121 results about "Viewing cone" patented technology

An autostereoscopic display device has both a barrier arrangement and a lens arrangement. A plurality of views are provided to different lateral viewing directions. At least a portion of the field of view has autostereoscopic output, and the portion having autostereoscopic output has no repetition of individual 2D views and comprises at least three individual 2D views. This means there is no reversal of the stereo views (“pseudo stereo views”) at viewing cone boundaries as there are no viewing cone boundaries.

The invention discloses a three-dimensional object detection method based on a view cone point cloud. The RGB image and the point cloud image are collected and annotated, the annotated RGB image dataset and the point cloud image data set are obtained, and the data set is used as a training sample and a test sample. It mainly consists of three networks: cone point cloud extraction network based ontwo-dimensional object detection, three-dimensional object segmentation network based on Point net and three-dimensional bounding box regression network based on offset residuals. The three networksare trained in turn, and the training output of the previous network is used as the input of the next network. The super parameters of each neural network model are set, and the three-dimensional object detection network model is trained by Tensorflow. When the cost loss function is reduced to the ideal degree and the training reaches the required maximum iteration number, the training model is generated. The point cloud image data set is inputted into the trained model, and the 3D object detection result is output. The method of the invention is more perfect and more difficult than the existing three-dimensional object detection method.

This invention relates to reflective cholesteric displays without using Bragg reflection. The helical pitch of the cholesteric liquid crystals may or may not be tuned in a visible wavelength, but its Bragg reflection will never be displayed. An elliptical polarizer with properly designed retardation and lamination positioned at the front of the display will be able to eliminate the Bragg reflection over a large viewing cone. The function of the display cell structure is merely a light shutter to switch the incident light ON and OFF. In the black-and-white display mode, the white state is achieved from the metal reflection in the cholesteric planar texture area; and the black state is obtained by cholesteric's depolarization effect and polarizer's filtration effect in the cholesteric focal conic texture area. In the full color mode, the full color state is created by the metal reflector and the micro-color filter in the cholesteric planar texture area; and the black state is realized in the cholesteric focal conic texture area. The reflective cholesteric display, remaining its long-term bistability but not applying the narrow-band Bragg reflection, initiates novel displays with a neutral wide-band reflection.

The invention discloses a method for mass model data dynamic scheduling and real-time asynchronous loading under virtual reality. The method includes the following steps of firstly, preprocessing 3D model scene data, secondly, cutting and partitioning a whole model scene, thirdly, conducting multithreading parallel distribution loading and fourthly, clipping a view cone. Based on a view cone clipping algorithm for acquiring the point of intersection of the view cone and a topographic region, real-time clipping based on the topographic region in dynamic scheduling is achieved, by the adoption of a multithreading processing mechanism of a model data partitioning scheduling and rending pipeline, a dynamic scheduling and drawing pipeline and a data processing scheduling pipeline between different mediums can be asynchronously loaded, and therefore dynamic balance of loading and performance efficiency of the mass scene model data is achieved in the hardware environment with a limited memory, a limited processor and the like.

A film construction (330) includes a broad band reflective polarizing film (312) that may be immersed in an ultra low refractive index medium (332, 334). The reflecting polarizing film is characterized by a pass axis and a block axis, and its reflectivity for white light of the pass state polarization increases with increasing incidence angle to provide a compressed or narrowed viewing cone selectively in one plane of incidence. In some embodiments, the plane of incidence associated with the compressed viewing cone is aligned with the pass axis. In other embodiments it is aligned with the block axis.

The invention provides an autostereoscopic display device in which a light blocking arrangement is provided for selectively blocking light which has or would pass between lenses. Elements are provided between adjacent lens locations, and the display can be configured so that light reaching these elements is either allowed to reach the viewer or is blocked from reaching the viewer. This means that a public (multiple cone) viewing mode can be chosen or a private (single narrow viewing cone) viewing mode.

