Disaster emergency surveying and mapping method based on remote sensing data

Abstract

The invention relates to a disaster emergency surveying and mapping method based on remote sensing data. The method comprises the steps of collecting a remote sensing image of a geological disaster high-risk area, identifying a shielding area in the remote sensing image, complementing the shielding area, and obtaining a complemented remote sensing image; and abnormal information and damage information are extracted from the complemented remote sensing image, earthquake disaster emergency mapping is carried out according to the abnormal information and the damage information, and a disaster emergency mapping map is generated. According to the invention, high-precision and high-timeliness requirements of disaster emergency rescue on geographic information are met.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of disaster emergency mapping, in particular to a disaster emergency mapping method based on remote sensing data. BACKGROUND

[0002] Natural disasters, especially geological disasters such as earthquakes, landslides, debris flows, and rock avalanches, pose an increasing threat to human society. These disasters not only cause significant casualties and property losses, but also have a profound impact on infrastructure, ecological environment, and social stability. With global climate change and the intensification of human activities, the frequency and intensity of geological disasters are on the rise, making the importance of disaster emergency mapping even more prominent.

[0003] Limitations of existing disaster emergency mapping technology:

[0004] Limitations of ground surveys: Although traditional ground survey methods can obtain detailed on-site information, they are time-consuming and inefficient, making it difficult to provide comprehensive geographic information in the first moments after a disaster occurs. Especially in disaster-stricken areas with disrupted transportation and complex terrain, ground surveys are extremely difficult, severely affecting the timeliness of emergency rescue.

[0005] Delay of remote sensing data: Satellite remote sensing is an important means of obtaining large-scale geographic information, but there is a time delay in data acquisition. For example, the revisit period of optical remote sensing satellites is usually several days or even weeks, while synthetic aperture radar (SAR) satellites can penetrate clouds, but data acquisition and processing also require a long time. This delay may result in rescue decisions based on outdated information, affecting rescue effectiveness.

[0006] Existence of occluded areas: In complex terrain and vegetation-covered areas, remote sensing images often have occlusion phenomena such as mountain shadows and vegetation occlusion. These occluded areas can result in missing critical information, such as road damage conditions and building damage levels. Traditional methods are difficult to effectively fill in these missing information, affecting the comprehensiveness and accuracy of disaster assessment.

[0007] Limitations of data resolution: Although the resolution of existing remote sensing data is constantly improving, in some local areas, especially in densely populated urban areas and complex terrain areas, it is still difficult to meet the needs of detailed disaster assessment. For example, high-resolution optical imagery can clearly show buildings and roads, but in densely vegetated mountainous areas, its resolution advantage may be weakened.

[0008] Limitations of traditional methods: Traditional remote sensing image processing methods mainly rely on manual visual interpretation or simple image processing algorithms. These methods often struggle to accurately extract abnormal information and damage information when faced with complex disaster scenarios. For example, in earthquake disasters, subtle changes such as road cracks and building tilting may be overlooked, leading to inaccurate disaster assessment results.

[0009] Lack of automation and intelligent means: Traditional methods lack automated and intelligent processing procedures, making it difficult to quickly respond to disaster emergency needs. For example, after a large-scale disaster occurs, a large amount of remote sensing data needs to be processed within a short period of time, and manual interpretation and simple algorithms cannot meet this demand.

[0010] Single data type: Existing disaster emergency mapping methods often rely on only a single type of remote sensing data, such as optical imagery or radar imagery, while ignoring other important data sources such as geological exploration data, forest and grassland distribution data, and population distribution data. These data play an important role in disaster risk assessment but have not been fully utilized.

[0011] In summary, existing disaster emergency mapping technologies have many shortcomings in terms of data acquisition timeliness, data coverage completeness, information extraction accuracy, and multi-source data fusion. These problems seriously hinder the efficiency and effectiveness of disaster emergency response. Therefore, developing an emergency mapping method that can quickly, comprehensively, and accurately obtain geographic information in disaster areas is of great significance to improving disaster emergency response capabilities and rescue efficiency. SUMMARY

[0012] To solve the problems existing in the prior art, the purpose of the present application is to provide a disaster emergency mapping method based on remote sensing data, which meets the high-precision and high-timeliness requirements of geographic information for disaster emergency rescue.

[0013] To achieve the above-mentioned purpose, the present application provides the following solutions:

[0014] A disaster emergency mapping method based on remote sensing data, comprising:

[0015] Collecting remote sensing images of a high-risk geological disaster area, identifying occluded areas in the remote sensing images, completing the occluded areas, and obtaining completed remote sensing images;

[0016] Extracting abnormal information and damage information from the completed remote sensing images, conducting earthquake emergency mapping according to the abnormal information and the damage information, and generating a disaster emergency mapping map.

