121 results about "Aerial imagery" patented technology

The invention discloses an unmanned aerial vehicle aerial image target detection method based on PSO-DETR, and belongs to the technical field of unmanned aerial vehicle aerial image target detection. Firstly, a parallel patch perception attention feature extraction module is constructed, and an efficient multi-branch backbone network C3KCSPnet is designed by fusing a CSPDarknet53 structure; according to the network, gradient flow is improved through deep optimization, and the capturing capacity of high-level semantic information is enhanced. And secondly, an enhanced channel offset hybrid operator is provided, the dependency relationship between channels is enhanced through a channel shuffling mechanism, and cross-channel interaction of local space information is realized in combination with channel offset operation, so that the feature recovery quality and fusion efficiency in an up-sampling stage are improved, and the problem of missed detection of a shielded target is further relieved. And finally, a re-parameterization hierarchical aggregation network is designed, effective integration of shallow details and deep semantics is realized through an efficient hierarchical fusion mechanism on the premise of ensuring controllable calculation complexity, and the detection performance of the small target is further enhanced.

The invention relates to the technical field of reservoir storage capacity analysis, in particular to an intelligent reservoir storage capacity analysis method and system based on multi-source data fusion, and the method comprises the following steps: collecting ground water level sensor data, meteorological station data, LiDAR point cloud data and aerial image data, synchronizing the data through a unified timestamp, integrating the data, summarizing the integrated data, and carrying out the analysis of the reservoir storage capacity. And generating a comprehensive multi-source data set. According to the invention, through real-time updating and accurate tracking of daily reservoir capacity changes, the accuracy of reservoir capacity estimation is significantly improved, daily cumulative statistics of water level, rainfall and evaporation capacity data is integrated, dynamic and detailed management is promoted, seasonal changes and long-term trends of the reservoir water level are analyzed, and the reservoir capacity estimation accuracy is improved. The method can accurately reveal the direct effect of manual intervention and the influence of the manual intervention on the natural supply mode of the reservoir, can make a scientific decision based on a comprehensive data view, integrates short-term prediction and real-time data flow monitoring to enable an early warning system to quickly respond, and optimizes the efficiency and effectiveness of disaster management.

Computer vision systems and methods for detecting structures using aerial imagery and heightmap data are provided. The system receives aerial imagery and at least one heightmap, and merges the aerial imagery and the heightmap to create a combined image. The system determines one or more structures of the land property based at least in part on the combined image and a computer vision model, which can detect one or more objects in the combined image. The system can generate and place a bounding box or a polygon around each of the detected objects, and generate and assign a structure classification to the bounding box or the polygon to indicate the structure of the object. The system can also determine a geographic location of each structure using the two-dimensional (2D) spatial information of the aerial imagery and the depth information of the heightmap.

The invention discloses a method for generating an aviation lifelike tree sample and identifying a tree species by using an image diffusion model. The method comprises the following steps: constructing a multi-modal data set; training an image-text comparison model and a Unet denoising network model; outputting predicted noise in the trained Unet denoising network model, and finally generating final image reconstruction which is in semantic alignment with the input text description; the obtained single tree crown images are synthesized; and constructing a YOLOv11 detection model, and training the YOLOv11 detection model by using the synthesized forest image to realize tree species identification. According to the method, language semantics and a diffusion model are combined to generate a vivid tree sample for capturing specific features of species and seasonal phenological changes, and the vivid tree sample is synthesized into a high-fidelity forest image, so that tree species identification in an aerial image is enhanced, and effective crown detection and accurate tree species identification are realized.

