105 results about "Direction vector" patented technology

The present application relates to the technical field of computer vision, and discloses a smoke key point detection method based on a multi-scale heat map, which solves the technical problems that smoke detection has unstable direction representation and weak interpretability, the global and local collaborative modeling ability is insufficient in a complex scene, resulting in poor robustness, lacks existence constraint, and has low calculation efficiency, and it is difficult to balance accuracy and real-time deployment; the present application captures the overall diffusion trend and spatial long-range relationship of smoke through self-attention mechanism and global dependency modeling, extracts fine features such as smoke edge and texture through local multi-scale context aggregation; the existence judgment is introduced as a pre-constraint to ensure that the direction vector calculation is only performed when it is confirmed that there is a smoke target, thereby avoiding the false triggering problem in the smoke-free scene; the direction angle is converted into the heat map prediction of the two specific key points of the starting point and the ending point, which effectively suppresses the output fluctuation caused by the semi-transparent and fuzzy boundary of smoke.

This invention relates to the field of industrial visual inspection technology and discloses a method for detecting machined holes in a mobile phone frame based on visual inspection. The method includes: acquiring an image of the subject to be inspected and extracting a set of sub-pixel edge point clouds; calculating the observed gradient direction vector and curvature description parameters of the edge points; matching theoretical direction vectors from a structured database; establishing geometric consistency weights by calculating the dot product of the direction vectors; constructing a weighted spatial distribution; constructing a weighted residual functional and solving for the minimum value; and outputting the detection result. This invention establishes a manifold alignment mechanism between the physical field and the theoretical field, thereby constraining the visual features by the design semantics, effectively attenuating tool mark artifacts, and ensuring that the detection parameters have deterministic convergence to the ideal model under complex interference.

This invention relates to the field of artificial intelligence technology and discloses a linear attention mechanism method, device, and electronic device for visual tasks. The method includes: performing norm decomposition and direction decomposition on the input query vector and key vector respectively to obtain the magnitude and unit direction vector corresponding to the query vector and key vector; designing a query norm-driven dynamic entropy reduction control mechanism and a cosine suppression nonnegativity preservation mechanism to obtain the query kernel function feature vector and the key kernel function feature vector; and finally calculating the linear attention result based on the query kernel function feature vector and the key kernel function feature vector. This invention significantly improves the model's ability to process information of different intensities while maintaining linear computational complexity, enhances the model's performance in complex tasks, and thus achieves a balance between efficiency and expressive power, effectively solving the problem of the lack of correlation between norm and entropy value in traditional linear attention mechanisms.

The application discloses a model similarity judgment method and device, equipment and a storage medium, and relates to the technical field of computer vision, which comprises the following steps: determining each distribution direction vector according to three-dimensional point cloud data of a to-be-judged model; generating two-dimensional point cloud data according to each distribution direction vector and the three-dimensional point cloud data; generating a triangle set of the to-be-judged model according to each two-dimensional point cloud data; and if a triangle similar to each triangle in the triangle set of the to-be-judged model is searched in a triangle set of a target model, it is determined that the to-be-judged model is similar to the target model. In the foregoing manner, the three-dimensional model local coordinate system constructed with each distribution direction vector as the coordinate axis direction is used to perform dimension reduction processing on the three-dimensional point cloud data, so that the data calculation amount is reduced and the calculation efficiency is improved, and then the similarity between the models is judged through the similarity of the triangles, so that the calculation cost of judging the similarity of the models can be effectively reduced, and the efficiency of judging the similarity of the models is improved.

The application discloses a kind of fruit tree image acquisition method and system based on two-dimensional skeleton and viewpoint planning, it is related to wisdom agricultural equipment technical field, method includes: obtaining crown image obtains branch mask and depth information;To mask thinning pruning obtains initial skeleton, identifies feature point and calculates growth direction after artifact removal;Based on feature point and direction vector, topological reconstruction is carried out to broken skeleton, and complete two-dimensional skeleton graph is obtained;Based on two-dimensional skeleton graph, viewpoint planning is carried out in reference depth plane, generates viewpoint sequence and is mapped into three-dimensional pose, realizes autonomous image acquisition.The application solves the problem of information loss caused by occlusion and three-dimensional reconstruction calculation lag, and realizes real-time, safe and high-resolution image acquisition.

