75 results about "Epipolar geometry" patented technology

The invention discloses an unmanned aerial vehicle camera attitude estimation optimization method and device, and relates to the technical field of unmanned aerial vehicles. The method comprises the following steps: acquiring data by using an unmanned aerial vehicle carrying various sensors, preprocessing, carrying out feature extraction and matching on preprocessed image data, removing mismatching points, and based on an initial point cloud, carrying out sparse reconstruction in an incremental expansion and bundle adjustment optimization stage; integrating the re-projection error term, the flight path constraint, the epipolar geometric constraint and the triangulation constraint into a target function; constructing a model for predicting the pose correction based on the deep residual convolutional neural network, and taking the initial pose of the camera and the corresponding local image as input; and adjusting the weight of each constraint in the target function based on the predicted pose correction and pose confidence, and minimizing the target function to improve the precision of unmanned aerial vehicle camera pose estimation. The problem of reconstruction scene deviation caused by inaccurate camera attitude estimation in the prior art is solved.

A dynamic vision SLAM (Simultaneous Localization and Mapping) method based on an extended Bayesian model belongs to the technical field of autonomous robot navigation and computer vision crossing, and mainly comprises the following steps: performing prior dynamic object recognition on a current frame image by using an improved PWt-YOLO network, and outputting a binary mask and semantic probability distribution; oRB feature extraction is carried out by using a hierarchical quadtree algorithm and combining an adaptive threshold value; establishing an extended Bayesian probability model, fusing semantic prior, optical flow residual and epipolar geometric constraints of two-dimensional Gaussian distribution constructed based on dynamic object boundaries, realizing time sequence transmission of dynamic feature probabilities through a Markov chain, and executing feature point filtering according to joint dynamic probabilities; and robust pose estimation is realized through RANSAC-PnP, a sliding window and the like. According to the method, the problem of SLAM system positioning drift in a dynamic environment can be effectively solved by constructing a multi-modal fusion dynamic feature discrimination system.

The invention discloses a binocular X-ray guide wire three-dimensional positioning method, a binocular X-ray guide wire three-dimensional positioning device and binocular X-ray guide wire three-dimensional positioning equipment based on epipolar geometry and a storage medium. The method comprises the steps that two to-be-processed images collected by two X-ray scanning devices in the biplane digital subtraction angiography system for a target object in a guide wire intervention state are acquired, and the to-be-processed images comprise intervention guide wires; conducting guide wire positioning on the two to-be-processed images based on projection matrixes which are obtained through pre-calibration and respectively correspond to the two X-ray scanning devices, and determining path information of the interventional guide wire, the path information comprising first position information of the interventional guide wire; obtaining an expected path of the interventional guide wire, and determining an offset error of the interventional guide wire based on expected position information of the interventional guide wire in the expected path and the first position information; the expected path of the intervention guide wire is updated based on the offset error, accurate positioning of the intervention guide wire and updating of the expected path are achieved, and accurate data support is provided for subsequent treatment.

The invention relates to a multi-modal image matching method and system based on learning features and epipolar geometric constraints. The method comprises the following steps: carrying out edge enhancement processing on an input image through wavelet transform; extracting a multi-scale dense feature map based on the transformed convolutional neural network, and generating a feature descriptor with rotation and scale invariance in combination with principal direction normalization; adopting an FLANN algorithm and dynamic distance constraint to realize preliminary feature matching; and introducing a basic matrix construction and epipolar geometric consistency verification mechanism, and eliminating mismatching point pairs in combination with an RANSAC affine constraint model. According to the method, image enhancement, deep learning and geometric verification strategies are fused, the problems of radiation nonlinearity and geometric distortion caused by imaging mechanism differences among multi-modal images are effectively solved, the matching precision and robustness are improved, and the method is suitable for remote sensing application scenes such as optical-SAR registration, multi-source image fusion and earth surface change detection.