The invention belongs to the technical field of geological prospecting evaluation, and discloses an ore body positioning method and a rapid sampling device for niobium-tantalum ore target region prediction and evaluation. According to the method, based on existing geological data, various geologic bodies in a 1: 250,000-1: 10,000 geological map are analyzed, distribution, form, invasion contact relation, formation times and other characteristics of alkaline rock masses are found out, and a favorable metallogenic area is delineated; a linear structure, an annular structure and alteration information are recognized through satellites and aerial images, and target area positioning is assisted; the method comprises the following steps: exploring and detecting ore-bearing rock mass and fracture structures through a magnetic method, delineating potential positions of ore bodies, and inferring geological attributes of abnormal sources in combination with geological data; measuring the total amount of gamma at different positions in the niobium-tantalum prediction target area by using radioactive detection equipment, and preliminarily judging the position of the niobium-tantalum ore according to a gamma abnormal result; and carrying out systematic sampling and assay analysis on the target area by using the rapid sampling device. By means of the method and device, the metallogenic geological condition and the prospecting prospect can be evaluated, and the exploration network degree is optimized.

High-definition map data is generated from a combination of disparate data sources, such as aerial imagery, vehicle object detections, and / or vehicle telemetry data. The resulting high-definition map is sufficiently granular to support higher levels of autonomous driving systems, such as L2 and L3. In particular, specialized, expensive data sources such as LiDAR are not required for the generation of the high-definition map. Map objects are detected within aerial imagery, and vehicle object detections are clustered to filter spurious detections. The resulting detected objects are used to generate HD layers on top of other map data.

To support operations related to evaluation of the aptitude for the location of a plant.SOLUTION: A location aptitude evaluation system stores a location aptitude evaluation model that is a machine learning model trained by using, as learning data, a data set formed with an aerial image as an explanatory variable and information on a feature used for evaluation of the aptitude for the location of a predetermined plant as an objective variable, and inputs an aerial image capturing an evaluation target that is the land or region to be evaluated in terms of the aptitude for the location of the predetermined plant and the periphery environment thereof, to the location aptitude evaluation model as an explanatory variable, to generate information based on a value output from the location aptitude evaluation model as an evaluation result.SELECTED DRAWING: Figure 2

A platform for automatically detecting and analyzing defects in infrastructure assets uses aerial imagery, machine learning models and an interactive geographic information system (GIS) interface. Drone or aircraft images are processed by deep neural networks to identify and quantify surface and subsurface defects. Defects are mapped onto 3D models with precise geospatial coordinates. A web platform allows engineers to visualize defects from individual or multiple time periods, filter and query based on attributes, analyze density heatmaps, highlight changes, and perform side-by-side comparative analysis. The unique interface enables rapid condition assessments to support proactive maintenance.

An approach for determining regulatory status of water resources under the Clean Water Act includes receiving input data layers for a geographic location, preprocessing the layers into a multi-channel image, feeding the image into a trained convolutional neural network, obtaining an output score indicating a probability of regulated waters, and determining regulatory status based on the score. The input layers may include aerial imagery, wetland / stream data, soil / elevation / climate data, land cover classification, and regulatory boundary data. The neural network may be trained on approved jurisdictional determinations. The method enables efficient, consistent predictions of Clean Water Act jurisdiction across diverse water resources and geographic regions. By incorporating regulatory information alongside environmental data, the approach accounts for variations in Clean Water Act implementation across jurisdictions and regulatory regimes. The scalable method supports rapid evaluation of large areas for policy analysis and environmental planning.

The invention discloses a lightweight multi-scale target detection method and system for an aerial image of an unmanned aerial vehicle, and relates to the field of target detection of the aerial image of the unmanned aerial vehicle. The method comprises the following steps: acquiring an unmanned aerial vehicle aerial image data set, and dividing the data set into a training set and a test set; carrying out improvement on the basis of YOLOv8-X to generate an LMS-YOLO network; in the design of the LMS-YOLO network, a C3-EMA module and a Ghost module are integrated; performing iterative training on the LMS-YOLO network through the training set to obtain a trained LMS-YOLO network model; and performing target detection on the aerial image of the unmanned aerial vehicle to be identified by using the trained LMS-YOLO network model. According to the invention, multi-scale target detection can be carried out on the aerial image of the unmanned aerial vehicle, and the accuracy and efficiency of target detection are improved.

