Unmanned aerial vehicle image compensation slicing method, device and equipment and storage medium

Abstract

The invention relates to the technical field of unmanned aerial vehicle image processing, and discloses an unmanned aerial vehicle image compensation slicing method, device and equipment and a storage medium. The method comprises the following steps: acquiring flight mark data, inspection parameter data and a region image when an unmanned aerial vehicle inspects a long-strip-shaped target region, and extracting various types of flight parameter data; performing superposition calculation on the coordinates of the unmanned aerial vehicle based on the positioning compensation data, and constructing a camera imaging view cone model in combination with the inspection parameter data; performing intersection calculation on the view cone model and the strip-shaped target area to obtain an effective image coverage area, and determining image slice parameter data; and performing non-uniform tile cutting on the inspection area image based on the slice parameters to generate image slice data, and performing multi-resolution structure storage of non-uniform spatial index according to the inspection main axis direction to obtain an inspection image compensation slice result. According to the invention, the adaptive slicing of the unmanned aerial vehicle image is realized, the redundancy of the long-strip-shaped target slice is reduced, and the three-dimensional visual loading efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of UAV image processing, and particularly to a UAV image compensation slicing method, device, equipment and storage medium. Background Art

[0002] In the technical field of geospatial data processing, the processing and visualization of UAV images are key links in fields such as infrastructure inspection, smart city construction, and environmental monitoring. The image slicing and 3D modeling of long-strip geographical targets (such as pipelines, roads, transmission lines, etc.) are the technical difficulties. Efficiently processing the image data of such special-shaped targets is crucial for inspection enterprises, geographic information system development institutions, and remote sensing data processing centers, which directly affects the system response speed, data storage cost, and overall user experience. Therefore, an efficient image slicing method for long-strip targets is of great significance for improving the operation efficiency and resource utilization rate of geographic information systems.

[0003] Currently, the commonly used slicing methods in the industry include traditional rectangular grid slicing (such as dividing by a fixed size of 512×512 pixels), vector mask clipping (relying on manually pre-drawn boundaries), intelligent slicing based on content recognition, and streaming slicing and other technologies. However, these methods still face serious challenges when processing long-strip targets: traditional rectangular slicing will result in more than 60% of the invalid areas in the slices of long-strip targets (such as pipelines, roads), significantly increasing the storage and processing costs; vector mask clipping relies on manual intervention and cannot dynamically adapt to the perspective changes of the UAV during the inspection process; although content recognition slicing can reduce redundancy, it requires a large amount of memory resources and has poor real-time performance when processing large TIFF images; streaming slicing only solves the problem of front-end loading delay and does not fundamentally reduce the number of underlying slices. These methods generally ignore the dynamics of UAV flight parameters and the image acquisition process, and fail to make full use of the compensation data provided by the RTK positioning system to optimize the slicing strategy, resulting in low slicing efficiency and excessive resource consumption of the final long-strip targets. Summary of the Invention

[0004] The main purpose of the present invention is to solve the problem that the existing UAV image slicing methods generally ignore the dynamics of UAV flight parameters and the image acquisition process, and fail to make full use of the compensation data provided by the RTK positioning system to optimize the slicing strategy, resulting in low slicing efficiency and excessive resource consumption of the final long-strip targets.

[0005] The first aspect of the present invention provides a method for compensating and slicing UAV images. The method for compensating and slicing UAV images includes: obtaining UAV flight marker data, flight inspection parameter data, and inspection area images collected by a target UAV during inspection of a long strip-shaped target area, and extracting various flight parameter data from the UAV flight marker data, where the flight parameter data includes UAV coordinate data, positioning compensation data, and positioning status data; based on the positioning compensation data, performing coordinate superposition calculation on the UAV coordinate data to obtain corrected geographical coordinates, and based on the flight inspection parameter data and the corrected geographical coordinates, constructing a camera imaging frustum model of the target UAV; performing geometric calculation on the intersection area of the camera imaging frustum model and the long strip-shaped target area to obtain an effective image coverage area, and based on the effective image coverage area and the positioning status data, determining the image slicing parameter data of the target UAV; based on the image slicing parameter data, performing non-uniform tile cutting on the inspection area images to generate inspection image slice data, and based on the inspection main axis direction corresponding to the long strip-shaped target area, performing multi-resolution structure storage of non-uniform spatial indexing on the inspection image slice data to obtain the inspection image compensation slice result of the target UAV.

[0006] Optionally, in the first implementation manner of the first aspect of the present invention, the extracting various flight parameter data from the UAV flight marker data, where the flight parameter data includes UAV coordinate data, positioning compensation data, and positioning status data, includes: based on a preset data format specification and field structure, performing preliminary parsing on each record data in the UAV flight marker data to obtain the preliminarily parsed flight marker data; extracting various coordinate parameter data from the preliminarily parsed flight marker data to obtain UAV coordinate data, and extracting multi-directional compensation parameter data from the preliminarily parsed flight marker data to obtain positioning compensation data, and extracting the positioning status Q value corresponding to each record data in the preliminarily parsed flight marker data to obtain a positioning accuracy index, and extracting the northward standard deviation and eastward standard deviation corresponding to each record data in the preliminarily parsed flight marker data to obtain horizontal accuracy data, and generating positioning status data based on the positioning accuracy index and the horizontal accuracy data.

[0007] Optionally, in the second implementation manner of the first aspect of the present invention, the positioning compensation data includes a northward compensation value, an eastward compensation value, and an elevation compensation value. Based on the positioning compensation data, performing coordinate superposition calculation on the UAV coordinate data to obtain corrected geographic coordinates includes: determining the type of coordinate system corresponding to the UAV coordinate data, and based on the type of coordinate system, performing unified coordinate system conversion on the UAV coordinate data to obtain coordinate data in a unified geographic coordinate system, and performing preliminary correction calculation on the northward compensation value, the eastward compensation value, and the elevation compensation value to obtain a preliminarily corrected northward compensation value, a preliminarily corrected eastward compensation value, and a preliminarily corrected elevation compensation value; algebraically adding the corresponding latitude value in the coordinate data in the unified geographic coordinate system to the preliminarily corrected northward compensation value to obtain a compensated latitude value, and algebraically adding the corresponding longitude value in the coordinate data in the unified geographic coordinate system to the preliminarily corrected eastward compensation value to obtain a compensated longitude value, and algebraically adding the elevation value in the unified geographic coordinate system to the preliminarily corrected elevation compensation value to obtain a compensated elevation value; combining the compensated latitude value, the compensated longitude value, and the compensated elevation value to form the corrected geographic coordinates, and performing comparison calculation on the corrected geographic coordinates and the coordinate data in the unified geographic coordinate system to generate corrected geographic coordinates along the inspection path corresponding to the strip-shaped target area.