The invention relates to a sparse view angle pose-free scene reconstruction method and system based on 3DGS, belongs to the field of computer vision and three-dimensional reconstruction, and solves the problems of easy failure and poor geometric consistency in a sparse view angle or weak texture environment due to dependence on accurate camera pose priori. Constructing a double-flow sensing module containing semantic flow and geometric flow, extracting semantic features by using a visual basic model, and extracting an explicit geometric corresponding relation by using a dense feature matching network, so as to regress relative camera pose under pose-free priori; a geometric guidance depth refinement module combining a potential diffusion model architecture and Pluecker ray coding is introduced, and scale fuzziness of monocular depth estimation is eliminated through a depth residual prediction mechanism; and based on the micronizable Gaussian rasterization, performing end-to-end optimization by using a self-supervised loss function including rendering consistency, reprojection and epipolar geometric constraint. According to the method, high-fidelity three-dimensional reconstruction is realized without supervision of external parameter true values, and geometric stability and rendering quality in a complex scene are improved.

The invention relates to a visual inertial fusion positioning method and system based on feature optimization, and belongs to the technical field of visual inertial navigation of autonomous robots. According to the method and the system, on the basis of a Geometric Feature Track Selection Visual Native Navigation System (GFT-VINS), through a feature trajectory selection strategy based on normal epipolar geometry, high-quality feature trajectories are screened out, low-quality trajectory interference is avoided, the positioning precision is improved, error accumulation caused by environmental changes and feature anomalies is reduced, and the system stability is enhanced. Meanwhile, in combination with a multi-state Constraint Kalman filter (MSCKF) framework, the efficient calculation capability is kept, the requirements of the autonomous robot for high precision, high stability and efficient calculation in a complex environment are met, and wide application of the autonomous robot is promoted.

The invention discloses an underwater target identification and positioning method based on fusion of a physical model and deep learning, and belongs to the technical field of computer vision, and the method comprises the steps: carrying out the transmissivity estimation and physical restoration of a collected underwater monocular image according to an underwater light propagation model, and carrying out the image correction; key feature points are extracted, and high-quality matching point pairs are screened in combination with the joint similarity; further deriving a basic matrix through the high-quality matching point pairs meeting the epipolar geometric constraint relation, solving an essential matrix, and obtaining relative attitude parameters between the cameras in combination with weighted re-projection-LM optimization; and finally, refraction correction triangulation is carried out according to the Snell's law, pixel-level fusion is carried out after scale normalization and space alignment are carried out on the refraction correction triangulation and dense depth output by the MiDaS, three-dimensional space coordinates of the target are inverted, and high-precision recognition and positioning of the underwater target are achieved. The system is light in structure, efficient in calculation, suitable for being integrated on various underwater autonomous or remote control robot platforms and used for tasks such as target recognition, tracking and positioning.

The invention discloses a dynamic scene visual SLAM method based on deep learning, and the method comprises the following steps: obtaining an original scene image, and extracting ORB feature points in an image scene; based on the improved target detection network, performing dynamic target detection on the original scene image to obtain a dynamic target frame in the image scene; associating the dynamic target frame and the ORB feature points by adopting a key frame synchronization mechanism, and removing the ORB feature points corresponding to the dynamic target frame; obtaining a feature scene image after primary elimination; performing secondary elimination on unmatched ORB feature points in adjacent feature scene images by adopting epipolar geometric constraint to obtain a static feature scene image; and performing pose estimation and / or map construction based on the ORB feature points corresponding to the static feature scene image. According to the method, the dynamic feature points of the dynamic target are extracted by improving the target detection network, and the dynamic feature points are eliminated twice by adopting the epipolar geometric constraint, so that the positioning precision is improved, and accurate map construction is facilitated.