The invention relates to an intelligent decision-making method and system for post-disaster emergency repair of a road infrastructure network. The method comprises the following steps: acquiring an aerial image of a post-disaster road infrastructure by using an unmanned aerial vehicle; features of the road infrastructure damage area are automatically identified and interpreted, and a multi-dimensional damage feature vector is generated; obtaining a first-aid repair team coordinate and existing first-aid repair resource information, calculating a spatial geometric distance between each damage area and the first-aid repair team, generating a distance matrix, performing normalization processing, and integrating the normalized distance matrix and the multi-dimensional damage feature vector into a three-dimensional tensor; the three-dimensional tensor is mapped into an importance degree matrix in a dimensionality reduction mode, an emergency dispatching scheme of a first-aid repair team is formulated through an optimization algorithm, and the first-aid repair team is assigned; and when the first-aid repair of any damaged area is completed, the steps are repeated until the first-aid repair of all the areas is completed. Compared with the prior art, according to the dynamic planning of the emergency repair real-time process, the optimal matching of the emergency repair task and the resource is ensured, and the resource use efficiency is maximized.

Described are systems and techniques for visualization and manipulation of non- composite oblique aerial imagery. Information can be obtained for an annotation feature for overlay on a displayed oblique image with an oblique viewpoint of a scene within a geographic area. Three-dimensional (3D) model data can be obtained for a portion of the geographic area and including 3D information of object surfaces within the scene. A 3D representation of the annotation feature can be generated using the 3D model data and georeferenced coordinates determined for the annotation feature. An oblique projection of the 3D representation of the annotation feature into the oblique viewpoint of the scene can be generated to include portions of the annotation feature unobstructed by the object surfaces within the scene. The oblique projection of the annotation feature can be output as a two-dimensional (2D) overlay on the displayed oblique image. Unlocking insights from Geo-Data, the present invention further relates to improvements in sustainability and environmental developments: together we create a safe and liveable world.

In some embodiments, an unmanned aerial vehicle (UAV) is provided. The UAV comprises one or more processors; a camera; one or more propulsion devices; and a computer-readable medium having instructions stored thereon that, in response to execution by the one or more processors, cause the UAV to perform actions comprising: receiving at least one image captured by the camera; generating labels for pixels of the at least one image by providing the at least one image as input to a machine learning model; identifying one or more landing spaces in the at least one image based on the labels; determining a relative position of the UAV with respect to the one or more landing spaces; and transmitting signals to the one or more propulsion devices based on the relative position of the UAV with respect to the one or more landing spaces.

A method for farmland path recognition and planning based on aerial imagery is proposed. First, a data set (DOM) is constructed, cropped, and scale-normalized to form a dataset for recognition and planning. Next, multi-scale identification of the inter-row crop passage area is performed to obtain passable areas. Then, path planning is performed based on the inter-row structure: the centerline of the passable area is extracted to obtain a skeleton; principal direction constraints and structural consistency analysis are applied to the skeleton; branch removal and break repair are performed to obtain continuous, directional, and consistent inter-row centerline paths. The paths are clustered, grouped, and sorted to automatically distinguish different ridges and determine the walking order within the plot. Based on the endpoint relationships of adjacent inter-row paths, cross-row connection paths are generated at the ridge ends, forming a global navigation path combining straight and turning routes, and determining the entrance and exit points. Finally, the results are converted into standard path data for output, providing navigation for ground-based equipment. This invention effectively improves the stability and consistency of planning.

A platform for automatically detecting and analyzing defects in infrastructure assets uses aerial imagery, machine learning models and an interactive geographic information system (GIS) interface. Drone or aircraft images are processed by deep neural networks to identify and quantify surface and subsurface defects. Defects are mapped onto 3D models with precise geospatial coordinates. A web platform allows engineers to visualize defects from individual or multiple time periods, filter and query based on attributes, analyze density heatmaps, highlight changes, and perform side-by-side comparative analysis. The unique interface enables rapid condition assessments to support proactive maintenance.