[0008] Optionally, in the third implementation manner of the first aspect of the present invention, constructing the camera imaging frustum model of the target UAV based on the flight inspection parameter data and the corrected geographic coordinates includes: extracting the corresponding pitch angle, roll angle, and heading angle data in the flight inspection parameter data to obtain the three-dimensional attitude parameters of the target UAV, and reading the corresponding camera focal length, field of view angle, and sensor size parameters in the flight inspection parameter data to obtain the camera imaging geometric parameters, and performing coordinate system conversion on the corrected geographic coordinates and the three-dimensional attitude parameters to obtain the camera coordinate system corresponding to the origin position of the target UAV; performing rotation matrix calculation on the three-dimensional attitude parameters to obtain the camera optical axis direction vector and multiple frustum corner point vectors, and performing frustum geometric parameter calculation on the camera imaging geometric parameters to obtain the near plane and far plane parameters of the frustum, and constructing spatial equations for the near plane and far plane parameters of the frustum and the frustum corner point vectors to obtain the four side equations of the frustum; integrating the geometric relationships of the camera optical axis direction vector, the frustum corner point vectors, and the four side equations of the frustum to obtain the topological structure data of the frustum vertex and the surface, and integrating the topological structure data and the camera coordinate system information for inspection space positioning to construct the camera imaging frustum model.

[0009] Optionally, in the fourth implementation manner of the first aspect of the present invention, the geometric calculation of the intersection area between the camera imaging frustum model and the long strip target area to obtain the effective image coverage area includes: performing triangular mesh division on the surface elevation information data corresponding to the long strip target area to generate a regional triangular mesh model, and performing a three-dimensional geometric intersection operation on the camera imaging frustum model and the regional triangular mesh model to obtain the regional space intersection line between the frustum and the ground surface; extracting the projection contour of the regional space intersection line to generate a frustum projection polygon corresponding to the long strip target area on the ground surface, and performing a unified coordinate system conversion on the frustum projection polygon based on the original image coordinate system corresponding to the inspection area image to obtain a geospatial polygon; calculating various geometric parameter values of the geospatial polygon to obtain the spatial distribution characteristics of the long strip target area, and performing vector boundary extraction and inspection pixel-level mask conversion on the spatial distribution characteristics to generate the effective image coverage area.

[0010] Optionally, in the fifth implementation manner of the first aspect of the present invention, the determining the image slice parameter data of the target UAV based on the effective image coverage area and the positioning status data includes: marking the positioning status data with a fixed solution status when the positioning status Q value is equal to a preset positioning status accuracy threshold to obtain a status marking result, and detecting whether the northward standard deviation and the eastward standard deviation in the positioning status data are greater than a preset standard deviation critical value to obtain a standard deviation detection result; determining the slice resolution corresponding to each inspection shooting position in the effective image coverage area based on the status marking result and the standard deviation detection result, and determining whether the aspect ratio of the area corresponding to the effective image coverage area is greater than a preset aspect ratio threshold; if the aspect ratio of the area corresponding to the effective image coverage area is greater than the preset aspect ratio threshold, determining that the effective image coverage area is a long strip image area, and adjusting the value of the first image tile parameter corresponding to the effective image coverage area to generate a first slice adjustment result for the long strip image area; if the aspect ratio of the area corresponding to the effective image coverage area is not greater than the preset aspect ratio threshold, determining that the effective image coverage area is a non-long strip image area and determining the value of the second image tile parameter corresponding to the effective image coverage area to generate a second slice adjustment result for the non-long strip image area; integrating and generating the image slice parameter data of the target UAV based on the status marking result, the slice resolution, the first slice adjustment result and the second slice adjustment result.

[0011] Optionally, in the sixth implementation manner of the first aspect of the present invention, the method of non-uniform tile cutting the inspection area image based on the image slice parameter data to generate inspection image slice data, and performing multi-resolution structure storage of non-uniform spatial indexing on the inspection image slice data based on the inspection main axis direction corresponding to the long strip target area to obtain the inspection image compensation slice result of the target unmanned aerial vehicle (UAV) includes: cutting and non-uniformly slicing the invalid area of the inspection area image based on the image slice parameter data to obtain a plurality of image slice tile data; performing identification coding of geographical location and zoom level on each of the image slice tiles to obtain the encoded image slice tile data, and identifying the inspection main axis direction corresponding to the long strip target area, and determining non-uniform spatial indexing in the index direction corresponding to the long strip target area based on the inspection main axis direction; performing multi-resolution structure organization and tile data storage on the encoded image slice tile data based on the non-uniform spatial indexing to obtain the inspection image compensation slice result of the target UAV.

[0012] The second aspect of the present invention provides a UAV image compensation slicing device, which includes: a data acquisition module, configured to acquire UAV flight marker data, flight inspection parameter data, and inspection area images collected by a target UAV during inspection of a long strip target area, and extract various flight parameter data in the UAV flight marker data, where the flight parameter data includes UAV coordinate data, positioning compensation data, and positioning status data; a visual cone construction module, configured to perform coordinate superposition calculation on the UAV coordinate data based on the positioning compensation data to obtain corrected geographical coordinates, and construct a camera imaging visual cone model of the target UAV based on the flight inspection parameter data and the corrected geographical coordinates; a region calculation module, configured to perform geometric calculation of the intersection region between the camera imaging visual cone model and the long strip target area to obtain an effective image coverage area, and determine the image slice parameter data of the target UAV based on the effective image coverage area and the positioning status data; a slice storage module, configured to perform non-uniform tile cutting on the inspection area image based on the image slice parameter data to generate inspection image slice data, and perform multi-resolution structure storage of non-uniform spatial indexing on the inspection image slice data based on the inspection main axis direction corresponding to the long strip target area to obtain the inspection image compensation slice result of the target UAV.

[0013] The third aspect of the present invention provides a UAV image compensation slicing device, including: a memory and at least one processor, where instructions are stored in the memory; the at least one processor invokes the instructions in the memory to enable the UAV image compensation slicing device to execute each step of the above-mentioned UAV image compensation slicing method.

[0014] The fourth aspect of the present invention provides a computer-readable storage medium, in which instructions are stored, and when they run on a computer, the computer is made to execute the various steps of the above-mentioned UAV image compensation slicing method.

[0015] The above-mentioned UAV image compensation slicing method, device, equipment and storage medium. In the embodiments of the present invention, by extracting and processing the flight marker data and parameter data when the target UAV inspects a long strip-shaped target, basic flight parameters such as UAV coordinates, positioning compensation, and status are obtained; then, based on the positioning compensation data, the coordinates are superimposed and calculated to obtain corrected geographical coordinates, and an accurate camera imaging frustum model is constructed in combination with the flight inspection parameters; then, the frustum model is geometrically intersected with the long strip-shaped target area to obtain an effective image coverage area, and the adaptive slicing parameters are determined according to the coverage area characteristics and positioning status; finally, non-uniform tile cutting is performed on the inspection area image according to the slicing parameters, and multi-resolution structure storage with non-uniform spatial indexing is performed in the main axis direction of the target area, and the final inspection image compensation slicing result is output. Through dynamic adaptive data processing and geometric analysis, the contradiction between the slicing efficiency and resource consumption of long strip-shaped geographical space data is solved, especially for the 3D modeling and visualization scenarios of linear facilities such as pipelines and roads; a dynamic slicing method based on UAV pose parameters is adopted, which not only makes full use of the accurate compensation data provided by the RTK positioning system, but also realizes the matching of the slicing shape and the target form; in addition, through non-uniform spatial indexing and multi-resolution storage strategies, the redundant data storage volume and processing time are effectively reduced, and the efficient slicing processing of long strip-shaped target image data is realized as a whole.

