670results about "Picture interpretation" patented technology

The invention discloses a robot gait self-adaptive adjustment method and system combined with topographic feature recognition, and relates to the technical field of robot gaits, and the method comprises the steps: collecting a topographic image in front of a robot advancing path, carrying out the deblurring processing, and extracting ground feature data through a lightweight neural network; matching the ground characteristic data with a preset gait control parameter database to obtain a gait control parameter set, and generating a control signal through a proportional-differential control method according to the gait control parameter set; correcting the control signal based on a reinforcement learning algorithm, and generating a target control instruction for driving a joint actuator; in the process of executing the target control instruction, collecting state feedback data in real time, and comparing and optimizing the feedback data with the currently executed target control instruction to realize feedback adjustment; the problems of insufficient terrain recognition precision caused by image blurring, response lag caused by sensing and control decoupling and lack of adaptive ability are effectively solved.

The invention relates to the technical field of goods shelf detection, in particular to a self-detection method and device for goods shelf settlement, and provides the following scheme: obtaining a top view image through an image sensor arranged right above the top of a goods shelf, dividing the image into a plurality of grid units, and positioning a rectangular geometric shape by utilizing Hough transform; and screening a plurality of to-be-detected areas in combination with the edge features. For an area to be measured, homographic registration and ortho-rectification are carried out based on a reference image, a displacement field is obtained by adopting sub-pixel-level dense registration, and a geometric parallax component field corresponding to imaging parameters is obtained through robust estimation. And under the hypothesis of small deformation, inverting the parallax into a pixel normal distance, and carrying out weighted aggregation on the local region to obtain a local distance measurement result. And by iteratively combining adjacent grids, determining a settlement area boundary, and finally outputting a settlement detection result. Millimeter-level settlement quantification can be realized under a single-frame image, hardware transformation is avoided, and the method is suitable for automatic detection and long-term monitoring of multi-specification goods shelves.

The invention belongs to the technical field of glacier identification, and provides a system for monitoring glacier dynamics by an unmanned aerial vehicle low-altitude flight photogrammetry method, which comprises the following steps that: a glacier area unmanned aerial vehicle low-altitude photogrammetry module selects an unmanned aerial vehicle as required, and plans a route and parameters; the image preprocessing module is used for extracting key feature points from each image by using Pi * 4Dcap; fine three-dimensional information, an ortho-image and a glacier digital surface model DSM are obtained through point cloud generation and densification, and glacier dynamic information can be identified by utilizing the glacier digital surface model DSM after correction and interpolation processing are carried out on the three-dimensional information, the ortho-image and the glacier digital surface model DSM; calculating the elevation, volume, speed and mass change of the glacier based on a digital surface model (DSM) of the glacier; the glacier fissure is identified by utilizing the orthoimage and the glacier digital surface model DSM, and the system can comprehensively and accurately monitor the dynamic information of the glacier.

The invention relates to the technical field of water conservancy projects, in particular to a water conservancy earthwork dynamic monitoring method and system based on multi-source point cloud data fusion, and the method comprises the steps: carrying out the identification and standardization processing of airborne LiDAR, ground scanning and oblique photography point clouds, and generating a standard point cloud data set with uniform spatial positions and balanced density; dividing an abrupt slope area, a gentle slope area and a flat area based on topographic gradients, and constructing a grid framework adaptive to topographic features; carrying out weighted fusion by combining the point cloud coverage rate and the normalized terrain gradient to generate high-precision grid elevation data; calculating an earthwork variation distribution matrix by comparing historical elevation models; dynamically analyzing the change trend of the excavation and filling area, and synchronously evaluating the construction progress deviation and the geological disaster risk. According to the method, the problems of insufficient multi-source point cloud fusion precision, poor terrain adaptability and dynamic evaluation splitting in the prior art can be solved, and the monitoring precision and the engineering safety early warning efficiency under the complex terrain condition are remarkably improved.

