Unmanned aerial vehicle emergency service space positioning method and system fusing multi-mode semantic model and address resolution

Abstract

The invention discloses an unmanned aerial vehicle emergency service space positioning method and system fusing a multi-mode semantic model and address resolution. The method comprises the following steps: acquiring disaster related data of a determined territorial range; performing audio extraction on the network short video data, and converting the network short video data into audio text data; audio text data are integrated into short video titles, place name keywords are extracted, POI point names are aligned, word frequencies and event types are counted, and batch collection of address coordinate data and coordinate system conversion are carried out; the satellite map data is identified, land classification is carried out, and town built-up area boundaries are integrated and drawn; performing spatial positioning on point-shaped, linear and planar place names, determining task allocation and paths, and controlling the unmanned aerial vehicle to execute tasks; and feeding an aerial image of the unmanned aerial vehicle back to the cloud platform for mapping, and generating a refined dynamic disaster map. According to the invention, the urban lifeline anti-rescue system is helped to quickly determine the specific position of the victim and delimit the emergency service space and task type of the unmanned aerial vehicle, so that targeted response is realized, a real-time disaster map is provided, and the ability of the urban lifeline system to deal with sudden disasters is enhanced.

Description

Technical Field

[0001] This invention relates to a spatial positioning method and system for UAV emergency services that integrates multimodal semantic models and address resolution, belonging to the field of spatial positioning technology. Background Technology

[0002] In recent years, frequent natural disasters and public accidents have posed a serious threat to people's lives and property, making urban disaster emergency management more complex and challenging. In disaster prevention and search and rescue operations, accurate spatial positioning technology and efficient search and rescue methods are key factors in improving rescue success rates and reducing losses. Traditional emergency response mechanisms rely on passively receiving reports from affected residents, lacking proactive information collection methods. This often leads to problems such as delayed information acquisition, inaccurate disaster area location, and untimely allocation of rescue resources, significantly impacting rescue efficiency and effectiveness. Therefore, spatial positioning and information processing methods based on modern technology, especially intelligent emergency response technology combined with multimodal data, have become important ways to improve disaster response efficiency and ensure rescue effectiveness.

[0003] Meanwhile, as drones are increasingly used in disaster relief, especially in complex disaster areas with poor communication, their flexibility and mobility have enabled them to play a crucial role that is difficult to replace with manual labor, such as detecting disaster information, searching for victims, and delivering emergency supplies. However, current applications of drones in urban disaster prevention also face problems such as inaccurate target positioning and poor targeting of search areas. In particular, the integration of multi-source information fusion and spatial positioning technology has not yet fully realized the potential of drones, often resulting in aimless operations in emergency disaster relief and prevention work that requires rapid response.

[0004] Short online videos provide a channel for the public to spontaneously publish various types of information, and can also serve as important clues to assist in disaster relief decision-making during disasters. Utilizing online platforms, disaster-related information can be proactively collected and extracted from these real-time audiovisual materials, such as the specific location of the disaster, changes in the surrounding environment, and even distress signals from affected individuals, thereby helping to assess the disaster situation. Therefore, integrating multimodal information sources, including real-time online video, audio, satellite imagery, and geographic information, and using intelligent analysis to achieve precise positioning and dynamic monitoring of disaster areas, has become crucial for improving urban disaster emergency response capabilities. Summary of the Invention

[0005] In view of this, the present invention provides a spatial positioning method, system, computer equipment, and storage medium for UAV emergency services that integrates multimodal semantic models and address resolution. It can identify the semantics of place names mentioned in short internet videos, convert the video modal data sequence into audio and text, realize multimodal data flow, and then extract place name keywords and event nature for parsing and positioning. This helps urban lifeline rescue systems quickly determine the specific location of disaster victims, thereby enabling targeted responses, improving task execution efficiency, and generating detailed dynamic disaster maps using UAV imagery to monitor disaster dynamics and enhance the ability of urban lifeline systems to respond to sudden disasters.

[0006] The first objective of this invention is to provide a spatial positioning method for UAV emergency services that integrates multimodal semantic models and address resolution.

[0007] The second objective of this invention is to provide a spatial positioning system for UAV emergency services that integrates multimodal semantic models and address resolution.

[0008] A third objective of this invention is to provide a computer device.

[0009] A fourth objective of this invention is to provide a storage medium.

[0010] The first objective of this invention can be achieved by adopting the following technical solution: A spatial positioning method for UAV emergency services that integrates multimodal semantic models and address resolution, the method comprising: Acquire disaster-related data within a defined geographical area, including online short video data, satellite map data, administrative boundary data, water system data, and road data; Extract audio from short online video data and convert the extracted audio into audio text data. Integrate audio and text data with short video titles, filter, classify, and identify place name keywords, align POI point names in public maps, count word frequency and event types, and perform batch collection and coordinate system transformation of address coordinate data; Identify satellite map data, classify construction land and non-construction land, and integrate and draw the boundaries of urban built-up areas; A unified coordinate system is established for administrative boundaries, urban built-up area boundaries, and point, line, and area place names. Point place names are spatially located, and line place names are spatially located based on water system data and road data. The scope of area place names is drawn, and area place names are spatially located. Based on spatial positioning information, determine task allocation and path, and control the UAV to execute the task; Aerial images taken by drones are fed back to a drone cloud platform for mapping, so as to correct them in real time and generate a detailed and dynamic disaster map.

[0011] Furthermore, the process of integrating audio text data with short video titles, filtering, classifying, and identifying place name keywords, aligning POI point names in public maps, statistically analyzing word frequencies and event characteristics, and performing batch collection and coordinate system transformation of address coordinate data specifically includes: The extracted audio text data and short video title data are integrated and merged into a text data file with a unified format; From the integrated text data, filter, identify and extract place name keywords in the relevant areas, and extract fragments of text information that mention place names; Align place name keywords with POI point names on public maps and merge place names with different names for the same item; The aligned place names are categorized and labeled as point-like, line-like, and area-like place names; Identify fragments of text information that mention place names and determine the event type; Collect point and area place names and addresses along with public map coordinates in batches and convert them into the WGS84 coordinate system.