[0016] Other features and advantages of the present invention will be described in the following specification, and part of them will become obvious from the specification, or be understood by implementing the present invention. The objectives and other advantages of the present invention are realized and obtained by the structures specifically pointed out in the specification, claims and drawings.

[0017] To make the above objectives, features and advantages of the present invention more obvious and understandable, the following specific preferred embodiments are given, and in conjunction with the accompanying drawings, the detailed description is as follows. Description of the Drawings

[0018] Figure 1 It is a schematic diagram of the first embodiment of the UAV image compensation slicing method in the embodiments of the present invention; Figure 2 It is a schematic diagram of an embodiment of the UAV image compensation slicing device in the embodiments of the present invention; Figure 3 It is a schematic diagram of an embodiment of the UAV image compensation slicing equipment in the embodiments of the present invention. Detailed implementation manners

[0019] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0020] The terms "including" and "having" and any variations thereof mentioned in the embodiments of the present invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units is not limited to the listed steps or units, but optionally further includes other unlisted steps or units, or optionally further includes other steps or units inherent to these processes, methods, products or devices.

[0021] For the convenience of understanding this embodiment, the specific process of the embodiments of the present invention will be described below. Please refer to Figure 1 , the first embodiment of the method for compensating and slicing UAV images in the embodiments of the present invention includes: 101. Obtain the UAV flight marker data, flight inspection parameter data, and inspection area images collected by the target UAV during the inspection of a long strip-shaped target area, and extract various flight parameter data from the UAV flight marker data. The flight parameter data includes UAV coordinate data, positioning compensation data, and positioning status data; The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) uses a digital computer or a machine controlled by a digital computer to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use the knowledge to obtain the best results in theory, method, technology, and application systems.

[0022] Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation / interaction systems, and mechatronics. Artificial intelligence software technologies mainly include several major directions such as computer vision technology, robotics, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0023] In this embodiment, based on a preset data format specification and field structure, each record data in the UAV flight marker data is preliminarily parsed to obtain the preliminarily parsed flight marker data; various coordinate parameter data in the preliminarily parsed flight marker data are extracted to obtain UAV coordinate data, and multi-direction compensation parameter data in the preliminarily parsed flight marker data are extracted to obtain positioning compensation data, and the positioning status Q values corresponding to each record data in the preliminarily parsed flight marker data are extracted to obtain positioning accuracy indicators, and the northward standard deviation and eastward standard deviation corresponding to each record data in the preliminarily parsed flight marker data are extracted to obtain horizontal accuracy data, and positioning status data is generated based on the positioning accuracy indicators and the horizontal accuracy data.

[0024] In practical applications, first, obtain the key data set collected by the target UAV during the inspection of a long strip-shaped target area (including UAV flight marker data, flight inspection parameter data, and inspection area images). Among them, the flight marker data is usually stored in the form of an MRK file, which contains detailed positioning and attitude information during the UAV flight. After obtaining this data, because the MRK file has a specific structured format, it needs to be correctly parsed according to the version information and data format defined in its file header. By initially parsing each record data in the MRK file based on the preset data format specifications and field structures, that is, first reading the file header of the MRK file to obtain the file version and data specifications, and then parsing the record data one by one according to these specifications to obtain the initially parsed flight marker data; furthermore, extract the longitude, latitude, and elevation values from the initially parsed flight marker data. These values together constitute the original coordinate data of the UAV, which record the spatial position information of the UAV during the inspection of the long strip-shaped target area; and extract the multi-directional compensation parameter data in the initially parsed flight marker data, specifically including the northward compensation value, eastward compensation value, and elevation compensation value. These compensation values are calculated by the UAV's RTK positioning system and are used to accurately correct the original GPS coordinates to obtain more accurate positioning compensation data; and extract the positioning status Q value corresponding to each record data (this value is an important indicator for evaluating the reliability of GPS positioning accuracy, that is, it is used to indicate the type and accuracy level of the current positioning solution). In the RTK system, when the Q value is equal to 50, it indicates that the RTK fixed solution state has been reached, with the highest positioning accuracy. This value will also be used as the key positioning accuracy indicator for judging the subsequent slice processing accuracy; and extract the northward standard deviation and eastward standard deviation data from each record to be used to evaluate the positioning accuracy in the horizontal direction, which together constitute the horizontal accuracy data (usually when these standard deviations are less than 0.01 meters, it indicates that the horizontal positioning has reached centimeter-level accuracy), and based on the extracted positioning accuracy indicator (Q value) and horizontal accuracy data (northward and eastward standard deviations), generate complete positioning status data through comprehensive analysis for use in subsequent judgment of the slice accuracy requirements at different positions. For example, when the Q value is 50 and the standard deviation is less than the critical value, a high-precision slicing strategy can be adopted; while in areas with lower accuracy, the slice resolution can be appropriately reduced to balance the processing efficiency and accuracy requirements.

[0025] 102. Based on the positioning compensation data, perform coordinate superposition calculation on the UAV coordinate data to obtain the corrected geographical coordinates, and based on the flight inspection parameter data and the corrected geographical coordinates, construct the camera imaging frustum model of the target UAV; In this embodiment, the positioning compensation data includes a northward compensation value, an eastward compensation value, and an elevation compensation value. By determining the type of coordinate system corresponding to the UAV coordinate data, and based on the coordinate system type, the UAV coordinate data is converted to a unified coordinate system to obtain the coordinate data in the unified geographic coordinate system, and the northward compensation value, the eastward compensation value, and the elevation compensation value are subjected to preliminary correction calculations to obtain the preliminarily corrected northward compensation value, the preliminarily corrected eastward compensation value, and the preliminarily corrected elevation compensation value; the latitude value corresponding to the coordinate data in the unified geographic coordinate system is algebraically added to the preliminarily corrected northward compensation value to obtain the compensated latitude value, and the longitude value corresponding to the coordinate data in the unified geographic coordinate system is algebraically added to the preliminarily corrected eastward compensation value to obtain the compensated longitude value, and the elevation value in the unified geographic coordinate system is algebraically added to the preliminarily corrected elevation compensation value to obtain the compensated elevation value; the compensated latitude value, the compensated longitude value, and the compensated elevation value are combined to form a corrected geographic coordinate, and the corrected geographic coordinate is compared with the coordinate data in the unified geographic coordinate system to generate a corrected geographic coordinate along the inspection path corresponding to the long strip-shaped target area; the corresponding pitch angle, roll angle, and heading angle data in the flight inspection parameter data are extracted to obtain the three-dimensional attitude parameters of the target UAV, and the corresponding camera focal length, field of view angle, and sensor size parameters in the flight inspection parameter data are read to obtain the camera imaging geometric parameters, and the corrected geographic coordinate and the three-dimensional attitude parameters are converted to the camera coordinate system corresponding to the origin position of the target UAV; the rotation matrix of the three-dimensional attitude parameters is calculated to obtain the camera optical axis direction vector and multiple frustum corner vectors, and the frustum geometric parameters of the camera imaging geometric parameters are calculated to obtain the near plane and far plane parameters of the frustum, and the near plane and far plane parameters of the frustum and the frustum corner vectors are used to construct a spatial equation to obtain the four side equations of the frustum; the geometric relationship of the camera optical axis direction vector, the frustum corner vectors, and the four side equations of the frustum is integrated to obtain the topological structure data of the frustum vertex and surface, and the topological structure data and the camera coordinate system information are integrated for inspection space positioning to construct a camera imaging frustum model.

