Visual interpretation auxiliary method for snow avalanche disaster based on multi-source remote sensing data fusion

Abstract

The application discloses an avalanche disaster visual interpretation auxiliary method based on multi-source remote sensing data fusion, obtains multi-temporal optical remote sensing images, a digital elevation model (DEM) and unmanned aerial vehicle high-resolution aerial images before and after the occurrence of an avalanche, carries out coordinate unification and registration pretreatment, then carries out normalized snow index, normalized vegetation index spectral feature information and slope, slope direction, profile curvature and plane curvature topographic feature information extraction, constructs an enhanced visualization layer containing a basic image layer, a comprehensive interpretation layer and a special subject checking layer through multi-source data fusion, integrates and displays spectral change features, terrain constraint conditions and high-resolution verification images, controls layer combination through interactive operation, carries out avalanche boundary sketching and attribute extraction, and carries out fine verification by using the unmanned aerial vehicle images. The application enhances the comprehensiveness and terrain correlation of interpretation basis, improves avalanche recognition precision and work efficiency, and is suitable for rapid investigation and evaluation of high-altitude mountainous area avalanche disasters.

Description

Technical Field

[0001] This invention relates to the fields of remote sensing applications and geological disaster monitoring technology, specifically to a visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion. Background Technology

[0002] Avalanches are a common and difficult-to-prevent natural disaster in high-altitude, complex mountainous areas, characterized by their suddenness, destructive power, tendency to trigger secondary disasters, and difficulty in rescue operations. Accurately identifying avalanche-prone areas and their threat range is crucial for avalanche disaster assessment, emergency response, and risk warning.

[0003] Remote sensing technology, due to its wide coverage, high identification efficiency, objective and accurate information, strong retrospective comparability, and ability to achieve long-term dynamic monitoring, has become an important means of avalanche risk assessment, dynamic monitoring, and early warning. Currently, avalanche remote sensing identification mainly relies on professionals to visually interpret high-resolution optical remote sensing images, making comprehensive judgments by identifying visual cues such as differences in snow color, surface texture features, and terrain morphology. This method has the following limitations:

[0004] 1) Limited information dimension: Relying solely on optical images makes it difficult to effectively distinguish between new and old avalanche traces. Cloud shadows and snow accumulation, as well as vegetation and avalanche accumulation areas, are easily confused, leading to misjudgment or missed detection.

[0005] 2) Lack of terrain information: Avalanches are closely related to terrain, but optical images lack three-dimensional information. Interpretation often requires manual comparison using digital elevation models (DEMs), which is inefficient and the spatial correspondence is not intuitive.

[0006] 3) Difficulty in detecting changes: Manually comparing multiple images to identify newly added avalanches is inefficient, highly subjective, and significantly affected by seasonal snow cover changes, resulting in insufficient stability.

[0007] 4) Insufficient verification methods: For small-scale or complex terrain areas, there is a lack of high-resolution data support, and existing technologies are unable to achieve effective fusion and collaborative verification of satellite and UAV imagery.

[0008] Although existing research has attempted to achieve automatic avalanche identification using interferometric synthetic aperture radar (InSAR), lidar (LiDAR), or deep learning algorithms, these methods still suffer from high costs, technical complexity, and limited reliability, making it difficult to meet the demands for high-precision, low-cost applications. Therefore, in disaster assessment and scientific research, human visual interpretation still plays an irreplaceable and crucial role. Currently, there is an urgent need to develop an intelligent assisted interpretation method that can integrate multi-source remote sensing data, enhance visual interpretation capabilities, and effectively improve interpretation efficiency and accuracy. Summary of the Invention

[0009] This invention provides a visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion, in order to solve the problems of single data source, weak terrain correlation, low change detection efficiency and verification difficulties in the current avalanche visual interpretation process.

[0010] According to the first aspect, one embodiment provides a visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion, the method comprising:

[0011] The system acquires multi-source information data of the target avalanche area, including optical remote sensing image data, digital elevation model data, and unmanned aerial photography data, and performs preprocessing.

[0012] Spectral and terrain features are extracted based on the preprocessed multi-source information data.