A method for surveying an environment by a movable surveying instrument with a progressional capturing of 2D-images by at least one camera and applying a visual simultaneous location and mapping algorithm (VSLAM) or a visual inertial simultaneous location and mapping algorithm (VISLAM) with a progressional deriving of a sparse evolving point cloud of at least part of the environment, and a progressional deriving of a trajectory of movement. The method comprises a progressional matching of the sparse evolving point cloud with a previously derived 3D-geometry, with a minimizing of a function configured to model a distance between the sparse point cloud and the previously derived 3D-geometry and deriving a spatial localization and orientation of the surveying instrument. At least one surveying measurement value of the environment by a spatial measurement unit is combined with the sparse point cloud or the previously derived 3D-geometry.

The invention discloses a multi-agent collaborative visual navigation reasoning enhancement method and system, and the method comprises the steps: an inference agent analyzes input data, and generates structured output containing a reasoning process and an initial answer; independently analyzing the input data by the evaluator agent, generating an evaluation answer and reviewing the output of the inference agent; comparing whether the initial answer is consistent with the evaluation answer, and if yes, outputting a final answer; if not, calling a bifurcation feedback agent, and generating a core bifurcation point abstract and a visual evidence verification plan; and calling a final decision maker agent to directionally retrieve the video evidence according to the visual evidence verification plan, and outputting a final answer. Four roles of an inference person, an evaluator, a divergence feedback person and a final decision maker are introduced, a complete'generation-evaluation-feedback-optimization 'decision link is constructed, and the cognition and inference accuracy in a complex scene is remarkably improved.

The invention relates to the technical field of geographic information engineering, in particular to a surveying and mapping geographic system based on a remote sensing technology, which is characterized in that a multi-modal remote sensing data acquisition module synchronously acquires satellite remote sensing images, aviation LiDAR point cloud, unmanned aerial vehicle oblique photography and ground Internet of Things sensing data; the data fusion collaboration module carries out geographic coordinate calibration and position coding, and generates a surveying and mapping data set with unified precision in combination with an improved Mamba fusion model; the dynamic terrain high-precision inversion module performs three-dimensional terrain reconstruction, and completes terrain deformation monitoring in combination with an InSAR interference measurement technology and a time sequence analysis algorithm; the real-time error correction module constructs a BP neural network optimized by an ISSA algorithm, and dynamically corrects a terrain three-dimensional model and a terrain deformation result; the intelligent visual application module constructs a digital twinborn simulation model and automatically generates a standardized surveying and mapping result report. Therefore, the problems of weak multi-source data fusion, low terrain inversion precision and the like in the prior art are solved.

The invention discloses a mining area subsidence monitoring method based on unmanned aerial vehicle photogrammetry and LiDAR technology, and the method comprises the steps: carrying out the flight shooting of a mining area through employing an unmanned aerial vehicle, and obtaining the ground surface image data of the mining area; processing the acquired image data through a motion recovery structure algorithm SfM to generate a digital surface model DSM, an orthographic map DOM and a digital elevation model DEM; the method comprises the following steps: scanning a mining area by using an airborne LiDAR system to obtain three-dimensional point cloud data of the surface of the mining area; the differential GNSS technology is adopted to correct LiDAR point cloud data and improve the precision of the LiDAR point cloud data; performing registration fusion on a digital elevation model (DEM) generated by unmanned aerial vehicle photogrammetry and the corrected LiDAR point cloud data in a unified coordinate system; the fused data are processed and analyzed, a mining area subsidence model is constructed, and the mining area surface subsidence condition is dynamically monitored. According to the method, the problems existing in a traditional mining area subsidence monitoring method are effectively solved, and powerful technical support is provided for mine geological disaster prevention and control, ecological environment protection and mining area sustainable development.