The invention discloses an SLAM dynamic point semantic filtering method based on DDMA-SAM, and belongs to the technical field of synchronous positioning and mapping. A decoupling distillation mechanism is introduced, an image encoder in an original SAM model is subjected to lightweight optimization, a DDMA-SAM semantic network integrating a multi-scale aggregation detection module and an efficient mask decoding module is constructed, and the SLAM dynamic point semantic filtering method based on DDMA-SAM is obtained. And the segmentation performance is improved while the model parameters are greatly compressed. Based on the semantic network, providing a semantic prior and geometric consistency combined-driven double filtering strategy; based on a semantic mask and a confidence threshold, carrying out preliminary dynamic point identification; in combination with the epipolar geometric constraint and the triangulation reprojection error of random sampling consistency estimation, fine elimination of dynamic feature points is realized, and only static points are reserved to participate in camera pose estimation. According to the method, the real-time performance of the system is kept, and meanwhile, the mapping quality and the track stability in a dynamic scene are effectively improved.

The invention provides a ship collision monitoring and early warning method and system for dams and navigation facilities, and relates to the technical field of shipping traffic. The method comprises the following steps: respectively acquiring images, messages and radar point cloud data through a camera, a ship automatic identification system and a millimeter wave radar at a main monitoring point and an auxiliary monitoring point, projecting the radar data to an image plane to form a radar image after time-space synchronization processing, sending the radar image to a ship identification network, and identifying a two-dimensional surrounding frame and a type of a ship; then, obtaining a three-dimensional bounding box and a type of the ship by utilizing epipolar geometric processing, comparing the three-dimensional bounding box and the type with message data, and issuing third-level early warning when the three-dimensional bounding box and the type are inconsistent with the message data or no related type exists; meanwhile, the future navigational speed and course of the ship are predicted according to a plurality of three-dimensional bounding boxes of the ship, and if the future navigational speed and course exceed a set threshold value or the risk of impacting a dam exists, secondary early warning is issued; if the ship triggers third-level and second-level early warning at the same time, first-level early warning is issued. According to the scheme, the accuracy of early warning and the completeness of an early warning mechanism can be greatly improved.

The invention belongs to the technical field of multi-target identity judgment among multiple passive sensors. The invention provides a space-time and track double-constraint multi-target identity judgment method for multiple passive sensors. According to the embodiment of the invention, the time-space criterion based on the continuous multi-frame epipolar geometric constraint and the track criterion based on the track feature cosine similarity are constructed, so that the identity judgment characteristics of multiple targets in different sensors can be described more accurately, and the accuracy and robustness of the identity judgment of the multiple targets are remarkably improved. The problem of weight optimization under different sensor configurations and scene characteristics is solved through an adaptive fusion weight mechanism, and performance degradation of a fixed weight strategy in a complex scene is avoided. The method has good scene adaptive ability and excellent real-time response characteristic, completely meets the rapid online processing requirement in the airborne environment with limited resources, maintains the flexibility and expansibility of cross-platform deployment, and provides reliable technical support for leading-edge applications such as multi-carrier collaborative awareness.

The application provides a feature matching method, system and device based on multi-geometry collaborative learning and a storage medium, and relates to the technical field of robot positioning and navigation. The steps include: acquiring an image pair, extracting two groups of sparse local features in the image pair and inputting the local features into a multi-geometry collaborative network model; enhancing the local features through an attention mechanism and iteratively updating the sparse network through a multi-layer perception and / or linear projection; generating and screening reliable matching points through affine transformation estimation, position coding and key point neighborhood expansion; constraining the matching process through pose information and adaptively adjusting the matching strategy for matching prediction; combining real matching information and homography geometry information for loss calculation to train the multi-geometry network model; and repeating the steps until a preset iteration number is reached. Through collaborative optimization of affine geometry, epipolar geometry and homography geometry, the application realizes mutual promotion of local feature discriminability and geometric consistency, and can achieve a better balance between speed and accuracy in feature matching.