[0017] Optionally, obtaining the high-risk geological disaster area comprises:

[0018] Obtaining multi-source data of a region to be detected, wherein the multi-source data comprises original remote sensing images, geological exploration data, forest and grassland distribution data, personnel settlement distribution data, disaster history data, railway network data and highway network data;

[0019] Performing correction processing on the multi-source data, extracting key geological features and geological disaster hidden danger features from the corrected multi-source data, and combining the disaster history data to divide buildings and surrounding environments into different fire disaster risk levels;

[0020] According to the different fire disaster risk levels, a region corresponding to a fire disaster risk level exceeding a target threshold is obtained as the geological disaster high-risk region.

[0021] Optionally, identifying the occluded region in the remote sensing image comprises:

[0022] Inputting the remote sensing image into a pre-constructed occluded region identification channel to identify the occluded region in the remote sensing image, wherein the occluded region identification channel is pre-trained and constructed based on a convolutional neural network, and training data comprises sample remote sensing images and sample occluded regions.

[0023] Optionally, obtaining the completed remote sensing image comprises:

[0024] Analyzing the remote sensing image to obtain an average obviousness of the remote sensing image, identifying the occluded region in the remote sensing image, configuring a collection range of a UAV according to the average obviousness, and collecting the occluded region by using the UAV;

[0025] Optimizing a collection position of UAV image collection in the occluded region to obtain an optimal collection position distribution, and merging the collection positions according to a repetition degree of the collection positions during the optimization of the collection position;

[0026] Collecting UAV images according to the optimal collection position distribution and the collection range, and obtaining the completed remote sensing image by combining the UAV images with the remote sensing image.

[0027] Optionally, configuring the collection range of the UAV comprises:

[0028] Obtaining a maximum collection range of UAV image collection, and multiplying the average obviousness by the maximum collection range of the UAV image collection to obtain the collection range.

[0029] Optionally, generating the disaster emergency surveying and mapping map comprises:

[0030] Extracting abnormal information and damage information from the completed remote sensing image to determine spatial distribution features of a fracture zone and road facility damage.

[0031] processing remote sensing images before and after disasters to obtain surface deformation characteristics;

[0032] According to the spatial distribution characteristics and the surface deformation characteristics, an earthquake disaster emergency mapping is performed to generate the disaster emergency mapping map.

[0033] Optionally, the surface deformation characteristics include satellite line-of-sight surface deformation characteristics and target distances of offset and deformation.

[0034] Optionally, obtaining the satellite line-of-sight surface deformation characteristics includes:

[0035] The remote sensing image before the disaster is selected as the main image, the remote sensing image after the disaster is selected as the auxiliary image, and the elevation data corresponding to the geological disaster high-risk area is selected, the elevation data is used to assist image registration, the influence of spectral offset and Doppler centroid difference is eliminated, and a filtering and coherence map is generated:

[0036] The filtering and coherence map is optimized and phase unwrapping, and a target number of control points are selected in a region far from the geological disaster high-risk area and without high-frequency residual terrain phase and unwrapping error, for correcting the interference phase and unwrapping phase;

[0037] The unwrapped phase is converted into deformation, the converted phase is geocoded, and the satellite line-of-sight surface deformation characteristics and the offset are obtained.

[0038] Optionally, obtaining the target distance of offset and deformation includes:

[0039] According to the remote sensing image, a preliminary registration offset file is obtained, the least square method is used to process the preliminary registration offset file, and a bilinear registration offset polynomial is determined;

[0040] The bilinear registration offset polynomial is used to estimate the disaster fault offset based on SLC data correlation; a median filter is used to filter noise information in the disaster fault offset, and the target distance is obtained.

[0041] The beneficial effects of the present application are:

[0042] The present application can quickly obtain complete geographic information by collecting remote sensing images of the geological disaster high-risk area and using a UAV to complete the occluded area, and solves the problem of incomplete data coverage in traditional methods.

[0043] The present application can more accurately determine the spatial distribution characteristics of the disaster rupture zone and road facility damage by identifying the occluded area with the help of a convolutional neural network and extracting abnormal information and damage information from the completed remote sensing image, and provides high-precision data support for disaster emergency mapping.

[0044] The present application comprehensively considers various data sources such as original remote sensing images, geological exploration data, forest and grassland distribution data, personnel settlement distribution data, disaster history data, railway network data and highway network data, and comprehensively evaluates disaster risks through correction processing and risk level division, so as to provide a more scientific basis for disaster prevention and emergency rescue.

[0045] The present application configures the collection range of the unmanned aerial vehicle according to the average apparentness of the remote sensing image, and optimizes the collection position, so as to avoid repeated collection and improve the efficiency and quality of unmanned aerial vehicle image collection.