[0026] In practical applications, first, since the GPS data recorded by the UAV during flight may adopt different coordinate systems, such as WGS84, CGCS2000 or local coordinate system, etc., it is necessary to determine the type of coordinate system corresponding to the UAV coordinate data, that is, by checking the coordinate system identifier in the MRK file or through coordinate value range analysis and combining with additional information in the flight inspection parameter data to determine. And based on the coordinate system type, perform a unified coordinate system conversion on the UAV coordinate data, convert all coordinate points to the same geographic coordinate system, usually select WGS84 or CGCS2000 as the standard reference system to ensure the consistency of spatial calculations in subsequent processing, and finally obtain the coordinate data in the unified geographic coordinate system; and process the positioning compensation data extracted from the MRK file, where the positioning compensation data includes northward compensation value, eastward compensation value and elevation compensation value. These values reflect the correction amount of the RTK system to the original GPS positioning result, that is, through preliminary correction calculations according to the specific hardware device characteristics or local environmental factors (for example, it may be necessary to consider factors such as antenna phase center offset, geomagnetic declination correction or system deviation in a specific area, and adjust the original compensation value through corresponding formulas or look-up table methods), so as to obtain the preliminarily corrected northward compensation value, preliminarily corrected eastward compensation value and preliminarily corrected elevation compensation value to ensure the consistency and accuracy of the compensation value under different flight conditions and regions; furthermore, after obtaining the coordinate data in the unified geographic coordinate system and the preliminarily corrected compensation value, perform an algebraic addition operation on the corresponding latitude value in the coordinate data in the unified geographic coordinate system and the preliminarily corrected northward compensation value (the northward coordinate calculation formula is: the compensated latitude value (i.e., the northward coordinate) = the original north coordinate + northward compensation value / 1000 (where the northward compensation is usually a linear displacement in meters, while the latitude is an angular unit, so the northward compensation value needs to be converted to the corresponding latitude increment before addition, and the specific conversion coefficient is related to the latitude. Generally, on the earth, every 111320 meters corresponds to about 1 degree of latitude change)), so as to obtain the compensated latitude value (which is more accurate than the latitude recorded by the original GPS); similarly, perform an algebraic addition on the corresponding longitude value in the coordinate data in the unified geographic coordinate system and the preliminarily corrected eastward compensation value to obtain the compensated longitude value (the eastward coordinate calculation formula is: the compensated longitude value (i.e., the eastward coordinate) = the original east coordinate + eastward compensation value / 1000 (where the eastward compensation value also needs to be converted to the longitude increment. Different from the latitude, the conversion coefficient of the longitude changes with the latitude. Generally, at the latitude φ, every (111320 × cosφ) meters corresponds to about 1 degree of longitude change)).And algebraically add the elevation value in the unified geographic coordinate system to the preliminarily corrected elevation compensation value (the elevation coordinate calculation formula is: elevation value in the unified geographic coordinate system = ellipsoidal height + elevation compensation value / 1000 (where the elevation compensation is usually directly in meters and can be directly added to the original elevation value, but the conversion between different elevation reference planes (such as ellipsoidal height, altitude, etc.) may need to be considered)) (where the MRK file contains ellipsoidal height data) to obtain the compensated elevation value; then combine the compensated latitude value, compensated longitude value, and compensated elevation value to form a complete corrected geographic coordinate to represent the actual position of the UAV during the inspection, which has higher accuracy and reliability. And to evaluate the compensation effect, compare and calculate the corrected geographic coordinate with the original coordinate data in the unified geographic coordinate system, analyze the correction amount of each point and the overall correction trend, and generate a sequence of corrected geographic coordinates along the corresponding inspection path of the long strip-shaped target area based on these analysis results. Through this high-precision coordinate correction based on RTK compensation data, the positioning accuracy of geospatial data can be greatly improved, thus laying a solid foundation for subsequent non-uniform slicing and 3D visualization.;

[0027] Secondly, extract the three-dimensional attitude information of the UAV (including pitch angle, roll angle, and heading angle data) from the flight inspection parameter data. These angle data are usually recorded in the UAV flight log or stored as additional information in the MRK file. Sometimes, it is also necessary to obtain them in combination with the NAV navigation file. Among them, the pitch angle describes the rotation of the UAV around the transverse axis (the nose up or down), the roll angle represents the rotation around the longitudinal axis (the fuselage tilts left or right), and the heading angle represents the rotation around the vertical axis (the flight direction), so as to obtain the three-dimensional attitude parameters describing the UAV's attitude in space (to determine the actual shooting direction of the camera. Especially when inspecting long-strip targets such as pipelines and roads, the UAV often needs to adjust its attitude to obtain the best coverage perspective); and read the camera parameters in the flight inspection parameter data, including camera focal length, field of view angle, and sensor size. These parameters are usually provided by the UAV manufacturer or recorded in the EXIF data of the camera. Among them, the focal length determines the field of view range of the camera, usually in millimeters; the field of view angle defines the horizontal and vertical angle ranges that the camera can capture, usually in degrees; the sensor size describes the physical size of the camera sensor, which affects the coverage area of the imaging, so as to obtain the camera imaging geometric parameters, which determine the perspective projection characteristics in the camera imaging process; and perform coordinate system conversion on the corrected geographical coordinates and three-dimensional attitude parameters, convert the position information in the global geographical coordinate system (such as WGS84) to the local camera coordinate system with the UAV position as the origin, that is, the conversion from geographical coordinates to ECEF (Earth-Centered, Earth-Fixed) coordinates, and then the conversion from ECEF coordinates to the local northeast-up (ENU) coordinate system. Finally, rotate the ENU coordinate system to the camera coordinate system according to the UAV attitude angle. And in this camera coordinate system, define the z-axis along the camera optical axis direction, the x-axis to the right, the y-axis downward, and the origin at the projection center of the camera, which is convenient for subsequent frustum geometry calculations; furthermore, perform rotation matrix calculation on the three-dimensional attitude parameters, that is, construct a 3×3 rotation matrix according to the pitch angle, roll angle, and heading angle. This matrix describes the rotation relationship from the geographical coordinate system to the camera coordinate system. By applying this rotation matrix, determine the direction vector of the camera optical axis, that is, the unit vector of the camera's orientation, and the four corner vectors of the frustum (where the frustum corner vectors are the directions from the camera position to the four corners of the field of view. These vectors are calculated through the camera's field of view angle and the rotation matrix).For example, if the horizontal field of view angle of the camera is 60 degrees and the vertical field of view angle is 45 degrees, the angles of the four corners of the field of view relative to the optical axis can be calculated, and then these angles can be converted into corresponding direction vectors through a rotation matrix); and the geometric parameters of the frustum are calculated based on the camera imaging geometric parameters to determine the near plane and the far plane of the frustum (where the near plane is usually set at a certain distance from the camera, and the far plane can be set on the ground or at a sufficient distance. The parameters of these two planes include their normal vectors (usually consistent with the optical axis direction) and the distances to the origin), that is, through the focal length, sensor size, and field of view angle of the camera, combined with the flight altitude of the UAV, the specific positions and sizes of the near plane and the far plane are calculated; furthermore, the sides of the frustum are represented by calculating the plane equations of the camera optical axis direction vector, the frustum corner point vectors, and the four side equations of the frustum, that is, the normal vector of the plane is calculated using the vector cross product, and then the standard equation of the plane is determined by combining the point coordinates on a plane, thus obtaining four plane equations, which together define the four side boundaries of the frustum (each side can be regarded as a plane determined by the camera position and two adjacent corner point vectors); furthermore, all the above geometric elements are integrated to construct a complete frustum model, that is, the camera optical axis direction vector, the frustum corner point vectors, and the frustum side equations are organized into a unified data structure to describe the topological relationship between the vertices and faces of the frustum, and this topological structure data of the frustum vertices and faces is combined with the camera coordinate system information to ensure the accurate positioning of the frustum in the inspection space, thereby realizing the construction of the camera imaging frustum model. This model accurately describes the imaging range of the UAV camera at a specific position and attitude, laying a foundation for subsequent intersection calculations with the ground surface and determining the effective image coverage area. Especially for the inspection scenarios of long-strip targets such as pipelines and roads, this accurate frustum model can significantly reduce redundant data in the subsequent slicing process and improve the processing efficiency and resource utilization rate of the entire system.