The present invention discloses a visual odometry method and system based on image depth prediction and monocular geometry. The method includes inputting two consecutive image frames, detecting and describing local features of the images using the Scale-Invariant Feature Transform (SIFT) algorithm, and then using the Fast Library for Approximate Nearest Neighbors (FLANN) algorithm to match corresponding feature point pairs between the two frames; solving for the essential matrix using epipolar geometry constraints to obtain the relative pose transformation of the camera; constructing and training a monocular depth prediction model to predict dense depth information for each input image frame; if the number of valid depth information pairs formed by two consecutive image frames is greater than a given threshold, using triangulation to estimate the scale factor to obtain the corrected relative pose transformation, otherwise using a combination of the Perspective-n-Point (PnP) projection algorithm, the Random Sample Consensus (RANSAC) algorithm, and local non-linear optimization to solve for the absolute pose transformation. The present invention effectively integrates the advantages of deep learning and traditional geometric methods, can adapt to dynamic environments, and improves the robustness and accuracy of monocular visual odometry.

The invention discloses a road target object visual positioning system and method based on urban streetscape. The method comprises the following steps: acquiring a positioning picture and a plurality of query pictures; performing feature matching and extraction on each query picture and the positioning picture; solving the pose parameters based on a deep learning network or a visual algorithm to obtain a solving result; and determining the corresponding moving direction and distance between the adjacent query pictures in the plurality of query pictures, and calculating the absolute coordinate position of the query picture by combining the solving result, the GPS coordinate of the positioning picture and the epipolar geometrical relationship. By implementing the method provided by the invention, the problem of precision limitation of a traditional visual positioning method based on picture feature matching under the condition that the shooting distance of the positioning picture is relatively large or the distance between the query picture and the positioning picture is relatively long can be solved.

The invention relates to a dynamic SLAM method and system based on lightweight YOLOv8 and an adaptive key frame strategy, and belongs to the technical field of computer vision. According to the method, a dynamic region in an image frame is detected through an improved lightweight network, and extracted dynamic feature points are removed in combination with an epipolar geometry constrained motion consistency detection method; extracting line features from the static region based on an improved LSD algorithm, and compensating the loss of feature points; designing a key frame selection strategy of adaptive weight to delete redundant key frames, and constructing a joint optimization objective function by combining static point features and static line features to perform pose estimation and construction of a point cloud map; and finally, performing closed-loop matching by using visual and semantic features between the key frames, verifying a closed loop, establishing a pose constraint, and performing global graph optimization. According to the method, the influence of a dynamic object in a scene on the mapping quality of the SLAM algorithm can be effectively reduced, and the robustness of the algorithm in a dynamic environment is enhanced.

The application discloses a kind of circular stripe three-dimensional reconstruction methods based on epipolar Markov modeling, steps are as follows:First, the method is aimed at the problem that existing concentric circle stripe projection measurement method is strongly dependent on camera-projection device geometry arrangement, in the camera-projection device system that has been calibrated, the epipolar geometry is combined with concentric circle equal phase model, and the intersection of the projector epipolar line corresponding to camera pixel and equal phase circle is obtained to obtain candidate projection point;Again, candidate point selection modeling is one-dimensional Markov chain sequence energy minimization problem, and candidate branch disambiguation is realized by smoothing constraint, branch switching penalty and gradient consistency constraint;Finally, according to the corresponding relationship of screened camera-projection device, triangulation is carried out to obtain three-dimensional point cloud.The application does not require camera and projector to meet strict parallel optical axis or single direction displacement assumption, and can improve the matching stability and three-dimensional reconstruction robustness of circular stripe structured light under the condition of complex curved surface and geometric configuration change.

This invention relates to the field of image analysis technology, specifically a deep learning-based method for detecting surface defects in tab adapters. The method includes acquiring the three-dimensional geometric curvature parameters of the bending region through multi-view polarization imaging; decoupling specular reflection and diffuse structural features using a photometric distribution model to resolve high-brightness reflection interference; nonlinearly projecting the surface texture onto a two-dimensional plane using a local tangent space mapping matrix to generate an isotropic corrected texture map to compensate for geometric distortion; identifying suspected defect points that disrupt flow continuity through singular value decomposition of the anisotropic structure tensor; and introducing stress distribution priors and epipolar geometric constraints for multi-view spatial consistency verification to effectively filter lighting artifacts. The verified region and corrected texture are fused and input into a deep recognition network with a spatial attention mechanism to analyze the surface defect image. This invention achieves accurate analysis of surface defect images through deep coupling of three-dimensional geometry and a photometric physical model.