[0046] The present application obtains the surface deformation characteristics including the satellite line-of-sight surface deformation characteristics and the target distance of the displacement and deformation through processing of remote sensing images before and after disasters, can more intuitively reflect the influence of disasters on the ground, provides more abundant information for earthquake disaster emergency mapping, and improves the scientificity and practicality of disaster emergency mapping. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description only represent some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor under the premise of the drawings.

[0048] Figure 1 A flow chart of a disaster emergency mapping method based on remote sensing data according to an embodiment of the present application. DETAILED DESCRIPTION

[0049] The technical solutions in the embodiments of the present application will be described clearly and completely with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments only represent some embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0050] In order to make the above-mentioned purposes, features and advantages of the present application more apparent and easy to understand, the present application will be further described in detail with reference to the drawings and specific embodiments.

[0051] As Figure 1As shown, the embodiment discloses a disaster emergency mapping method based on remote sensing data, comprising: collecting remote sensing images of a high-risk geological disaster area, identifying the occluded area in the remote sensing image, completing the occluded area, and obtaining the completed remote sensing image; extracting abnormal information and damage information from the completed remote sensing image, and generating a disaster emergency mapping map according to the abnormal information and damage information.

[0052] Specifically, the embodiment discloses a disaster emergency mapping method based on remote sensing data, comprising: in the monitoring work of a high-risk geological disaster area, the application first collects image data of the area by using remote sensing technology. These remote sensing images can provide the application with a wide range of surface information, helping the application quickly understand the topography, land use situation and potential geological disaster hazards in the area. However, due to the complexity of the natural environment, there may be some parts of the remote sensing image that are occluded by vegetation, clouds, buildings, etc. These occluded areas will interfere with the accurate identification and analysis of geological disaster hazards by the application.

[0053] In order to overcome this problem, the application uses advanced image processing technology to identify the occluded area in the remote sensing image. By analyzing the texture, color, brightness and other characteristics of the image, it accurately locates which parts are occluded. After identifying the occluded area, the application further uses an image completion algorithm to complete these areas. The completion algorithm will reasonably infer the possible topography and landform features of the occluded part based on the image information of the surrounding non-occluded areas, and fill it in completely, thereby generating a complete and continuous remote sensing image. This process not only needs to consider the continuity of the terrain, but also needs to ensure that the completed image is consistent with the surrounding area in terms of visual and geographical features, so that the subsequent analysis work can proceed smoothly.

[0054] After obtaining the completed remote sensing image, the application can extract abnormal information and damage information from it. Abnormal information may include surface cracks, ground subsidence, landslide signs, etc. These information often is a precursor to the occurrence of geological disasters. Damage information may involve road damage, building tilt, vegetation damage, etc. These information can help the application understand the actual impact of geological disasters on human society and the natural environment. Through comprehensive analysis of these abnormal and damage information, the application can more comprehensively grasp the distribution range, harm degree and possible development trend of geological disasters.

[0055] Based on the extracted abnormal information and damage information, the present application further carries out earthquake disaster emergency mapping work. Emergency mapping is a key link in the disaster response process, which can provide intuitive and accurate disaster information for rescue personnel, government departments and disaster-stricken people. The information extracted by the present application is converted into symbols, colors and notes on the map, and a disaster emergency mapping map is generated. This map clearly marks the high-risk areas of geological disasters, the disaster points that have occurred, the damaged infrastructure and the possible rescue channels and other important information. It can not only help rescue personnel quickly understand the situation at the disaster site and develop a reasonable rescue plan, but also provide a scientific basis for government departments' emergency decision-making to timely allocate resources and organize rescue forces, so as to minimize the loss caused by geological disasters. At the same time, the disaster emergency mapping map can also enable disaster-stricken people to understand their environment and cooperate with rescue work to improve their self-rescue and mutual rescue capabilities.

[0056] Further, obtaining the high-risk area of geological disasters comprises: obtaining multi-source data of a to-be-detected area; the multi-source data comprises: original remote sensing images, geological exploration data, forest and grassland distribution data, personnel settlement distribution data, disaster history data, railway network data, and highway network data; performing correction processing on the multi-source data, extracting key geological features and geological disaster hidden danger features from the corrected multi-source data, combining the disaster history data, and dividing buildings and surrounding environments into different fire risk levels; according to the different fire risk levels, obtaining an area corresponding to a fire risk level exceeding a target threshold as the high-risk area of geological disasters.

[0057] Specifically, when carrying out the identification and monitoring of the high-risk area of geological disasters, first, multi-source data of the to-be-detected area need to be obtained. These data are widely sourced and rich in variety, which can provide comprehensive support for accurate identification of geological disasters. Specifically, the multi-source data includes the following aspects:

[0058] Original remote sensing images: high-resolution images obtained through satellite remote sensing technology, which can intuitively reflect the information of landforms, terrain, vegetation coverage, and land use status. These images can provide basic data for macro-identification of geological disasters, helping the present application to quickly understand the overall situation in the region.