[0013] Based on the feature extraction results, a multi-layered visualization fusion layer is constructed by fusing multi-source data.

[0014] Based on multi-layered visualization fusion layers, visual interpretation of avalanche disasters is performed through human-computer interaction, and interpretation result files are output.

[0015] Furthermore, multi-source information data of the target avalanche area, including optical remote sensing image data, digital elevation model data, and unmanned aerial photography data, is acquired and preprocessed, specifically including:

[0016] Acquire optical remote sensing images of the target avalanche area at different time points before and after the avalanche, and collect high-resolution optical image data; acquire digital elevation model data of the target avalanche area; acquire UAV aerial images of the target avalanche area after the avalanche disaster; and perform coordinate system I, geometric correction and registration processing on all data.

[0017] Furthermore, spectral and terrain features are extracted based on the preprocessed multi-source information data, specifically including:

[0018] Normalized snow index (NDSI) and normalized vegetation index (NDVI) are calculated based on optical remote sensing images to obtain the status and extent of snow-covered and vegetation-covered areas; the NDSI difference ΔNDSI between images before and after an avalanche is calculated and a change detection map is generated to identify areas of cover change caused by the avalanche.

[0019] Based on digital elevation model data, topographic factors including slope, aspect, profile curvature, and planar curvature are calculated, and mountain shadow maps are generated to enhance the three-dimensionality of the terrain.

[0020] Furthermore, based on the feature extraction results, a multi-layered visualization fusion layer is constructed through multi-source data fusion, specifically including:

[0021] A multi-layered visualization fusion layer is constructed, comprising a base image layer, a comprehensive interpretation layer, and a thematic verification layer. The base image layer is obtained by overlaying a mountain shadow map with preset transparency onto a high-resolution optical image as the base map. The comprehensive interpretation layer is obtained by overlaying layer elements on the base image layer, including a ΔNDSI difference map displayed as a pseudo-color heatmap, the optimal avalanche slope range displayed with color fill, curvature feature areas displayed as outlines or semi-transparent overlays, and low-vegetation areas displayed with highlighted annotations. The thematic verification layer is obtained by overlaying the base background layer and the comprehensive interpretation layer with an independent layer of original optical remote sensing imagery from the post-avalanche period for verifying false cloud shadow information and an independent layer of UAV aerial photography for detailed verification of suspected avalanche points.

[0022] Furthermore, a multi-layered visualization fusion layer is constructed, comprising a basic image layer, a comprehensive interpretation layer, and a thematic verification layer, specifically including:

[0023] The optimal avalanche slope range is 25° to 60°; the curvature feature region includes a negative profile curvature region indicating the avalanche trough and a positive plane curvature region indicating the confluence topography; the low vegetation region is determined by areas with NDVI values ​​below a set threshold.

[0024] Furthermore, based on multi-layered visualization fusion layers, visual interpretation of avalanche disasters is performed through human-computer interaction, and interpretation result files are output, specifically including:

[0025] The multi-layered visualization fusion layer is loaded into professional geographic information software. The layer display is controlled through human-computer interaction, suspected avalanche areas are interpreted and marked, and detailed verification is performed using UAV aerial photography. The interpretation results are then output as a file.

[0026] Furthermore, based on multi-layered visualization fusion layers, visual interpretation of avalanche disasters is performed through human-computer interaction, and interpretation result files are output, specifically including:

[0027] Load the generated multi-layered visualization fusion layer into the geographic information software;

[0028] Use the interactive interface to complete the following operations:

[0029] Flexible control over the on / off state and transparency of each layer allows for free combination of different information dimensions for observation;

[0030] Use the mouse to directly circle, outline, and mark the suspected avalanche initiation area, movement path, accumulation area range, and boundaries on the blended layer;

[0031] Based on the calculation and recording function of geographic information software, it automatically records and calculates spatial information including the boundary, area, and center coordinates of each marked area, and associates it with attribute information including the slope, aspect, and elevation of its location.

[0032] Based on the original optical remote sensing images, cloud interference and false information about clouds are removed through visual interpretation.

[0033] Switch to the drone aerial photography layer to conduct detailed verification and validation of key suspected avalanche points;

[0034] The final output includes structured results such as avalanche boundary vector files, attribute tables, and interpretation reports.

