38 results about "Map matching" patented technology

This application relates to an offline map matching method, apparatus, computer device, and storage medium. The method includes: starting from the starting position of the complete GPS trajectory to be matched, selecting a sliding window according to a set sliding window length, and searching for the shortest path set for each sliding window in the road network based on candidate starting and ending points; matching the local GPS trajectory corresponding to each shortest path to the corresponding road segment in the road network to generate the shortest matching path without backtracking, calculating the score of each matching path, and retaining the top k matching paths with the highest scores; performing local non-shortest path travel detection on the top k matching paths, and using a split matching algorithm to correct the travel trajectories of existing local non-shortest paths. The embodiments of this application effectively enhance the robustness of road network matching for local non-shortest path travel, improving the overall efficiency and trajectory matching accuracy of offline map matching.

The application discloses a kind of incremental mapping method, equipment and storage medium, the incremental mapping method includes: if it is detected that mobile device is in priori map range outside, stop using priori map to locate, and locate based on the incremental map constructed;If it is detected that mobile device moves from priori map range outside to priori map range, enable priori map to locate, and the environment perception data of current time and priori map are matched based on branch and bound algorithm, and the matching score corresponding to the pose of matching success is obtained;Add monomial edge constraint to the pose corresponding to the environment perception data of matching score greater than preset score threshold;The accumulated error optimization is carried out to the pose calculated using monomial edge constraint, to make incremental map and priori map align. By adding monomial edge matched with priori map to constrain mapping pose, the alignment effect between incremental map and priori map can be improved, to avoid the appearance of map ghosting and fault.

The application provides a fusion road feature information adaptive map matching optimization method and device, can adaptively adjust a window of a to-be-matched trajectory according to current inertial trajectory features, fully utilizes road feature information, and can further improve positioning effect of a map matching device.

The application provides a railway map matching method and system based on a Beidou navigation system and an odometer, relates to the technical field of heavy-haul railway transportation, and comprises the following steps: acquiring first state data of a train based on the Beidou navigation system, and acquiring second state data of the train based on the odometer; performing fusion processing on the first state data and the second state data, determining an observation state of the train at each moment and corresponding state transition probability and state emission probability; and determining an optimal matching path of the train at each moment based on the observation state of the train at each moment and the corresponding state transition probability and state emission probability. The method and system provided by the application improve the positioning accuracy and stability of heavy-haul railway vehicles in complex environments.

This application provides a real-time trajectory data processing method, apparatus, and system. The method includes: acquiring trajectory data of a moving object and performing denoising processing on the trajectory data to obtain valid trajectory point data; acquiring M candidate road network points of the current trajectory point from the valid trajectory point data; where M is a positive integer; determining a variable sliding window corresponding to the current trajectory point; wherein the length of the variable sliding window is K, the variable sliding window includes N first transition probabilities of the previous trajectory point and a first candidate road network point combination corresponding to each first transition probability, each first candidate road network point combination including K historical candidate road network points, where N and K are positive integers; and performing map matching processing on the M candidate road network points based on the variable sliding window to obtain the road network points of the current trajectory point. This achieves high-precision and low-latency real-time map matching of trajectory data.

This invention relates to the field of robot autonomous navigation technology, and provides a laser SLAM localization method and apparatus for scenes lacking geometric features. The method includes: acquiring an original point cloud collected in a target scene; creating a geometric point cloud map of the target scene based on the original point cloud; and creating a marked point cloud map based on the marked point cloud in the original point cloud; matching the real-time point cloud currently scanned by the robot with the geometric point cloud map to determine a geometric residual; matching the real-time marked point cloud in the real-time point cloud currently scanned by the robot with the marked point cloud map to determine a marked residual; determining a total residual based on the geometric residual, the marked residual, and their respective weights; and optimizing the total residual, determining the robot's current position as the position corresponding to the minimum total residual. This invention accurately determines the robot's current position by optimizing the total residual through a weighted average of the geometric residual and the marked residual.