[0012] Furthermore, the events include events pending investigation, events awaiting rescue, and temporary shelter events; The events to be investigated refer to those where text feedback indicates that flooding has occurred in certain locations; The events requiring rescue refer to those where text feedback indicates that residents are trapped at a location or that there are distress signals. The term "temporary refuge event" refers to an event in which a location becomes a temporary refuge or escape destination for residents, as indicated by text feedback.

[0013] Furthermore, the process of identifying and classifying satellite map data into built-up land and non-built-up land, and integrating and drawing the boundaries of urban built-up areas, specifically includes: Identify satellite map data, define typical land use categories, create classification feature files, and classify different land use categories using the maximum likelihood method; The system organizes and merges fragmented boundaries and plots in land use classification, connecting similar land uses into areas, and ultimately classifying and merging them into construction land and non-construction land. Based on the identification of construction land and non-construction land, the satellite layer of satellite map data is overlaid, and the boundaries of urban built-up areas are drawn along the connected construction land areas.

[0014] Furthermore, the process of unifying the coordinate system for administrative boundaries, urban built-up area boundaries, and point, line, and area place names, spatially locating point place names, spatially locating line place names based on water system and road data, and drawing the scope of area place names for spatial location, specifically includes: The coordinate system for administrative boundaries, urban built-up area boundaries, and point, line, and area place name event tables is unified as WGS84. Spatial positioning of point-like place names is achieved by displaying XY data, where X data represents the longitude field and Y data represents the latitude field. Link the linear place names with the road layer of road data and the water system layer of water system data in the attribute table for spatial positioning; The address coordinates of the area place name are used to locate the address point using the XY data display method. Combined with the land use classification layer, satellite layer and public map, the area of ​​the area place name is drawn for spatial positioning. Generate a spatial location map of disaster prevention and relief information with point, line, and area elements and event types.

[0015] Furthermore, the step of determining task allocation and path based on spatial positioning information and controlling the UAV to execute the task specifically includes: Based on spatial positioning information and event type, establish a buffer zone for the maximum mission radius, and select a launch point in areas where conditions permit. Obtain the mission objectives set by the drone cloud platform, as well as the corresponding waypoint routes, strip routes, and area routes; Based on the urgency of the event type, different mission priorities are assigned to the drones, and the drones are loaded with corresponding expansion modules and rescue supplies according to the mission. Control the drone to perform tasks based on the flight path and mission priority.

[0016] Furthermore, the step of feeding back drone aerial imagery to the drone cloud platform for mapping, in order to correct and generate a detailed dynamic disaster map in real time, specifically includes: Based on spatial positioning information, a buffer zone is established and superimposed with the urban built-up area boundary to form a map of potential disaster inspection areas. The drone orthophotos of the potential disaster-stricken inspection area are acquired and fed back to the drone cloud platform so that the drone cloud platform can complete image stitching and georegulation to generate a real-time map of the potential disaster-stricken inspection area. The system acquires a real-time map of the potential disaster inspection area generated by the drone cloud platform, replaces the corresponding location in the satellite map, and determines whether to expand the inspection scope based on the actual disaster situation captured in the footage, thereby generating a refined dynamic disaster map.

[0017] The second objective of this invention can be achieved by adopting the following technical solution: A spatial positioning system for UAV emergency services that integrates multimodal semantic models and address resolution, the system comprising: The acquisition unit is used to acquire disaster-related data within a defined geographical area, including online short video data, satellite map data, administrative boundary data, water system data, and road data. The extraction unit is used to extract audio from short online video data and convert the extracted audio into audio text data. The statistical unit is used to integrate audio and text data into short video titles, filter, classify and identify place name keywords, align POI point names in public maps, count word frequency and event nature, and perform batch collection and coordinate system transformation of address coordinate data. The identification unit is used to identify satellite map data, classify construction land and non-construction land, and integrate and draw the boundaries of urban built-up areas; Spatial positioning units are used to unify the coordinate system of administrative boundaries, urban built-up area boundaries, and point, line, and area place names, to spatially locate point place names, to spatially locate line place names based on water system data and road data, and to draw the scope of area place names and spatially locate area place names. The control unit is used to determine task allocation and path based on spatial positioning information, and to control the UAV to perform tasks; The generation unit is used to feed back drone aerial images to the drone cloud platform for mapping, so as to correct and generate a detailed dynamic disaster map in real time.

[0018] The third objective of this invention can be achieved by adopting the following technical solution: A computer device includes a processor and a memory for storing processor-executable programs, wherein when the processor executes the program stored in the memory, it implements the above-described UAV emergency service spatial positioning method.

[0019] The fourth objective of this invention can be achieved by adopting the following technical solution: A storage medium storing a program, which, when executed by a processor, implements the above-described spatial positioning method for unmanned aerial vehicle (UAV) emergency services.

[0020] The present invention has the following advantages over the prior art: This invention employs multimodal data fusion technology to effectively integrate information such as short videos, audio, text, and satellite images. Through intelligent processing of disaster information, it accurately identifies the geographical location of disaster areas, avoiding the waste of rescue resources caused by information lag and inaccurate positioning in traditional emergency responses. Combined with spatial positioning technology and a UAV cloud platform, this invention enables UAVs to perform more targeted and efficient rescue missions, allowing them to play a greater role in urban disaster prevention and relief. The generation of real-time, detailed, and dynamic disaster maps also provides strong decision support and judgment criteria for disaster prevention actions. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0022] Figure 1 This is a flowchart of the UAV emergency service spatial positioning method that integrates multimodal semantic model and address resolution according to Embodiment 1 of the present invention.

[0023] Figure 2 This is a land classification map of Embodiment 1 of the present invention.

[0024] Figure 3 This is a spatial location map of disaster prevention and relief information with point, line, and surface elements and event types, as described in Embodiment 1 of the present invention.

[0025] Figure 4 This is a map showing the potential disaster-affected inspection area in Embodiment 1 of the present invention.