The invention provides a method and a device for installing, positioning and monitoring a laminated slab in a beam. Relates to the technical field of building construction. According to the specific implementation scheme, the method comprises the steps that reverse modeling is conducted on a prefabricated laminated slab component through laser scanners placed at multiple target positions, and a point cloud model is generated; inputting preset data of the prefabricated laminated slab component into the building information model to generate a target preset model; and the point cloud model is matched with a target preset model, and when it is detected that the matching degree of the point cloud model and the target preset model is smaller than a first preset threshold value, multiple installation point positions for beam-entering installation of the prefabricated laminated slab component and normal state parameters of the multiple installation point positions are determined based on the point cloud model and the target preset model. According to the technical scheme, the accuracy and construction efficiency of installation and positioning of the laminated slab into the beam can be improved, structural hidden dangers caused by installation dislocation are avoided, it is guaranteed that the laminated slab and the beam form a reliable overall stress system, and the stability and safety of a building structure are enhanced.

The invention discloses an unmanned aerial vehicle multispectral system error calibration and correction method based on a Boolean model, and belongs to the technical field of photogrammetry and remote sensing. Comprising the following steps: acquiring a multispectral image, acquiring original POS data, and performing band registration preprocessing on the image to generate a first orthoimage to confirm a system error magnitude; carrying out bundle method strict adjustment by using ground control points to obtain reference POS data, and solving seven-parameter conversion parameters; correcting the original POS data by using seven parameters and generating a high-precision orthoimage; and generating a simplified multispectral camera calibration and verification formula according to the difference value of the POS data before and after correction. Complex coordinate conversion is simplified into fixed value addition and subtraction operation, one-button calibration and batch correction of system errors of the multi-spectral camera of the unmanned aerial vehicle are achieved, the positioning precision is improved to the centimeter level from the meter level, dependence on ground control points is greatly reduced, and an efficient standardized solution is provided for large-scale engineering data processing.

This invention relates to a system and method for non-invasive detection of gravesites and archaeological features using multimodal remote sensing and machine learning. Remotely sensed datasets, including RGB, multispectral, hyperspectral, LiDAR, and thermal imagery, are orthorectified, mosaicked, and subdivided into tiled image segments. Features are labeled through manual annotation of visible markers and environmental signatures and expanded via iterative augmentation. A supervised pipeline trains computer vision models, such as YOLO-based detectors, in parallel with tabular models derived from spectral indices (NDVI, NDRE), LiDAR elevation derivatives, and thermal anomalies. Inference outputs are cross-validated against thresholded evidence layers to reject false positives and upgraded when spectral, spatial, and thermal evidence align. Validated detections are exported as GIS-compatible layers with confidence scores and metadata. The system provides a scalable, replicable tool supporting archaeologists, Indigenous communities, and planners in cemetery investigations, cultural resource management, and humanitarian searches for unmarked or clandestine graves.

The invention relates to the technical field of highway engineering safety monitoring, in particular to a highway subgrade three-dimensional deformation monitoring system based on unmanned aerial vehicle oblique photography, and the system comprises an unmanned aerial vehicle oblique photography module which adopts a five-lens synchronous shooting strategy to realize the omni-directional coverage of a subgrade; the three-dimensional point cloud generation engine generates a high-precision three-dimensional point cloud model through feature extraction, registration and triangulation strategies; the deformation field analysis platform models a roadbed surface into a parameterized Riemannian manifold based on a differential geometry theory, captures deformation dynamic characteristics through multi-scale curvature flow analysis, and identifies potential risks by using topological evolution characteristics of deformation singular points; and the risk evaluator evaluates the risk level based on the singular point density, the change rate and the topological stability and generates early warning information. According to the invention, millimeter-level accurate monitoring of roadbed deformation is realized, various deformation modes from micro-cracks to large-scale settlement can be captured, the potential landslide risk is early warned 30-45 days in advance, and the false alarm rate is significantly reduced.

The invention discloses an image processing and measuring method and system based on unmanned aerial vehicle radar and oblique photography, relates to the technical field of oblique photography image measurement, and is used for solving the problem that the accuracy of region division and anomaly detection results is reduced. The method comprises the following steps: calculating a vegetation change gradient in combination with a radar echo amplitude to generate vegetation distribution characteristics; extracting an under-forest ground and generating uniform extrapolation points; arranging a clinometer to obtain an elevation and a coordinate; calculating a local slope and a crown fluctuation amplitude; according to the method, the area vegetation coverage state is evaluated, abnormal areas are screened, extrapolation ground points of the abnormal areas are marked and output, high-precision exploration of the forest area is achieved, terrain complexity and vegetation distribution characteristics are accurately reflected, reliable data support is provided for forest management and terrain analysis, and exploration efficiency and precision are improved.

