23 results about "Orthogonal plane" patented technology

The invention discloses a sparse view angle 3D-DSA reconstruction method based on a three-dimensional Poisson generative model. The method comprises the following steps: obtaining pairing data of a sparse view angle 2D-DSA projection drawing and a 3D-DSA reconstruction image; extracting features of a projection image by using a projection domain encoder, then converting features of two-dimensional projection into a three-dimensional image domain according to a geometrical relationship of the cone beam CT, and then obtaining a prior image by using an image decoder; the method comprises the following steps: constructing a three-dimensional Poisson generation model, adding noise to a three-dimensional image patch in a training stage to obtain a disturbance image, outputting a network reconstruction image by taking a prior image as a condition and the three-dimensional image patch before noise addition as a target image, and calculating loss of the output image and the target image to update network parameters; the mean square error loss and the mean square error loss of the maximum intensity projection images of the three orthogonal planes are used during loss calculation; in the sampling stage, random noise is used as input, a prior image is used as a condition, noise of a noise image is continuously denoised within a limited step length, and finally a reconstructed 3D-DSA image is obtained.

The invention relates to a point cloud target detection method and system fusing 2D features, and electronic equipment. The method comprises the following steps: performing voxelization processing on collected original point cloud data to generate a three-dimensional voxel feature tensor; respectively projecting the three-dimensional voxel feature tensors to three orthogonal plane view angles of front view, bird's-eye view and side view to generate a multi-view angle feature map; performing cross-view feature alignment and splicing on the multi-view feature map to generate a composite feature map; optimizing the composite feature map by using a channel attention weighting mechanism to obtain an optimized composite feature map; and target detection is carried out based on the optimized composite feature map, so that the technical problem that the detection recall rate of the target is reduced due to three-dimensional space information loss in a BEV perspective in a traditional method is solved, and the technical effect of improving the detection recall rate of the target in a sparse point cloud scene is achieved.

This invention discloses a three-dimensional modeling method for the true shape of boulders in a soil-rock fill slope, comprising: Step 1, acquiring the two-dimensional projections of the boulders of the actual soil-rock fill slope onto three orthogonal planes; Step 2, obtaining the coordinates of the two-dimensional projection contour points, calculating the centroid of the contour line, moving the centroid to the origin of the coordinate system, and recalculating the coordinates of the contour points; Step 3, segmenting the contour line into spatial regions based on the projections of the boulders on the XOZ and YOZ planes, and randomly reconstructing a contour line; Step 4, performing compatibility analysis on the XOY plane and correcting the reconstructed contour line; Step 5, randomly constructing several contour lines within each spatial region; Step 6, generating a three-dimensional geometric model with a true shape based on the contour lines. The method of this invention constructs a three-dimensional realistic model based on the true shape of the boulders, resulting in higher accuracy in subsequent finite element analysis of the stability of the soil-rock fill slope from a heterogeneous perspective.

The invention belongs to the technical field of block chain and artificial intelligence crossing, and particularly relates to a fair federal consensus method and system based on gradient projection. The objective of the invention is to solve the problem that the fairness of a global model is reduced due to unfair client contribution evaluation and improper projection target selection in a gradient conflict resolution mechanism under the condition of non-independent identically distributed data (Non-IID). The method comprises the following steps: initializing a client key and parameters; randomly selecting m clients for local training in each round; calculating an effort value according to the loss change and the accuracy rate, and preferentially assigning a block creator; the server detects gradient conflicts and performs selective projection resolution, and a gradient with a large length is projected to an orthogonal plane with a gradient with a small length; aggregating the conflict-free gradient to update the global model; and circulating to a preset round number and outputting a final model and a creator sequence. According to the scheme, model performance and participation fairness are considered, and consensus efficiency and excitation compatibility of federated learning in a heterogeneous environment are improved.