[0059] Geological exploration data: geological exploration data is obtained by field exploration and analysis of geological structures, rock layer distribution, fault location, etc. These data can help the present application to deeply understand the geological structure in the region and identify potential geological hazards that may trigger geological disasters, such as active faults and weak rock layer distribution.

[0060] Forest and grassland distribution data: The distribution of forests and grasslands is of great significance for the identification and assessment of geological disasters. On the one hand, the root systems of forests and grasslands can enhance soil stability and reduce the probability of landslides and other geological disasters. On the other hand, large areas of forests and grasslands can also become potential risk areas for secondary disasters such as fires. Therefore, obtaining forest and grassland distribution data can help the invention better assess disaster risks in the region.

[0061] Personnel settlement distribution data: Personnel settlements are the areas most directly and severely affected by geological disasters. Understanding the distribution of personnel settlements, including the location, size, and density of towns, villages, and residential areas, can help the invention determine the scope and extent of the impact of geological disasters on human society, thereby providing important basis for disaster warning and emergency rescue.

[0062] Disaster history data: Disaster history data is an important reference for assessing geological disaster risks. By collecting past geological disaster events in the region, including disaster type, occurrence time, impact range, and disaster severity, the invention can analyze the spatiotemporal distribution patterns and triggering factors of geological disasters, thereby providing experience and reference for current and future disaster risk assessment.

[0063] Railway network data: As an important transportation infrastructure, the safe operation of railways is crucial for regional economic and social development. Obtaining railway network data can help the invention understand the distribution of railway lines in high-risk geological disaster areas, assess the potential threats of geological disasters to railway transportation, and develop preventive measures and emergency rescue plans in advance.

[0064] Highway network data: Similar to railways, highways are also important transportation arteries. Highway network data can help the invention determine the orientation and location of highways in high-risk geological disaster areas, analyze the potential impact of geological disasters on highways, such as traffic disruptions and road damage caused by landslides, debris flows, and other disasters, so as to take timely measures to ensure smooth and safe transportation.

[0065] After obtaining the above-mentioned multi-source data, it is necessary to correct these data. Due to the differences in acquisition methods, time, and accuracy of different data sources, it is necessary to use data fusion, coordinate conversion, error correction, and other technical means to unify multi-source data to the same coordinate system and accuracy level, ensuring the compatibility and consistency of data. Corrected multi-source data can provide more accurate and reliable geological disaster risk assessment basis for the invention.

[0066] Next, key geological features and geological disaster hidden danger features are extracted from the corrected multi-source data. Key geological features include terrain slope, rock layer inclination, fault location, soil type, etc., which are the basic conditions for geological disasters to occur. While geological disaster hidden danger features include surface cracks, ground subsidence, landslide signs, debris flow valleys, etc., which are direct manifestations of impending or ongoing geological disasters. By extracting and analyzing these features, combined with disaster history data, potential geological disaster hidden danger points can be further identified.

[0067] Based on the identification of geological disaster hidden danger points, buildings and their surrounding environments are divided into different fire risk levels. The division of fire risk levels needs to consider factors such as building type, building material, building density, surrounding vegetation coverage, and distance from flammable materials. For example, a wooden structure building located at the edge of a forest may have a higher fire risk level; while a building located in the city center, surrounded by fire facilities, and made of reinforced concrete may have a relatively lower fire risk level. Through this division, the fire risk levels of different areas in the region can be clearly determined.

[0068] Finally, according to the divided fire risk levels, the areas corresponding to the fire risk levels that exceed the target threshold are determined as geological disaster high-risk areas. The target threshold is set comprehensively according to the geological conditions, population density, economic development level, and disaster response capacity of the region. Areas exceeding the target threshold mean that their fire risk is high and may cause more serious secondary disasters such as fire spread, casualties, etc. when a geological disaster occurs. Therefore, by determining these areas as geological disaster high-risk areas, clear focus areas can be provided for subsequent disaster monitoring, early warning, and emergency rescue work, improving the efficiency and scientific nature of disaster response.

[0069] Further, the method for identifying the occlusion area in the remote sensing image comprises: inputting the remote sensing image into a pre-constructed occlusion area identification channel to identify the occlusion area in the remote sensing image, wherein the occlusion area identification channel is pre-trained based on a convolutional neural network, and the training data comprises sample remote sensing images and sample occlusion areas.

[0070] Specifically, in the process of processing remote sensing images in geological disaster high-risk areas, identifying occlusion areas is a key step. Occlusion areas can be caused by various factors such as cloud cover, vegetation occlusion, building shadows, etc., which can interfere with accurate identification and analysis of geological disaster hidden dangers. In order to efficiently and accurately identify the occlusion area in the remote sensing image, the present invention adopts an advanced method based on convolutional neural network (CNN) and constructs a special occlusion area identification channel.

