessment data is transformed into easily understandable visual information, facilitating rapid decision-making and the development of effective response strategies. Quantitative analysis using the disaster assessment model generates a precise disaster situation map, effectively integrating and analyzing multi-source heterogeneous data. This significantly improves the accuracy and timeliness of disaster situation assessments, providing important technical support for disaster emergency management.

[0060] In summary, the multi-source data fusion disaster situation dynamic assessment method provided by this application has the following advantages:

[0061] Beneficial effects:

[0062] The method collects multi-source disaster parameters from multiple data sources to obtain a first disaster dataset; uses a time alignment algorithm to unify the time scale of the first disaster dataset to obtain a second disaster dataset; uses a spatial interpolation algorithm to map the second disaster dataset to a grid spatial framework to obtain a third disaster dataset; performs multi-source data correction and fusion based on the third disaster dataset to obtain a disaster fusion dataset; inputs the disaster fusion dataset into a dynamic disaster situation assessment model to obtain a dynamic disaster situation assessment result; and constructs a disaster situation map based on the dynamic disaster situation assessment result. In other words, by collecting disaster-related data from various data sources, unifying the data scale using time alignment and spatial interpolation algorithms, processing the data using an environmental factor correction model, and analyzing the fused dataset using the dynamic disaster situation assessment model, an accurate disaster situation map is generated. This achieves effective integration and analysis of multi-source heterogeneous data, significantly improving the accuracy and timeliness of disaster situation assessment.

[0063] Example 2: Based on the same inventive concept as the method for dynamic assessment of disaster situation by multi-source data fusion in the above-mentioned Example 1, this application also provides a dynamic assessment system for disaster situation by multi-source data fusion, please refer to the attached Figure 2 The multi-source data fusion disaster situation dynamic assessment system includes:

[0064] The data acquisition module 11 is used to collect multi-source parameters of disasters based on multiple data sources to obtain a first disaster data set; the time alignment module 12 is used to unify the time scale of the first disaster data set using a time alignment algorithm to obtain a second disaster data set; the spatial mapping module 13 is used to map the second disaster data set to a grid space framework using a spatial interpolation algorithm to obtain a third disaster data set; the data correction module 14 is used to perform multi-source data correction and fusion based on the third disaster data set to obtain a disaster fusion data set; the dynamic assessment module 15 is used to input the disaster fusion data set into a disaster situation dynamic assessment model to obtain a disaster situation dynamic assessment result; the situation map construction module 16 is used to construct a disaster situation map based on the disaster situation dynamic assessment result.

[0065] Furthermore, the data acquisition module 11 in the multi-source data fusion disaster situation dynamic assessment system is further used to:

[0066] The multiple data sources include a tower system, a communication network system, a population thermal system, a traffic monitoring system, an online public opinion system, and a video large model analysis system.

[0067] Furthermore, the time alignment module 12 in the multi-source data fusion disaster situation dynamic assessment system is further configured to:

[0068] According to a predetermined time scale, a unified time axis is constructed; a plurality of update frequencies corresponding to the plurality of data sources are obtained; based on the plurality of update frequencies, the first disaster data set is time-aligned according to the unified time axis to generate the second disaster data set.

[0069] Furthermore, the spatial mapping module 13 in the multi-source data fusion disaster situation dynamic assessment system is further used to:

[0070] According to the standard grid units, the grid space framework is constructed; data mapping constraints are set; based on the data mapping constraints, the second disaster dataset is spatially interpolated and projected according to the grid space framework to generate the third disaster dataset.

[0071] Furthermore, the data correction module 14 in the multi-source data fusion disaster situation dynamic assessment system is further used to:

[0072] The third disaster data set is corrected for environmental interference to obtain a fourth disaster data set; weights are assigned to the multiple data sources to obtain weight values ​​of each data source; and multi-source data fusion is performed on the fourth disaster data set based on the weight values ​​of each data source to obtain the disaster fusion data set.

[0073] Furthermore, the data correction module 14 in the multi-source data fusion disaster situation dynamic assessment system is further used to:

[0074] Taking the tower system as the benchmark data source, a data quality assessment model is used to perform reliability scoring on the multiple data sources to obtain a multi-source reliability score; feature identification is performed on the fourth disaster data set to obtain disaster scene features; based on the multi-source reliability score and the disaster scene features, weights are assigned to the multiple data sources to obtain weight values ​​for each data source.

[0075] Furthermore, the data correction module 14 in the multi-source data fusion disaster situation dynamic assessment system is further used to:

[0076] Extract the video data in the third disaster data set; use the environmental factor correction model to preprocess the video data to obtain corrected video data; update the third disaster data set according to the corrected video data to generate the fourth disaster data set.