[0026] Figure 5 This is a refined dynamic disaster map of Embodiment 1 of the present invention.

[0027] Figure 6 This is a structural block diagram of the UAV emergency service spatial positioning system that integrates multimodal semantic model and address resolution according to Embodiment 2 of the present invention.

[0028] Figure 7 This is a structural block diagram of the computer device according to Embodiment 3 of the present invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. Example

[0030] This embodiment provides a spatial positioning method for UAV emergency services that integrates multimodal semantic models and address resolution. This method acquires multiple data sources related to disasters, such as short online videos, satellite maps, administrative boundaries, water systems, and roads. It uses AI tools to perform text conversion and processing of the short video audio data, then extracts place name keywords and identifies events. Specific place names are categorized into point, line, and polygon types based on their spatial characteristics. The Baidu coordinate picking system and plugins are then used to parse and convert address coordinate information, generating a spatial location map of disaster prevention and relief information with point, line, and polygon elements and event types. Based on this map, the UAV's mission path is set to execute search and rescue missions in disaster areas. An ArcGIS buffer zone is established, overlaid with a base map of urban built-up areas, forming a map of potential disaster-affected areas. The refined disaster map is updated in real-time through aerial photography and cloud mapping, ultimately confirming the disaster-affected areas and their status, thus assisting in disaster prevention decision-making. Figure 1 As shown, it includes the following steps: S101. Obtain disaster-related data within a defined geographical area.

[0031] The disaster case identified in this embodiment is "XX Flood". Disaster-related data within a defined geographical area is obtained, including short video data, satellite maps with spatial location information, administrative division boundaries, water system data, and road data, in tif and shp formats, respectively.

[0032] In this embodiment, the extracted short video data was obtained from the Xindou data analysis platform. In the video search of "short video", the place name was entered as keyword 1, and the disaster names "flood" and "water disaster" were entered as keyword 2 respectively. The "casual shooting" category was selected for the search, and relevant videos on the day the disaster occurred were selected. Finally, a total of 509 data entries were filtered and downloaded.

[0033] This embodiment loads satellite map data, administrative boundary data, water system data, and road data into ArcGIS, generating "Satellite Map.tif", "Administrative Boundary.shp", "Water System.shp", and "Road.shp" layers respectively. An Excel spreadsheet is edited with "Provincial Administrative Division - County / District Administrative Division" as two columns to generate a "Research Towns.xls" file, which is then imported into ArcGIS to form a "Research Towns Table" file. The "Administrative Boundary.shp" file is then joined with the "Research Towns Table" file, with the join field being "name" and the "County / District Administrative Division" field. The data with valid associated fields is stored as "Research Town.shp" and used as the research scope. The "Data Management Tool - Raster - Raster Processing - Crop" tool is used to crop the "Satellite Map.tif" raster, and the output scope is "Research Town.shp", outputting the raster dataset "Research Town Satellite.Image". The "Crop (Analysis)" tool is used to crop the vector data of "Water System.shp" and "Road.shp", with the cropping feature selected as "Research Town.shp", and the feature classes "Research Town Water System.shp" and "Research Town Road.shp" are output respectively.

[0034] S102. Extract audio from short video data from the network and convert the extracted audio into audio text data.

[0035] This embodiment utilizes extraction and recognition tools to extract audio from short video content and convert it into text data for export. Specifically, the audio extraction tool used is the Magic Sound Workshop AI tool. Using the "Magic Sound Workshop - Efficiency Tools - Text Extraction" tool, the collected short video link or file is uploaded, the audio is extracted, and it is converted into a text modal txt file for export.

[0036] S103. Integrate audio and text data into short video titles, filter, classify, and identify place name keywords, align POI point names in public maps, count word frequency and event types, and perform batch collection and coordinate system transformation of address coordinate data.

[0037] Furthermore, step S103 specifically includes: S1031. Integrate the extracted audio text data with the short video title data and merge them into a text data file with a unified format.

[0038] In this embodiment, the content of the "Description of Work" column in the data table downloaded from the Xindou platform is copied and edited into a "Title.txt" file, with the same format as the audio text.

[0039] S1032. Filter, identify and extract place name keywords in relevant areas from the integrated text data, and extract fragments of text information that mention place names.

[0040] In this embodiment, after integrating the text data, an AI tool is used to clean up invalid data, removing duplicate advertisements, meaningless voiceovers and background music, etc. The AI ​​tool used is Doubao, and it is also used to assist in filtering and identifying place name keywords from the text modality file. Specifically, the AI ​​tool is instructed to: "Identify, analyze, and list the frequency of all specific place names appearing in the text, remove advertising content, and collect the original text paragraphs, summarizing them in a table. Requirements: Each column in the table should be a place name, the original text, and the frequency." The AI ​​processing results are then summarized and created in a "place name.xls" file.

[0041] S1033. Align place name keywords with POI point names on public maps and merge place names with different names.

[0042] This embodiment uses statistical results from AI tools to filter and discard some place names with unclear references. It also manually corrects and aligns place names according to POI data from Baidu Maps, merges identical place names, and manually unifies the place names in the "place name.xls" file into the standard place names in Baidu Maps.

[0043] S1034. Classify and label the aligned place names as point-like, line-like, and area-like place names.

[0044] This embodiment classifies and labels place names according to the rule of "specific shops and small public places - point place names, rivers and roads - line place names, residential areas, large public places, mountains and lakes - area place names". A "Place Name Classification" column is created in the "Place Name.xls" file, and the classified point, line and area place names are respectively created in the "Place Name.xls" file as "point place names", "line place names" and "area place names".

[0045] S1035. Identify fragments in the text information that mention place names and determine the event type.