[0028] 103. Geometric calculations of the intersection area between the camera imaging frustum model and the long-strip target area are performed to obtain the effective image coverage area, and based on the effective image coverage area and the positioning state data, the image slicing parameter data of the target UAV is determined; In this embodiment, the surface elevation information data corresponding to the long strip target area is divided into triangular meshes to generate a regional triangular mesh model, and a three-dimensional geometric intersection operation is performed between the camera imaging frustum model and the regional triangular mesh model to obtain the regional spatial intersection line between the frustum and the ground surface; the projection contour of the regional spatial intersection line is extracted to generate a frustum projection polygon corresponding to the ground surface of the long strip target area, and based on the original image coordinate system corresponding to the inspection area image, the frustum projection polygon is subjected to a unified coordinate system conversion to obtain a geospatial polygon; various geometric parameter values of the geospatial polygon are calculated to obtain the spatial distribution characteristics of the long strip target area, and the vector boundary of the spatial distribution characteristics is extracted and the mask conversion at the inspection pixel level is performed to generate an effective image coverage area; the positioning state data with the positioning state Q value equal to the preset positioning state accuracy threshold is marked with fixed solution state data to obtain a state marking result, and it is detected whether the northward standard deviation and the eastward standard deviation in the positioning state data are greater than the preset standard deviation critical value to obtain a standard deviation detection result; based on the state marking result and the standard deviation detection result, the slice resolution corresponding to each inspection shooting position in the effective image coverage area is determined, and it is judged whether the aspect ratio of the area corresponding to the effective image coverage area is greater than the preset aspect ratio threshold; if the aspect ratio of the area corresponding to the effective image coverage area is greater than the preset aspect ratio threshold, it is determined that the effective image coverage area is a long strip image area, and the value of the first image tile parameter corresponding to the effective image coverage area is adjusted to generate a first slice adjustment result for the long strip image area; if the aspect ratio of the area corresponding to the effective image coverage area is not greater than the preset aspect ratio threshold, it is determined that the effective image coverage area is a non-long strip image area and the value of the second image tile parameter corresponding to the effective image coverage area is determined to generate a second slice adjustment result for the non-long strip image area; based on the state marking result, the slice resolution, the first slice adjustment result and the second slice adjustment result, the image slice parameter data of the target UAV is integrally generated.

[0029] In practical applications, first, since elevation data usually comes from Digital Elevation Model (DEM) or Digital Surface Model (DSM), which store elevation values of the earth's surface in a regular grid form. By performing triangular mesh division on the surface elevation information data corresponding to the long and narrow target area, that is, using the Delaunay triangulation algorithm for processing, triangular meshes that meet specific mathematical properties are generated, avoiding the appearance of overly long and narrow triangles, thereby improving the stability and accuracy of subsequent geometric calculations. Thus, a regional triangular mesh model representing the long and narrow target area and its surrounding terrain is generated (this model accurately describes the three-dimensional geometric shape of the earth's surface). Then, a three-dimensional geometric intersection operation is performed between the camera imaging frustum model and the regional triangular mesh model, that is, by detecting the intersection points of the four sides, near plane, and far plane of the frustum with each triangle in the triangular mesh (such as using spatial partitioning techniques (such as octree or KD tree) to accelerate this process and avoid exhaustive detection of all triangles), the regional spatial intersection lines between the frustum and the earth's surface are obtained (these intersection lines form one or more closed contours, representing the projection boundary of the frustum on the earth's surface); furthermore, the regional spatial intersection lines are projected onto the horizontal plane, and then the closed polygon contour is reconstructed according to the connection relationship, thereby generating the frustum projection polygon corresponding to the long and narrow target area on the earth's surface, accurately representing the ground area range that the camera can observe at the current position and attitude. Based on the original image coordinate system corresponding to the inspection area image, a unified coordinate system transformation is performed on the frustum projection polygon, that is, a transformation from the local coordinate system to the global geographic coordinate system, and considering the conversion relationship between map projection and coordinate system, a geographic space polygon represented by geographic coordinates is obtained (this geographic space polygon can be directly spatially registered with the image data); furthermore, geometric parameters such as area, perimeter, minimum bounding rectangle, and aspect ratio are calculated for the geographic space polygon. In particular, the aspect ratio parameter is a key indicator for judging whether the target is long and narrow. For example, when the aspect ratio is greater than 5, the target is usually determined to be a long and narrow target such as a road or a pipeline. Through the calculation of these geometric parameters, the spatial distribution characteristics of the long and narrow target area are obtained; and vector boundary extraction is performed on the spatial distribution characteristics to obtain an accurate boundary line describing the effective area. At the same time, a mask conversion at the inspection pixel level is performed to generate a binary mask image with the same resolution as the original image, where pixels with a value of 1 represent the effective coverage area, and pixels with a value of 0 represent the invalid area. The effective image coverage area finally generated by this mask image will be used as the spatial basis for subsequent non-uniform slicing, guiding the system to only perform high-quality slicing on the effective area, thereby significantly reducing redundant data, improving the processing efficiency and storage utilization rate of the entire system, and being particularly suitable for three-dimensional modeling and visualization application scenarios of long and narrow targets such as pipelines and roads.