The invention discloses a mandibular canal segmentation method and system, electronic equipment and a storage medium. The method comprises the following steps: acquiring an oral maxillofacial region image; cutting the oral maxillofacial region image to obtain a plurality of local images of the mandible; inputting the plurality of local images into a mandibular canal segmentation model, and outputting to obtain a segmentation result; the mandibular canal segmentation model comprises an encoder, a decoder and a feature fusion module, three two-dimensional convolution branches are configured on three orthogonal projection planes in the feature fusion module, and each two-dimensional convolution branch can obtain boundary information, center trend and local bending features of the mandibular canal on different orthogonal planes and fuse the boundary information, the center trend and the local bending features to obtain multi-view features; and fusing the multi-view feature and the local texture feature to obtain an enhanced feature, fusing the local texture feature and the enhanced feature by a decoder, and outputting to obtain a segmentation result. And through the feature fusion module, comprehensive coverage of multi-view geometric features is realized, and the problems of segmentation fracture and boundary blur caused by single feature extraction are effectively avoided.

The present application belongs to the technical field of industrial detection, and particularly relates to a forging round steel straightness online detection method based on point cloud data, which comprises the following steps: firstly, obtaining original point cloud and multi-dimensional physical characteristic data of the round steel in a transmission state; then, slicing the point cloud, and combining local geometric missing features and thermal deformation driving features to calculate a compensation step, and correcting a primary centroid to obtain a real three-dimensional physical axis; subsequently, projecting the axis to an orthogonal plane, and using a restriction relationship between local curvature and material bending stiffness to obtain a pseudo-boundary penalty factor, and then correcting an envelope angle and calculating a comprehensive straightness error. The present application eliminates the interference of point cloud missing, thermal eccentricity and surface microscopic defects, and improves the accuracy of the forging round steel straightness detection.

The application relates to a three-dimensional extended target tracking method of an adaptive Gaussian process under unknown measurement noise. Firstly, a forgetting factor is adaptively updated by recursively calculating the change rate of the Cramer-Rao lower bound trace, so that the dynamic optimization of filter parameters is realized; then, a projection method based on GP is used to represent a three-dimensional cross-shaped target, three-dimensional point clouds acquired by a sensor are projected to three orthogonal planes, and the three-dimensional extended shape of the target is reconstructed by using two-dimensional contour information; finally, a robust state estimator is constructed by fusing variational Bayesian inference and MCC, the joint posterior distribution of unknown measurement noise and system state is solved by variational approximation, and a cost function is constructed based on the maximum correlation entropy criterion, so that the influence of abnormal values is inhibited in a reweighted manner. The tracking method provided by the application can significantly improve the tracking precision and robustness of a three-dimensional extended target under unknown strong non-Gaussian measurement noise.

The application relates to the technical field of image processing, and provides a face micro-expression recognition method and device, electronic equipment and a storage medium, the method comprising the following steps: acquiring a face micro-expression image sequence to be recognized; pre-processing the face micro-expression image sequence; extracting HCTP features of three orthogonal planes of the pre-processed face micro-expression image sequence; determining feature histograms in three dimensions based on the HCTP features; the HCTP features are image features obtained by combining CTP features on the basis of Haar features; the feature histograms in the three dimensions are standardized; the three feature histograms after the standardization are concatenated into a histogram vector; the histogram vector is input into a trained neural network model based on deep learning for classification; and a corresponding expression category is obtained. In this way, by considering the face micro-expression feature information of the three orthogonal planes and the central pixel information, comprehensive features are extracted, and the accuracy of face micro-expression recognition is improved.

The application provides a video rain removal method and system based on a space-time difference state space model, and belongs to the technical field of deep learning. The application solves the problem of how to ensure accurate rain removal while retaining the background texture to the greatest extent and breaking the trade-off between "rain removal completeness" and "background fidelity". The method comprises the following steps: S1: extracting shallow space-time features from an input video; S2: constructing anchor points and reference features by using multi-plane scanning, generating a gating map by calculating geometric consistency, and removing rain information in the features by using a difference mechanism; S3: during training, capturing local space-time details by using orthogonal plane convolution, during inference, enhancing the video background texture by using a single-stream convolution after reparameterization, and outputting a rain-removed video. The system comprises a 3D convolution layer module, a consistency-guided difference gating module, and a reparameterized three-plane feedforward network.