[0046] This embodiment utilizes AI tools to identify fragments of text information mentioning place names and determine the event type. If necessary, a second manual check is performed. Depending on the specific disaster situation, the event type is categorized into those awaiting investigation, those awaiting rescue, and temporary shelter, and recorded in a table. Specifically, the AI ​​tool identifies fragments of place names mentioned in the "Original Text" column of the "Place Name.xls" file. Based on the actual flood disaster situation of "XX Flood," the event situation described in the text is divided into those awaiting investigation, those awaiting rescue, and temporary shelter. The classification rule is: events awaiting investigation... "Event" refers to a text message indicating a location experiencing flooding; "Awaiting Rescue" refers to a text message indicating a location where residents are trapped or there are distress signals; "Temporary Shelter" refers to a text message indicating a location becoming a temporary shelter or escape destination for residents. Create a "Event Type" column in the "Point Names," "Line Names," and "Area Names" worksheets within "Place Names.xls" to count the types of events appearing in each place name. Additionally, you can attach instructions to the AI ​​tool to mark text that is difficult to accurately identify, making it easier for manual secondary checks and identification.

[0047] S1036. Collect point and area place names, addresses, and public map coordinate information in batches and convert them into the WGS84 coordinate system.

[0048] This embodiment utilizes the Baidu coordinate picking system and a browser plugin to batch crawl and convert place names, addresses, and coordinates. Using the Chrome browser, download and install the Baidu Maps coordinate picking system batch processing tool. Open the plugin tool on the Baidu Maps coordinate picking system website. In the address latitude and longitude batch extraction column of the tool, input the place names whose addresses and coordinates need to be extracted in batches for batch acquisition. Similarly, use the same tool to convert the Baidu Maps coordinate system to the WGS84 coordinate system. Finally, summarize them into a "place name.xls" file. Edit and create "BAT longitude", "BAT latitude", "WGS84 longitude", and "WGS84 latitude" columns in the "point place names" and "area place names" worksheets.

[0049] S104. Identify satellite map data, classify construction land and non-construction land, and integrate and draw the boundaries of urban built-up areas.

[0050] Furthermore, step S104 specifically includes: S1041. Identify satellite map data, outline typical land use categories, form a classification feature file, and classify different land use categories using the maximum likelihood method.

[0051] This embodiment utilizes ArcGIS's image classification to identify satellite map data, extracting typical land use categories to create a classification feature file. The maximum likelihood method is then used to classify different land use categories such as buildings, roads, mountains and forests, farmland, and water bodies. Specifically, in ArcGIS, the "Study Town Satellite.Image" file is opened. Using the "Draw Polygon" tool in the "Image Classification" tool, multiple representative areas in the satellite image are identified, and polygons representing different land use categories such as buildings, roads, mountains and forests, farmland, and water bodies are drawn. In the "Training Sample Manager" of the "Image Classification" tool, polygon samples of the same land use category are merged, and the exported training samples are saved as a feature file named "Training Sample.gsg". In "Spatial Analysis Tools - Multivariate Analysis - Maximum Likelihood Method", "Study Town Satellite.Image" is input as the raster band, "Training Sample.gsg" is the feature file, and the classified raster data is output as "Land Classification.Image".

[0052] S1042. Regulate and merge the fragmented boundaries and patches in the land use classification, so that similar land uses can be connected into areas and ultimately classified and merged into construction land and non-construction land.

[0053] This embodiment, based on the maximum likelihood method for classifying different land use categories such as buildings, roads, forests, farmland, and water areas, uses filtering tools to smooth and merge fragmented boundaries and patches in the land use classification, connecting similar land uses into patches as much as possible, and finally classifying and merging them into construction land, forests, farmland, and water areas, etc. Among them, buildings and roads are construction land, while forests, farmland, and water areas are non-construction land. Specifically, the "Spatial Analysis Tools - Raster Generalization - Mode Filter" tool in ArcGIS is used to smooth and merge fragmented boundaries and patches in the land use classification. The tool is entered as "Land Use Classification.Image", the number of adjacent features is set to "FOUR", the replacement threshold is selected as "MAJORITY", and the output raster is "Land Use Classification Filter.Image". Observe the filtering results and try to connect similar land uses into patches. Repeat the filtering process as needed, and finally use the "3D Analysis Tools - Raster Reclassification - Reclassification" tool to classify and merge all land uses into construction land, mountain / forest, farmland, and water area, outputting "Land Reclassification.Image". Figure 2 As shown, buildings and roads are designated as construction land, while mountains, forests, farmland, and water areas are designated as non-construction land.

[0054] S1043. Based on the identification of construction land and non-construction land, and by overlaying the satellite layer status of satellite map data, draw the boundary of urban built-up area along the connected construction land areas.

[0055] This embodiment, based on the identification of built-up land and non-built-up land in "Land Use Reclassification.Image", combined with the observation of the actual situation in "Study Town Satellite.Image", uses ArcGIS editor tools to correct and draw the urban built-up area boundary in "Study Town.shp", creates a new "Built-up Area" field, sets the type to double precision, sets the urban built-up area patches in this field to 1 and the non-built-up area to 0, and then uses different symbols to distinguish the urban built-up area and non-built-up area in "Settings - Symbol System" of this layer.

[0056] S105. Unify the coordinate system for administrative boundaries, urban built-up area boundaries, and point, line, and area place names; spatially locate point place names; spatially locate line place names based on water system data and road data; and draw the scope of area place names and spatially locate area place names.

[0057] Furthermore, this step specifically includes: S1051. Unify the coordinate system of administrative boundaries, urban built-up area boundaries, and point, line, and surface place name event table information to WGS84.

[0058] In this embodiment, administrative boundaries and urban built-up area boundaries have been unified into the WGS84 coordinate system. The "point place names" and "area place names" worksheets in "place name.xls" have also been converted to latitude and longitude fields in the WGS84 coordinate system. Create "point place names", "line place names", and "area place names" tables in the ArcGIS research database, and load the "point place names", "line place names", and "area place names" worksheets from "place name.xls" into these tables respectively. This forms a database with administrative boundaries, urban built-up area boundaries, and point, line and area place name tables.

[0059] S1052. Spatial positioning of point-like place names using XY data display method.

[0060] In this embodiment, the "point-like place names" table is spatially located by displaying XY data. The X data is the "WGS84 longitude" field, and the Y data is the "WGS84 latitude" field. An appropriate symbol system is selected, and the location is displayed on the map. The data is then exported and saved as a "point-like place names.shp" file.