The application provides a four-legged robot adaptive terrain recognition and navigation method for an explosion-proof environment, and relates to the field of high-definition positioning technology.The application quantifies endogenous disturbance and sensing uncertainty of an actuator by synchronously extracting electromagnetic torque pulsation entropy and point cloud shielding entropy rate;cross-modal feature space alignment is achieved by using bilinear interpolation, and matching is performed based on Mahalanobis distance and a high-precision semantic map;by using semantic space consistency score and a left multiplication disturbance model in a Lie group space, a pose correction vector is output to perform manifold projection correction, thereby eliminating nonlinear drift caused by skidding;high-definition navigation or a tactile stepping strategy is triggered according to positioning reliability, and driving gain is dynamically optimized in combination with energy evolution attenuation gradient.The application effectively solves the problems of sensing non-determinacy and positioning misalignment under explosion-proof working conditions, and improves the navigation robustness and sub-centimeter positioning resolution of the robot in an extreme environment.

A method and system for evaluating planning performance of an autonomous driving vehicle (ADV) are disclosed. The method includes the following operations: at an analysis application, receiving a log file recorded by the ADV for a driving scenario in a region, and a high-definition map matching the region; extracting planning messages and perception messages from the log file; and aligning the planning messages and the perception messages based on their timestamps. The method further includes calculating a single performance score for each planning cycle of the ADV for the driving scenario based on the planning messages; calculating a weight for each planning cycle based on the perception messages and the high-definition map; and then calculating a weighted score for the driving scenario based on the single performance scores and their corresponding weights.

The invention relates to a computer-implemented map matching method for mapping a source polyline (1) onto a digital map (2). The computer-implemented method according to the invention is characterized by the following process steps: A1: Transferring the source polyline (1) into the digital map (2) taking into account the geoposition information; A2: Identify all line segments (4) of map polylines (5) that are within a specified local proximity to the source polyline (1) and take these remaining line segments (4) into account for the further procedure; B1: For each line segment (4) of the remaining map polylines (5): Projection of the respective line segment (4) onto the source polyline (1), wherein sections of the respective line segment overlapping with the source polyline (1) are referred to as map segment interval (5.Int) and sections of the source polyline (1) overlapping with the respective map segment intervals (5.Int) are referred to as source segment interval (1.Int); B2: For each map segment interval (5.Int) of the remaining map polylines (5): Calculation of a quality parameter (6) dependent on a length and / or angle ratio between corresponding map segment intervals (5.Int) and source segment intervals (1.Int); and C1: For each remaining map polyline (5): Sum the quality parameters (6) assigned to the individual map segment intervals (5.Int); and C2: Determine the map polyline (5) with the lowest summed quality parameter as the target polyline.

The application provides a road network matching method and device, comprising: obtaining a first road network of a target geographic area in a high-definition map and a second road network of the target geographic area in a standard-definition map; wherein the first road network and the second road network comprise intersection nodes, road lines and the relationship between the intersection nodes and the road lines; matching the intersection nodes in the first road network with the intersection nodes in the second road network to obtain a matching relationship of the intersection nodes; and matching the road lines in the first road network with the road lines in the second network according to the matching relationship of the intersection nodes, thereby achieving strong universality through geometric matching of the road network, and being suitable for road network matching in various scenarios and meeting the matching requirements of the standard-definition map and the high-definition map. In addition, the above scheme calculates the context relationship to deduce the matching, and has a rigorous structure and high accuracy.

Computer-implemented method for map matching of GNSS positions (G1, G2, G3) of a vehicle (1) with location information (P1a, P1b, P1c, P2a, P2b, P3a, P3b) of a digital road map (K), comprising the steps: Acquisition (S1) of GNSS positions (G1, G2, G3) of the vehicle (1) along a route of the vehicle (1); Recording (S2) a distance (E1, E2) determined using a vehicle sensor (10), in particular a wheel speed sensor, between each pair of detected, consecutive GNSS positions (G1, G2, G3) of the vehicle (1); and Map matching (S3) of the determined distance (E1, E2) between the two recorded consecutive GNSS positions (G1, G2, G3) of the vehicle (1) with a route length (L1, L2) between the two recorded consecutive GNSS positions (G1, G2, G3) of the vehicle (1) on the digital road map (K).