The invention relates to a three-dimensional target detection method and system based on point cloud weighted channel characteristics, and the method comprises the steps: carrying out the extraction of a target in a two-dimensional image through a pre-trained deep convolutional neural network, and obtaining a plurality of target objects; based on each target object, determining a point cloud viewcone in the corresponding three-dimensional point cloud space; segmenting the point cloud in the view cone based on a segmentation network of the point cloud to obtain an interested point cloud; And based on the network with the weighted channel characteristics, performing 3D Box parameter estimation on the point cloud of interest to obtain 3D Box parameters, and performing three-dimensional target detection. According to the method, the characteristics of the image can be learned more accurately through the deep convolutional neural network. Based on a network with weighted channel characteristics, 3D Box parameter estimation is carried out on the point cloud of interest, the weight of characteristic drop of unimportant points can be reduced, the weight of key points can be increased, interference points can be restrained, the key points can be enhanced, and therefore the precision of 3D Box parameters can be improved.

The invention provides a 3D target detection method for point cloud screening based on image semantic features. The method comprises the following steps: segmenting image data by a 2D semantic segmentation method to obtain semantic prediction; projecting the generated semantic prediction to an LIDAR point cloud space through a known projection matrix, thereby enabling each point in the point cloudto obtain the semantic category attribute of the corresponding image position; extracting related points of vehicles, pedestrians and riders from the original point cloud and forming a view cone; taking the view cone as the input of a depth 3D target detector, and designing a loss function conforming to the characteristics of the view cone to perform network training. According to the invention,a 3D target detection algorithm for point cloud screening based on image semantic features is designed, so that the 3D detection time is greatly shortened and the calculation requirements are greatlyreduced. Finally, the performance of the method on a reference data set KITTI of 3D target detection shows that the method has good real-time target detection performance.

A liquid crystal display (LCD) includes an LCD panel, an orienting base, an imaging unit, an image processing unit, and a control unit. The LCD panel is configured for displaying information readable from any vantage point inside the viewing cone thereof. The orienting base is configured for supporting and orienting the LCD panel. The imaging unit is configured for capturing an image of an immediate area in front of the LCD panel. The image processing unit is configured for processing the captured images to detect whether the eyes of the viewer are within the immediate area but are outside the viewing cone of the LCD panel, and, if yes, to locate the eyes of the viewer. The control unit is configured for controlling the orienting base to orient the LCD panel to aim the viewing cone of the LCD panel to include the located eyes of the viewer.

A film construction (320) includes a broad band reflective polarizing film (312) that may be immersed in an ultra low refractive index medium (322, 324). The reflecting polarizing film is characterized by a pass axis and a block axis, and its reflectivity for white light of the pass state polarization increases with increasing incidence angle to provide a compressed or narrowed viewing cone in one or two planes.

Systems and methods for electronically controlling the viewing angle of a display using liquid crystal optical elements are provided. Each liquid crystal optical element may be associated with a respective scattering module and may selectively steer a device generated light beam to one of two or more scattering regions of its associated scattering module. When a scattering region receives a steered light beam, the steered light beam may be scattered into a viewing cone having at least one viewing angle defined by a characteristic of that scatter region. Each liquid crystal optical element may be made from one or more suitable liquid crystal materials that can be controlled electronically to vary the effective index of refraction of one or more different regions of the liquid crystal optical element, thereby steering incoming light towards a particular one of two or more scattering regions of an associated scattering module.

A projection screen includes a base sheet, a surface roughness structure, a reflective layer, and a light absorption layer. The base sheet has a first side and a second side opposite to the first side. A plurality of first surfaces and second surfaces are formed on the first side, each first surface faces an optical projection system, and each second surface is disposed between two adjacent first surfaces and forms an angle with respect to a neighboring first surface. The surface roughness structure is formed on at least the first surfaces and capable of diffusing a light beam to a limited extent. The reflective layer is formed on the surface roughness structure and capable of reflecting most of the light beam diffused by the surface roughness structure to a limited extent towards a limited viewing cone, and the light absorption layer is formed on the second surfaces.

A display processing system for providing video signals for displaying an image is provided. The display processing system comprises input channels for receiving a plurality of component images, each component image being a portion of the complete image for display, the image data of each component image being defined by a view frustum having a clipping plane, and a combiner for combining the image data of the component images. According to the present invention, at least two clipping planes of two of the view frustums are non-coplanar. Such display processing system corrects for the defect of driving angular resolution to the edges of the field of view of a display system, in particular for large field-of-view display systems, and balances the overall system angular resolution by allowing each image generator channel to render to an optimal view frustum for its portion of the complete image.