[0035] Furthermore, based on multi-layered visualization fusion layers, visual interpretation of avalanche disasters is performed through human-computer interaction, and interpretation result files are output, specifically including:

[0036] The attribute table contains information on the area, perimeter, center point coordinates, average elevation, average slope, and dominant slope aspect of each avalanche polygon region.

[0037] According to the second aspect, one embodiment provides a visual interpretation assistance system for avalanche disasters based on multi-source remote sensing data fusion, the system comprising:

[0038] The data acquisition and processing module is used to acquire multi-source information data of the target avalanche area, including optical remote sensing image data, digital elevation model data, and unmanned aerial photography data, and to perform preprocessing.

[0039] The feature extraction module is used to extract spectral and terrain features based on the preprocessed multi-source information data;

[0040] The multi-source fusion module is used to construct a multi-layered visualization fusion layer based on feature extraction results and multi-source data fusion.

[0041] The visual interpretation module is used to perform visual interpretation of avalanche disasters based on multi-layered visual fusion layers through human-computer interaction, and output interpretation result files.

[0042] According to a third aspect, one embodiment provides an electronic device, the device comprising: a processor and a memory;

[0043] The memory is used to store one or more program instructions;

[0044] The processor is configured to run one or more program instructions to perform the steps of a visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion as described in any of the preceding claims.

[0045] According to a fourth aspect, one embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of a visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion as described in any of the preceding claims.

[0046] This invention provides a visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion, which has the following beneficial effects:

[0047] 1) Achieve deep fusion of multi-source information: Integrate multi-dimensional information such as spectrum, topography, temporal variation and high-resolution imagery to provide a comprehensive interpretation basis and significantly improve the comprehensiveness and reliability of the interpretation basis;

[0048] 2) High-efficiency change detection and localization: The ΔNDSI is used to quickly locate the area of ​​change, and the terrain factors are combined to achieve preliminary screening, which greatly improves the interpretation efficiency;

[0049] 3) Multi-level verification mechanism: Supports multi-scale verification from satellite imagery to UAV imagery, effectively reducing the probability of missed and false positives;

[0050] 4) User-friendly human-computer interaction: It adopts visual layer combinations and interactive operations, which conforms to the working habits of professionals and lowers the threshold for use;

[0051] 5) Output standardization: The interpretation results can be directly generated into a standard geographic information format, which facilitates subsequent integration and application;

[0052] 6) Focus on supporting decision-making: Emphasize the "human-in-the-loop" interpretation model, give full play to expert knowledge, and avoid the uncertainty in fully automated methods.

[0053] 7) This invention is highly efficient, easy to implement, and highly operable. The Landsat 8 / 9 or Sentinel-2 data used for the large-scale avalanche analysis in the early stage are both open-source and free data, which are easy to obtain and are not limited by the topography, weather, climate, or site environment of the survey area. They can be completed in the indoor preparation stage. Attached Figure Description

[0054] Figure 1 A flowchart illustrating a visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion, as provided in one embodiment of the present invention;

[0055] Figure 2 A flowchart illustrating a visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion, as provided in one embodiment of the present invention;

[0056] Figure 3 The present invention provides a logical structure diagram of a visual interpretation auxiliary system for avalanche disasters based on multi-source remote sensing data fusion, as an embodiment of the present invention. Detailed Implementation

[0057] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of the invention. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to the present invention are not shown or described in the specification. This is to avoid obscuring the core parts of the invention with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0058] Furthermore, the features, operations, or characteristics described in the specification can be combined in any suitable manner to form various embodiments. At the same time, the steps or actions in the method description can be rearranged or adjusted in a manner obvious to those skilled in the art. Therefore, the various orders in the specification and drawings are only for the clear description of a particular embodiment and do not imply a necessary order, unless otherwise stated that a particular order must be followed.

[0059] The first embodiment of this invention provides a visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion. The following is a combination of... Figure 1 and Figure 2 A detailed explanation will be provided.