[0071] Construction of the occlusion region identification channel: The occlusion region identification channel is constructed based on a pre-trained convolutional neural network (CNN) in deep learning. CNN is a powerful image processing tool that can automatically learn features in images, enabling the recognition of complex image patterns. When constructing the occlusion region identification channel, the invention first needs to prepare a large amount of training data. These training data include sample remote sensing images and corresponding sample occlusion region annotation information.

[0072] Sample remote sensing images: These sample images cover actual remote sensing data of various geological disaster high-risk areas, containing different terrains, landforms, vegetation types, and possible occlusion conditions. The diversity of sample images is crucial to ensure the model's wide applicability.

[0073] Sample occlusion region annotation: For each sample remote sensing image, the invention needs to manually annotate the occlusion region. These annotation information as "ground truth" are used to train the CNN model, so that it can learn the features and patterns of the occlusion region.

[0074] During the training of the occlusion region identification channel, the CNN model learns the features of the occlusion region by learning a large number of sample remote sensing images and their corresponding occlusion region annotation information. The training process of the CNN model can be divided into the following steps:

[0075] Feature extraction: CNN extracts features from input remote sensing images through convolutional layers and pooling layers. Convolutional layers can capture local features in images, such as edges, textures, etc.; pooling layers are used to reduce the spatial dimension of features while preserving important information.

[0076] Feature fusion and classification: After multiple convolution and pooling operations, the CNN model fuses the extracted features and classifies them through fully connected layers. Finally, the model can output the probability of each pixel in the image belonging to the occlusion region.

[0077] Loss function and optimization: During the training process, there is a difference between the model's output and the "ground truth" manually annotated. By defining a loss function (such as cross-entropy loss function), the model can quantify this difference and adjust the network parameters through the backpropagation algorithm to minimize the value of the loss function. After multiple iterations of training, the model is gradually optimized and can accurately identify the occlusion region.

[0078] Identification process: In practical applications, the remote sensing image to be processed is input into the pre-trained occlusion region identification channel. The CNN model automatically identifies the occlusion regions in the image and outputs a mask of the occlusion regions. This mask is a two-dimensional array with the same size as the input image, where each pixel value represents whether the pixel belongs to the occlusion region. For example, a pixel value of 1 indicates that it belongs to the occlusion region, and a pixel value of 0 indicates that it does not belong to the occlusion region.

[0079] Application of occlusion region identification: After identifying the occlusion regions, the present application can further analyze and process these regions. For example, for regions obscured by clouds, interpolation can be performed using time series data or the surrounding information can be used to complete the region; for regions obscured by vegetation, vegetation distribution data can be combined for analysis to evaluate its potential impact on geological disaster identification. In this way, the present application can effectively reduce the interference of occlusion regions on geological disaster monitoring and analysis, and improve the accuracy and reliability of disaster identification.

[0080] Further, obtaining the completed remote sensing image includes: analyzing the remote sensing image to obtain the average distinctness of the remote sensing image, identifying the occlusion region in the remote sensing image, configuring the collection range of the unmanned aerial vehicle according to the average distinctness, collecting the occlusion region by the unmanned aerial vehicle; optimizing the collection position of the unmanned aerial vehicle image collection in the occlusion region to obtain the optimal collection position distribution, and merging the collection positions according to the repetition degree of the collection positions during the optimization of the collection positions; collecting the unmanned aerial vehicle image according to the optimal collection position distribution and the collection range, and combining the unmanned aerial vehicle image with the remote sensing image to obtain the completed remote sensing image.

[0081] Further, configuring the collection range of the unmanned aerial vehicle includes: obtaining the maximum collection range of the unmanned aerial vehicle for image collection; multiplying the average distinctness by the maximum collection range of the unmanned aerial vehicle image collection to obtain the collection range.

[0082] Specifically, remote sensing image analysis and occlusion region identification:

[0083] The collected remote sensing image is analyzed to calculate its average distinctness. The average distinctness is obtained by analyzing the brightness, contrast, texture and other visual features of the image, and can reflect the overall clarity and information richness of the image. Specifically, the average distinctness can be quantified by calculating the standard deviation or variance of the gray value of each pixel in the image. A lower average distinctness usually means that there are more occlusions or blurred regions in the image, while a higher average distinctness indicates that the image is clearer.

[0084] Next, advanced image processing algorithms are used to identify the occluded regions in the remote sensing images. These algorithms can precisely locate the occluded parts by analyzing the texture, color, brightness, and other features of the images. For example, cloud-occluded regions usually have lower brightness and higher contrast, while vegetation-occluded regions may exhibit specific texture patterns. By analyzing these features, the occluded regions can be distinguished from the non-occluded regions, and a mask of the occluded regions can be generated.