This invention discloses a long-term robust localization method accommodating scene changes, relating to the field of autonomous navigation and localization technology for mobile robots. The method comprises three stages: a front-end LiDAR inertial odometry stage, which updates the initial pose and distortion-free point cloud through IMU motion integration and ESIKF updates; a fusion map construction stage, which uses map matching scores and pose consistency as dual criteria, combined with incremental octrees to dynamically insert and delete point clouds, generating a fusion map; and a LiDAR matching optimization stage, which extracts ground point, corner point, and area point features, and outputs high-precision localization results through KNN search and L-M optimization. This invention can simultaneously address changes in the environment both inside and outside the prior map area, improving long-term localization robustness and accuracy while reducing computational resource consumption, making it suitable for scenarios such as logistics delivery and autonomous driving.

The invention discloses a priori knowledge-based map matching method and device, and the method comprises the steps: determining a plurality of road network sections and a plurality of pieces of seed data according to an obtained to-be-matched section; aiming at each network section, calculating to obtain an observation probability logarithmic sum, an initial state probability and a state transition probability logarithmic sum; calculating a feature function value and a plurality of feature weights corresponding to each piece of seed data based on the road network section and the to-be-matched section, and further calculating to obtain a feature score sum of all the seed data; according to the logarithmic sum of the observation probability, the logarithmic sum of the initial state probability and the state transition probability and the feature score sum of all the seed data, the matching score of the road section to be matched and the road network road section is calculated; and taking the road network road section with the maximum matching score as the optimal matching road section of the to-be-matched road section. According to the method, by introducing prior seed data, the matching precision in a complex scene is effectively improved, and the robustness and applicability of a map matching algorithm are improved.

This invention relates to the field of prefabricated building cost estimation technology, and discloses an intelligent matching cost estimation method for design parameters in interior prefabricated construction. The method includes constructing a dynamically coupled spatiotemporal evolution simulation network for prefabricated components to carry and simulate the interaction and change process of design parameters throughout the entire assembly cycle. Based on this network, spatiotemporal trajectory simulation is performed on historical project data to generate a sequence of component evolution paths containing cost correlations. A semantic mapping channel from the evolution path sequence to cost elements is established, and the sequence is parsed into a structured cost influence factor map. For real-time input design schemes, their descriptions are injected into the network for simulation to obtain the evolution path of the target component. The cost influence factor map is used to perform map matching and factor mapping on the target path, automatically outputting refined cost estimation entries. This invention achieves accurate and automated estimation of costs dynamically affected by design parameters.

The application provides a train monitoring turntable control method, system, device and medium, and belongs to the field of rail transit monitoring and control. The method comprises the following steps: acquiring real-time GNSS positioning data of a train; performing map matching on the positioning data by using a pre-established track prior database to obtain accurate matching mileage of the train on a track; querying curvature information of a front track section according to the matching mileage, and dynamically calculating a forward-looking distance based on the curvature; determining a target point in the database according to the matching mileage and the dynamic forward-looking distance, and generating a control instruction to drive the turntable to follow. The automatic tracking control of the turntable is realized under high-speed conditions, and the turntable is all-weather, high-precision and smooth and stable.