[0060] like Figure 1 As shown, in step S100, multi-source information data of the target avalanche area, including optical remote sensing image data, digital elevation model data, and unmanned aerial photography data, is acquired and preprocessed.

[0061] The above steps specifically include:

[0062] S110 acquires optical remote sensing images of the target area at different time phases before and after the avalanche, such as shortwave infrared channel Landsat 8 / 9 and Sentinel-2 satellite images. Images with minimal rain and fog coverage should be selected as much as possible. These images are mainly used for calculating the Normalized Snow Index (NDSI) and Normalized Difference Vegetation Index (NDVI) of the target area. At the same time, conventional optical images with a spatial resolution better than 1m, such as Jilin-1 and Gaofen-7, are collected for use as the basic images for conventional visual interpretation.

[0063] S120 acquires digital elevation model (DEM) data of the target area, such as 5m grid WorldDEM™ data or 12.5m grid ALOS PALSAR DEM data, for extracting terrain-related factors such as slope and aspect.

[0064] S130: Acquire drone aerial images of avalanche disasters in the target area, with a resolution of 5-15cm, for detailed verification of key avalanche areas.

[0065] S140, perform coordinate system I, geometric correction and registration processing on all the above data to ensure pixel-level spatial alignment.

[0066] In this embodiment, the Galongla Snow Mountain area, where an avalanche occurred in early March 2024, is selected as the target area for data collection and preprocessing. The specific operations are as follows:

[0067] ① Optical satellite imagery: Acquire one Sentinel-2A L2A-level image each before the avalanche (late February 2024) and after the avalanche (mid March 2024) (cloud cover <10% in both cases), with a spatial resolution of 10m. Simultaneously, acquire one panchromatic fusion image of the area from Jilin-1 in April 2024 (resolution 0.75m) as a high-resolution base map.

[0068] ②DEM data: Acquire ALOS PALSAR DEM data of the area with a spatial resolution of 12.5m.

[0069] ③ Aerial photography data from drones: Weather permitting, in late March 2024, a DJI M300 drone equipped with a Zenmuse P1 camera was used to conduct aerial photography of key suspected areas. The terrain-following flight technology was used to obtain orthophotos (DOM) and digital surface models (DSM) with a resolution of approximately 6cm.

[0070] ④ Data Preprocessing: In a professional geographic information software platform, all data are unified to the WGS 84 coordinate system and UTM projection (zone number 46N). Based on the Gaofen-7 image, the ALOS PALSAR DEM and UAV DOM of the Sentinel-2 image are precisely registered to ensure that all data layers achieve pixel-level alignment.

[0071] like Figure 1 As shown, in step S200, spectral and terrain features are extracted based on the multi-source information data obtained through preprocessing.

[0072] The above steps specifically include:

[0073] S210, Spectral Feature Calculation: Based on Landsat 8 / 9 or Sentinel-2 optical imagery, calculate the Normalized Snow Index (NDSI) and Normalized Vegetation Index (NDVI) to obtain the status and extent of snow-covered and vegetation-covered areas.

[0074] S220, Topographic Factor Derivation: Based on DEM data, calculate the slope, aspect, profile curvature and planar curvature of the target area to generate a mountain shadow map and enhance the three-dimensionality of the terrain;

[0075] S230, Temporal Change Detection: Calculate the NDSI difference (ΔNDSI) between images before and after an avalanche, generate a change detection map, and identify areas of coverage change caused by the avalanche (e.g., fresh avalanche areas often show a significant increase in NDSI value).

[0076] The specific steps are as follows:

[0077] Spectral feature calculation: Based on two Sentinel-2 images, NDSI and NDVI were calculated respectively.

[0078] NDSI = (B3 - B12) / (B3 + B12), where B3 is the green light band and B12 is the SWIR shortwave infrared band; calculate the difference between the two periods of NDSI to generate the ΔNDSI raster layer.

[0079] NDVI = (B8 - B4) / (B8 + B4), where B8 is the near-infrared band and B4 is the red band.

[0080] Topographic factor derivation: Based on the registered ALOS PALSAR DEM data, the slope, aspect, profile curvature, and plan curvature are extracted using the "Surface Parameters" tool in professional geographic information software (taking SuperMap iDesktop as an example). Using the "Mountain Shadow" tool, the solar azimuth angle (e.g., 300°) and elevation angle (e.g., 40°) are set to generate a terrain-enhanced mountain shadow map.