[0085] Configuring the collection range of the UAV:

[0086] According to the average visibility of the remote sensing images, the collection range of the UAV is configured. This process includes the following steps:

[0087] (1) Obtain the maximum collection range of the UAV: The maximum collection range of the UAV refers to the largest geographical area that the UAV can cover in a single flight mission. This range is limited by factors such as flight time, battery endurance, flight altitude, and sensor resolution. For example, a high-performance UAV may be able to cover several square kilometers in a single flight, while a low-performance UAV may only be able to cover a smaller area.

[0088] (2) Calculate the collection range: To ensure that the UAV's collection can effectively supplement the occluded regions in the remote sensing images, the collection range of the UAV needs to be adjusted according to the average visibility of the remote sensing images. The specific method is to multiply the average visibility by the maximum collection range of the UAV to obtain the final collection range. Areas with lower average visibility mean more occlusions, and more detailed collection is needed, so the collection range can be appropriately expanded; areas with higher average visibility can appropriately reduce the collection range to improve collection efficiency.

[0089] where the formula is: Collection Range = Average Visibility × Maximum Collection Range of UAV.

[0090] Optimization of UAV collection position: In the occluded regions, the collection position of the UAV image collection is optimized. The goal of optimization is to obtain the optimal distribution of collection positions to ensure that the UAV's collected images can maximize the supplement of the occluded regions in the remote sensing images, while avoiding unnecessary repeated collection.

[0091] (1) Analyze the repetition of collection positions: In the optimization process, first analyze the repetition of collection positions. If there is a large overlap between multiple collection positions, it will lead to low collection efficiency and may introduce redundant information. By calculating the spatial distance between collection positions and the coverage range of the occluded region, the collection positions can be reasonably adjusted to remove repeated or redundant collection points.

[0092] (2) Merging collection positions: According to the collection range, the repeated collection positions are merged. For example, if the distance between two collection points is less than the effective coverage range of the UAV sensor, these two collection points can be merged into one collection point. In this way, the number of unnecessary collections can be reduced, and the collection efficiency can be improved.

[0093] UAV image collection and remote sensing image completion: According to the optimized collection position distribution and the configured collection range, the UAV is used for image collection. The images collected by the UAV have high resolution and flexible viewing angles, which can provide more detailed ground information, especially in the occluded area. After the collection is completed, the UAV images are fused with the remote sensing images to obtain the completed remote sensing images.

[0094] (1) Image fusion: During the fusion process, the UAV images need to be geometrically corrected and color adjusted to ensure their spatial and visual consistency with the remote sensing images. Through image stitching, fusion algorithms and other technical means, the effective information in the UAV images is supplemented to the occluded area of the remote sensing images, thereby generating a complete and continuous remote sensing image.

[0095] (2) Completed remote sensing image: The completed remote sensing image not only clearly reflects the actual situation of the ground, but also provides more accurate data support for the identification, analysis and early warning of geological disasters. This image can be used in various application scenarios, such as identification of geological disaster hidden points, disaster risk assessment, emergency rescue path planning, etc.

[0096] Through the above method, the invention can effectively solve the problem of occluded areas in remote sensing images and improve the accuracy and reliability of geological disaster monitoring. The completed remote sensing image can be used in various application scenarios, such as identification of geological disaster hidden points, disaster risk assessment, emergency rescue path planning, etc. In addition, combined with the flexibility and high-resolution collection capability of the UAV, this method can also provide real-time data support for dynamic monitoring of geological disasters, helping relevant departments to timely grasp the development trend of disasters and formulate scientific and reasonable response measures.

[0097] In summary, this comprehensive monitoring method combining remote sensing images and UAV images optimizes the collection strategy of UAVs, realizes effective supplementary collection of occluded areas, and provides an efficient and accurate solution for monitoring and early warning of high-risk areas of geological disasters.

[0098] Further, the generation of the disaster emergency mapping map includes: extracting abnormal information and damage information from the completed remote sensing image, determining the spatial distribution characteristics of the rupture zone and the damage of road facilities; processing the remote sensing images before and after the disaster to obtain the surface deformation characteristics; the surface deformation characteristics include: satellite line-of-sight surface deformation characteristics and target distance of displacement and deformation; according to the spatial distribution characteristics and the surface deformation characteristics, the earthquake disaster emergency mapping is carried out, and the disaster emergency mapping map is generated.

[0099] Further, obtaining satellite line-of-sight surface deformation characteristics includes: selecting remote sensing images before the disaster as main images, remote sensing images after the disaster as auxiliary images, and selecting elevation data corresponding to the high-risk area of geological disasters, using the elevation data to assist image registration, eliminating the influence of spectral shift and Doppler centroid difference, and generating filter and coherence map; optimizing and phase unwrapping the filter and coherence map, and selecting target number, distribution uniformity control points in the area far from the high-risk area of geological disasters, and without high-frequency residual terrain phase and unwrapping error, for correcting interference phase and unwrapping phase; convert the unwrapped phase into deformation, and geocode the converted phase to obtain satellite line-of-sight surface deformation characteristics and displacement.