This invention discloses a vehicle-mounted real-time localization and mapping method and system based on a domain controller and a lidar, belonging to the field of vehicle-mounted localization and mapping technology. The method includes: receiving and preprocessing 3D point cloud data acquired by a lidar; performing scan-to-scan matching between the current frame and the previous frame, outputting a relative pose transformation matrix and determining the initial pose; calculating the displacement and angle changes between the current frame and the previous keyframe, storing the data in a keyframe database when a final translation threshold or a preset rotation threshold is reached; retrieving the first and second sets of keyframe point clouds, stitching them together to generate a multi-scale local sub-map; performing scan-to-map matching, outputting a globally optimized pose transformation matrix, transforming the point cloud to the world coordinate system and overlaying it onto the global map, outputting a six-degree-of-freedom pose. The system includes a lidar and a domain controller, connected via a high-speed Ethernet cable and a synchronization signal line. This invention achieves high-precision, high-real-time, and computationally adaptive vehicle-mounted localization and mapping.

The present application belongs to the technical field of intelligent operation and maintenance of photovoltaic power stations, and particularly relates to a high-precision photovoltaic module defect positioning method based on a composite-wing unmanned aerial vehicle, comprising the following steps: acquiring image data streams collected by the unmanned aerial vehicle in real time during inspection; using a defect recognition model deployed on the aerial vehicle to perform online processing and target recognition on the image data streams to detect suspected defect targets on the photovoltaic module and output pixel-level bounding box information of the suspected defect targets; when a defect target is recognized, the unmanned aerial vehicle is controlled to enter an automatic hovering state, and high-definition zoom image collection is performed on the defect target; the present application can realize accurate mapping from aerial images to ground module assets by fusing centimeter-level RTK positioning, visual geometric solving and high-precision GIS map matching technology, achieve a positioning effect of "what you see is what you get", and improve the decision-making efficiency of power station operation and maintenance through intelligent management of defect data.

The application discloses a kind of based on sparse cellular signaling data map matching method, equipment and storage medium, the map matching method includes the base station trajectory sequence of acquisition cellular signaling data;Trajectory division is carried out to base station trajectory sequence, obtain each sub base station trajectory sequence;Each sub base station trajectory sequence is spatially coded, and the corresponding space coding sequence is obtained;Each space coding sequence is data enhanced, and the dense representation sequence of sparse cellular signaling data in corresponding sub base station trajectory sequence is obtained;Each dense representation sequence is input into map matching model, and the map matching result of corresponding sub base station trajectory sequence is obtained, and then the map matching result of base station trajectory sequence is obtained.The application improves the accuracy of map matching.

The application discloses a kind of road network trajectory matching method, device and equipment, the method can include: based on the corresponding relation of single-trip trajectory data and road network data, in road network data, determine the sub-road network data corresponding to single-trip trajectory data;Sub-road network data is split based on split length threshold value, and topological relationship reconstruction is carried out to the sub-road network data after splitting;Based on hidden Markov model, the trajectory point included in single-trip trajectory data and the reconstructed sub-road network data are matched to realize road network trajectory matching.The method starts from road network data itself, reconstructs road network structure and topological relationship, and then constructs the topological relationship of sub-road network data more suitable for trajectory point state, so as to improve the upper limit of HMM algorithm matching ability, improve the accuracy of map matching.

The invention discloses a map matching and forward and reverse projection wide remote sensing image segmentation and sample automatic generation method. The method comprises the following steps: step 1, taking an original image, metadata and coastline vector data as an input triple; data cleaning and normalization are carried out through an image preprocessing module; entering a geometric correction and registration module to generate a corrected image with an Alpha channel; 2, performing spatial indexing and matching on the coastline vector data and the corrected image; using a mask generation algorithm to automatically align and rasterize the vector data, and constructing a final true value mask of pixel-level alignment; 3, initializing a reverse mapping grid through a coordinate transformation processing module; calculating a nonlinear mapping relation from geographic coordinates to original sensor pixel coordinates by using a coordinate inverse transformation calculation module; and through resampling processing, accurately backfilling the semantic tag to an original image space, and outputting a final inverse transformation mask. According to the invention, high precision is ensured, and meanwhile, the order of magnitude of processing speed is improved.