This disclosure relates to a method for solving the fatigue failure probability function of a turbine shaft based on extended dimensionless integration. The method includes: standardizing the input variables under a first distribution parameter and determining the fatigue failure limit state function; obtaining multiple unit direction vectors in the standard normal space and determining a first limit state surface based on the fatigue failure limit state function; filtering the unit direction vectors in the standard normal space corresponding to the first distribution parameter to obtain initial samples and effective unit direction vectors; transforming the initial samples on the first limit state surface to a second limit state surface in the standard normal space corresponding to the second distribution parameter; using the information from the transformed samples, determining the target distance between the origin of the coordinate system in the standard normal space along the effective unit direction vector and the second limit state surface using an interpolation strategy; and using dimensionless integration to determine the failure probability function value corresponding to the second distribution parameter based on the target distance.

An artificial intelligence capsule positioning system is provided. An artificial intelligence capsule positioner of the artificial intelligence capsule positioning system collects multiple pieces of measured data related to a capsule. The multiple pieces of measured data include a direction vector and a spatial coordinate of the capsule, frequencies and strengths of signals transmitted by a signal transmitter on the capsule, and strengths of the signals received by antennas. The artificial intelligence capsule positioner uses an artificial intelligence machine learning algorithm to learn the multiple pieces of measured data from a database to analyze rules for how a position of the capsule changes with the multiple pieces of measured data to establish a training model. When a new capsule flows in a human body, the artificial intelligence capsule positioner uses the training model to position the new capsule.

A beamforming design method for intelligent metasurface arrays based on orthogonal time-frequency-space technology includes: aggregating multipath channels under direct illumination into cascaded channels with near-field direction vectors; randomly initializing the phase shift of intelligent metasurface units; establishing an effective channel matrix based on the cascaded channels; then establishing a multi-user and rate objective function; maximizing the objective function through phase shift optimization using a hierarchical search strategy; repeatedly executing the hierarchical search strategy to search and optimize subsequent blocks until the number of intelligent metasurface units in each block reaches a preset value; then performing an exhaustive search in the final block, thus completing one iteration; repeating the iteration until the maximum number of iterations is reached; outputting the final optimal phase shift; and implementing beamforming design based on the final optimal phase shift. This invention can reduce the complexity of beamforming design and ensure reliable transmission of high capacity during high-speed movement.

This invention discloses a geological disaster early warning method based on remote sensing technology, belonging to the field of geological disaster early warning technology. The method includes: acquiring remote sensing image data with at least two different imaging principles within the same geographic unit and the same observation time window; performing geometric correction on the remote sensing image data under a unified coordinate system; and converting the remote sensing image data into corresponding physical quantity data and completing scale consistency processing through sensor response conversion and radiometric calibration; using the remote sensing image data and digital elevation data under the unified coordinate system, calculating the slope and aspect of each pixel, and constructing the light propagation path based on the solar altitude angle and solar azimuth angle. This invention models the influence range of environmental factors from the perspective of propagation structure by constructing light propagation paths and heat diffusion propagation paths under a unified coordinate system and calculating the direction vector field formed by the gravity direction of the slope. This allows anomaly identification to be based not only on numerical changes but also on the analysis of propagation mechanisms.

This invention relates to a method for solving the inverse kinematics of a six-DOF offset serial manipulator based on dual quaternions, belonging to the field of manipulator design technology. The method is as follows: Six-DOF offset serial manipulators are classified into Type I and Type II. A reference coordinate system and an end-effector coordinate system are established for each type. The link lengths, position coordinates, unit direction vectors, and line spacings of each joint are defined, thus obtaining the Plück coordinates of each joint. Inverse kinematics analysis and solutions are performed for both Type I and Type II manipulators. The solution method is then validated. This invention achieves efficient inverse kinematics solving, meeting the real-time requirements for inverse kinematics solving speed in industrial scenarios and improving the control performance and application efficiency of manipulators in high-precision operation scenarios.