[0061] S1053. Link the attribute tables of linear place names with the road layer of road data and the water system layer of water system data to perform spatial positioning.

[0062] In this embodiment, the "name" field in "Water System.shp" and "Road.shp" is connected with the "Place Name" field in the "Linear Place Names" table. An appropriate symbol system is selected, the location is displayed on the map, and the file is exported and saved as "Linear Place Names.shp".

[0063] S1054. Use the XY data display method to locate the address coordinates of the area place name, and combine the land use classification layer, satellite layer and public map to draw the scope of the area place name and perform spatial positioning.

[0064] In this embodiment, the "area place names" table is used to locate address points by displaying XY data. The X data is the "WGS84 longitude" field, and the Y data is the "WGS84 latitude" field. Combining the land classification layer, satellite layer, Baidu online map and their address points, the area of ​​the area place names is drawn by outlining using the editor tools in ArcGIS. An appropriate symbol system is selected, and the location is displayed on the map. The data is then exported and saved as a "area place names.shp" file.

[0065] S1055. Generate a spatial location map of disaster prevention and relief information with point, line, and area elements and event types.

[0066] This embodiment displays "point place names.shp", "line place names.shp", and "area place names.shp" together on the "study city.shp" base map. The operation involves selecting "Attributes - Symbol System - Category - Unique Value" for the point, line, and area place name layer, setting the value field to "Event Category", adding all types, and unifying the symbol system format for the same event type. This ultimately generates a spatial location map of disaster prevention and relief information with point, line, and area features and event types, as shown below. Figure 3 As shown.

[0067] S106. Based on spatial positioning information, determine task allocation and path, and control the UAV to execute the task.

[0068] Furthermore, this step specifically includes: S1061. Based on spatial positioning information and event type, establish a buffer zone for the maximum mission radius and select a launch point in areas where conditions permit.

[0069] This embodiment establishes a buffer zone with the maximum task radius in ArcGIS based on the spatial location map of disaster prevention and relief information, which includes point, line, and area features and event types, as well as event types such as pending investigation, pending rescue, and temporary shelter. A launch point is selected in areas where conditions permit. The maximum distance for pending investigation tasks does not exceed half of the aircraft's maximum single flight distance. For pending rescue and temporary shelter tasks, the distance is appropriately reduced based on equipment load and task execution redundancy. Specifically, based on the spatial location map of disaster prevention and relief information, the disaster event types and their spatial location information (including "pending investigation," "pending rescue," and "temporary shelter") marked in the map database are retrieved. In ArcGIS, "Spatial Analysis Tools - Neighborhood Analysis - Buffer Zone" is opened, and point, line, and area features are input in batches according to the task nature. The buffer distance is set to the maximum task radius between the launch point and each task objective, generating the maximum flight range of the UAV.

[0070] For "pending investigation" tasks, the maximum mission radius is set to half of the aircraft's maximum single-flight range to ensure safe return. In this embodiment, the UAV model used is the Matrice 350 RTK, with a maximum flight speed of 23 m / s and a maximum flight time of approximately 55 minutes (at a speed of 8 m / s). It can carry various expansion modules. For "pending rescue" and "temporary refuge" tasks, the flight range is appropriately shortened based on the UAV's currently mounted equipment (such as infrared thermal imagers, loudspeakers, emergency supplies, etc.) and additional mission redundancy (search time). When the payload reaches 650g, the flight range is approximately 81.8%; when the payload reaches 1kg, the flight range is approximately 76%. Therefore, when selecting a starting point, depending on the mission difficulty, its straight-line distance from the mission target must be at least less than 80% of the maximum flight range.

[0071] Finally, taking into account factors such as buildings, terrain obstacles, and no-fly zones, candidate launch sites are selected within the buffer zone of the maximum straight-line radius of the corresponding mission, using suitable drone parking locations or temporarily deployed mobile takeoff platforms (such as the roof of an emergency vehicle or a temporary helipad). When meteorological data shows wind speeds > 8 m / s, additional launch sites with wind protection (such as the leeward side of a building) must be selected to ensure the safety of drone takeoff and landing.

[0072] S1062. Obtain the mission objectives set by the UAV cloud platform and the corresponding waypoint routes, strip routes, and area routes.

[0073] This embodiment sets mission objectives in the drone cloud platform (DJI Swift 2) and creates corresponding waypoint routes, strip routes, and area routes. Route settings may include: safe takeoff altitude, tilt climb to the starting point switch, speed, altitude, aircraft yaw angle mode, gimbal pitch angle control mode between waypoints, and completed actions. Specifically, mission objectives are set in the drone cloud platform, and corresponding waypoint routes, strip routes, and area routes are created. Each route objective corresponds to point, line, and area elements. Key parameters to set for each route include: safe takeoff altitude (avoiding low-altitude obstacles), flight speed and gimbal pitch angle mode, actions between waypoints (taking photos, hovering, releasing supplies, etc.), and emergency return-to-home conditions (low battery, loss of contact, etc.).

[0074] S1063. Assign different task priorities to drones based on the urgency of the event type.

[0075] This embodiment prioritizes specific tasks based on the urgency of the event type. For example, "awaiting rescue" tasks take precedence over "awaiting investigation" tasks, and "temporary shelter" tasks in high-risk disaster areas take precedence over "awaiting investigation" tasks. The specific rules for the required modules and their internal priority order for each event type are as follows: the drone loads the corresponding expansion modules and rescue supplies according to the task.

[0076] Tasks pending investigation: These refer to events where text feedback indicates flooding in certain locations; they mainly target elements marked on the map that need to be investigated (such as a flooded area or an area with suspected house damage), requiring drones to conduct a comprehensive scan to confirm the disaster situation; the priority of these tasks is sorted by the area of ​​the built-up area and the population density, with areas with larger built-up areas and denser populations being given priority.