This invention discloses a method and system for localizing the calibration of carbon emission factors for motor vehicles, comprising the following steps: S1, collecting high-spatiotemporal precision trajectory data of motor vehicles at the second-by-second level and performing hierarchical completion or elimination processing; S2, projecting the preprocessed discrete trajectory data points onto the road network of a traffic geographic information system and performing map matching to obtain a continuous driving path; S3, dividing the journey into multiple driving units based on the road grade and traffic service level of the matched road segments; S4, extracting features, standardizing them, and then selecting typical operating conditions in the feature space; S5, matching the typical operating conditions with a benchmark emission factor model operating condition library, inheriting or interpolating emission factors for matching operating conditions, and dynamically correcting non-matching operating conditions to obtain emission factors; S6, constructing a localized carbon emission factor library; S7, measuring exhaust gas concentration and calculating the measured emission factors, thereby achieving a high-precision, high-reliability, and verifiable localized motor vehicle carbon emission factor library at a low cost.

This application belongs to the field of UAV positioning technology, specifically involving a UAV positioning method based on two-stage image matching and IMU multi-source fusion, including: Step 1, Initializing UAV pose: Before the UAV enters a complex environment, initial positioning is completed using satellite signals, and the UAV pose is calculated by integration using acceleration and angular velocity information provided by the IMU; Step 2, Aerial image acquisition: Visible light video streams are acquired through the UAV control pod, and aerial images are acquired in real time; Step 3, Real-time map matching: High-precision positioning of the UAV is achieved through two steps: coarse search and fine positioning; Step 4, IMU data fusion: Using the map matching result as the real-time observation value, the IMU integral data and the map matching positioning result are fused based on the Extended Kalman Filter (EKF) algorithm to correct the UAV positioning deviation in real time.

The application discloses a kind of distributed road weight training methods.The method is first obtained target area motor vehicle travel trajectory and is converted into learning target section sequence using map matching algorithm;Subsequently, the road section characteristic vector containing multi-dimensional objective attributes such as road grade, type and real-time traffic speed is constructed, and the training set containing positive and negative samples is constructed in combination with temporarily generated temporary coefficient;The training set is hashed into fragments using distributed computing architecture, the local update result of each training node is summarized by parameter server, and the global road weighting coefficient is generated based on road network topology constraint to perform gradient arbitration;Before outputting the final coefficient, construct virtual traffic to perform road network load balancing deduction, and use the penalty mask fed back to correct the feature dimension that causes abnormal aggregation by weighting reduction.The application realizes the accurate inversion of large-scale road network traffic cost by deep ordering learning and distributed parallel computing, effectively improves the rationality of path planning and the balance of road network flow.

This invention discloses a multi-scale lining image stitching method and computer program product for tunnel detection parallax scenarios, relating to the field of computer vision technology. It solves the technical problem of poor adaptability and difficulty in ensuring stitching quality in traditional lining image stitching methods. The method includes: performing feature map matching on a first image and a second image to obtain an initial global homography matrix; performing multi-scale matrix iterative optimization and generating a global aligned feature map to obtain an optimal global homography matrix; performing local non-rigid adjustment on the second image using a TPS interpolation algorithm; extracting initial masks from the registered first and second images and generating seam masks; finding the optimal seam within the seam masks to generate a content mask; and fusing the registered first and second images based on the content mask to output a panoramic image. This invention exhibits good adaptability to scenarios with varying degrees of parallax, significantly improving the stitching quality of lining images.

The application discloses a path planning method, an electronic device and a computer readable storage medium. The method comprises the following steps: obtaining a plurality of to-be-unlocked tugs forming a deadlock in a map; obtaining available avoidance points of the plurality of to-be-unlocked tugs from a set of avoidance points matched with the map and sorting the available avoidance points to generate an avoidance point sequence; determining a candidate tug corresponding to a candidate avoidance point and a blocked tug other than the candidate tug from the plurality of to-be-unlocked tugs; determining a first detour distance of the candidate tug and a second detour distance of the blocked tug based on the candidate avoidance point; obtaining a maximum allowed detour distance; selecting the candidate avoidance point corresponding to the candidate tug as a target avoidance point based on the first detour distance, the second detour distance and the maximum allowed detour distance; and determining an avoidance path of the plurality of to-be-unlocked tugs based on the target avoidance point. The above scheme can improve the operation efficiency of the tug.