This invention belongs to the technical field of material physical property analysis, and relates to a method and system for predicting the crack propagation trend of heat sinks using fractal geometry. The method includes the following steps: generating a two-dimensional original matrix of the infrared temperature field; inverting the stress source deformation kernel matrix using a preset empirical Green's space derivative function to generate a six-component moment tensor; generating a principal stress azimuth vector; generating an anisotropic rectangular metric grid; generating a multifractal singular extremum clustering image of the thermal field; generating the steepest fractal gradient vector; and coupling the normalized principal stress azimuth vector and the steepest fractal gradient vector into a preset multi-field parameter feature weighted superposition model to output the current crack propagation amount and crack propagation direction vector. This invention solves the problem that existing nondestructive testing technologies suffer from uncertainty in predicting crack propagation paths and rates due to the failure to effectively integrate microscopic mechanical stress field and macroscopic heat dissipation field information.

The application discloses a method for establishing and detecting a quick arbitrary shape bounding box of a mechanical arm and relates to the field of collision detection of the mechanical arm. The method comprises the following steps: obtaining an i-th coordinate system of an i-th rod in the mechanical arm, calculating a first homogeneous transformation matrix of two adjacent coordinate systems; calculating a second homogeneous transformation matrix between the coordinate system and a base coordinate system; calculating an origin of the i-th rod and obtaining orthogonal direction vectors of three axes in the coordinate system of the i-th rod; the i-th rod in the mechanical arm has n end points, the maximum distance in all distances of the n end points in the positive and negative directions of the orthogonal direction vectors is obtained; a first bounding box is obtained based on the maximum distance and the origin; when the mechanical arm moves, an updated second bounding box is obtained based on the second homogeneous transformation matrix and a homogeneous transformation matrix between the base coordinate system and a world coordinate system, and the application further discloses a collision detection method based on the updated bounding box, which greatly improves the calculation efficiency and meets the high-frequency real-time requirement.

The present disclosure provides an animation processing method and device, a computer storage medium and an electronic device, and relates to the technical field of computers. The method comprises the following steps: determining a rotation angle on a target direction axis according to a first horizontal plane orientation vector of a reference skeleton posture and a second horizontal plane orientation vector of a skeleton posture to be evaluated, and aligning the two to obtain an aligned skeleton posture according to the rotation angle; for the same skeleton segment, determining an included angle between a first skeleton segment direction vector of the skeleton segment in the reference skeleton posture and a second skeleton segment direction vector of the skeleton segment in the aligned skeleton posture to obtain a skeleton direction deviation of the skeleton segment; generating a comprehensive deviation based on the skeleton direction deviations of the skeleton segments; obtaining a maximum deviation value in the skeleton direction deviations, and applying threshold constraint to the comprehensive deviation according to a comparison result of the maximum deviation value and at least one deviation threshold to generate a motion evaluation result. The method can be applied to motion comparison between skeletons of different character body types, and can improve the motion evaluation quality.

The present application relates to a kind of topological perception spray trajectory planning methods for large size multi-curvature spacecraft spraying, comprising obtaining spacecraft surface point cloud data and constructing undirected topological graph and constructing the high-dimensional fusion feature vector of each vertex;Spacecraft surface point cloud data is segmented into several spray topological regions based on graph convolutional neural network;Global gravity direction vector is introduced to construct joint optimization objective function on each spray topological region and solve, construct global geodesic distance scalar field to generate smooth spray curve, construct the local spray trajectory of each spray topological region;Global path planning is carried out to spray topological region, and the global optimal region traversal sequence is obtained, and then global spray trajectory is obtained.The present application is suitable for spacecraft complex curved surface heat protection coating spraying scene, effectively eliminates the process defects of large curvature transition zone, significantly improves the thickness uniformity, quality consistency and process reliability of special coating manufacturing of spaceflight.

The application belongs to the technical field of shield tunnel construction, and discloses a shield segment bolt hole alignment deviation detection method based on feature matching, which comprises the following steps: S1, collecting point cloud data; S2, three-dimensional point cloud data set; S3, point cloud data subset; S4, identifying the circular hole area existing on the plane; S5, using the Euclidean clustering algorithm to divide the point cloud data of each circular hole area into independent point cloud clusters; S6, outputting the direction vector of the cylindrical center axis of each bolt hole and the three-dimensional coordinates of the cylindrical bottom center point; S7, finding the bolt hole closest to the bolt hole on the segment to be assembled and approximately parallel in the axial direction for pairing; S8, calculating the alignment deviation of each pair of bolt holes, comparing the alignment deviation, and judging whether each pair of bolt holes is aligned and qualified. The application is intelligent detection, high in efficiency, avoids the influence of manual intervention on the accuracy of data, and is easy to store and check the detection record.