An optical system that reduces reflections from an electronic display is provided herein. The optical system may include an electronic display; and a light directing medium configured to steer a plurality of viewing cones into an eye motion box facing said display so that said viewing cones overlap at least partially at the eye motion box plane, wherein a viewing cone is a geometric location only within which a viewer can see light coming from said electronic display.

The invention discloses an organization and scheduling method based on a large-scale oblique photography model, and the method comprises the steps: carrying out the model data hierarchical division ofoblique photography model data, determining the simplified model data and model material files of all hierarchies through combining with the simplification degrees of all hierarchies, and generatingmodel files of different hierarchies; partitioning the model file of each hierarchy to generate partitioned model data of different hierarchies, and recording tiles by adopting OCT coding to generatea Json index file; generating a quadrangular view cone region according to the selected view angle and viewpoint, and calling block model data in the view cone region in combination with the index file; and matching the current view cone area according to the viewpoint and view angle changes, and updating the model data of the corresponding view cone area. According to the method, the data index and the tree-shaped data structure are established after the data is layered and blocked, so that quick positioning and effective management of the data are facilitated; through scheduling strategies of pre-access, three-level caching and the like, the model data loading and rendering efficiency is improved, and the data loading and visualization capacity of the three-dimensional GIS is expanded.

A liquid crystal display is provided which comprises a front panel comprising a front alignment layer having an alignment direction, a rear panel comprising a rear alignment layer having an alignment direction, and a liquid crystal layer between the front and rear alignment layers. The liquid crystal layer has a rotational twist angle of about 90° and a pre-tilt angle of not more then 2°. The alignment, material and thickness of the liquid crystal layer are such that at the mid-point of the rotational twist, the direction of liquid crystal directors coincide with an off-normal viewing direction of the liquid crystal display.

An alignment system for side view mirrors featuring a mirror with regions of diverse curvature, with an upright alignment line in a region of greater curvature. The side view mirror is set up so that the side of a user's vehicle lies entirely on the side of the alignment line closest to the vehicle. This generates a cone of vision to the rear of the vehicle which overlaps with the cone of vision from a rear view mirror such that blind spots are greatly reduced or eliminated.

A multi-view display device comprises a pixelated display panel and a backlight comprising an arrangement of light sources (30), wherein each light source, when turned on, illuminates an associated region of pixels of the display panel. A display controller is adapted to control the pixelated display panel and the arrangement of light sources such that a partial display output is provided comprising simultaneously a set of at least three 2D views with no repetition of individual 2D views.This arrangement provides an output with controlled illumination direction of the pixels so that view repetitions are avoided. The output can be a single cone of views, and the location from which the cone of views can be viewed depends on the relationship between the light sources of the backlight which are activated and the display panel.

The embodiment of the invention provides a bimodal bionic vision sensor with a view cone and a view rod. The sensor comprises a first preset number of voltage mode active pixel sensor circuits and a second preset number of current mode active pixel sensor circuits, wherein each voltage mode active pixel sensor circuit comprises a first type photosensitive device, and each current mode active pixelsensor circuit comprises a second type photosensitive device. Through the voltage mode active pixel sensor, the target voltage signal representing the light intensity information in the target lightsignal can be output, the precision of representing the light intensity information by the obtained target voltage signal is higher, and an image with higher quality can be obtained, that is, the image has a higher image signal-to-noise ratio. Through the current mode active pixel sensor, a specified digital signal representing light intensity gradient information in a target light signal can be output, the performance indexes such as an image dynamic range and a shooting speed of the bionic vision sensor are ensured, and the stability and robustness of the bionic vision sensor are stronger.

A projection screen includes a base sheet, a surface roughness structure, a reflective layer, and a light absorption layer. The base sheet has a first side and a second side opposite to the first side. A plurality of first surfaces and second surfaces are formed on the first side, each first surface faces an optical projection system, and each second surface is disposed between two adjacent first surfaces and forms an angle with respect to a neighboring first surface. The surface roughness structure is formed on at least the first surfaces and capable of diffusing a light beam to a limited extent. The reflective layer is formed on the surface roughness structure and capable of reflecting most of the light beam diffused by the surface roughness structure to a limited extent towards a limited viewing cone, and the light absorption layer is formed on the second surfaces.

