28 results about "View plane" patented technology

The invention belongs to the technical field of automatic driving, and relates to a Transform attention mechanism-fused point cloud target detection method and system, and the method comprises the steps: 1, point cloud preprocessing: mapping an irregular sparse point set into a cylindrical voxel tensor of a fixed structure; 2, PillarVFE feature coding is carried out, and point-channel joint feature optimization is realized while the lightness of the network is maintained; 3, generating a pseudo image by the BEV backbone network, and mapping the pilar features to a two-dimensional aerial view plane after feature coding enhancement to form a dense BEV feature map; 4, outputting by a detection head, completing multi-scale feature extraction and context sampling fusion, and finally predicting a three-dimensional bounding box by the detection head; according to the method, the three-dimensional detection performance is remarkably improved in sparse, small-target and long-distance scenes while high reasoning efficiency is kept, and the method has good engineering practicability and research and popularization value.

The invention relates to the technical field of building models, in particular to a BIM-based building construction drawing design model construction method. The method comprises the following steps: extracting three-dimensional space coordinates and construction attributes of a component from a preset BIM building design model, wherein the construction attributes comprise layered material information of the component and the material thickness of each layer; projecting the three-dimensional space coordinates of the component to a two-dimensional view plane to generate a two-dimensional projection contour; calculating the vertical distance from the component to the view plane, and determining the display level of the component according to the size of the vertical distance; and identifying a construction node position on the two-dimensional projection contour, matching the construction node position with the construction attribute, and generating a detailed drawing index mark at the construction node position according to a matching result. According to the space distribution and the sideline direction of the blank area, the size marking lines are automatically generated, the reasonable spacing is kept, and the problems of shielding and overlapping are avoided.

An Artificial Intelligence Gaze Interface Sensor (AI GIS) system and method for precisely tracking human gaze on a digital or projected view plane without head-mounted hardware. A context camera detects a user's face and matches it against stored biometric profiles using AI facial recognition, automatically loading personalized eye geometry, inter-pupillary distance and gaze calibration data. A conditional activation controller initializes a second Peregrine Foveal Camera (PFC), modeled after the dual-fovea retina of the Falco peregrinus, only upon confirmed eye detection. An AI algorithm triggers visual effects, hidden content, animations, gaze-based engagements and semantically mapped audio on the digital or projected plane based on gaze coordinate outputs, from a dedicated computational pipeline, correlated to the digital or projected plane. This invention serves as the sensing layer for U.S. application Ser. No. 18 / 395,290.

The invention provides a troposphere minimum scattering angle calculation method based on an earth reference ellipsoid, and the method comprises the steps: constructing a troposphere scattering path geometrical relationship, and calculating a spherical distance between a transmitting antenna and a receiving antenna; calculating the prime circle curvature radius of the earth reference ellipsoid at the latitude of the transmitting antenna and the latitude of the receiving antenna, the spherical distance between the highest obstacle of the transmitting antenna and the transmitting antenna, the included angle between the connecting line of the transmitting antenna and the receiving antenna and the transmitting view plane, and the spherical distance between the highest obstacle of the receiving antenna and the receiving antenna; and the included angle between the antenna connecting line and the receiving view plane line is received, so that the minimum troposphere scattering angle is calculated. According to the method, the geometrical relationship of the troposphere scattering path is constructed through the geodetic coordinates of the transmitting antenna, the receiving antenna and the electric wave ray encountering the highest obstacle, the mathematical model for calculating the minimum scattering angle of the troposphere is provided, the troposphere scattering signal transmission requirement is better met, and accurate calculation of the minimum scattering angle of the troposphere can be achieved.

Systems and methods for automatic correction of view foreshortening in cardiac echo using view synthesis. A trained image segmentation method is used to infer a 2D pose from the input image. The pose is transferred to a new view representing a non-foreshortened view plane. A non-foreshortened image is then rendered.

The invention relates to a product size detection method and system, terminal equipment and a storage medium, and the product size detection method is realized through a detection table, a main camera and an auxiliary camera. The detection table is used for placing a detected product; the light of the main camera is perpendicular to the detection table top and is used for collecting a top view image of a detected product; the light of the auxiliary camera forms a certain inclination angle with the detection table surface, and is used for collecting the shadow and three-dimensional contour image of the detected product; the product size detection method comprises the following steps: obtaining a top view image collected by a main camera and a shadow and three-dimensional contour image collected by an auxiliary camera; determining the plane size of a top view angle of the detected product based on the top view image, and determining the height and the plane size of an inclined view angle of the detected product based on the shadow and the three-dimensional contour image; and performing weighted fusion on the plan view plane size and the inclined view plane size by using a predetermined scale factor to obtain the final plane size of the detected product.