The invention discloses a multi-scale semantic enhanced point supervision algorithm for counting multiple types of crop plants in an aerial image, and the algorithm comprises the following steps: data preprocessing: carrying out the marking of a crop image through a point; the RcsNet50 backbone network is used as a feature extractor to encode the input image; a CSGEM module is used for enhancing crop plant semantics and background information semantics such as soil and shadow in a semantic weight adjustment mode; an RDDCM module is used to adapt to scale changes of different crop plants through a parallel nonlinear convolution mechanism; performing dual-path decoding by using an attention map path AMP and a density map path DMP, and introducing a CoT module in the decoding process; and applying element-by-element multiplication to AMP and DMP decoding results to obtain the predicted number of crop plants. The influence of complex background information and scale change of crops on counting precision can be reduced, the MCPCNet is greatly improved in the aspects of precision, calculation efficiency and generalization ability, and the MCPCNet has superiority and practicability in the aspect of coping with complex agricultural challenges.

An improved living area estimation system is described herein that can obtain aerial images of structures and property data corresponding to structures. The improved living area estimation system can then perform image processing on the aerial images to estimate a living area of each structure. The improved living area estimation system can group property records for the various structures based on one or more criterion and, for each group of property records, determine the average difference between living area calculated using aerial images and living area included in the property data. For a particular structure, the improved living area estimation system can process the aerial image(s) to estimate living area and then modify this value by the average difference determined for the group to which the structure belongs to determine an estimated living area of the structure.

The invention provides a multispectral target detection method based on prompt learning. The multispectral target detection method is characterized by comprising the following innovative steps: step 1, preprocessing aerial image data; and step 2, combining a deep learning model with a prompt learning model Global-Aware Prompt Block (GAP-Block) provided by the invention to extract image features. And step 3, fusing the features of the multi-modal image by using an adaptive Multi-Feature Fusion (AMFF) method provided by the invention. And a fourth step of enhancing the image features by using a feature pyramid network (FPN). And 5, inputting the image features into a target detector for detection. The method is combined with prompt learning to solve the problems that the training cost is high and current scale features among different modalities are mainly concerned when an existing method is used for downstream task training, multi-modal feature graphs can be dynamically integrated, and the feature expression ability is further improved; and an efficient and flexible method is provided for a multi-modal target detection task.

The invention discloses an unmanned aerial vehicle aerial image-based wetland waterfowl automatic identification and statistics method and system, and belongs to the technical field of software. The objective of the invention is to solve the problem of inaccurate population scale and spatial distribution statistical results caused by missed detection, false detection and uncertainty of an unmanned aerial vehicle aerial photographing waterfowl recognition result in the prior art. Therefore, the invention provides a method which comprises the following steps of: dividing a grid and constructing a multi-frame observation effort tensor; performing waterfowl detection and trajectory aggregation based on multi-frame images; modeling based on the observation effort, the scene features and the detection probability of manual checking; performing layered inference on the real number of grids based on the detection model; fusing real intensity field inference of spatial adjacency and ecological factors; and updating the time sequence state space under multi-period monitoring. Therefore, statistical inference of the real number and spatial distribution of the waterfowls is realized, and the method has the advantages of high statistical reliability and accurate uncertainty quantification.

Methods, non-transitory computer-readable storage media, and computer or computer systems directed to detecting, analyzing, and tracking stages of housing construction using satellite or aerial imagery in combination with a machine learned model are described.