This invention relates to a clustering method for observation targets based on single-satellite continuous observation constraints, comprising the following steps: at each moment, determining whether the observation target is physically potentially visible relative to the satellite using off-axis angle constraints and detection distance constraints; when the observation target is physically potentially visible relative to the satellite, using direction vector similarity analysis to extract the spatial nearest neighbor targets of the observation target, constructing a set of satellite line-of-sight candidate directions pointing towards the geometric center of the observation target group; selecting the optimal satellite line-of-sight candidate direction that covers the most observation targets from the set of satellite line-of-sight candidate directions, and obtaining the optimal coverage target set under the optimal satellite line-of-sight candidate direction; and taking the intersection of the optimal coverage target sets at all moments within the sliding window, selecting the intersection target group that is always located in the optimal field of view corresponding to the optimal satellite line-of-sight candidate direction within the sliding window as the continuous observation target set.

This invention provides a method, system, and device for controlling a character on a sphere's surface. A coordinate system dynamic correction module calculates the direction from the sphere's center to the character's current position in real time to determine a local ascent vector. Based on this local ascent vector, the character's rotational posture is corrected so that the character's feet always face the sphere's center, and a local coordinate system is constructed using the local ascent vector. A vector projection module acquires the camera's direction vector and performs orthogonal projection processing on it using the local coordinate system to generate a tangential movement vector parallel to the tangential plane of the ground surface. This tangential movement vector is then used to drive the character to perform displacement operations. A terrain adsorption module responds to the displacement operations by emitting a probe ray based on the local ascent vector direction determined by the local coordinate system to perform sampling processing on the ground grid, acquiring the intersection point between the probe ray and the ground grid, determining the character's target height based on the intersection point, and correcting the character's position using the target height so that the character adheres to the ground grid in real time during movement.

The application relates to a fast prediction method and device for sound scattering characteristics of a complex target under limited beam incidence, which comprises the following steps: constructing a three-dimensional geometric model of the complex target, and dividing the surface into a triangular facet grid; based on the set incidence parameters of the limited beam, calculating the included angle between the direction vector of the triangular facet to the sound source and the beam direction vector, and screening the illuminated facets in response to the included angle being less than half of the beam angle; calculating the scattering contribution of all the screened illuminated facets in the frequency domain, coherently superimposing the scattering contribution, obtaining the frequency domain scattering sound field of the complex target under the limited beam incidence, and interpolating the frequency domain scattering sound field in the wideband frequency range to obtain the wideband scattering response, and calculating the wideband time domain scattering waveform of the complex target under the limited beam incidence. The application can be applied to the near-field incidence condition of the limited beam, can consider the calculation accuracy and efficiency, and can process the underwater target sound scattering characteristics of the complex target geometric shape.

This invention discloses a GPU-based method, device, and medium for rapid 3D orebody inversion, relating to the field of 3D orebody inversion technology. The method includes importing multi-source exploration data, generating observation data in a unified coordinate system, establishing a 3D voxel grid, and acquiring a data acquisition configuration package; performing multiphysics forward modeling to generate normalized prediction vectors and normalized residual vectors, establishing the forward and transpose interfaces of sensitivity operators, constructing update direction vectors, and generating a pre-projection parameter field; sequentially performing box-constrained projection, total variation projection, and wavelet sparse projection on the pre-projection parameter field to obtain a sparse parameter field and generate the final parameter field; generating a 3D grade voxel array, calculating grade segmentation thresholds, extracting the orebody connected body set, generating a orebody triangular grid and an orebody existence probability voxel array, and acquiring a digital twin base data package. This method shortens computation time, reduces storage pressure, and improves the accuracy and practicality of the inversion.

This invention relates to the field of energy storage dispatching technology, specifically disclosing an end-to-end energy storage dispatching strategy generation system. The system includes a load forecasting module, a dynamic programming optimization module, and a rolling closed-loop dispatching module. The load forecasting module constructs historical daily load data into load feature vectors in a high-dimensional space, retrieves similar historical days by vector distance, and generates a baseline load forecast vector by weighted summation. The dynamic programming optimization module maps the dispatching problem to a two-dimensional state space of time period and state of charge, encodes charging and discharging actions by using the direction vector and amplitude vector of the state transition edges, and solves for the optimal path to obtain the charging and discharging power command sequence. The rolling closed-loop dispatching module constructs the deviation between the actual load and the forecast into a deviation vector, transforms it in the two-dimensional state space, projects it to correct the forecast for the remaining time period, and then calls the dynamic programming optimization module to resolve the problem and extract the current command for issuance. This invention achieves full-process vectorized processing of energy storage dispatching.