[0030] Secondly, perform accuracy evaluation on the positioning status data. That is, start from the positioning status Q value in the extracted flight marker data and compare it with the preset positioning status accuracy threshold (i.e., Q = 50. Since in the RTK positioning system, a Q value equal to 50 usually indicates the achievement of the RTK fixed solution state, which is the highest-precision positioning state). That is, mark the records with the positioning status Q value equal to 50 as fixed solution state data (these marking results will be used to judge which areas require high-precision slicing processing) and achieve this by traversing the positioning status data and adding a Boolean flag or a specific status code to each record, thereby obtaining the complete status marking results. And since the northward standard deviation and eastward standard deviation reflect the accuracy fluctuations in the horizontal direction, a smaller standard deviation indicates higher positioning stability and reliability. By detecting whether the northward standard deviation and eastward standard deviation in the positioning status data exceed the preset standard deviation critical value (such as 0.01 meters), traverse the positioning status data, compare the northward standard deviation and eastward standard deviation in each record with the preset critical value. If any standard deviation exceeds the critical value, mark the record as "insufficient accuracy", otherwise mark it as "qualified accuracy", and finally obtain the complete standard deviation detection results (so that subsequently, the images exceeding the critical value will have their slicing resolution reduced (such as from 1 cm / pixel to 5 cm / pixel)). Furthermore, based on the status marking results and standard deviation detection results, jointly determine the slicing resolution corresponding to each inspection shooting position in the effective image coverage area. That is, for areas that simultaneously meet "RTK fixed solution" and "qualified accuracy", the highest slicing resolution can be allocated, such as 1 centimeter / pixel; for areas that only meet "RTK fixed solution" but have an excessive standard deviation, the resolution can be appropriately reduced, such as 5 centimeters / pixel; and for areas that are not in the RTK fixed solution state, a lower resolution, such as 10 centimeters / pixel, is adopted. This accuracy-based hierarchical slicing strategy can ensure high quality in key areas while reducing the overall data processing burden. Then, judge the shape characteristics of the effective image coverage area, especially whether the aspect ratio is greater than the preset aspect ratio threshold (usually set the aspect ratio to 5 as the threshold for judging long and narrow targets, which is determined based on the morphological characteristics of typical long and narrow targets such as roads and pipelines). Thus, by calculating the minimum bounding rectangle of the effective image coverage area (i.e., the geospatial polygon generated in the previous steps), and then calculating the ratio of its long side to its short side to obtain the aspect ratio of the area. If the aspect ratio is greater than the preset threshold of 5, determine the effective image coverage area as a long and narrow image area, and adjust the standard tile parameters (such as tile width = 512 pixels, tile height = effective area height + 10% buffer) to generate a slicing strategy more suitable for long and narrow targets (such as traditional tile slicing usually uses squares, such as 512×512 pixels, but this will generate a large amount of redundant space for long and narrow targets.Therefore, adjust the values of the first image tile parameters. For example, fix the tile width at 512 pixels, but dynamically adjust the tile height according to the actual height of the effective area, and add an edge buffer of about 10% to ensure visual continuity. Such non-uniform tiles can better adapt to the spatial distribution of long and narrow targets, significantly reducing storage and processing redundancy. This adjustment process generates the first slice adjustment result for the long and narrow image area. On the contrary, if the aspect ratio of the effective image coverage area is not greater than the preset threshold, it is determined as a non-long and narrow image area. Such areas are suitable for using the traditional uniform slicing strategy (i.e., cutting according to the standard 512×512, for example, determining the values of the second image tile parameters corresponding to the non-long and narrow image area, usually using the standard 512×512 pixel uniform grid slicing method) to generate the second slice adjustment result for the non-long and narrow image area, so as to be able to adopt the most suitable slicing strategy for different-shaped targets and ensure the efficient use of system resources. Furthermore, based on all the intermediate results obtained above, including the status marking result, the slice resolution of each shooting position, the first slice adjustment result of the long and narrow area, and the second slice adjustment result of the non-long and narrow area, comprehensive integration is carried out to generate the complete target UAV image slice parameter data (where these parameter data include the slice strategy, resolution, tile size, and boundary information of each area, etc.) to guide the subsequent non-uniform tile cutting process, ensuring that the entire system can achieve efficient slice processing for long and narrow geospatial data, significantly reducing data redundancy, and improving the loading efficiency of 3D visualization, which is particularly suitable for 3D modeling and visualization application scenarios of long and narrow targets such as pipelines and roads.

[0031] 104. Based on the image slice parameter data, perform non-uniform tile cutting on the inspection area image to generate inspection image slice data, and based on the inspection main axis direction corresponding to the long and narrow target area, perform multi-resolution structure storage of non-uniform spatial indexing on the inspection image slice data to obtain the inspection image compensation slice result of the target UAV.

[0032] In this embodiment, based on the image slice parameter data, crop the invalid area and perform non-uniform slicing on the inspection area image to obtain multiple image slice tile data; perform identification coding of geographical location and zoom level on each image slice tile to obtain the encoded image slice tile data, and identify the inspection main axis direction corresponding to the long and narrow target area, and based on the inspection main axis direction, determine the non-uniform spatial index of the index direction corresponding to the long and narrow target area; based on the non-uniform spatial index, perform multi-resolution structure organization and tile data storage on the encoded image slice tile data to obtain the inspection image compensation slice result of the target UAV.

[0033] In practical applications, the original high-resolution images of the inspection area are loaded into the memory of the processing system. Since there is usually a large amount of useless information on both sides of the images when the drone inspects along pipelines or roads, the invalid areas are cropped by using the mask of the effective image coverage area in the image slicing parameter data, that is, the spatial intersection calculation is performed between the original image and the geographical space mask of the effective coverage area, the pixel data within the mask is retained, and the redundant data outside the mask is discarded. In addition, the cropped images are non-uniformly sliced according to the slicing strategy in the image slicing parameter data. For the areas determined to be strip-shaped, long-strip-shaped tiles are used for cutting, such as a fixed width of 512 pixels but the height is dynamically adjusted according to the height of the effective area; for non-strip-shaped areas, standard square tiles are used for cutting. Through this targeted slicing strategy, a series of image slice tile data with different shapes and sizes are generated. The total amount of these tile data is significantly reduced compared with traditional uniform slicing, and the data volume can be reduced by 30%-70%. Then, each image slice tile is identified and encoded. By combining the geographical location of the tile (usually the coordinates of the center point or the upper left corner point of the tile) and the zoom level (indicating the resolution level of the tile), a unique identification code is generated (this code adopts a quadtree or a similar spatial index structure, which can support fast spatial queries and multi-resolution loading). In addition, the inspection main axis direction corresponding to the strip-shaped target area is identified, that is, by analyzing the geometric shape of the effective coverage area and calculating its main direction vector. For targets such as pipelines or roads, this main axis direction usually follows the extension direction of the target. Thus, based on the identified inspection main axis direction, a non-uniform spatial index corresponding to the index direction of the strip-shaped target area is determined. This index structure has a finer division in the main axis direction and is relatively rough in the vertical direction, so as to better adapt to the spatial distribution characteristics of strip-shaped targets. Then, based on the constructed non-uniform spatial index, the encoded image slice tile data is organized into a multi-resolution structure, and a pyramid structure with different resolutions is generated for each tile to support the zoom operation in the 3D visualization system. And this pyramid structure maintains a higher detail retention rate in the main axis direction and can be more aggressively downsampled in the non-main axis direction, so as to further reduce the data volume while ensuring the visual effect. Then, these data are stored according to the organization method of the spatial index, and the file system or the database system can be selected for storage, generating the final target drone inspection image compensation slice result, which can significantly improve the processing efficiency and storage utilization rate of strip-shaped geographical space data.