The application relates to the technical field of digital processing of architectural engineering drawings, and discloses a method and system for matching an axis net positioning drawing and a large-scale detailed drawing based on feature points, wherein the method comprises the following steps: acquiring axis net positioning drawing data and large-scale detailed drawing data of the same single item of work on the same floor, and when the large-scale detailed drawing describes cross-floor information, the data can be applied to the axis net positioning drawing of the target floor; obtaining a matched room pair based on the geometric feature similarity of the number of room contour line segments and the length of each line segment; extracting three types of feature point pairs, i.e. room center points, door center points and the most distant corner points from the matched room pair; solving an affine transformation matrix based on the three types of feature point pairs; verifying the effectiveness of the affine transformation matrix through an intersection-over-union ratio; and transforming the lines and components in the large-scale detailed drawing to the corresponding positions in the axis net positioning drawing by using the effective affine transformation matrix.

This invention discloses a task-adaptive trajectory representation learning method, relating to the field of traffic big data analysis technology. The method includes: acquiring GPS trajectories and corresponding road network data; obtaining road segment trajectories through map matching; constructing an alignment matrix based on the GPS trajectories and road segment trajectories; inputting the two types of trajectories into a pre-trained GPS trajectory encoder and a road segment trajectory encoder, respectively, to obtain GPS trajectory embedding sequences and road segment trajectory embedding sequences; constructing a task cue vector based on the target downstream task, using it as task condition information; fusing the two types of embedding sequences through a hierarchical fusion module to obtain a task-specific fusion representation; finally, inputting the fusion representation into a prediction module to output the downstream task prediction result, and updating parameters based on the task loss. This invention achieves adaptive fusion of multimodal trajectory features under different downstream task requirements, improving the task specificity and adaptability of trajectory representation.

The application discloses an end-to-end automatic driving long-tail scene data collection and automatic labeling method. First, multi-modal time series data containing images, point clouds and odometer information are collected; then, an occupancy map is constructed, and perception uncertainty and prediction uncertainty are fused to calculate a comprehensive long-tail scene judgment coefficient, so that a long-tail scene with high risk and high uncertainty is automatically identified; then, only for the data judged as a long-tail scene, a trajectory true value generation process based on particle filtering is started, the process fuses the motion prediction of the odometer and the map matching of the laser radar point cloud, and through state prediction, observation update, resampling and weighted average, a high-quality self-vehicle trajectory point sequence conforming to the kinematic constraint of the vehicle is generated as a true value label. The application realizes accurate screening and automatic generation of true value of the long-tail scene, improves the data collection efficiency and labeling quality, and enhances the safety and generalization ability of the end-to-end automatic driving model in complex scenes.

This application discloses a method, apparatus, medium, device, and vehicle for evaluating a map matching algorithm module. The map matching algorithm module is used to divide two target maps into multiple sub-maps of the same number and match the sub-maps in the two target maps. The method includes: obtaining the relative transformation relationship between each first sub-map matching pair output by the map matching algorithm module and its corresponding relative transformation relationship truth value, wherein the first sub-map matching pair includes a pair of successfully matched sub-maps; based on the difference between the relative transformation relationship between each first sub-map matching pair and its corresponding relative transformation relationship truth value, counting the number of first sub-map matching pairs that meet a preset difference condition; determining whether the number of targets is the same as the number of sub-maps of the target map; and determining that the two target maps are successfully matched if the number of targets is the same as the number of sub-maps of the target map. This scheme can evaluate the map matching effect of the map matching algorithm module.