[0077] Rescue-Awaiting Missions: Rescue-awaiting events refer to events where text feedback indicates that residents are trapped or that there are distress signals (such as trapped individuals reporting their location or accident sites). These missions primarily target rescue-awaiting elements on the map and require drones to carry drop modules, rescue supplies (such as first-aid kits, communication equipment, and food) or guide rescue personnel to the location. They also need to be equipped with thermal imaging cameras (infrared sensing) and loudspeakers to search for trapped individuals. Mission priorities are ranked according to the urgency of the emergency and the number of people trapped, with priority given to locations where life safety is directly threatened.

[0078] Temporary refuge mission: This refers to events where text feedback indicates a location has become a temporary refuge or escape destination for residents; it mainly targets temporary refuge elements on the map (such as temporary shelters and evacuation routes), requiring drones to be equipped with loudspeakers, and in some cases, carry drop modules and rescue supplies to survey the safety of temporary sites, the availability of supplies, and the accessibility of routes. If there are potentially dangerous sites, it is necessary to reassure the public to wait for rescue, or guide them to the correct escape routes to new refuge sites; priority is ranked according to the number of refugees and the maximum capacity of the site, with priority given to ensuring the normal functioning of refuge sites where people are concentrated.

[0079] S1064. Control the UAV to perform tasks according to the flight path and mission priority.

[0080] In this embodiment, after confirming the mission route, the route and mission priority are sent to the UAV. During the execution, the ground station receives flight status and image data in real time. If new obstacles or changes in the disaster situation are found, the route is adjusted in time or an alternative start point is activated. For missions that cannot be completed in a single flight, multiple batches of UAVs take turns or work in parallel to complete the mission plan and integrate the results.

[0081] S107. Feed back the drone aerial images to the drone cloud platform for mapping, so as to correct and generate a detailed dynamic disaster map in real time.

[0082] Furthermore, step S107 specifically includes: S1071. Based on spatial positioning information, establish a buffer zone and overlay it with the boundaries of urban built-up areas to form a map of potential disaster-affected areas.

[0083] This embodiment utilizes a spatial location map of disaster prevention and relief information based on point, line, and polygon features and event types. A buffer zone is established in ArcGIS. Using the "Spatial Analysis Tools - Neighborhood Analysis - Buffer Zone" tool, point, line, and polygon features with event types "Pending Investigation" and "Pending Rescue" are input in batches. The buffer distance is set to 150m. The buffer zone is then overlaid with the boundary of the urban built-up area to generate a map of potential disaster-affected areas for inspection. Figure 4 As shown in the figure, the marked buffer zones are the potential disaster inspection areas.

[0084] S1072. Acquire aerial images of potential disaster-affected inspection areas collected by UAV orthophotos, and feed these images back to the UAV cloud platform so that the UAV cloud platform can complete image stitching and georegistration to generate a real-time map of the potential disaster-affected inspection areas.

[0085] This embodiment uses a UAV cloud platform for cloud mapping. Entering the "Mapping Aerial Photography" task creation interface, select the "Orthophoto" mode. Considering disaster prevention needs, the ground resolution is set to 2-5 cm / pixel for densely populated urban areas and 5-10 cm / pixel for suburban and open areas. The overlap is set to 65%-75% forward overlap and 50%-60% lateral overlap to ensure image stitching quality. The flight altitude is calculated based on the ground resolution and UAV camera parameters, using the formula GSD = (H * p) / f, where GSD is the ground resolution, H is the flight altitude, p is the sensor pixel size (image width in pixels / sensor physical width), and f is the sensor focal length.

[0086] For example, when the drone is equipped with the H20 module with the lowest weight load, the CMOS size is 1 / 1.7", the variable focal length is 6.83-119.94 mm, the effective pixel is 20 million, and the maximum photo size is 5184×3888 pixels. At this time, the sensor pixel size is calculated to be approximately 1.465 μm / px. Therefore, when the ground resolution for data collection in most urban areas is set to 5 cm / px, the flight altitude should be around 232.9 meters. See Table 1 below for specific examples.

[0087] If mission regulations or safety requirements limit flight altitude (e.g., ≤120m in some areas), then maintaining an accuracy of 5cm / px and acquisition speed requires the use of modules with shorter focal lengths or larger sensor pixel sizes.

[0088] In this embodiment, the aerial images of the potential disaster-affected inspection area collected by the UAV are fed back to the UAV cloud platform. The UAV cloud platform completes image stitching and georegistration to generate a real-time map of the potential disaster-affected inspection area.

[0089] Table 1 Focal Length-GSD-Height Comparison Table

[0090] S1073. Obtain a real-time map of the potential disaster inspection area generated by the drone cloud platform, replace the corresponding location in the satellite map, and determine whether to expand the inspection scope based on the actual disaster situation captured in the video, thereby generating a refined dynamic disaster map.

[0091] This embodiment downloads the real-time map of the potential disaster-affected inspection area generated by the UAV cloud platform and overwrites the corresponding location on the original satellite map. During the data acquisition process, based on the actual transmitted images, the aerial imagery is compared with the original buffer zone. If obvious signs of disaster (such as collapsed buildings, flooded roads, fire traces, etc.) are found outside the buffer zone, the inspection area is immediately expanded and a new buffer zone task is generated. The corrected map is stored as a "refined dynamic disaster-affected urban map." Figure 5 As shown, as the disaster spreads or is contained, real-time photography can be used to update the replacement buffer zone, allowing for continuous monitoring of the disaster dynamics and supporting disaster prevention decisions.