The invention relates to a ground station system (38a-c) for georeferencing aerial images, in particular for calibrating and / or surveying drone photogrammetry data, of an area (10a-c), such as, for example, a construction site, an archaeological find and / or excavation site or an infrastructure facility, with a plurality of control point units (12a-c, 14a-c, 16a-c) that can be arranged distributed over the area (10a-c), wherein each of the control point units (12a-c, 14a-c, 16a-c) has an optical marker (18a-c) on at least one surface that is uniquely identifiable from the air. It is proposed that each of the control point units (12a-c, 14a-c, 16a-c) is associated with a Global Navigation Satellite System (GNSS) module (20) for acquiring its instantaneous geographic position and a Wide Area Network (WWAN) module (28a-c) for at least wirelessly transmitting the acquired instantaneous geographic position, wherein at least one of the control point units (12a-c, 14a-c, 16a-c) is integrated into a functional object (22a-c, 24a-c, 26a-c) that is at least temporarily assigned to the area (10a-c) and performs a main function within the area (10a-c) that is different from georeferencing.

An autonomous system and method for capturing event-driven aerial imagery utilizes a drone equipped with advanced sensors and navigation subsystems to operate without human intervention. Upon detecting predefined target events, such as structure fires, explosions, gunshots, emergency sirens, a specific license place, or a specific face, the drone autonomously launches and employs direction-finding triangulation to pinpoint the event's latitude, longitude, and elevation. The drone autonomously executes optimized flight profiles. Data transmission utilizes bonded and blended communication channels to ensure reliable video streaming to users, such as first responders or news agencies. Compliance with FAA altitude regulations is enforced, and the system can be controlled remotely via internet or cellular connections. Continuous operation is enabled through tethered power or automated battery replacement stations. Applications include law enforcement, emergency services, and news gathering, providing immediate aerial reconnaissance without requiring human operators to be present at unpredictable event locations.

An embodiment of this specification discloses a satellite image super-resolution reconstruction method assisted by drone images. The method includes obtaining aerial images collected by a drone, performing aerial triangulation based on image control points, and orthorectifying the aerial images; selecting first images in different terrain areas, downsampling the first images to obtain second images, and constructing first training data to train a generative adversarial network; obtaining a target image to be reconstructed, reading the target image in blocks as at least two sub-blocks, generating reconstruction results of each sub-block based on the trained generative adversarial network, and stitching the reconstruction results based on the edge gradient fusion method to obtain a reconstructed image. The embodiment of this specification can obtain higher-resolution images without upgrading or replacing hardware, saving costs and being easy to implement technically.

An autonomous system and method for capturing event-driven aerial imagery utilizes a drone equipped with advanced sensors and navigation subsystems to operate without human intervention. Upon detecting predefined target events, such as structure fires, explosions, gunshots, emergency sirens, a specific license place, or a specific face, the drone autonomously launches and employs direction-finding triangulation to pinpoint the event's latitude, longitude, and elevation. The drone autonomously executes optimized flight profiles. Data transmission utilizes bonded and blended communication channels to ensure reliable video streaming to users, such as first responders or news agencies. Compliance with FAA altitude regulations is enforced, and the system can be controlled remotely via internet or cellular connections. Continuous operation is enabled through tethered power or automated battery replacement stations. Applications include law enforcement, emergency services, and news gathering, providing immediate aerial reconnaissance without requiring human operators to be present at unpredictable event locations.

Described are systems and techniques for visualization and manipulation of non-composite oblique aerial imagery. Information can be obtained for an annotation feature for overlay on a displayed oblique image with an oblique viewpoint of a scene within a geographic area. Three-dimensional (3D) model data can be obtained for a portion of the geographic area and including 3D information of object surfaces within the scene. A 3D representation of the annotation feature can be generated using the 3D model data and georeferenced coordinates determined for the annotation feature. An oblique projection of the 3D representation of the annotation feature into the oblique viewpoint of the scene can be generated to include portions of the annotation feature unobstructed by the object surfaces within the scene. The oblique projection of the annotation feature can be output as a two-dimensional (2D) overlay on the displayed oblique image.