[0077] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. Figure 1 The multi-source data fusion dynamic disaster situation assessment method and specific examples in Example 1 are also applicable to the multi-source data fusion dynamic disaster situation assessment system in this embodiment. Through the above detailed description of the multi-source data fusion dynamic disaster situation assessment method, those skilled in the art can clearly understand the multi-source data fusion dynamic disaster situation assessment system in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here.

[0078] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

[0079] Obviously, for those skilled in the art, several improvements and modifications can be made to the present application without departing from the principles of the present application, and these improvements and modifications also fall within the scope of protection of the present application.

Claims

1. A dynamic disaster situation assessment method based on multi-source data fusion, characterized by: include: Collect disaster multi-source parameters based on multiple data sources to obtain the first disaster data set; Using a time alignment algorithm to unify the time scale of the first disaster dataset to obtain a second disaster dataset; Mapping the second disaster dataset to a grid space framework using a spatial interpolation algorithm to obtain a third disaster dataset; Perform multi-source data correction and fusion according to the third disaster data set to obtain a disaster fusion data set; Inputting the disaster fusion data set into a disaster situation dynamic assessment model to obtain a disaster situation dynamic assessment result; A disaster situation map is constructed based on the dynamic assessment results of the disaster situation.

2. The method for dynamic disaster situation assessment based on multi-source data fusion according to claim 1, characterized in that: The time scale of the first disaster dataset is unified by using a time alignment algorithm to obtain a second disaster dataset, including: Construct a unified timeline based on the predetermined time scale; Obtaining multiple update frequencies corresponding to the multiple data sources; Based on the multiple update frequencies, the first disaster dataset is time-aligned according to the unified time axis to generate the second disaster dataset.

3. The method for dynamic disaster situation assessment based on multi-source data fusion according to claim 1, characterized in that: The second disaster dataset is mapped to a grid space framework using a spatial interpolation algorithm to obtain a third disaster dataset, including: constructing the grid space framework according to standard grid units; Set data mapping constraints; Based on the data mapping constraints, spatial interpolation projection is performed on the second disaster dataset according to the grid space framework to generate the third disaster dataset.

4. The method for dynamic disaster situation assessment based on multi-source data fusion according to claim 1, wherein: Performing multi-source data correction and fusion according to the third disaster dataset to obtain a disaster fusion dataset includes: performing environmental interference correction on the third disaster dataset to obtain a fourth disaster dataset; Performing weight assignment on the multiple data sources to obtain a weight value for each data source; Multi-source data fusion is performed on the fourth disaster dataset based on the weight values ​​of each data source to obtain the disaster fusion dataset.

5. The method for dynamic disaster situation assessment based on multi-source data fusion according to claim 4, characterized in that: Performing weight assignment on the multiple data sources to obtain a weight value for each data source includes: Taking the tower system as a benchmark data source, a data quality assessment model is used to perform reliability scoring on the multiple data sources to obtain a multi-source reliability score; Performing feature recognition on the fourth disaster data set to obtain disaster scene features; Based on the multi-source reliability scores and the disaster scenario characteristics, weights are assigned to the multiple data sources to obtain weight values ​​for each data source.

6. The method for dynamic disaster situation assessment based on multi-source data fusion according to claim 4, characterized in that: Performing environmental interference correction on the third disaster dataset to obtain a fourth disaster dataset includes: Extracting video data from the third disaster data set; Preprocessing the video data using an environmental factor correction model to obtain corrected video data; The third disaster data set is updated according to the corrected video data to generate the fourth disaster data set.

7. The method for dynamic disaster situation assessment based on multi-source data fusion according to claim 1, wherein: The multiple data sources include a tower system, a communication network system, a population thermal system, a traffic monitoring system, an online public opinion system, and a video large model analysis system.

8. A multi-source data fusion disaster situation dynamic assessment system, characterized by: The method for dynamically assessing a disaster situation by fusion of multi-source data according to any one of claims 1 to 7 is implemented, wherein the system for dynamically assessing a disaster situation by fusion of multi-source data comprises: A data collection module is used to collect disaster multi-source parameters based on multiple data sources to obtain a first disaster data set; A time alignment module, configured to unify the time scale of the first disaster dataset using a time alignment algorithm to obtain a second disaster dataset; A spatial mapping module, configured to map the second disaster dataset to a grid spatial framework using a spatial interpolation algorithm to obtain a third disaster dataset; A data correction module, configured to perform multi-source data correction and fusion based on the third disaster data set to obtain a disaster fusion data set; A dynamic assessment module, configured to input the disaster fusion data set into a disaster situation dynamic assessment model to obtain a disaster situation dynamic assessment result; The situation map construction module is used to construct a disaster situation map according to the dynamic assessment result of the disaster situation.

Images