[0092] It should be noted that although the method operations of the above embodiments are described in a specific order, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the described steps may be performed in a different order. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps. Example

[0093] like Figure 6 As shown, this embodiment provides a UAV emergency service spatial positioning system that integrates multimodal semantic model and address resolution. The system includes an acquisition unit 601, an extraction unit 602, a statistics unit 603, an identification unit 604, a spatial positioning unit 605, a control unit 606, and a generation unit 607. The specific functions of each unit are as follows: The acquisition unit 601 is used to acquire disaster-related data within a defined geographical area, including online short video data, satellite map data, administrative boundary data, water system data, and road data; Extraction unit 602 is used to extract audio from short video data from the network and convert the extracted audio into audio text data; The statistical unit 603 is used to integrate audio text data into short video titles, filter, classify and identify place name keywords, align POI point names in public maps, count word frequency and event nature, and perform batch collection and coordinate system transformation of address coordinate data. The identification unit 604 is used to identify satellite map data, classify construction land and non-construction land, and integrate and draw the boundaries of urban built-up areas; Spatial positioning unit 605 is used to unify the coordinate system of administrative boundaries, urban built-up area boundaries, and point, line, and area place names, to spatially locate point place names, to spatially locate line place names based on water system data and road data, and to draw the scope of area place names and spatially locate area place names. The control unit 606 is used to determine the task allocation and path based on spatial positioning information and control the UAV to perform the task; The generation unit 607 is used to feed back the drone aerial images to the drone cloud platform for mapping, so as to correct and generate a detailed dynamic disaster map in real time.

[0094] It should be noted that the system provided in this embodiment is only an example of the above-described division of functional units. In practical applications, the above functions can be assigned to different functional units as needed, that is, the internal structure can be divided into different functional units to complete all or part of the functions described above. Example

[0095] This embodiment provides a computer device, such as... Figure 7 As shown, it includes a processor 702, a memory, an input device 1203, a display device 704, and a network interface 705 connected via a system bus 701. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium 706 and internal memory 707. The non-volatile storage medium 706 stores an operating system, computer programs, and a database. The internal memory 707 provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. When the processor 702 executes the computer programs stored in the memory, it implements the UAV emergency service spatial positioning method of Embodiment 1 described above, as follows: The process involves acquiring disaster-related data within a defined geographical area, including online short video data, satellite map data, administrative boundary data, water system data, and road data. Audio is extracted from the online short video data and converted into audio-text data. This audio-text data is then integrated with the short video titles, and place name keywords are filtered, categorized, and identified. Points of Interest (POIs) in public maps are aligned, word frequencies and event types are statistically analyzed, and address coordinate data is collected in batches and converted to a new coordinate system. Satellite map data is identified, categorized into built-up land and non-built-up land, and urban built-up area boundaries are integrated and drawn. Administrative boundaries, urban built-up area boundaries, and point, line, and area place names are unified into a single coordinate system. Point place names are spatially located, and line place names are spatially located based on water system and road data. Area place name ranges are drawn, and area place names are spatially located. Based on the spatial location information, task allocation and paths are determined, and the UAV is controlled to execute the tasks. Aerial images captured by the UAV are fed back to the UAV cloud platform for mapping, allowing for real-time correction and the generation of a refined, dynamic disaster map. Example

[0096] This embodiment provides a storage medium, which is a computer-readable storage medium, storing a computer program. When the computer program is executed by a processor, it implements the UAV emergency service spatial positioning method of Embodiment 1 above, as follows: The process involves acquiring disaster-related data within a defined geographical area, including online short video data, satellite map data, administrative boundary data, water system data, and road data. Audio is extracted from the online short video data and converted into audio-text data. This audio-text data is then integrated with the short video titles, and place name keywords are filtered, categorized, and identified. Points of Interest (POIs) in public maps are aligned, word frequencies and event types are statistically analyzed, and address coordinate data is collected in batches and converted to a new coordinate system. Satellite map data is identified, categorized into built-up land and non-built-up land, and urban built-up area boundaries are integrated and drawn. Administrative boundaries, urban built-up area boundaries, and point, line, and area place names are unified into a single coordinate system. Point place names are spatially located, and line place names are spatially located based on water system and road data. Area place name ranges are drawn, and area place names are spatially located. Based on the spatial location information, task allocation and paths are determined, and the UAV is controlled to execute the tasks. Aerial images captured by the UAV are fed back to the UAV cloud platform for mapping, allowing for real-time correction and the generation of a refined, dynamic disaster map.

[0097] It should be noted that the computer-readable storage medium in this embodiment can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0098] In this embodiment, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this embodiment, the computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable program. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable storage medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable storage medium can be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination thereof.

[0099] The computer-readable storage medium described above can be used to write computer programs for executing this embodiment in one or more programming languages ​​or combinations thereof. These programming languages ​​include object-oriented programming languages—such as Java, Python, and C++—and conventional procedural programming languages—such as C or similar programming languages. The program can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0100] In summary, this invention employs multimodal data fusion technology to effectively integrate information such as short videos, audio, text, and satellite images. Through intelligent processing of disaster information, it accurately identifies the geographical location of disaster areas, avoiding the waste of rescue resources caused by information lag and inaccurate positioning in traditional emergency responses. Combined with spatial positioning technology and a UAV cloud platform, this invention enables UAVs to perform more targeted and efficient rescue missions, allowing them to play a greater role in urban disaster prevention and relief. The generation of real-time, detailed, and dynamic disaster maps also provides strong decision support and judgment criteria for disaster prevention actions.

[0101] The above description is merely a preferred embodiment of the present invention. It will be apparent to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects. The scope of the present invention is defined by the appended claims rather than the foregoing description, and thus all variations falling within the meaning and scope of the equivalents of the claims are intended to be included within the present invention.

Claims

1. A spatial positioning method for UAV emergency services that integrates multimodal semantic models and address resolution, characterized in that, The method includes: Acquire disaster-related data within a defined geographical area, including online short video data, satellite map data, administrative boundary data, water system data, and road data; Extract audio from short online video data and convert the extracted audio into audio-text data. Integrate audio and text data with short video titles, filter, classify, and identify place name keywords, align POI point names in public maps, count word frequency and event types, and perform batch collection and coordinate system transformation of address coordinate data; Identify satellite map data, classify construction land and non-construction land, and integrate and draw the boundaries of urban built-up areas; A unified coordinate system is established for administrative boundaries, urban built-up area boundaries, and point, line, and area place names. Point place names are spatially located, and line place names are spatially located based on water system data and road data. The scope of area place names is drawn, and area place names are spatially located. Based on spatial positioning information, determine task allocation and path, and control the UAV to execute the task; Aerial images taken by drones are fed back to a drone cloud platform for mapping, so as to correct them in real time and generate a detailed and dynamic disaster map.