The invention relates to the technical field of point cloud models, in particular to a river topography three-dimensional point cloud dynamic reconstruction method, which comprises the following steps of: acquiring original three-dimensional point cloud data of a target river, and constructing a manifold triangular mesh based on the original three-dimensional point cloud data; according to the method, the complex three-dimensional river form is collapsed into the one-dimensional point column set, the central axis skeleton line is extracted, the topological trend and the central trend of the river can be captured, the orthogonal plane is generated in the tangential direction of the central axis skeleton, the section contour is extracted through the Alpha-Shape algorithm, and parameterized dimension reduction expression of the geometric form of the river is achieved; according to the method, the contour of each section is connected by using a spline interpolation algorithm by taking the central axis skeleton as a scanning path, and the three-dimensional surface model of the river topography with a continuous topological structure is generated, so that the precision and availability of the model in hydrodynamic analysis and engineering calculation are improved, and the technical problem that unstructured scattered point clouds are difficult to be directly used for hydrological feature extraction is solved.

The present invention discloses a method for detecting the grasping state of a flexible manipulator in different postures and motion states, and belongs to the field of tactile signal analysis and processing based on flexible manipulators. The present invention conducts grasping state classification and recognition experiments in five static scenarios: horizontally upward, horizontally downward, horizontally left, horizontally right, and horizontally; and in six dynamic scenarios: translational reciprocating motion in three axial directions, front-back, left-right, and up-down, and semicircular arc reciprocating motion in three orthogonal planes in space. The present invention greatly broadens the experimental scenarios, making the experimental results highly generalizable and universal. The present invention proposes a self-adaptive standardized high-dimensional fuzzy system, and for the first time uses a fuzzy system to solve the problem of grasping state detection of a flexible manipulator. Compared with the grasping state detection results using other machine learning methods, the classification accuracy of the results is higher.

[Problem] To increase the reliability and throughput of measurement of the properties of a film formed on a substrate. [Solution] A substrate is illuminated with polarization components consisting of p-polarization and s-polarization of terahertz waves with mutually orthogonal planes of vibration. Time waveforms of the electric field intensities of the p-polarization and the s-polarization are acquired from a time waveform of the electric field intensity of a polarization component reflected from the substrate. A spectrum of an actually measured amplitude ratio of the p-polarization and the s-polarization, and a spectrum of an actually measured phase difference between the p-polarization and the s-polarization are acquired. A spectrum of a calculated amplitude ratio of the p-polarization and the s-polarization, and a spectrum of a calculated phase difference between the p-polarization and the s-polarization are acquired through simulation using an optical model of the substrate. The properties of a film formed on the substrate are measured by fitting involving at least alignment of the spectrum of the calculated amplitude ratio with the spectrum of the actually measured amplitude ratio. Among a plurality of measurement values used in the measurement of the properties of the film, those with high reliability are employed.

The invention discloses a parametric modeling method based on three-view adaptive projection, which is suitable for a task of automatically generating a structured modeling instruction from three-dimensional point cloud data and can realize CAD (Computer Aided Design) structure reconstruction with high geometric consistency and semantic interpretability. The method comprises the following steps: firstly, extracting global features by using a point cloud encoder to predict a rotation attitude in an Euler angle form, and performing rotation alignment on a point cloud; then projecting the rotated point clouds to three groups of orthogonal planes respectively, extracting a high-quality projection point set by combining adaptive grid sampling of normal filtering and density sensing, extracting structured two-dimensional features by adopting a hierarchical K-nearest neighbor aggregation mechanism based on a three-view projection result, and further guiding a parameter decoder to generate a sketch and a stretching modeling instruction; the method has the remarkable advantages in the aspects of structure alignment precision, complex geometry adaptability and program editability, and is suitable for application scenes such as reverse design, industrial part modeling, building sketch reconstruction and intelligent auxiliary modeling.

The present invention proposes a method for automatically measuring the length of the mitral valve leaflets based on 3D to 2D morphological transformation. The method first obtains three-dimensional medical image data of the patient's heart and extracts six partitions of the mitral valve and related heart structures. By fitting the mitral valve structure point set, the orthogonal plane is calculated, and the cKDTree algorithm is used to determine the connecting areas of each partition, and the corresponding projection plane is established. The contour of the projected point set is then extracted by the Alpha shapes algorithm and corrected to a linear three-dimensional point set, and finally the length of each partitioned leaflet is calculated. This method improves the accuracy, consistency and efficiency of measurement through automation technology, reduces human errors, is applicable to a variety of three-dimensional medical imaging data, and provides a more reliable basis for the clinical diagnosis and treatment of mitral valve disease.