The application discloses a dynamic scene SLAM positioning method based on a target detection algorithm YOLOv4 and geometric constraints. Firstly, an image is collected through a visual sensor, a potential moving object area in the image sequence is recognized by using the YOLOv4 target detection algorithm, feature points of a part outside the area are extracted, and the matching feature points are calculated by using an optical flow method. After a transformation matrix of two frames is calculated, whether the feature points are dynamic points is judged by using an epipolar geometry epipolar constraint. When the epipolar constraint is invalid due to degenerate motion, whether the feature points are dynamic feature points is judged by using a boundary constraint. Finally, a dynamic object area is determined in combination with a target detection result, and after the corresponding area is removed, tracking, local mapping and loop detection threads are sequentially performed, a SLAM algorithm for a dynamic scene is realized, and the existing visual SLAM algorithm is inaccurate in pose estimation in a dynamic environment and is prone to be lost. The application significantly improves the visual SLAM pose precision in a dynamic scene while ensuring real-time performance, and can smoothly run on an embedded system.

The present invention discloses a method for multi-node collaborative navigation of a flexible lander under limited navigation information, belonging to the field of deep space exploration technology. The present invention is implemented as follows: based on the current altitude and the associated observation information between nodes, an indirect observation equation for navigation landmarks is constructed based on homography transformation or epipolar geometry to build a transmission path for observation information between nodes; the Fisher information matrix is ​​used to analyze node observability, and multi-node information interaction conditions are established with node observability and the flexible lander altitude as indicators; the information interaction conditions are used to determine the information transmission direction between nodes that makes the positions of all nodes observable and the minimum transmitted observation information in the corresponding information transmission direction; a single node uses the observation information obtained by the node itself and the observation information transmitted by other nodes to estimate the node state through a nonlinear filtering algorithm, and all nodes synchronously perform state estimation according to the single node state estimation method to achieve multi-node collaborative navigation under limited navigation information.

The present invention discloses an image template matching method and a monitoring system, device, and storage medium, including: acquiring a captured image and a template image; extracting each feature point of the captured image respectively through at least a first feature matching model and a second feature matching model; normalizing the feature point data and merging them into a captured feature data collection; pairing each feature point in the captured feature data collection with each feature point in the template feature data collection according to a homography matrix H to obtain a plurality of matching point groups, wherein the matching point groups include feature points in the captured feature data collection and feature points in the template feature data collection that meet epipolar geometry constraints; calculating the degree of difference between the feature points in the captured feature data collection and the feature points in the template feature data collection in each matching point group; and deriving the defect position based on the degree of difference information. The present design reduces the computing load, improves the analysis efficiency, and shortens the processing time.

Techniques are described for identifying correspondences between images to generate a fundamental matrix for the camera positions related to the images. The resultant fundamental matrix enables epipolar geometry to correlate common features among the images. Correspondences are identified by confirming feature matches across images by applying a homography to data representing features across images. Further techniques are described herein for generating a representation of a boundary of a feature of a structure based on a digital image. In one or more embodiments, generating a representation of a boundary of a particular feature in a digital image comprises determining a portion of the image that corresponds to the structure, and determining a portion of the image that corresponds to the particular feature. One more vanishing points are associated with the portion of the image corresponding to the particular feature. The one or more vanishing points are used to generate a set of bounding lines for the particular feature, based on which the boundary indicator for the feature is generated.