[0081] Feature region extraction: Using professional geographic information software (taking SuperMap iDesktop as an example), the "reclassification" and "raster calculator" tools are used to extract the optimal avalanche slope range of 25°~60° from the slope data; negative profile curvature regions (concave slopes, avalanche troughs) and positive plane curvature regions (convex slopes, runoff topography) are extracted from the curvature data; and low vegetation cover regions (NDVI<0.2) are extracted from the post-avalanche NDVI data.

[0082] like Figure 1As shown, in step S300, based on the feature extraction results, a multi-layer visualization fusion layer is constructed by fusing multi-source data.

[0083] The above steps specifically include:

[0084] S310, Basic Image Layer Construction: The latest optical images of Jilin-1 and Gaofen-7 with a resolution better than 1 meter are used as the base map, and a semi-transparent mountain shadow map is superimposed to form a basic background layer with terrain enhancement effect.

[0085] S320, Comprehensive Interpretation Layer Construction: Overlay the following independently controllable thematic layer elements on the base background layer:

[0086] a. The ΔNDSI difference map is presented in the form of a pseudo-color heatmap, highlighting areas of significant change, such as red areas indicating significant changes, i.e., potential new avalanche areas.

[0087] b. The optimal slope range for avalanche development (25°~60°) is shown using color-filled areas;

[0088] c. Negative profile curvature areas (concave slopes, indicating avalanche troughs) and positive plane curvature areas (indicating confluence topography) are displayed with outlines or semi-transparent overlays;

[0089] d. Low NDVI areas (indicating bare ground or avalanche paths) are highlighted, as these areas are often the scraping zones of avalanches;

[0090] S330, Thematic Image Layer Construction: The following independently controllable thematic check layer elements are overlaid on top of the base background layer and the comprehensive interpretation layer:

[0091] a. Overlay the original Landsat 8 / 9 or Sentinel-2 multispectral images (such as RGB true color or false color composite) for visual interpretation to remove avalanche areas misjudged due to cloud cover;

[0092] b. High-resolution aerial image embedding: High-resolution UAV aerial imagery is used as an independent layer and "added" to the corresponding satellite image location after georegistration. Interpreters can access this layer at any time to zoom in and observe details of suspected avalanche points, verifying features such as avalanche cracks, deposition patterns, and avalanche boundaries.

[0093] The specific steps are as follows:

[0094] Constructing multi-layered thematic maps in professional geographic information software (using SuperMap iDesktop as an example):

[0095] Constructing the base image layer: Using the Gaofen-7 0.8m panchromatic fusion image (RGB true color) as the base map, the mountain shadow image is overlaid on it with 30%~40% transparency to form a base background layer with a three-dimensional terrain effect.

[0096] Constructing a comprehensive interpretation layer: On top of the base image layer, the following thematic layers are layered sequentially, with independent display controls set for each layer:

[0097] a. ΔNDSI change layer: The ΔNDSI difference map is rendered in pseudo-color using gradient color bands such as "red-yellow-green". Red indicates a significant increase in NDSI value (i.e., a high probability of new snow accumulation / avalanche), and green indicates a decrease.

[0098] b. Optimal Slope Layer: The slope range of 25° to 60° is displayed with a semi-transparent blue fill.

[0099] c. Topographic curvature layer: The negative profile curvature area is highlighted with a cyan outline, and the positive plane curvature area is highlighted with a crimson outline.

[0100] d. Low vegetation layer: Areas with NDVI < 0.2 are filled with a semi-transparent gray.

[0101] Construct a thematic verification layer: Overlay the following auxiliary verification layers on top:

[0102] a. Multispectral Image Layer: Add true-color composite of Sentinel-2 images from the post-avalanche period (B4, B3, B2) as an independent layer, which is turned off by default and used to quickly check for false information such as cloud shadows during interpretation.

[0103] b. Drone Image Layer: The 6cm resolution orthophoto (DOM) generated by the drone is used as an independent layer and added to the corresponding position after precise registration. It is turned off by default and is used for detail verification.