The present invention contributes to suppressing misinsertion of a locking mechanism. The first assembly member (10) has: an orthogonal wall (11) provided with an orthogonal plane (11a); a through-hole (12) provided in the orthogonal wall; a partition wall (13) as a first locking part (31) that divides the through-hole into a first through-hole (12a) and a second through-hole (12b); and a protruding portion (14) protruding from the orthogonal plane, the second assembly member (20) having an abutting wall (22) with a distal end abutting against the orthogonal plane of the first peripheral edge portion (17a), the second locking portion having: a locking piece (32a) inserted into the first through-hole; and a locking claw (32b) which is provided at the free end of the locking piece extending from the first through-hole and which hooks and locks the first locking part, and a protruding part (14) which protrudes from an orthogonal plane of the third peripheral part (17c) closer to the second through-hole than the first through-hole, and locks the tip of the contact wall when the second locking part is inserted into the second through-hole, and the protruding part (14) protrudes from the orthogonal plane of the third peripheral part (17c) and locks the tip of the contact wall when the second locking part is inserted into the second through-hole. The second assembly member is made to float when inserted into the first through hole relative to the second locking portion.

A system and method for generating sample data for analysis provide an image capture unit configured to capture a stack of images in image layers through a thickness of a sample, each image layer comprising pixel data in two orthogonal planes across the sample at a given sample depth; a processing unit configured: a) to process the captured pixel data to determine therefrom a pixel value of the energy of each pixel of the image, b) to select from each group of pixels through the stack of images the pixel having a highest energy; and c) to generate an output image file comprising a set of pixel data obtained from the selected pixels, wherein the output image file comprises for each pixel, the pixel position in the two orthogonal planes, the pixel vale and the depth position of the pixel in the image stack. The depth position of the selected pixel is provided in a fourth channel of the output image file, which represents a topography of a sample. An analysis unit comprises an input for receiving the output image file and to determine therefrom sample data, including identification of constituents in the sample and / or quantity of said constituents in the sample; and preferably comprises an artificial intelligence.

To contribute to preventing incorrect insertion of the locking mechanism. [Solution] The first assembly member 10 has an orthogonal wall 11 on which an orthogonal plane 11a is provided, a through hole 12 provided in the orthogonal wall, a partition wall 13 as a first locking part 31 that divides the through hole into a first through hole 12a and a second through hole 12b, and a projection 14 that protrudes from the orthogonal plane, and the second assembly member 20 has a contact wall 22 that contacts the orthogonal plane of the first peripheral edge 17a at its tip, and the second locking part is inserted into the first through hole The mounting part has a locking piece 32a that can be inserted, and a locking claw 32b provided at the free end of the locking piece that comes out of the first through hole, which hooks onto the first locking part to lock it in place. The projection 14 protrudes from the orthogonal plane on the second through hole side of the first through hole at the third peripheral edge 17c, and when the second locking part is inserted into the second through hole, it locks the tip of the contact wall, causing the second mounting member to float higher than when the second locking part is inserted into the first through hole.

The embodiment of the invention provides an intelligent driving method and device based on a world model, equipment and a storage medium, and belongs to the technical field of intelligent driving. The method comprises the following steps: performing feature coding processing on each 3D occupancy map of a target vehicle to obtain coding features, and performing axial pooling processing on the coding features along three coordinate axes to obtain plane features corresponding to three orthogonal planes; performing resampling processing on a plurality of plane features under the same orthogonal plane, predicting plane change information of the current orthogonal plane at a future moment, and integrating the plane change information of the three orthogonal planes to obtain predicted change information; and generating a query vector based on the plane feature corresponding to the current moment and the predicted change information, performing feature coding and decoding processing on the executed action sequence based on the query vector, and predicting and outputting a target execution action of the target vehicle at the future moment. According to the invention, the precision of the output result can be improved while the lightweight calculation of the world model is improved.