The invention provides a data processing method and device based on a smart city, and relates to the technical field of data processing.The method comprises the steps that a city-level spatial data base is constructed through oblique photography of an unmanned aerial vehicle, and a unified and real spatial foundation is formed by generating a live-action three-dimensional model, a digital orthoimage and a digital elevation model; on this basis, a building base and a building white model are extracted, spatial logic verification is carried out on multi-source business basic data, a standard address system stably associated with a building three-dimensional entity is further constructed, precise spatial anchoring of governance objects such as population, houses and units is achieved, and the method is suitable for mass production. And finally, the associated data is uniformly converged to a city information model platform, and reliable and updatable spatial data support is provided for smart city application. According to the invention, the timeliness of smart city data processing in a dynamic earth surface deformation scene can be improved.

The invention provides a cable-stayed bridge cross brace hoisting monitoring and early warning method, and the method comprises the steps: building a finite element model based on the design parameters of a cable-stayed bridge and a cross brace, and calculating the safety envelope of the displacement, rotation angle and stress of a control point; displacement, strain and cable force sensors and industrial cameras are arranged in a cross brace and an anchoring area, and data and images are collected. Inputting the image into a first convolutional neural network model, extracting attitude parameters such as vertical and lateral offsets and torsion angles of each control point, and adapting to different imaging scales through a scale adaptive convolution module; and aligning and splicing the attitude parameters with sensor displacement, strain and cable force indexes according to time to form a feature sequence, inputting the feature sequence and safety envelope parameters into a second convolutional neural network risk assessment model, outputting a comprehensive risk index reflecting deviation from safety envelope, and generating construction treatment instructions such as hoisting speed reduction and hoisting point stress adjustment. According to the monitoring and early warning method, the construction safety is greatly improved, and monitoring and early warning of the whole process of cross brace hoisting are achieved.

The invention discloses a visual imaging method and system combined with unmanned aerial vehicle multi-temporal automatic comparison, and the system comprises an unmanned aerial vehicle aerial survey module which is used for carrying out the repeated aerial survey of the same target region at different time points, and obtaining multi-period image data; the data transmission and storage module is used for transmitting the multi-period image data acquired by the unmanned aerial vehicle aerial survey module in real time; the method has the beneficial effects that the aerial survey of the unmanned aerial vehicle supports dynamic route adjustment, multi-sensor cooperative acquisition is realized, the overlap ratio of multiple aerial survey paths is high, the image data integrity and consistency are high, and the manual intervention and data supplementary acquisition cost is greatly reduced; multi-stage image automatic registration, feature extraction and change detection are realized through deep learning, the processing precision is high, a large amount of manual operation is not needed, and the efficiency is remarkably improved; three-dimensional modeling and multi-mode comparison are supported, dynamic changes of ground objects can be clearly displayed, interaction functions are rich, and the requirements for single-person detail checking and multi-person collaborative rechecking can be met.

The invention relates to a target tracking optimization method and system based on image recognition laser ranging. According to the method, when the intensity of a signal returned by laser ranging operation is lower than a preset threshold value, a candidate target is determined as a non-target and a non-target mark is added, and in subsequent image processing, when a candidate target associated with the non-target mark is identified again, the identification confidence of the candidate target is reduced, so that the identification confidence is lower than a threshold value for triggering ranging. According to the method and the system, repeated laser ranging operation on a non-target object can be avoided, so that unnecessary computing resource consumption is reduced, the efficiency and the accuracy of the system for tracking a real target are improved, particularly, tracking interruption can be effectively avoided when the target moves highly dynamically, and the robustness and the reliability of the system are improved.