2. The spatial positioning method for UAV emergency services according to claim 1, characterized in that, The process of integrating audio and text data into short video titles, filtering, classifying, and identifying place name keywords, aligning POI names in public maps, statistically analyzing word frequencies and event characteristics, and performing batch collection and coordinate system transformation of address coordinate data specifically includes: The extracted audio text data and short video title data are integrated and merged into a text data file with a unified format; From the integrated text data, filter, identify and extract place name keywords in the relevant areas, and extract fragments of text information that mention place names; Align place name keywords with POI point names on public maps and merge place names with different names for the same item; The aligned place names are categorized and labeled as point-like, line-like, and area-like place names; Identify fragments of text information that mention place names and determine the event type; Collect point and area place names and addresses along with public map coordinates in batches and convert them into the WGS84 coordinate system.

3. The spatial positioning method for unmanned aerial vehicle (UAV) emergency services according to claim 2, characterized in that, The events include those pending investigation, those requiring rescue, and temporary shelter events; The events to be investigated refer to those where text feedback indicates that flooding has occurred in certain locations; The events requiring rescue refer to those where text feedback indicates that residents are trapped at a location or that there are distress signals. The term "temporary refuge event" refers to an event in which a location becomes a temporary refuge or escape destination for residents, as indicated by text feedback.

4. The spatial positioning method for UAV emergency services according to claim 1, characterized in that, The process of identifying and classifying satellite map data into built-up land and non-built-up land, and integrating and drawing the boundaries of urban built-up areas, specifically includes: Identify satellite map data, define typical land use categories, create classification feature files, and classify different land use categories using the maximum likelihood method; The system organizes and merges fragmented boundaries and plots in land use classification, connecting similar land uses into areas, and ultimately classifying and merging them into construction land and non-construction land. Based on the identification of construction land and non-construction land, the satellite layer of satellite map data is overlaid, and the boundaries of urban built-up areas are drawn along the connected construction land areas.

5. The spatial positioning method for unmanned aerial vehicle (UAV) emergency services according to claim 1, characterized in that, The process of unifying administrative boundaries, urban built-up area boundaries, and point, line, and area place names into a unified coordinate system for spatial positioning of point place names, spatial positioning of line place names based on water system and road data, and spatial positioning of area place names by drawing the scope of area place names specifically includes: The coordinate system for administrative boundaries, urban built-up area boundaries, and point, line, and area place name event tables is unified as WGS84. Spatial positioning of point-like place names is achieved by displaying XY data, where X data represents the longitude field and Y data represents the latitude field. Link the linear place names with the road layer of road data and the water system layer of water system data in the attribute table for spatial positioning; The address coordinates of the area place name are used to locate the address point using the XY data display method. Combined with the land use classification layer, satellite layer and public map, the area of ​​the area place name is drawn for spatial positioning. Generate a spatial location map of disaster prevention and relief information with point, line, and area elements and event types.

6. The spatial positioning method for unmanned aerial vehicle (UAV) emergency services according to claim 1, characterized in that, The process of determining task allocation and path based on spatial positioning information and controlling the UAV to execute the task specifically includes: Based on spatial positioning information and event type, establish a buffer zone for the maximum mission radius, and select a launch point in areas where conditions permit. Obtain the mission objectives set by the drone cloud platform, as well as the corresponding waypoint routes, strip routes, and area routes; Based on the urgency of the event type, different mission priorities are assigned to the drones, and the drones are loaded with corresponding expansion modules and rescue supplies according to the mission. Control the drone to perform tasks based on the flight path and mission priority.

7. The spatial positioning method for unmanned aerial vehicle (UAV) emergency services according to claim 1, characterized in that, The process of feeding drone aerial imagery back to the drone cloud platform for mapping, in order to correct and generate a detailed dynamic disaster map in real time, specifically includes: Based on spatial positioning information, a buffer zone is established and superimposed with the urban built-up area boundary to form a map of potential disaster inspection areas. The drone orthophotos of the potential disaster-stricken inspection area are acquired and fed back to the drone cloud platform so that the drone cloud platform can complete image stitching and georegulation to generate a real-time map of the potential disaster-stricken inspection area. The system acquires a real-time map of the potential disaster inspection area generated by the drone cloud platform, replaces the corresponding location in the satellite map, and determines whether to expand the inspection scope based on the actual disaster situation captured in the footage, thereby generating a refined dynamic disaster map.

8. A spatial positioning system for UAV emergency services that integrates multimodal semantic modeling and address resolution, characterized in that, The system includes: The acquisition unit is used to acquire disaster-related data within a defined geographical area, including online short video data, satellite map data, administrative boundary data, water system data, and road data. The extraction unit is used to extract audio from short online video data and convert the extracted audio into audio text data. The statistical unit is used to integrate audio and text data into short video titles, filter, classify and identify place name keywords, align POI point names in public maps, count word frequency and event nature, and perform batch collection and coordinate system transformation of address coordinate data. The identification unit is used to identify satellite map data, classify construction land and non-construction land, and integrate and draw the boundaries of urban built-up areas; Spatial positioning units are used to unify the coordinate system of administrative boundaries, urban built-up area boundaries, and point, line, and area place names, to spatially locate point place names, to spatially locate line place names based on water system data and road data, and to draw the scope of area place names and spatially locate area place names. The control unit is used to determine task allocation and path based on spatial positioning information, and to control the UAV to perform tasks; The generation unit is used to feed back drone aerial images to the drone cloud platform for mapping, so as to correct and generate a detailed dynamic disaster map in real time.

9. A computer device comprising a processor and a memory for storing a processor-executable program, characterized in that, When the processor executes the program stored in the memory, it implements the UAV emergency service spatial positioning method according to any one of claims 1-7.

10. A storage medium storing a program, characterized in that, When the program is executed by the processor, it implements the UAV emergency service spatial positioning method according to any one of claims 1-7.

Images