A method of remote image data set presentation includes dividing an image viewing space of a client device into a set of three-dimensional (3D) volumes, wherein the image viewing space presents a two-dimensional (2D) viewing plane of data and determining a set of 3D volumes of the image viewing space intersecting the 2D viewing plane. The method further includes retrieving, from a data server, image data associated with each of the 3D volumes intersecting the 2D viewing plane and presenting a portion of the retrieved image data corresponding to a position of the 2D viewing plane within the image viewing space.

A viewfinder for viewing an object under magnification, the viewfinder comprising: an objective lens for producing an image of an object located in an object plane and having a radial extent defined by an aperture stop; a partially transmissive reflector for allowing light to be transmitted therethrough from the objective lens to a mirror assembly and for providing for the reflection of light returned thereto from the mirror assembly; a mirror assembly receiving a light component from the partially transmissive reflector and arranged to produce a focused image of the object in the mirror assembly and the light received from the mirror assembly being reflected back to the partially transmissive reflector and retransmitted to produce an image of the object; a viewing lens arrangement for producing an optical image of the object which isvisible to an observer at an exit pupil in a viewing plane; wherein the objective lens has a beam path angle (α) as defined by a distance from the object plane to the objective lens and a radial extent of the aperture stop, wherein the viewfinder has a viewing angle (β) as defined by a distance along an optical axis from the mirror arrangement in the viewing plane and a radial extent of the exit pupil in the viewing plane; wherein the viewfinder is configured such that a ratio of displacement of the beam path angle (α) to the viewing angle (β) is at least 3:1, thereby providing the observer with a greater change in the perception of the depth of the object being viewed relative to an extent of displacement of the observer's head.

The application discloses a standing long jump measuring method and device, and the method comprises the following steps: calibrating a received standing long jump mat image and obtaining a perspective conversion matrix of an original plane coordinate system and a new view plane coordinate system; sending a take-off preparation prompt information after identifying that the face information of a tester is consistent with the pre-stored information; converting the original plane image including the real-time acquired tester bone data into a new view plane image through the perspective conversion matrix; receiving a first image of the tester in a standing long jump mat take-off area in the new view plane image, determining that the tester does not step / cross the take-off line, detecting the take-off condition of the tester, and detecting the landing image of the tester after determining that the tester takes off; determining whether the tester commits a rule violation through the landing image; if no rule violation is committed, obtaining the long jump distance of the tester based on the landing image; and if a rule violation is committed, sending a rule violation prompt signal. The application can accurately identify the rule violation actions of the tester in the standing long jump process, and improve the measuring accuracy.

A method of remote image data set presentation includes dividing an image viewing space of a client device into a set of three-dimensional (3D) volumes, wherein the image viewing space presents a two-dimensional (2D) viewing plane of data and determining a set of 3D volumes of the image viewing space intersecting the 2D viewing plane. The method further includes retrieving, from a data server, image data associated with each of the 3D volumes intersecting the 2D viewing plane and presenting a portion of the retrieved image data corresponding to a position of the 2D viewing plane within the image viewing space.

This disclosure provides an airborne display system for a vehicle and a corresponding method. The system (100) comprises a display device (102) equipped with a height adjustment mechanism. The display device (102) includes at least one screen coated with a micro-flap layer (202), an optical arrangement (104) configured to move between a retracted position and an extended position, and a sensor arrangement (106) for detecting the movement of a vehicle occupant's fingers. When the optical arrangement (104) is in the extended position, it is designed to focus the light emitted by the micro-flap layer (202) onto a floating viewing plane (204) located between the optical arrangement (104) and the occupant.The sensor arrangement (106) enables the occupant to interact with the display device (102) by detecting the movement of the occupant's fingers as they cross the floating viewing plane (204).