The invention relates to the technical field of intelligent inspection of highway facilities, in particular to an unmanned aerial vehicle aerial photography highway separation gate damage identification method, which comprises the following steps of: firstly, generating an imitated flight route matched with linear distribution of a separation gate based on GIS (Geographic Information System) geographic information and a digital elevation model of the highway separation gate; then the unmanned aerial vehicle is controlled to autonomously fly along a route, a high-resolution sequence image containing an isolation fence stand column and a mesh is collected at a preset equidistant shooting point, and positioning and attitude determination data are synchronously recorded; a pre-training deep learning model is used for detecting the inclination angle of the stand column and the mesh damage area, and abnormity is marked; and finally, fusing the positioning and attitude determination data and the identification result, calculating the three-dimensional geographic coordinates of the abnormal points through a space forward intersection algorithm, and generating an isolation gate anomaly report with the geographic coordinates. The problems of low efficiency and poor precision of traditional inspection are solved, the abnormity identification and positioning precision is improved, and a precise basis is provided for isolation gate maintenance.

The invention discloses a space-based double-satellite asynchronous detection multi-target correlation positioning system and method, the system comprises space-based optical satellites and a central processing unit, each space-based optical satellite is equipped with a large-view-field optical electro-telescope, and the large-view-field optical electro-telescope is used for carrying out image acquisition on a space target; and the central processing unit comprises a computer, the computer is in communication connection with the space-based optical satellite, and an image processing module, an angle measurement data generation module, a data association module and a target positioning module are integrated in the computer. According to the characteristics of space-based double-star asynchronous observation, a complete process from image processing, angle measurement data generation, time series data association to final positioning is designed, and the method is particularly suitable for monitoring and positioning a large number of space non-cooperative moving targets; by introducing time registration and sequence geometric consistency judgment, the association problem caused by asynchronous double-star observation time is effectively solved, and the accuracy and robustness of data association in a multi-target scene are improved.

The invention provides an underwater terrain inversion method based on a height measurement satellite, belongs to the technical field of remote sensing geoscience application, and constructs an automatic processing flow integrating space screening, elevation correction, signal denoising, double-strategy surface and bottom recognition and physical correction. The method is suitable for underwater topography generation in complex water body environments such as lakes and reservoirs, and the efficiency of underwater topography data acquisition is greatly improved. A strategy of combining adaptive kernel density estimation and clustering analysis is adopted to remarkably improve the water surface and water bottom signal recognition capability, and a multi-temporal collaborative analysis and refractive index physical correction model is introduced to realize high-precision underwater terrain extraction of water bodies such as inland reservoirs. The method has obvious advantages in the aspects of automation degree, application range and result stability, and has important engineering application and popularization value.

The invention relates to the technical field of geological disaster monitoring, and discloses an unmanned aerial vehicle aerial survey intelligent identification and grading method and system for geological disaster hidden danger points, and the method comprises the steps: converting a pixel-level identification result into hierarchical data which retains a spatial topological structure through constructing a confidence quad-tree; establishing a weighted decision grid based on aerial survey parameters to realize coupling modeling of observation conditions and confidence uncertainty, calculating topology difference to generate a spatial difference field to quantify the deviation degree of a confidence field and a standard grading decision surface, and finally executing self-adaptive grading judgment to output a risk grade result with a confidence label; according to the method, the problem of confidence field and decision boundary space topology mismatch caused by truncation of confidence continuous measurement information in a grading stage in a traditional serial architecture is solved, systematic underestimation of misjudgment risks in observation condition limited areas such as alpine and gorge areas is avoided, and consistency of actual credibility of a risk quantification result and a nominal grade is realized.