[0104] like Figure 1 As shown, in step S400, based on a multi-layered visualization fusion layer, visual interpretation of avalanche disaster is performed through human-computer interaction, and the interpretation result file is output.

[0105] The above steps specifically include:

[0106] Interpreters used computers to load the multi-source fusion thematic layers generated by this method into geographic information software platforms such as SuperMap, and used the interactive interface to complete the following operations:

[0107] ① You can flexibly control the on / off state and transparency of each layer, and freely combine different information dimensions for observation.

[0108] ② Use the mouse to directly circle, outline, and mark the suspected avalanche initiation area, movement path, accumulation area range, and boundaries on the fused image.

[0109] ③Based on the calculation and recording function of the geographic information software system, the spatial information such as the boundary, area, and center coordinates of each marked area is automatically recorded and calculated, and associated with the slope, aspect, elevation and other attribute information of its location.

[0110] ④ Based on the original Landsat 8 / 9 or Sentinel-2 optical images, remove "cloud" interference and delete "false" cloud information through visual interpretation.

[0111] ⑤ Switch to the drone image layer to conduct detailed verification and validation of key suspected points.

[0112] ⑥ The output includes structured results such as avalanche boundary vector files, attribute tables, and interpretation reports.

[0113] The specific steps are as follows:

[0114] The interpreters performed the following operations using professional geographic information software (taking SuperMap iDesktop as an example):

[0115] Interactive Interpretation: The base image layer, ΔNDSI layer (40%~50% opacity), and optimal slope layer were enabled. Through visual observation, three suspected avalanche deposition areas were initially identified within the red ΔNDSI area and the optimal slope range. Subsequently, the terrain curvature layer was enabled, revealing that the upstream initiation zones of these areas were all located within negative profile curvature (concave slopes), and the paths matched positive plane curvature (convex slopes, ridges), further confirming the avalanche paths. By adjusting the opacity and combination of each layer, the complete range of the avalanche from the initiation zone to the deposition zone was clearly and three-dimensionally displayed.

[0116] Detailed verification: For one suspected area with blurred boundaries, other layers were turned off, and the drone DOM image of that location was viewed separately and zoomed in. Typical features such as avalanche cracks, snow accumulation, and scraping marks were clearly observed, and it was finally confirmed that this was a medium-sized slope avalanche.

[0117] Cloud removal: For another area with significant ΔNDSI changes, after enabling the Sentinel-2 true color layer, it was found that the area was covered by clouds. The changes in ΔNDSI were caused by the clouds, so it was marked as false information and excluded.

[0118] Results Generation: Using the editing tools of professional geographic information software, polygonal vector features are drawn along the confirmed avalanche boundary. The system automatically records the spatial attributes of each polygon, such as area, perimeter, and center point coordinates. Using tools like "Value Extraction to Point," information such as the average elevation, slope, and aspect of each polygon's location is linked as attribute fields to the vector file. Finally, the interpretation results (avalanche boundary vector file (e.g., Shapefile), attribute table, and screenshots of the interpretation process) are exported and an interpretation description file is generated.

[0119] This embodiment demonstrates that the method of the present invention significantly improves the efficiency and accuracy of avalanche interpretation, increasing the daily workload of interpreters from approximately 10 square kilometers using traditional methods to approximately 50 square kilometers, while significantly reducing the misjudgment rate.

[0120] Corresponding to the aforementioned method for visual interpretation assistance of avalanche disasters based on multi-source remote sensing data fusion, this invention also discloses a system for visual interpretation assistance of avalanche disasters based on multi-source remote sensing data fusion, which specifically includes:

[0121] The data acquisition and processing module is used to acquire multi-source information data of the target avalanche area, including optical remote sensing image data, digital elevation model data, and unmanned aerial photography data, and to perform preprocessing.

[0122] The feature extraction module is used to extract spectral and terrain features based on the preprocessed multi-source information data;

[0123] The multi-source fusion module is used to construct a multi-layered visualization fusion layer based on feature extraction results and multi-source data fusion.

[0124] The visual interpretation module is used to perform visual interpretation of avalanche disasters based on multi-layered visual fusion layers through human-computer interaction, and output interpretation result files.