The application relates to a monocular vision inertial odometer method based on feature point depth, which comprises the following steps: acquiring input images of two adjacent frames and extracting feature points in the images respectively; matching the extracted feature points; predicting first camera pose changes by using acceleration measurement data and angular rate measurement data output by an inertial measurement sensor; obtaining the depth of each pair of matched feature points based on the first camera pose changes; if the number of feature points meets a preset condition, constructing a weight for each pair of feature points and obtaining a weight matrix; constructing a weighted observation equation set based on an epipolar geometry constraint equation and the weight matrix; constructing a quadratic form based on the weighted observation equation set and solving the quadratic form to obtain an essential matrix of the matched feature points in the least square sense; obtaining a second rotation matrix and a second translation vector by singular value decomposition based on the essential matrix and the second rotation matrix and the second translation vector, and obtaining the representation of actual camera pose changes.

The invention discloses a large-viewing-angle visual positioning method based on geometric constraints of dynamic and static objects, and relates to the technical field of computer vision. The method aims at solving the problem that the pose estimation precision is reduced due to feature matching failure under the large view angle change in the existing feature point-based positioning technology. The method comprises the following steps: identifying an object from an image pair synchronously acquired by a multi-camera system, fitting the contour into an ellipse, and distinguishing a static ellipse from a dynamic ellipse; reconstructing an object motion vector by using the projection change of the dynamic ellipse, establishing a rotation constraint based on the invariance of the object motion vector, and solving the initial estimation of the relative pose between the cameras in combination with a non-rolling constraint; constructing an epipolar geometric space by using the initial estimation, and establishing a corresponding relation by defining an ellipse distance function and matching a static ellipse; and finally, combining geometric constraints of dynamic and static ellipses to construct an optimization model, and performing nonlinear optimization on the relative pose parameters of the camera to obtain a high-precision estimation result. According to the method, robust geometric constraints provided by dynamic and static objects in a scene are utilized, so that the robustness and precision of visual positioning under large visual angle difference are remarkably improved.

The invention discloses a low-altitude unmanned aerial vehicle closed dynamic flight environment semantic modeling method based on deep learning, and belongs to the technical field of low-altitude unmanned aerial vehicles. The method mainly comprises the following four steps: (1) acquiring an airborne RGB-D image and IMU data of the unmanned aerial vehicle, and constructing an IMU pre-integration model; (2) carrying out dynamic and static judgment and dynamic feature elimination on feature points in a scene by utilizing a target detection network in combination with epipolar geometric constraints, and realizing front-end anti-interference; (3) carrying out pixel-level semantic segmentation on the screened key frame based on a DeepLab v3 + network to obtain an environment semantic mask; and (4) through closed-loop detection and global pose map optimization, fusing semantic information and geometric information, and constructing a globally consistent closed dynamic flight environment three-dimensional dense semantic map. According to the method, the problems that the positioning of the low-altitude unmanned aerial vehicle in the closed dynamic flight environment is easily interfered and the mapping accumulative error is large are effectively solved, and high-precision and strong-robustness environment semantic modeling is realized.

The invention discloses a fire safety monitoring system based on computer vision, particularly relates to the field of computer vision, and is used for solving the problems of low fire source identification precision, large illumination interference and difficulty in unified analysis of multi-view monitoring images in existing fire monitoring. A candidate hot spot extraction module is arranged to perform binarization processing on a monitoring image to extract a candidate hot spot region, a morphological enhancement module is utilized to adaptively adjust a nuclear parameter fusion hot spot region, and a standard view angle correction module is utilized to realize perspective correction based on feature point detection and homography matrix calculation; when a multi-view-angle image exists, a view angle fusion module is used for calculating three-dimensional point cloud data of a hot spot based on epipolar geometry, and a real hot spot is identified through time sequence and space consistency verification; and finally, the monitoring output module completes spatial positioning and risk extension trend prediction of real hot spots, outputs fire risk position and trend information, and realizes high-precision fire source identification and real-time early warning in a complex scene.