The invention discloses a method and a system for constructing a flight three-dimensional model in ground disaster monitoring in a mountain and valley region, and belongs to the technical field of flight three-dimensional model modeling. The construction method comprises the steps of preparation before aerial photo image data acquisition, aerial photo image data acquisition, aerial photo image data processing and aerial photo image data aerial triangulation operation, in the preparation before aerial photo image data acquisition, a measurement area is divided into a plurality of area blocks, image control points are arranged in the area blocks, the adjacent area blocks have common image control points, and the common image control points are arranged in the area blocks. In aerial photo image data processing, aerial photo image data of each region block is stored as an independent data set, data set attention grading is carried out on each data set according to the attention degree of the corresponding region block, and a resampling method with higher image sampling quality is adopted for the data set with higher attention grading. According to the scheme, the three-dimensional modeling efficiency can be effectively improved, and ground disaster monitoring based on the three-dimensional model can be carried out as soon as possible.

The invention relates to the technical field of photogrammetry, and discloses a photogrammetry space reconstruction method based on multi-source image fusion, which comprises the following steps: extracting an interior orientation element and an initial exterior orientation element, and establishing an image beam intersection constraint intensity model in an object space to calculate a constraint weight; monitoring a weight space change rate to identify a heterogeneous image connection area and extracting a high-weight image bundle as a geometric anchor point; a differential regularization item is applied to geometric anchor points to construct a stable geometric skeleton, a weight damping factor is utilized to restrain a low-weight image beam to execute conformal mapping correction, the image beam constraint strength is represented, the condition number of an adjustment method equation matrix is improved, numerical oscillation caused by difference of different-source sensors is restrained, geometric step of a fusion interface is eliminated, and the conformal mapping correction accuracy is improved. And geometric closing and seamless stitching of a multi-source image under a unified framework are realized.

The present application relates to a kind of based on unmanned aerial vehicle photogrammetry suspension bridge sling cable force fast identification method, by being laid on the three-dimensional laser scanner of suspension bridge and collecting the point cloud data of suspension bridge, the intersection coordinates of main cable and sling and the form of main cable are obtained after the point cloud data collected by point cloud splicing and point cloud processing, and the intersection coordinates obtained are temperature corrected, then the intersection coordinates after correction are extended according to the relationship of pixel accuracy and the distance of unmanned aerial vehicle from main cable, obtain unmanned aerial vehicle flight point, to complete the planning of unmanned aerial vehicle route, the orthographic image of main cable section is collected by unmanned aerial vehicle, then image processing method is used to obtain the intersection coordinates of main cable and sling edge line, obtain the intersection coordinates of main cable and sling, obtain cable force calculation parameter, calculate sling force according to cable force calculation formula, based on unmanned aerial vehicle photogrammetry suspension bridge sling cable force fast identification method can make the daily cable force detection of suspension bridge more quickly, accurately.

The invention relates to a three-dimensional map surveying and mapping method, device and equipment based on an unmanned aerial vehicle and a medium. The method comprises the following steps: acquiring a global remote sensing image of a to-be-mapped area, and performing scene semantic analysis on the global remote sensing image to obtain scene semantic information; the global remote sensing image is collected by a remote sensing satellite; generating an initial flight path according to the scene semantic information, and obtaining three-dimensional surveying and mapping data collected by the unmanned aerial vehicle along the initial flight path; the three-dimensional surveying and mapping data comprises multi-view image data and corresponding positioning and attitude determination data; generating an initial point cloud of the to-be-mapped area based on the three-dimensional mapping data; carrying out ground feature category division on the initial point cloud to generate a hierarchical semantic point cloud; and generating a three-dimensional map model corresponding to the to-be-mapped area based on the hierarchical semantic point cloud. By adopting the method, the layered three-dimensional map with surface feature semantic information can be generated, and dynamic optimization can be realized.

The invention discloses a substation operation scheme compilation method based on a three-dimensional model, and relates to the technical field of power operation maintenance, and the method comprises the steps: collecting multi-source data, constructing a three-dimensional digital twinborn body, accessing real-time data, and forming a digital twinborn platform; querying an overhaul task and performing text structured processing to form an overhaul equipment set; carrying out special analysis and providing a special analysis result; the maintenance scene is displayed in the digital twin platform, and a maintenance operation scheme is generated and output; and after the maintenance operation is finished, field maintenance data are returned, and the digital twin platform data are updated. According to the method, the three-dimensional digital twinborn body which is synchronously mapped with the transformer substation and dynamically updated is constructed, so that a maintenance operation scheme is compiled based on a real-time field state, the problem that a traditional three-dimensional visualization system is subjected to static display and is disjointed with the field is solved, traditional tedious and error-prone steps depending on artificial experience are converted into automatic algorithm calculation, and the maintenance efficiency is improved. The time required for generating special analysis is greatly reduced, and the operation and maintenance cost is reduced.