This invention belongs to the field of UAV autonomous positioning and navigation technology, and relates to a UAV map construction method for environments where satellite imagery is unavailable. Specifically, it provides a method for constructing a UAV work area map based on aerial imagery, addressing issues such as image distortion and edge misalignment in existing area map construction methods. This invention orthogonally divides the UAV's aerial photography work area into a grid according to the same geographic coordinate system. The UAV performs a coverage flight mission within the work area according to a planned flight path, maintaining an interval of 50% of the field of view width between adjacent flight paths. During aerial photography, the UAV simultaneously acquires image and geographic positioning information (GPS information and yaw angle information of the center point) as image location points. Based on UAV sensor data and the UAV's flight trajectory and pose, keyframes of geographic location information are extracted from the aerial video, and then used to fill the corresponding grid content to obtain a global map of the work area.

This invention provides a method for generating 3D objects from images of buildings taken from the air using an aerial videography device. [Solution] The method includes a calculation unit that generates an initial spatial object which is a 3D object corresponding to a building based on multiple aerial images, and a mapping candidate selection step, defining an object surface corresponding to the exterior surface of the building and a building surface normal vector perpendicular to the exterior surface of the building on the building surface. At least one of the multiple aerial images includes an initial texture corresponding to the exterior surface of the building, and a shooting direction vector from the shooting device toward the exterior surface of the building is defined for each of the multiple aerial images. The method also includes an intermediate texture selection step in which the calculation unit selects an aerial image for which the dot product of the shooting direction vector and the building surface normal vector is negative as a mapping candidate aerial image, and selects the initial texture with the largest area from among multiple initial textures corresponding to the multiple mapping candidate aerial images as an intermediate texture.

Computer vision systems and methods for automatic parcel alignment are provided. The system receives a geotagged aerial image, parcel information, and semantic information where each of the parcel and semantic information are overlaid on the geotagged aerial image. The system cleans the parcel information and the semantic information. The system optimizes the parcel information by grouping geo-registered parcels present in the geotagged aerial image into a plurality of islands and generating a plurality of parcel alignment solutions for each island of the plurality of islands. The system refines the plurality of parcel alignment solutions for each island and regularizes each island. The system generates a composite parcel alignment solution based on the refined plurality of parcel alignment solutions for each regularized island to align the geo-registered parcels of each regularized island with the geotagged aerial image.

The present application belongs to the technical field of geological prospecting and evaluation, and discloses a kind of ore body positioning method and rapid sampling device for niobium-tantalum ore target area prediction and evaluation.The present application is based on existing geological data, through the analysis of various geological bodies in 1:250,000-1:10,000 geological map and the identification of the distribution, shape, intrusive contact relationship and formation age of alkaline rock mass, the favorable metallogenic area is delineated;Through satellite and aerial image identification of linear structure, ring structure and alteration information, target area positioning is assisted;Through magnetic exploration, ore-bearing rock mass and fault structure are detected, and potential position of ore body is delineated, combined with geological data, geological properties of anomaly source are inferred;Using radioactive detection equipment, the total amount of gamma at different positions in the niobium-tantalum prediction target area is measured, and the position of niobium-tantalum ore is preliminarily judged according to the gamma anomaly result;Using rapid sampling device, target area is systematically sampled and analyzed.Through the method and device, metallogenic geological conditions and prospecting potential can be evaluated, and exploration grid is optimized.

Described are systems and techniques for visualization and manipulation of non- composite oblique aerial imagery. One or more user inputs can be obtained by an oblique image visualization system associated with a plurality of non-composite oblique aerial images of a geographic area of interest. View perspective information indicative of a user-requested view of a scene or area within the geographic area of interest can be determined from the user inputs. One or more parameters of the view perspective information can be compared to a corresponding one or more parameters associated with each oblique image. A best match oblique image can be selected from the plurality of non-composite oblique aerial images based on the comparison. The best match oblique image can be output for display by the oblique image visualization system in response to the user-requested view. Unlocking insights from Geo-Data, the present invention further relates to improvements in sustainability and environmental developments: together we create a safe and liveable world.