The application discloses a kind of based on Gaussian splashing air-ground scene three-dimensional reconstruction method, comprising: S1, obtains the multi-source air-ground scene image containing large pitch angle aerial photograph and ground view and camera pose;S2, based on the scene sparse point cloud generated by motion recovery structure, instantiation generates initial three-dimensional space element set;S3, introduce multi-measure adaptive density control mechanism to realize the adaptive matching of Gaussian density in different scale space;S4, by camera Jacobian matrix, three-dimensional space element is projected to two-dimensional view plane, after depth ordering, along the direction of ray extraction color and opacity, by Alpha mixing formula integration output predicted two-dimensional rendering image;S5, introduce multi-dimensional geometric constraint to guarantee the consistency of physical surface;S6, after S2-S5 closed loop circulation to convergence, extract and output high-precision three-dimensional reconstruction point cloud and new view two-dimensional rendering image.

This application relates to a semantic grid map construction method, system, and medium with automatic obstacle removal function, belonging to the field of lawnmower technology. The semantic grid map construction method includes: acquiring original point cloud data and original image data of the lawn according to the target projection resolution, and converting them to the same view plane; temporally synchronizing the original point cloud data and original image data to obtain a three-dimensional point cloud and a two-dimensional image, and transmitting them to a data queue; spatially transforming the three-dimensional point cloud, and constructing fused image data by combining object semantic labels with the two-dimensional image in a layered fusion mechanism; generating a lawn grid based on the fused image data, and marking obstacle nodes using a ray tracing mechanism to obtain a lawn grid map; accurately distinguishing obstacle types (static / dynamic, passable / impassable) through category information; and dynamically clearing grids in obstacle-free areas through a ray tracing mechanism, reducing false positive obstacles caused by limited detection field of view.

The invention provides a reliability evaluation method and system for an unmanned driving perception system, and relates to the technical field of data processing, and the method comprises the steps: synchronously obtaining mine card detection data based on a multi-source sensor; after image data and point cloud data in the mine card detection data are preprocessed, target detection is carried out on the image data and the point cloud data, and candidate area images and candidate area point clouds containing a target mine card are obtained; extracting multi-scale visual features from the candidate region images through a convolutional neural network, extracting spatial features from the candidate region point clouds, and performing inter-frame association and fusion on the multi-scale visual features and the spatial features based on coordinate transformation to generate a frame-level reference true value; the multi-scale visual features comprise bounding boxes, confidence coefficients and category labels; the spatial features comprise a central point thermodynamic diagram of a target mine card on an overlook plane, the size of the mine card and orientation regression; and checking the mine card detection data of each sensor and the frame-level reference truth value, and determining the detection confidence coefficient.

The present disclosure provides concepts for curating medical images of an anatomical structure acquired during an examination. Curating the medical images includes obtaining a user preference indicating desired anatomical sub-structures in a target view plane, generating a segmentation score for each medical image depicting the target view plane, and identifying primary medical images based on the segmentation scores. The segmentation score indicates a confidence of the desired anatomical sub-structures being present in the respective medical image. The invention enables customized curation of medical images based on user-specified anatomical features, improving efficiency and relevance of selected medical images from an examination, such that subsequent medical image analysis may be more effective and efficient.

Systems and techniques are described for redirecting light. For example, an apparatus for redirecting light can include a display positioned along a first optical axis. The first optical axis passes through a viewing plane of the display and intersects with a viewing position. The apparatus can include a first light redirecting element positioned along the first optical axis. The first light redirecting element is configured to redirect light from a scene toward a second optical axis. The apparatus includes an image sensor. The image sensor is configured to receive the light from the scene. The first light redirecting element is included along an optical path between the scene and the image sensor.

The application discloses a multi-view target detection and tracking method based on view-angle-space joint attention, and the steps are as follows: acquiring multi-view images aligned at the same time, extracting two-dimensional features and projecting the two-dimensional features to a bird's eye view plane; calculating effective observation regions of each view angle; weighting and fusing multi-view features by using a view-angle-space joint attention module and an uncertainty perception fusion module, and gain modulating in combination with a region prior heat map; inputting the fused features into a decoding network to obtain detection results, and then inputting the detection results into a tracker to complete target association and track state updating; in a training stage, a teacher network and a student network are adopted, and the network is trained by using a supervision constraint and a cross-view self-supervised consistency constraint. The application can improve the stability of multi-view target tracking in a complex scene, and has the advantages of strong robustness and convenience for engineering deployment.