The invention relates to the technical field of unmanned aerial vehicle photogrammetry, and discloses a three-dimensional model generation method based on unmanned aerial vehicle oblique photogrammetry. The method comprises the steps of obtaining an original oblique photography image set of an unmanned aerial vehicle and a terrain parameter set of a to-be-modeled area; performing multi-scale preprocessing on the original image set to obtain a first scale feature image set and a second scale feature image set; a textural feature evolution sequence is extracted from the second scale feature image set, shooting constraint initialization is completed in combination with the topographic parameter set, and an initial control constraint group is output; performing spatial feature reconstruction on the first scale feature image set according to the constraint group to obtain a three-dimensional topological feature vector set; performing multi-dimensional registration constraint on the vector set based on a similarity threshold value, and positioning an optimal registration path; performing topology correction by combining the path and the topographic parameter set, and outputting a corrected three-dimensional feature set; and if a preset condition is not met, incrementally updating the texture feature evolution sequence, and if the preset condition is met, outputting a target three-dimensional grid model.

The application discloses a high-precision earthwork measurement algorithm and system for constructing a three-dimensional model based on unmanned aerial vehicle aerial survey, and the method comprises the following steps: arranging image control points and check points in a to-be-measured area in advance; starting unmanned aerial vehicle aerial survey image data collection, and performing supplementary measurement on an obstacle-shielded area to obtain supplementary data; importing the aerial survey image data to generate a real scene three-dimensional model; finding an image control point mark position on the three-dimensional model and corresponding puncture points of actual measured image control points, converting model latitude and longitude coordinates into project coordinates after matching all the image control points, regenerating a three-dimensional model, and importing check point coordinate data to verify the model precision. Seed points are added to remove obstacles and restore the ground terrain, and then drawings and design elevations are imported to perform high-precision earthwork calculation, so that the problems of complex calculation operation, easy errors and large time consumption of manual operation combined with other software are solved, accurate data support is provided for earthwork construction, and the project construction period is saved.

The invention discloses an unmanned aerial vehicle adaptive three-dimensional reconstruction method based on semantic and geometry collaborative perception, and the method comprises the steps: obtaining image data collected by a binocular camera in real time, executing semantic segmentation in parallel to generate a real-time semantic graph, and calculating a geometric complexity graph based on a depth graph; according to a preset association rule base, fusing the real-time semantic graph and the geometric complexity graph, and identifying a region of interest; dynamically generating an optimized flight path on the basic route based on the position and the attribute of the region of interest; and controlling the unmanned aerial vehicle to fly according to the optimized flight path, adjusting sensor parameters to execute high-density data acquisition, and transmitting the acquired data and metadata to a ground station. According to the method, adaptive optimization of three-dimensional reconstruction of the unmanned aerial vehicle is realized by integrating semantic segmentation, geometric analysis and dynamic path planning.

The invention relates to the technical field of three-dimensional modeling, and discloses a building height extraction method and system based on an oblique photogrammetry model, which can significantly improve the precision and efficiency of extracting the building height from an OSGB oblique photogrammetry model, and realize automatic accurate identification and height calculation of a building main body structure. Through fusion of space geometric feature analysis and a deep learning model, the problems of reflection interference of a glass curtain wall, misjudgment of a complex roof structure and misreckoning of accessory structures such as a balcony and an equipment room into the height of a main body are effectively solved, and the reliability of a main body height extraction result is ensured. Meanwhile, subjective links such as manual measurement, manual outline sketching and empirical parameter adjustment are completely abandoned, personal error accumulation is greatly reduced, and the processing speed is remarkably increased.