[0034] In the embodiments of the present invention, by extracting and processing the flight marker data and parameter data when the target UAV inspects a long strip target, basic flight parameters such as UAV coordinates, positioning compensation, and status are obtained; then, based on the positioning compensation data, the coordinates are superimposed and calculated to obtain corrected geographical coordinates, and an accurate camera imaging frustum model is constructed in combination with the flight inspection parameters; then, the frustum model is geometrically intersected with the long strip target area to obtain an effective image coverage area, and adaptive slicing parameters are determined based on the coverage area characteristics and positioning status; finally, the inspection area image is non-uniformly sliced according to the slicing parameters and stored in a multi-resolution structure with non-uniform spatial indexing in the main axis direction of the target area, and the final inspection image compensation slice result is output. Through dynamic adaptive data processing and geometric analysis, the contradiction between the slicing efficiency and resource consumption of long strip geographical space data is solved, especially for the 3D modeling and visualization scenarios of linear facilities such as pipelines and roads; a dynamic slicing method based on UAV pose parameters is adopted, which not only makes full use of the accurate compensation data provided by the RTK positioning system but also realizes the matching of the slice shape and the target form; in addition, through non-uniform spatial indexing and multi-resolution storage strategies, the redundant data storage amount and processing time are effectively reduced, and the efficient slicing processing of long strip target image data is overall realized.

[0035] The method for compensating and slicing UAV images in the embodiments of the present invention is described above. Next, the device for compensating and slicing UAV images in the embodiments of the present invention will be described. Please refer to Figure 2 , an embodiment of the device for compensating and slicing UAV images in the embodiments of the present invention includes: A data acquisition module 201, configured to acquire UAV flight marker data, flight inspection parameter data, and inspection area images collected by the target UAV when inspecting a long strip target area, and extract various flight parameter data from the UAV flight marker data, where the flight parameter data includes UAV coordinate data, positioning compensation data, and positioning status data; A frustum construction module 202, configured to perform coordinate superposition calculation on the UAV coordinate data based on the positioning compensation data to obtain corrected geographical coordinates, and construct a camera imaging frustum model of the target UAV based on the flight inspection parameter data and the corrected geographical coordinates; An area calculation module 203, configured to perform geometric calculation of the intersection area between the camera imaging frustum model and the long strip target area to obtain an effective image coverage area, and determine image slicing parameter data of the target UAV based on the effective image coverage area and the positioning status data; The slice storage module 204 is configured to perform non-uniform tile cutting on the inspection area image based on the image slice parameter data, generate inspection image slice data, and perform multi-resolution structure storage with non-uniform spatial indexing on the inspection image slice data based on the inspection main axis direction corresponding to the long strip target area, so as to obtain the inspection image compensation slice result of the target UAV.

[0036] In the embodiment of the present invention, by extracting and processing the flight mark data and parameter data when the target UAV inspects a long strip target, basic flight parameters such as UAV coordinates, positioning compensation, and status are obtained; then, based on the positioning compensation data, the coordinates are superimposed and calculated to obtain corrected geographical coordinates, and an accurate camera imaging frustum model is constructed in combination with the flight inspection parameters; then, the frustum model is subjected to geometric intersection calculation with the long strip target area to obtain an effective image coverage area, and adaptive slice parameters are determined according to the coverage area characteristics and positioning status; finally, non-uniform tile cutting is performed on the inspection area image according to the slice parameters, and multi-resolution structure storage with non-uniform spatial indexing is performed in the main axis direction of the target area, and the final inspection image compensation slice result is output. Through dynamic adaptive data processing and geometric analysis, the contradiction between the slice efficiency and resource consumption of long strip geographical space data is solved, especially for the 3D modeling and visualization scenarios of linear facilities such as pipelines and roads; a dynamic slicing method based on UAV pose parameters is adopted, which not only makes full use of the accurate compensation data provided by the RTK positioning system, but also realizes the matching between the slice shape and the target form; in addition, through non-uniform spatial indexing and multi-resolution storage strategies, the redundant data storage amount and processing time are effectively reduced, and the efficient slicing processing of long strip target image data is realized as a whole.

[0037] Above Figure 2 The UAV image compensation slicing device in the embodiment of the present invention is described in detail from the perspective of modular functional entities. Next, the UAV image compensation slicing device in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0038] Figure 3It is a schematic structural diagram of a drone image compensation slicing device provided by an embodiment of the present invention. The drone image compensation slicing device 300 may vary greatly due to configuration or performance differences, and may include one or more processors (central processing units, CPU) 310 (for example, one or more processors) and a memory 320, and one or more storage media 330 for storing application programs 333 or data 332 (for example, one or more mass storage devices). Among them, the memory 320 and the storage media 330 may be transient storage or persistent storage. The program stored in the storage media 330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the drone image compensation slicing device 300. Further, the processor 310 may be configured to communicate with the storage media 330 and execute a series of instruction operations in the storage media 330 on the drone image compensation slicing device 300.

[0039] The drone image compensation slicing device 300 may further include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input / output interfaces 360, and / or, one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art can understand that Figure 3 The shown structural diagram of the drone image compensation slicing device does not limit the drone image compensation slicing device, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0040] The present invention also provides a drone image compensation slicing device. The computer device includes a memory and a processor. When the computer-readable instructions stored in the memory are executed by the processor, the processor executes each step of the drone image compensation slicing method in the above embodiments.

[0041] The present invention also provides a computer-readable storage medium. The computer-readable storage medium may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions run on a computer, the computer executes each step of the drone image compensation slicing method.

[0042] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0043] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0044] This application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This application can also be practiced in a distributed computing environment, where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0045] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for compensating and slicing drone images, characterized in that: The drone image compensation slicing method comprises: Acquire the drone flight mark data, flight inspection parameter data and inspection area image collected by the target drone when inspecting the long strip target area, and extract multiple flight parameter data from the drone flight mark data, wherein the flight parameter data includes drone coordinate data, positioning compensation data and positioning status data; Based on the positioning compensation data, coordinate superposition calculation is performed on the coordinate data of the UAV to obtain corrected geographic coordinates, and based on the flight inspection parameter data and the corrected geographic coordinates, a camera imaging cone model of the target UAV is constructed; Performing geometric calculation of the intersection area of ​​the camera imaging cone model and the long strip target area to obtain an effective image coverage area, and determining image slice parameter data of the target UAV based on the effective image coverage area and the positioning state data; Based on the image slicing parameter data, the inspection area image is non-uniformly tiled to generate inspection image slicing data, and based on the inspection main axis direction corresponding to the long strip target area, the inspection image slicing data is stored in a multi-resolution structure with non-uniform spatial indexing to obtain the inspection image compensation slicing result of the target UAV.

2. The drone image compensation slicing method according to claim 1, characterized in that: The extracting of a plurality of flight parameter data from the UAV flight mark data, wherein the flight parameter data includes UAV coordinate data, positioning compensation data and positioning status data, includes: Based on the preset data format specification and field structure, each record data in the UAV flight mark data is preliminarily parsed to obtain the preliminarily parsed flight mark data; A variety of coordinate parameter data are extracted from the flight mark data after preliminary analysis to obtain UAV coordinate data, and multi-directional compensation parameter data are extracted from the flight mark data after preliminary analysis to obtain positioning compensation data, and the positioning state Q value corresponding to each record data in the flight mark data after preliminary analysis is extracted to obtain a positioning accuracy index, and the north standard deviation and east standard deviation corresponding to each record data in the flight mark data after preliminary analysis are extracted to obtain horizontal accuracy data, and positioning state data is generated based on the positioning accuracy index and the horizontal accuracy data.