The application provides a dam and navigation facility ship collision monitoring and early warning method and system, and relates to the technical field of shipping traffic. It comprises: at the main and auxiliary monitoring points, image, message and radar point cloud data are obtained through a camera, an automatic ship identification system and a millimeter wave radar respectively, after time and space synchronization processing, the radar data is projected to the image plane to form a radar image, and then sent to a ship identification network to identify the two-dimensional bounding box and type of the ship; then, the three-dimensional bounding box and type of the ship are obtained by using epipolar geometry processing, and compared with the message data, if they are inconsistent or have no related type, a three-level early warning is issued; at the same time, the future speed and heading of the ship are predicted according to the multiple three-dimensional bounding boxes of the ship, if the set threshold is exceeded or there is a risk of collision with the dam, a two-level early warning is issued; if the ship triggers the three-level and two-level early warnings at the same time, a one-level early warning is issued. The scheme can greatly improve the accuracy of early warning and the perfection of the early warning mechanism.

The present invention discloses an intelligent multi-view stereo matching method and device. The method includes: constructing a global context Transformer module for exploring the global context relationship between the reference view features and the source view features at the smallest scale within the view; constructing a three-dimensional perception Transformer module based on epipolar geometry to mine the three-dimensional consistency information between the globally perceived reference view features and the globally perceived source view features; using differentiable deformation operations to map the large-scale features and the three-dimensionally consistent features to multiple forward planes of the reference view to obtain the mapped multi-scale source view features, and performing group-level correlation operations on the multi-scale source view features to obtain a multi-scale matching cost volume; constructing a multi-scale cost regularization module to realize the mapping between the matching cost volume and the depth map of the reference view; and using an L1 loss function to constrain the difference between the multi-scale depth map and the true depth map to realize the training of the network. The device includes: a processor and a memory.

This invention proposes a robust adaptive VIO navigation and positioning method suitable for GNSS-denied environments. First, it utilizes an IMU and camera to acquire the acceleration, angular velocity, and surrounding environmental information of UGVs. Then, it establishes a measurement constraint model for VIO based on the epipolar geometry and trifocal tensor geometry relationships between multiple images. Next, in the filtering algorithm, the H∞ criterion is introduced into the CMSCKF to improve the robustness of the VIO filtering algorithm. Based on the system uncertainty, a three-stage robust adaptive VIO filtering algorithm is adopted, comprehensively utilizing the standard Kalman filter algorithm and the H∞ filter algorithm to complement each other's advantages, thereby improving the overall performance of the filtering algorithm. This results in VIO integrated navigation not only having high filtering accuracy but also good robustness, thus improving the navigation and positioning performance of UGVs in complex environments.

A method for generating a relief representation model using a plurality of N pairs of stereo images, each of the N pairs of stereo images being associated with a map of photogrammetric disparities which is obtained by correlation calculation based on the pair of stereo images in question, comprises: a calculation of maps of geometric disparities based on a predetermined relief representation model, by projection of the predetermined relief representation model into an epipolar geometry of the pairs of stereo images; a calculation of a cost function, representative of a difference between the geometric and photogrammetric disparities, associated with the predetermined relief representation model; and an updating of the predetermined relief representation model by optimization of the gradient descent type of the cost function, and iteration until a predefined endpoint criterion is reached. Also a computer program and a system.

The present invention discloses a method, system and terminal for implicit field reconstruction based on perspective matching and epipolar geometry constraints. The method comprises: obtaining an initial image group, calculating the photometric error between the initial image group and the multi-perspective rendering value of the target object, and training the three-dimensional implicit field; extracting and combining sampling point features to optimize the three-dimensional implicit field, outputting the target feature body and inputting it into a geometric network, outputting a signed distance field value, and extracting the geometric result of the target object from it; obtaining normal information based on the geometric result, inputting it into a texture network and outputting the rendering value of the camera ray. The present invention uses a set of images with inaccurate poses as input to reconstruct the three-dimensional implicit field, and uses the optimized camera pose to jointly optimize the three-dimensional implicit field. The camera rays passing through key points in the input image are used to achieve light optimization and match light consistency, which can effectively improve the pose optimization results. The three-dimensional implicit field and camera pose can be further optimized by enhancing point cloud features.