The invention provides a bird's-eye view construction method and device, a vehicle and a computer readable storage medium, and relates to the technical field of image processing, and the method comprises the steps: obtaining first bird's-eye views of a plurality of cameras in the vehicle, obtaining a plurality of first bird's-eye views, determining the corresponding index coordinates of the pixel point coordinates of each first bird's-eye view in an image index table, and obtaining a plurality of first bird's-eye views; on the basis of the camera identifier carried by the index coordinate, retaining one overlapping region included in each group of overlapping regions in the plurality of first aerial views, and filtering other overlapping regions to obtain a plurality of second aerial views, the camera identifier represents whether a camera corresponding to the camera identifier has an effective field angle in a plane area of an aerial view plane of the vehicle; and fusing the plurality of second aerial views to obtain a panoramic aerial view. According to the invention, repeated data carrying and storage operations during the construction of the vehicle panoramic aerial view can be avoided, the memory bandwidth occupation can be greatly reduced, and the construction efficiency of the vehicle panoramic aerial view can be improved.

The application discloses a virtual Grid arrangement method and system based on proton radiotherapy beam direction optimization. The method comprises the following steps: acquiring tumor target region contour data of a patient, and organ at risk and / or planning risk volume contour data, constructing structure binary images, internal shrinkage three-dimensional images and external expansion three-dimensional images; in a plurality of candidate beam directions, a beam direction robustness score is constructed according to water equivalent path length penetrating the target region, and a target beam direction is determined; the target beam direction is taken as the central axis direction of a virtual Grid cylinder, an initial Grid center coordinate set is generated in a two-dimensional beam eye view plane perpendicular to the target beam direction; a continuous available interval is searched along the target beam direction for the optimized center coordinate, and a finite length virtual Grid cylinder is generated; and a virtual Grid arrangement result is determined and output according to a Grid volume proportion, a minimum safety distance, a cylinder separation degree, a weighted normal tissue overlapping volume and a robust risk penalty term.

Systems and techniques are described for redirecting light. For example, an apparatus for redirecting light can include a display positioned along a first optical axis. The first optical axis passes through a viewing plane of the display and intersects with a viewing position. The apparatus can include a first light redirecting element positioned along the first optical axis. The first light redirecting element is configured to redirect light from a scene toward a second optical axis. The apparatus includes an image sensor. The image sensor is configured to receive the light from the scene. The first light redirecting element is included along an optical path between the scene and the image sensor.

This invention proposes an object positioning system and method. The object positioning system includes a processing device, a sensing camera, and a memory. The sensing camera is coupled to the processing device and mounted on a self-propelled device to generate image frames. The processing device executes computer-readable program code stored in the memory to: generate a mask of an entity in the image frame and determine its category using an instance segmentation model; project the mask onto a top view plane of a global coordinate system to generate a projection mask; identify the front edge of the projection mask relative to the sensing camera; determine a reference position corresponding to the front edge; and generate a measured position of the entity on the top view plane based on the reference position and the entity category.

The application relates to a camera measurement field establishment method and a target, which comprises the following steps: determining the theoretical coordinates of a measured point according to the digital model of a measured object; adjusting the size of a target mark according to an actual working distance; installing a target base on the measured point of the measured object; manually adjusting the angle of the target mark according to the image quality of the mark in the camera field of view, so that the target mark plane is parallel to the camera view plane; taking a photo to calculate the three-dimensional coordinates of the mark center point in the camera coordinate system; calculating the three-dimensional coordinates of the measured point in the camera coordinate system according to the size information of the base and the target and the conversion relationship formula of the target mark center and the measured point; repeating the above steps to record the three-dimensional coordinates of the measured point in the camera coordinate system and the known coordinates of the measured point; calculating the relationship between the measured point coordinates in the camera coordinate system and the known coordinates of the measured point by adopting an optimal fitting algorithm according to the recorded measured point coordinates, and then establishing the measurement field of the camera; and finally realizing the rapid establishment of the measurement field of the camera.

The invention provides an interaction system and method combining a locust approaching perception algorithm and an aerial view. The method comprises the following steps: acquiring a fisheye image; performing joint calibration on the four fisheye cameras by using a composite checkerboard calibration plate, performing distortion removal processing on fisheye images, and mapping the fisheye images to a unified aerial view plane through a perspective transformation matrix to obtain an aerial view; segmenting the aerial view according to a central area and an overlapping area, calculating an adaptive weight matrix in the overlapping area based on a distance between a pixel and an overlapping boundary, carrying out weighted stacking fusion to obtain a panoramic aerial view, and carrying out gray histogram equalization and linear adjustment on the panoramic aerial view; processing the panoramic aerial view by using a semi-global matching algorithm to generate a multi-view depth map, dividing the depth map into eight areas, and judging the static obstacle condition of each area in combination with continuous multi-frame data; and meanwhile, inputting the aerial view into the D-LGMD model, and judging the dynamic obstacle by monitoring the membrane potential. According to the invention, the comprehensive perception capability and reliability in a complex and changeable environment are obviously improved.