[0125] It should be noted that for a detailed description of the avalanche disaster visual interpretation assistance system based on multi-source remote sensing data fusion provided in the embodiments of the present invention, please refer to the relevant description of the avalanche disaster visual interpretation assistance method based on multi-source remote sensing data fusion provided in the embodiments of the present invention, which will not be repeated here.

[0126] Those skilled in the art will understand that all or part of the functions of the various methods in the above embodiments can be implemented by hardware or by computer programs. When all or part of the functions in the above embodiments are implemented by computer programs, the program can be stored in a computer-readable storage medium, which may include: read-only memory, random access memory, disk, optical disk, hard disk, etc., and the program is executed by a computer to achieve the above functions. For example, the program can be stored in the memory of a device, and when the program in the memory is executed by the processor, all or part of the above functions can be achieved. In addition, when all or part of the functions in the above embodiments are implemented by computer programs, the program can also be stored in a server, another computer, disk, optical disk, flash drive, or external hard drive, etc., and can be downloaded or copied to the memory of a local device, or the system of the local device can be updated. When the program in the memory is executed by the processor, all or part of the functions in the above embodiments can be achieved.

[0127] The above examples illustrate the present invention only to aid in understanding it and are not intended to limit the scope of the invention. Those skilled in the art can make various simple deductions, modifications, or substitutions based on the principles of this invention.

Claims

1. A visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion, characterized in that, The method includes: Acquire multi-source information data of the target avalanche area, including optical remote sensing image data, digital elevation model data, and UAV aerial photography data, and perform preprocessing. Based on the preprocessed multi-source information data, spectral and topographic features are extracted, specifically including: calculating the Normalized Difference in Snow Index (NDSI) and Normalized Difference in Vegetation Index (NDVI) based on optical remote sensing images to obtain the status and extent of snow-covered and vegetation-covered areas; calculating the NDSI difference ΔNDSI between images before and after an avalanche and generating a change detection map to identify areas of cover change caused by the avalanche; and calculating topographic factors including slope, aspect, profile curvature, and planar curvature based on digital elevation model data and generating a mountain shadow map to enhance the three-dimensionality of the terrain. Based on the feature extraction results, a multi-layered visualization fusion layer is constructed through multi-source data fusion. Specifically, this includes constructing a multi-layered visualization fusion layer comprising a base image layer, a comprehensive interpretation layer, and a thematic verification layer. The base image layer is obtained by overlaying a mountain shadow map with preset transparency onto a high-resolution optical image as the base map. The comprehensive interpretation layer is obtained by overlaying layer elements on the base image layer, including a ΔNDSI difference map displayed in pseudo-color heatmap form, the optimal avalanche slope range displayed in color fill mode, curvature feature regions displayed in outline or semi-transparent overlay form, and low-vegetation areas displayed in highlighted annotation form. The thematic verification layer is obtained by overlaying the base image layer and the comprehensive interpretation layer with an independent layer of original optical remote sensing imagery from the post-avalanche period used to verify false cloud shadow information and an independent layer of UAV aerial photography used for detailed verification of suspected avalanche points. Based on multi-layered visualization fusion layers, visual interpretation of avalanche disasters is performed through human-computer interaction, and interpretation result files are output.

2. The visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion as described in claim 1, characterized in that, The system acquires multi-source information data of the target avalanche area, including optical remote sensing imagery, digital elevation model data, and UAV aerial photography data, and performs preprocessing, specifically including: Acquire optical remote sensing images of the target avalanche area at different time points before and after the avalanche, and collect high-resolution optical image data; acquire digital elevation model data of the target avalanche area; acquire UAV aerial images of the target avalanche area after the avalanche disaster; and perform coordinate system I, geometric correction and registration processing on all data.

3. The visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion as described in claim 2, characterized in that, Constructing a multi-layered visualization fusion layer comprising a base image layer, a comprehensive interpretation layer, and a thematic verification layer, specifically including: The optimal avalanche slope range is 25° to 60°; the curvature feature region includes a negative profile curvature region indicating the avalanche trough and a positive plane curvature region indicating the confluence topography; the low vegetation region is determined by areas with NDVI values ​​below a set threshold.