This invention relates to a method and system for resisting satellite navigation spoofing using an array antenna based on integrated navigation, belonging to the field of integrated navigation technology. It can improve system robustness and solve the problem of satellite navigation receiver terminals being susceptible to external spoofing interference. The method includes: S1, calculating the current carrier attitude based on the angular rate, specific force information, and magnetic flux information output by the inertial measurement unit; S2, calculating the projection of the true direction vector of the currently visible satellite in the geographic system, then projecting it onto the array antenna coordinate system according to the carrier attitude, and calculating the elevation angle and azimuth angle of the visible satellite relative to the array antenna coordinate system; S3, sampling and down-converting the satellite signal of the current epoch using digital frequency storage technology, and continuously adjusting the zero-point direction using array antenna directional nulling technology to detect whether a spoofing signal enters; S4, determining the direction of arrival of the spoofing signal; S5, shaping the array antenna radiation pattern; S6, adjusting the beam pointing to suppress the entry of spoofing signals.

The present application relates to radar wind field data related technical field, specifically including an airport wake vortex radar wind field data enhancement method based on spectral width anomaly detection, the method comprises: based on the initial error matrix, the spatial error component of each detection partition, configure dynamic data enhancement matrix, under the constraint of wake vortex dissipation time iteration optimization enhancement parameter, carry out the enhancement decision of wind field data deviation. The technical problems of unable to accurately decompose the spatial error component of horizontal wind direction angle and vertical wind shear, easy to confuse the action law of different interference sources, and difficult to effectively improve the signal-to-noise ratio of wind field data are solved, the technical effects of introducing the spatial correlation mapping of the detection partition wind field direction vector and the wake vortex influence area, optimizing the enhancement coefficient, accurately splitting the horizontal and vertical direction spatial error component, extracting the environmental disturbance enhancement factor, efficiently decoupling the atmospheric turbulence, electromagnetic interference and equipment noise, improving the space-time consistency of wind field data, and reducing the wake vortex early warning false alarm rate are realized.

The application provides a tunnel construction equipment positioning method, which can be applied to the technical field of tunnel monitoring. The method comprises the following steps: collecting a three-dimensional point cloud of a target tunnel section where construction equipment is located, obtaining a cylindrical point cloud representing the surface profile of the target tunnel section, and forming a circular cross section of the cylindrical point cloud target tunnel section; fitting a direction vector of the central axis of the target tunnel section based on the cylindrical point cloud; obtaining a pose offset parameter of the construction equipment relative to the target tunnel section based on the spatial relationship between the direction vector and a preset reference point of the construction equipment; and correcting the pose of the construction equipment based on the pose offset parameter. The application also provides a tunnel construction equipment positioning device and a storage medium.

The present application belongs to the technical field of pipeline nondestructive testing, and relates to a kind of oil and gas pipeline elbow feature recognition method based on attitude data, step 1, through Savitzky-Golay filtering and resampling pretreatment attitude data;Step 2, calculate three-dimensional direction vector based on pitch angle and heading angle;Step 3, calculate the angle change of continuous point space vector;Step 4, combined with the candidate area preliminary screening of girth weld data;Step 5, elbow type division based on geometric features;Step 6, elbow local three-dimensional trajectory reconstruction;Step 7, calculate the radius of curvature by three-point circle determination method.The present application realizes elbow recognition in small-bore oil and gas pipeline through girth weld auxiliary positioning, space vector calculation and local three-dimensional reconstruction algorithms, and describes the spatial local geometric features of elbow area.