The application discloses a kind of multi-view tomato seedling branch angle analysis method and system, method includes: collecting tomato seedling orthographic projection image, through the same seedling multiple view image of unique identifier association, provide complete associated basic data.Based on the branch contour of annotation, train YOLO v8-seg instance segmentation model, separate branch contour, exclude background, leaf and other interference.Skeleton processing is carried out to the segmentation result, identifies branch branch point and two side endpoints, calculates the angle of double-view plane projection, provides core input for three-dimensional calculation.Fusion double-view data constructs three-dimensional direction vector, obtains actual angle by vector dot product, combined with growth environment parameters, builds growth evaluation model by machine learning, realizes seedling growth state quantitative evaluation and trend prediction.Provides reliable technical support for quantifying tomato seedling branch angle and predicting growth.

The invention discloses a full-sparse 3D target detection method based on cross-voxel displacement and explicit supervision, and belongs to the technical field of automatic driving environment perception. The method comprises the following steps: acquiring a laser radar point cloud and performing voxelization; sparse voxels are input into a backbone network, a cross-voxel displacement module is introduced in a feature extraction process, a feature interaction channel between long-distance voxels is constructed with linear complexity by using a reverse sequence pairing strategy based on a Manhattan distance, and the problem that a sparse convolution receptive field is limited is overcome; a voxel-guided virtual voxel generator is used for directly regressing offset to generate virtual voxels, a back-and-forth projection process is replaced, and positive samples are screened in combination with a label distribution strategy based on the physical centroid; an explicit center perception supervision module is introduced, virtual voxel features are compressed to an aerial view plane, and the virtual voxel features are directly constrained by using a Gaussian heat map of the center of a target object; and finally, outputting a 3D target detection result through a sparse detection head.

The application provides a method and device for registration and fusion of visible light images and infrared images, which are applied to a device provided with a visible light camera and an infrared camera, wherein the optical axis of the visible light camera is perpendicular to the front view plane of the device, the method comprising: establishing a registration model of coordinate positions of a visible light image of a target object collected by the visible light camera and an infrared image of the target object collected by the infrared camera according to the spatial relative positions of the visible light camera and the infrared camera, conversion parameters and a horizontal distance of the target object from the visible light camera; and registration and fusion of the visible light image and the infrared image of the target object according to the registration model. In the application, the data of the visible light image and the data of the infrared image in the device provided with the double cameras can be fused into one image through the established registration model.

The invention relates to a multi-modal information fusion-based low-altitude 3D target detection method. The method comprises the following steps of: aligning an enhanced point cloud and an RGB image and then respectively carrying out feature extraction; on the basis of the virtual points, the extracted image features and camera calibration parameters, point cloud features are complemented, and agent features are sampled; mapping the complemented point cloud features and the image features to an aerial view in a unified manner to obtain point cloud BEV features and image BEV features respectively; screening potential target areas in the aerial view plane based on a preset 2D detection clue, and respectively extracting basic features from each potential target area to construct a feature set, and performing three-stage cascade optimization on the feature set based on the fused BEV feature fusing the point cloud BEV feature, the image BEV feature and the 2D detection clue, the proxy feature and the complemented point cloud feature, outputting the region-of-interest feature, classifying the region-of-interest feature, and outputting an accurate 3D detection result.

A display system that comprises a hologram replicator 1210 arranged to output an array of replicas of a holographic wavefront, wherein angles of the holographic wavefront correspond to spatial coordinates of an image corresponding to the holographic wavefront; and an optical component 1200 arranged to receive the array of replicas, wherein the optical component has a primary surface 1202 arranged to direct a primary reflection of the array of replicas towards the viewing plane to form a primary array of composite holographic wavefronts at a viewing plane and a secondary surface 1204 arranged to direct a secondary reflection of the array of replicas towards the viewing plane to form a secondary array of ghost composite holographic wavefronts at the viewing plane. The hologram replicator and optical combiner are arranged such that primary and secondary arrays form a combined array in which the composite holographic wavefronts (1650, 1651, fig.16A) and the ghost holographic wavefronts (1650’, 1651’, fig.16A) are substantially interleaved. A corresponding method of holographic projection is also disclosed.