4. The visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion as described in claim 1, characterized in that, Based on multi-layered visualization fusion, visual interpretation of avalanche disasters is performed through human-computer interaction, and the interpretation results are output, including: The multi-layered visualization fusion layer is loaded into professional geographic information software. The layer display is controlled through human-computer interaction, suspected avalanche areas are interpreted and marked, and detailed verification is performed using UAV aerial photography. The interpretation results are then output as a file.

5. The visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion as described in claim 4, characterized in that, Based on multi-layered visualization fusion, visual interpretation of avalanche disasters is performed through human-computer interaction, and the interpretation results are output, including: Load the generated multi-layered visualization fusion layer into the geographic information software; Use the interactive interface to complete the following operations: Flexible control over the on / off state and transparency of each layer allows for free combination of different information dimensions for observation; Use the mouse to directly circle, outline, and mark the suspected avalanche initiation area, movement path, accumulation area range, and boundaries on the blended layer; Based on the calculation and recording function of geographic information software, it automatically records and calculates spatial information including the boundary, area, and center coordinates of each marked area, and associates it with attribute information including the slope, aspect, and elevation of its location. Based on the original optical remote sensing images, cloud interference and false information about clouds are removed through visual interpretation. Switch to the drone aerial photography layer to conduct detailed verification and validation of key suspected avalanche points; The final output includes structured results such as avalanche boundary vector files, attribute tables, and interpretation reports.

6. The visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion as described in claim 5, characterized in that, Based on multi-layered visualization fusion, visual interpretation of avalanche disasters is performed through human-computer interaction, and the interpretation results are output, including: The attribute table contains information on the area, perimeter, center point coordinates, average elevation, average slope, and dominant slope aspect of each avalanche polygon region.

7. A visual interpretation assistance system for avalanche disasters based on multi-source remote sensing data fusion, characterized in that, The system includes: The data acquisition and processing module is used to acquire multi-source information data of the target avalanche area, including optical remote sensing image data, digital elevation model data, and UAV aerial photography data, and to perform preprocessing. The feature extraction module is used to extract spectral and topographic features based on preprocessed multi-source information data. Specifically, it includes: calculating the Normalized Difference Snow Index (NDSI) and Normalized Difference Vegetation Index (NDVI) based on optical remote sensing images to obtain the status and extent of snow cover and vegetation cover areas; calculating the NDSI difference ΔNDSI between images before and after an avalanche and generating a change detection map to identify areas of cover change caused by the avalanche; and calculating topographic factors including slope, aspect, profile curvature, and planar curvature based on digital elevation model data and generating a mountain shadow map to enhance the three-dimensionality of the terrain. The multi-source fusion module is used to construct a multi-layered visualization fusion layer based on feature extraction results through multi-source data fusion. Specifically, it includes constructing a multi-layered visualization fusion layer comprising a base image layer, a comprehensive interpretation layer, and a thematic verification layer. The base image layer is obtained by overlaying a mountain shadow map with preset transparency onto a high-resolution optical image as the base map. The comprehensive interpretation layer is obtained by overlaying layer elements on the base image layer, including a ΔNDSI difference map displayed in pseudo-color heatmap form, the optimal avalanche slope range displayed in color fill mode, curvature feature regions displayed in outline or semi-transparent overlay form, and low-vegetation areas displayed in highlighted annotation form. The thematic verification layer is obtained by overlaying the base image layer and the comprehensive interpretation layer with an independent layer of original optical remote sensing imagery from the post-avalanche period used to verify false cloud shadow information and an independent layer of UAV aerial photography imagery used for detailed verification of suspected avalanche points. The visual interpretation module is used to perform visual interpretation of avalanche disasters based on multi-layered visual fusion layers through human-computer interaction, and output interpretation result files.

8. An electronic device, characterized in that, The device includes: a processor and a memory; The memory is used to store one or more program instructions; The processor is configured to run one or more program instructions to perform the steps of a visual interpretation aid method for avalanche disasters based on multi-source remote sensing data fusion as described in any one of claims 1 to 6.

Images