The application discloses a terrain occlusion fast calculation method based on GPU. The method takes a digital elevation model (DEM) as input, determines an observation point and a region to be calculated, generates a plurality of rays in an azimuth angle range covering the region, and performs ray traversal on a GPU in parallel: threads calculate a direction vector according to a ray number, perform step sampling at a preset step length, and map to a DEM grid, calculate an elevation angle or a slope parameter of a sampling point relative to the observation point, and maintain an extreme value, to complete a visible / occlusion determination, so that an initial occlusion matrix is obtained; further, morphological erosion post-processing is performed on the initial occlusion matrix to suppress holes and boundary discontinuity caused by discrete sampling, and a stable terrain occlusion distribution result is output.

The application discloses a multi-antenna receiving beam forming method of GMSK frequency hopping signals, which comprises the following steps: based on a known GMSK training sequence and by an approximate maximum posterior probability method, accurately estimating a real direction of an expected signal, so as to effectively correct a mismatched direction vector; on this basis, further adopting a covariance matrix reconstruction technology, removing an expected signal component from a sampling covariance matrix, effectively suppressing the expected signal cancellation phenomenon, and being able to perceive a spatial interference distribution and form a relatively accurate null in the interference direction, and significantly improving the output signal-to-interference-and-noise ratio performance of the system.

This invention relates to a clustering method for observation targets based on continuous observation constraints of two satellites, comprising the following steps: at each moment, determining whether the observation target is physically potentially visible relative to the satellite using off-axis angle constraints and detection distance constraints; when the observation target is physically potentially visible relative to the satellite, using direction vector similarity analysis to extract the spatial nearest neighbor targets of the observation target, constructing a set of satellite line-of-sight candidate directions pointing towards the geometric center of the observation target group; selecting the optimal satellite line-of-sight candidate direction that covers the most observation targets from the set of satellite line-of-sight candidate directions; and setting the sliding window length to seconds, selecting the optimal set of observation targets at each moment within the sliding window, and taking the intersection of all optimal sets of observation targets to select the intersection target group that is always located in the optimal field of view corresponding to the optimal satellite line-of-sight candidate direction within the sliding window as the continuous observation target set.

This application provides a path planning method, apparatus, electronic device, and medium for transfer tasks. The method includes: receiving a transfer instruction including a starting position, an ending position, a first direction vector from the starting position to its adjacent main road, and a second direction vector from the adjacent main road of the ending position to the ending position; determining a first straight path based on the first direction vector, determining a second straight path based on the second direction vector, and using the endpoints of the first and second straight paths located on the main roads as the first and second positions; starting from the first position, performing a path search with the straight-line movement cost being a preset value and the turning movement cost being N times the preset value, so that the searched path from the first position to the second position is the target path with the minimum total movement cost; and connecting the first straight path, the target path, and the second straight path end-to-end to obtain the final travel path. This method reduces the number of turns in the path.

This invention discloses a GPU-based method, device, and medium for rapid 3D orebody inversion, relating to the field of 3D orebody inversion technology. The method includes importing multi-source exploration data, generating observation data in a unified coordinate system, establishing a 3D voxel grid, and obtaining a data acquisition configuration package; performing multiphysics forward modeling to generate normalized prediction vectors and normalized residual vectors, establishing the forward and transpose interfaces of the sensitivity operator, constructing update direction vectors, and generating a pre-projection parameter field; sequentially performing box-constrained projection, total variation projection, and wavelet sparse projection on the pre-projection parameter field to obtain a sparse parameter field and generate the final parameter field; generating a 3D grade voxel array, calculating the grade segmentation threshold, extracting the orebody connected body set, generating a orebody triangular grid and an orebody existence probability voxel array, and obtaining a digital twin base data package. This method shortens computation time, reduces storage pressure, and improves the accuracy and practicality of the inversion.

The application provides a contour-based 3D glue coating anomaly detection method and device, wherein the method comprises the following steps: projecting three-dimensional point cloud data of a glue coating contour on a plane to obtain a projection data set, calculating a unit direction vector of two points based on random two points in the projection data set, dividing a maximum inner point set according to the normalized distance of all points in the projection data set to the unit direction vector, fitting a direction vector according to the maximum inner point set to obtain a reference straight line, correcting the projection data set through rotation and translation of the reference straight line to obtain a target point cloud, measuring the glue coating contour based on the target point cloud to obtain a measurement sequence, detecting the glue coating contour according to the measurement sequence to obtain a glue coating anomaly detection result. The application can correct the contour data, improve the accuracy and robustness of the detection result, and detect the contour data with multiple curve characteristics, thereby expanding the application range of the detection method.