[0100] Further, obtaining the target distance of displacement and deformation includes: obtaining a preliminary registration displacement file according to the remote sensing image, processing the preliminary registration displacement file using the least squares method to determine a bilinear registration displacement polynomial; using the bilinear registration displacement polynomial to estimate the disaster fault displacement based on the SLC data correlation; using a median filter to filter noise information in the disaster fault displacement to obtain the target distance.

[0101] Specifically, extracting abnormal information and damage information: extracting abnormal information and damage information from the completed remote sensing image is the basis of disaster emergency mapping. These information mainly includes the following aspects:

[0102] (1) Identification of rupture zone: by analyzing the linear features, surface cracks and terrain changes in the remote sensing image, the spatial distribution of the earthquake rupture zone is determined. The rupture zone is the direct evidence of the release of seismic energy, and its trend, length and width are crucial for evaluating the scale and influence range of the earthquake.

[0103] (2) Spatial distribution characteristics of road facility damage: using the high-resolution characteristics of remote sensing images, the damage of roads, bridges and other transportation facilities is identified. This includes road breakage, deformation, collapse and bridge damage, etc. By extracting this information, it can provide basis for the passage of rescue vehicles and traffic diversion.

[0104] Obtaining ground deformation features: Ground deformation features are an important manifestation of earthquake disasters, which can reflect the direct impact of earthquakes on the ground. Obtaining ground deformation features mainly includes the following steps:

[0105] (1) Satellite line-of-sight ground deformation features: Image registration and preprocessing: Select the remote sensing image before the disaster as the main image, and the remote sensing image after the disaster as the auxiliary image. At the same time, obtain the elevation data (such as digital elevation model, DEM) corresponding to the high-risk area of geological disasters. Use elevation data to assist image registration, eliminate the influence of spectral shift and Doppler centroid difference, and generate filter and coherence map.

[0106] Optimization and phase unwrapping of filter and coherence map: Optimize the filter and coherence map to improve image quality. Then perform phase unwrapping to convert the interference phase into continuous deformation phase. In the area far from the high-risk area of geological disasters and without high-frequency residual terrain phase and unwrapping error, select a target number of uniformly distributed control points to correct the interference phase and unwrapping phase.

[0107] Extraction of deformation features: Convert the unwrapped phase into deformation information, and associate the deformation information with geographic coordinates through geographic coding, to finally obtain the satellite line-of-sight ground deformation features and displacement.

[0108] (2) Target distance of displacement and deformation:

[0109] Obtaining of preliminary registration displacement file: Obtain the preliminary registration displacement file based on remote sensing images, which records the pixel displacement information between the images before and after the disaster.

[0110] Determination of bilinear registration displacement polynomial: Use the least squares method to process the preliminary registration displacement file to fit the bilinear registration displacement polynomial. This polynomial can more accurately describe the geometric deformation between images.

[0111] Estimation of disaster fault displacement: Based on synthetic aperture radar (SAR) single look complex (SLC) data, use the bilinear registration displacement polynomial and cross-correlation to estimate the displacement of the disaster fault. This process can quantify the horizontal and vertical displacement of the earthquake fault.

[0112] Obtaining of target distance: Use a median filter to filter noise information in the disaster fault displacement to obtain a more accurate target distance. The target distance refers to the relative displacement distance on both sides of the earthquake fault, which is an important parameter for evaluating the impact of earthquakes.

[0113] Earthquake disaster emergency mapping: Based on the extracted spatial distribution features and ground deformation features, perform earthquake disaster emergency mapping. The specific steps are as follows:

[0114] Data integration and analysis: Integrate the extracted spatial distribution characteristics of the rupture zone, road facility damage, and ground deformation. Analyze the correlation between these data, such as the relationship between the rupture zone and ground deformation, the relationship between road damage and seismic intensity, etc.

[0115] Map symbol and annotation design: Design map symbols and annotations according to the type and importance of disaster information. For example, use red lines to represent the rupture zone, use different colored symbols to represent different levels of road damage, and use terrain contour lines to represent the size of ground deformation, etc.

[0116] Map generation and output: Draw the integrated data on the map according to the designed symbols and annotations to generate the disaster emergency mapping map. The map should contain basic elements such as scale, compass, legend, etc. so that users can quickly understand and use it.

[0117] The generated disaster emergency mapping map can provide important geographic information support for earthquake disaster emergency response. It can help rescue personnel quickly understand the topography, road conditions, and ground deformation of the disaster area, so as to develop more effective rescue plans. At the same time, this map also provides a scientific basis for government departments' disaster assessment and post-disaster reconstruction, which helps to reasonably allocate resources and speed up the recovery process.

[0118] In summary, through the above detailed steps, the present application can generate high-quality disaster emergency mapping maps to provide strong support for earthquake disaster emergency response and post-disaster recovery.