The invention discloses a rock mass deformation monitoring method and a computer program product, and the method comprises the steps: carrying out the spherical target extraction of point cloud data in different periods, and obtaining the spherical center coordinates and the target radius of each spherical target in different periods; and determining the sphere center displacement offset of each spherical target in different periods according to the sphere center coordinates and the target radius, and calculating the displacement deviation spectrum of each point cloud section line in different periods by adopting a section line point cloud registration algorithm based on the sphere center coordinates of each spherical target in different periods. And three-orthogonal-plane point cloud data is constructed based on the sphere center target of the minimum sphere center displacement offset. And importing the three-orthogonal-plane point cloud data into the point cloud data of different periods, and introducing the selected displacement deviation spectrum as geometric constraint to carry out overall registration on the point cloud data of different periods. The displacement variation of the point cloud is calculated through the directed distances from the point cloud to the three orthogonal planes in different periods after overall registration, the overall deformation of the rock mass is reflected, and the overall deformation trend of the rock mass and the deformation trend of the key area are macroscopically reflected.

An apparatus includes an x-ray source configured to generate x-rays, at least some of which impinge a sample and at least one diffractor configured to concurrently diffract at least some of the x-rays from the sample in a first direction and in a second direction substantially orthogonal to the first direction. The apparatus further includes at least one two-dimensional (2D) position-sensitive x-ray detector configured to receive at least some of the diffracted x-rays and to concurrently generate first spectral information of the x-rays diffracted in the first direction and second spectral information of the x-rays diffracted in the second direction. The apparatus further includes circuitry configured to receive the first and second spectral information and to generate a single x-ray absorption spectroscopy (XAS) spectrum using the first and second spectral information.

To provide a shield tunnel construction evaluation method and a shield tunnel construction evaluation system capable of suppressing cost more than before.SOLUTION: The shield tunnel construction evaluation system 100 includes a three-dimensional measuring instrument 20 for measuring three-dimensional coordinates of an inner peripheral surface of a segment ring and an inner peripheral surface of a skin plate, and a shield tunnel construction evaluation device 40 for evaluating a construction state of the shield tunnel based on a measurement result of the three-dimensional measuring instrument 20. The shield tunnel construction evaluation device 40 obtains a tail clearance based on a first shape obtained by projecting a plurality of first measurement points on the inner peripheral surface of the skin plate onto an orthogonal plane orthogonal to the machine axis of the shield machine, a second shape obtained by projecting a plurality of second measurement points on the inner peripheral surface of the segment ring onto the orthogonal plane, and the thickness of the segment ring.SELECTED DRAWING: Figure 3

The application relates to the technical field of medical ultrasonic imaging, in particular to a palm ultrasonic 2.5-dimensional imaging method and system based on artificial intelligence guidance, which comprises the following steps: loading a reconstruction task package and recording the initial posture of a probe; collecting 2-dimensional ultrasonic images and motion data in real time, and obtaining the optimized posture of each frame of image through tight coupling optimization; mapping each frame of image to a three-dimensional voxel space based on the optimized posture to form a 2.5-dimensional original data set; performing spatial coverage and multi-dimensional quality scoring on the 2.5-dimensional original data set; generating a dynamic guidance interface based on the scoring; starting online incremental reconstruction in real time and displaying the optimal orthogonal plane; after scanning termination, performing global optimization and cavity repair on the complete data set to generate a structured examination report. Through the construction of a reconstruction task package and a tight coupling optimization algorithm, the application realizes closed-loop intelligent guidance and real-time reconstruction feedback of palm ultrasonic 2.5-dimensional imaging, effectively reduces the operation threshold, and improves the degree of diagnosis standardization.

The present invention discloses a volume electron microscope data visualization method, device and storage medium for generating instantiated projections, relating to the field of image processing technology, comprising: performing dot multiplication on a texture data block of volume electron microscope data and a segmentation mask block corresponding to the volume electron microscope data to obtain instantiated texture data filtered by the mask, and superimposing a marked subcellular structure on the instantiated texture data filtered by the mask to obtain instantiated texture data with superimposed marks; converting the instantiated texture data with superimposed marks on a plane at any angle into an orthogonal plane by data rotation, and using a projection function to obtain an instantiated projection image on the current orthogonal plane and normalize it; synthesizing the normalized projection image after spatial scale correction, thereby outputting a visualization result corresponding to the current ID; the volume electron microscope data visualization method can quickly generate a visualization format containing morphology and texture of multiple objects in the volume electron microscope data, thereby realizing rapid retrieval, verification and analysis of the volume electron microscope data.