3. The drone image compensation slicing method according to claim 2, characterized in that: The positioning compensation data includes a north compensation value, an east compensation value and an elevation compensation value. Based on the positioning compensation data, coordinate superposition calculation is performed on the coordinate data of the UAV to obtain corrected geographic coordinates, including: Determine the coordinate system type corresponding to the coordinate data of the drone, and based on the coordinate system type, perform a unified coordinate system conversion on the coordinate data of the drone to obtain coordinate data under a unified geographic coordinate system, and perform preliminary correction calculation on the north compensation value, the east compensation value, and the elevation compensation value to obtain a preliminary corrected north compensation value, a preliminary corrected east compensation value, and a preliminary corrected elevation compensation value; algebraically adding the latitude value corresponding to the coordinate data in the unified geographic coordinate system to the initially corrected north compensation value to obtain a compensated latitude value, algebraically adding the longitude value corresponding to the coordinate data in the unified geographic coordinate system to the initially corrected east compensation value to obtain a compensated longitude value, and algebraically adding the elevation value in the unified geographic coordinate system to the initially corrected elevation compensation value to obtain a compensated elevation value; The compensated latitude value, the compensated longitude value and the compensated elevation value are combined to form the corrected geographic coordinates, and the corrected geographic coordinates are compared and calculated with the coordinate data under the unified geographic coordinate system to generate the corrected geographic coordinates along the inspection path corresponding to the long strip target area.

4. The drone image compensation slicing method according to claim 1, characterized in that: The step of constructing a camera imaging cone model of the target UAV based on the flight inspection parameter data and the corrected geographic coordinates includes: Extracting the corresponding pitch angle, roll angle and heading angle data in the flight inspection parameter data to obtain the three-dimensional attitude parameters of the target UAV, and reading the corresponding camera focal length, field of view angle and sensor size parameters in the flight inspection parameter data to obtain the camera imaging geometric parameters, and performing coordinate system conversion on the corrected geographic coordinates and the three-dimensional attitude parameters to obtain the camera coordinate system of the corresponding origin position of the target UAV; Performing rotation matrix calculation on the three-dimensional posture parameters to obtain a camera optical axis direction vector and a plurality of cone corner point vectors, and performing cone geometry parameter calculation on the camera imaging geometry parameters to obtain cone near plane and far plane parameters, and constructing space equations for the cone near plane and far plane parameters and the cone corner point vectors to obtain four side equations of the cone; The geometric relationship between the camera optical axis direction vector, the view cone corner point vector and the four side face equations of the view cone is integrated to obtain the topological structure data of the view cone vertices and faces, and the topological structure data and the camera coordinate system information are integrated for inspection space positioning to construct a camera imaging view cone model.

5. The drone image compensation slicing method according to claim 1, characterized in that: The geometric calculation of the intersection area of ​​the camera imaging cone model and the long strip target area to obtain the effective image coverage area includes: Performing triangular mesh division on the surface elevation information data corresponding to the long strip target area to generate a regional triangular mesh model, and performing a three-dimensional geometric intersection operation on the camera imaging cone model and the regional triangular mesh model to obtain a regional spatial intersection line between the cone and the surface; Extracting the projection contour of the regional spatial intersection line to generate the corresponding cone projection polygon on the surface of the long strip target area, and converting the cone projection polygon into a unified coordinate system based on the original image coordinate system corresponding to the inspection area image to obtain a geographic space polygon; The numerical values ​​of various geometric parameters of the geographic space polygon are calculated to obtain the spatial distribution characteristics of the long strip target area, and the spatial distribution characteristics are subjected to vector boundary extraction and pixel-level mask conversion to generate an effective image coverage area.

6. The drone image compensation slicing method according to claim 2, characterized in that: The determining of the image slice parameter data of the target UAV based on the effective image coverage area and the positioning status data includes: The positioning state data whose positioning state Q value is equal to the preset positioning state accuracy threshold is marked with a fixed solution state data to obtain a state marking result, and the north standard deviation and the east standard deviation in the positioning state data are detected to determine whether they are greater than a preset standard deviation critical value to obtain a standard deviation detection result; Based on the status marking result and the standard deviation detection result, determining the slice resolution corresponding to each inspection shooting position in the effective image coverage area, and judging whether the area aspect ratio value corresponding to the effective image coverage area is greater than a preset aspect ratio threshold; If the area aspect ratio value corresponding to the effective image coverage area is greater than a preset aspect ratio threshold, determining that the effective image coverage area is a long strip image area, and adjusting the value of the first image tile parameter corresponding to the effective image coverage area to generate a first slice adjustment result of the long strip image area; If the area aspect ratio value corresponding to the effective image coverage area is not greater than a preset aspect ratio threshold, determining that the effective image coverage area is a non-strip image area and determining a second image tile parameter value corresponding to the effective image coverage area, and generating a second slice adjustment result of the non-strip image area; Image slice parameter data of the target UAV is integrated and generated based on the status marking result, the slice resolution, the first slice adjustment result, and the second slice adjustment result.

7. The drone image compensation slicing method according to claim 1, characterized in that: Based on the image slice parameter data, the inspection area image is non-uniformly tiled to generate inspection image slice data, and based on the inspection main axis direction corresponding to the long strip target area, the inspection image slice data is non-uniformly spatially indexed and stored in a multi-resolution structure to obtain the inspection image compensation slice result of the target UAV, including: Based on the image slice parameter data, the inspection area image is cropped and non-uniformly sliced ​​to obtain a plurality of image slice tile data; Performing identification encoding of the geographic location and zoom level on each of the image slice tiles to obtain encoded image slice tile data, and identifying the inspection main axis direction corresponding to the long strip target area, and determining the non-uniform spatial index of the index direction corresponding to the long strip target area based on the inspection main axis direction; Based on the non-uniform spatial index, the encoded image slice tile data is multi-resolution structured and stored to obtain the inspection image compensation slice result of the target UAV.

8. A drone image compensation slice device, characterized in that: The drone image compensation slicing device comprises: A data acquisition module is used to acquire the drone flight mark data, flight inspection parameter data and inspection area image collected by the target drone when inspecting the long strip target area, and extract a variety of flight parameter data from the drone flight mark data, wherein the flight parameter data includes drone coordinate data, positioning compensation data and positioning status data; A cone construction module, used to perform coordinate superposition calculation on the coordinate data of the UAV based on the positioning compensation data to obtain corrected geographic coordinates, and to construct a camera imaging cone model of the target UAV based on the flight inspection parameter data and the corrected geographic coordinates; An area calculation module is used to perform geometric calculation of the intersection area of ​​the camera imaging cone model and the long strip target area to obtain an effective image coverage area, and determine the image slice parameter data of the target UAV based on the effective image coverage area and the positioning state data; The slicing storage module is used to perform non-uniform tile cutting on the inspection area image based on the image slicing parameter data to generate inspection image slicing data, and based on the inspection main axis direction corresponding to the long strip target area, perform non-uniform spatial index multi-resolution structure storage on the inspection image slicing data to obtain the inspection image compensation slicing result of the target UAV.

9. A drone image compensation slice device, characterized in that: The drone image compensation slicing device comprises: a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instructions in the memory so that the drone image compensation slicing device performs each step of the drone image compensation slicing method as described in any one of claims 1-7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the various steps of the drone image compensation slicing method as described in any one of claims 1 to 7 are implemented.

Images