[0119] The above-described embodiments are only descriptions of the preferred modes of the present application and do not limit the scope of the present application. Without departing from the design spirit of the present application, various modifications and improvements to the technical solutions of the present application made by those skilled in the art shall fall within the protection scope determined by the claims of the present application.

Claims

1. A disaster emergency mapping method based on remote sensing data, characterized in that, The method comprises the following steps: Collecting remote sensing images of a high-risk area of geological disasters, identifying an occluded area in the remote sensing images, completing the occluded area, and obtaining a completed remote sensing image; Extracting abnormal information and damage information from the completed remote sensing image, and generating a disaster emergency mapping map according to the abnormal information and the damage information.

2. The remote sensing data-based disaster emergency mapping method according to claim 1, characterized in that, The high-risk area of geological disasters is obtained by: Obtaining multi-source data of the area to be detected; the multi-source data includes: original remote sensing images, geological exploration data, forest and grassland distribution data, personnel settlement distribution data, disaster history data, railway network data and highway network data; Correcting the multi-source data, extracting key geological features and geological disaster hidden danger features from the corrected multi-source data, and combining the disaster history data to divide buildings and their surrounding environment into different fire risk levels; According to the different fire risk levels, the area corresponding to the fire risk level exceeding the target threshold is obtained as the high-risk area of geological disasters. 3.The disaster emergency mapping method based on remote sensing data according to claim 1, characterized in that, Identifying the occluded area in the remote sensing images comprises: Inputting the remote sensing image into a pre-constructed occluded area identification channel to identify the occluded area in the remote sensing image, wherein the occluded area identification channel is pre-trained based on a convolutional neural network, and the training data includes sample remote sensing images and sample occluded areas. 4.The disaster emergency mapping method based on remote sensing data according to claim 1, characterized in that, Obtaining the completed remote sensing image comprises: Analyzing the remote sensing image to obtain the average apparentness of the remote sensing image, identifying the occluded area in the remote sensing image, configuring the collection range of the unmanned aerial vehicle according to the average apparentness, and collecting the occluded area by using the unmanned aerial vehicle; Optimizing the collection position of the unmanned aerial vehicle image collection in the occluded area to obtain the optimal collection position distribution, and merging the collection positions according to the repetition of the collection positions during the optimization of the collection positions; According to the optimal collection position distribution and the collection range, the unmanned aerial vehicle image is collected, and the unmanned aerial vehicle image is combined with the remote sensing image to obtain the completed remote sensing image.

5. The remote sensing data-based disaster emergency mapping method according to claim 4, wherein, Configuring the collection range of the unmanned aerial vehicle comprises: Obtaining the maximum collection range of the unmanned aerial vehicle for image collection; multiplying the average apparentness by the maximum collection range of the unmanned aerial vehicle image collection to obtain the collection range.

6. The remote sensing data-based disaster emergency mapping method according to claim 4, wherein, Generating the disaster emergency mapping map comprises: Extracting abnormal information and damage information from the completed remote sensing image, determining the spatial distribution characteristics of the rupture zone and the damage of road facilities; Processing the remote sensing images before and after the disaster to obtain the surface deformation characteristics; According to the spatial distribution characteristics and the surface deformation characteristics, the earthquake disaster emergency mapping is performed to generate the disaster emergency mapping map.

7. The remote sensing data-based disaster emergency mapping method according to claim 6, wherein, The surface deformation characteristics include: satellite line-of-sight surface deformation characteristics and target distance offset and deformation.

8. The remote sensing data-based disaster emergency mapping method according to claim 7, characterized in that, Obtaining the satellite line-of-sight surface deformation characteristics comprises: The remote sensing image before the disaster is selected as a main image, a remote sensing image after the disaster is selected as an auxiliary image, and elevation data corresponding to the high-risk geological disaster area is selected, the elevation data is used for auxiliary image registration, the influence of spectrum deviation and Doppler centroid difference is eliminated, and a filtering and coherence map is generated: The filtering and coherence map is optimized and phase unwrapping is performed, and a target number of control points that are uniformly distributed in a region far from the high-risk geological disaster area and free of high-frequency residual terrain phase and unwrapping error are selected to correct the interference phase and unwrapping phase; The unwrapped phase is converted into deformation, the converted phase is geocoded, and satellite line-of-sight surface deformation characteristics and offsets are obtained. 9.The disaster emergency mapping method based on remote sensing data according to claim 7, characterized in that, The target distance of the offset and deformation includes: According to the remote sensing image, a preliminary registration offset file is obtained, a least square method is used to process the preliminary registration offset file, and a bilinear registration offset polynomial is determined; The bilinear registration offset polynomial is used to estimate the disaster fault offset based on SLC data correlation, a median filter is used to filter noise information in the disaster fault offset, and the target distance is obtained.

Images