The invention relates to a multichannel LED dome screen geometric correction method. The method comprises the following steps: step 1, obtaining a frame buffer texture image A of an RTT camera S1; step 2, acquiring a frame buffer texture image B of the camera in the step 2; step 3, defining a front-view matrix camera as S3; defining S3 projection matrix as a front view matrix, the window size being full screen size, defining a S3 view matrix as a unit matrix, adding a quadrilateral plane model to a front view matrix camera node, pasting the texture cache image B on the quadrilateral plane model, performing point extraction operation on the image B by adopting shader language programming, and finally outputting a display image C. According to the invention, a traditional projector projection screen is replaced by the multi-channel LED dome screen, so that the brightness and resolution uniformity of the whole picture are consistent; geometric correction is carried out through shader language programming to realize spatial transformation from a view cone plane point to a dome screen point, spatial restoration correction of the image is rapidly realized, the delay characteristic can bebasically ignored, and the stability is high.

A method and a system for calibrating a camera in a surveillance system. The method and system use a mathematical rotation between a first coordinate system and a second coordinate system in order to calibrate a camera with a map of an area. In some embodiments, the calibration can be used to control the camera and / or to display a view cone on the map.

An apparatus for recognizing a pupillary distance for three-dimensional (3D) display includes a display configured to output a 3D image corresponding to a reference pupillary distance, a controller configured to control a viewing cone included in the 3D image, and a user inputter configured to receive a user feedback indicating whether an artifact is viewed in the 3D image in response to the controlling of the viewing cone. The controller may move the viewing cone within a margin corresponding to the reference pupillary distance, and change the reference pupillary distance or determine the reference pupillary distance to be a desired pupillary distance of the user based on the user feedback.

The invention relates to an automatic driving multi-sensor fusion sensing method and system under an extreme working condition. The method comprises the following steps: obtaining a smoke fog area in image data, obtaining corresponding laser point cloud data in a three-dimensional view cone range of the image data, projecting the laser point cloud data to an image domain corresponding to the image data, and obtaining smoke fog point cloud data of the image domain according to an intersection of the two; meanwhile, obtaining radar domain obstacle point cloud data based on the laser point cloud data. And according to the intersection of the smoke mist point cloud data of the image domain and the obstacle point cloud data of the radar domain, false obstacle fusion sensing data are obtained, and then the false obstacle fusion sensing data is combined with real obstacle data obtained by the millimeter wave radar to remove false obstacles. According to the method, a visual angle where smoke fog possibly exists is provided for the laser point cloud based on the image data, the laser point cloud data are mutually verified between the image domain and the radar domain, the laser point cloud data are combined with the millimeter wave radar, and finally high-precision automatic driving scene data are obtained.

The invention provides a light field display device, relates to the technical field of display, and aims to solve the problem of visual fatigue caused by inconsistent monocular focusing depth and binocular convergence depth and prevent a user from feeling dizzy. The light field display device comprises a display screen and a lens unit arranged on the light emitting side of the display screen, wherein the lens unit comprises a plurality of lenses arranged in an array; the display screen is arranged on the focal planes of the plurality of lenses; the display screen comprises a first substrate and a second substrate which are oppositely combined, wherein the first substrate is a light emitting side substrate, the second substrate comprises a plurality of pixel islands arranged in an array, and each pixel island comprises at least one sub-pixel; the plurality of pixel islands correspond to the plurality of lenses; light rays emitted by sub-pixels of the pixel island are transmitted to a view area formed by human eyes through the corresponding lenses to be smaller than or equal to a half pupil area, and light rays, emitted by different sub-pixels in the pixel island, of different viewpoints are emitted into different view cone cells through the lenses. The invention is suitable for manufacturing the light field display device.

The invention relates to a normal-based large-scale point cloud visualization method, and the method comprises the steps: S1, constructing a node point cloud balanced octree structure according to a space structure of point cloud data; S2, constructing an octree structure with normal information according to the balanced octree structure and the normal information of the point cloud, and constructing normal LOD visual nodes in a down-sampling mode; and S3, finally, determining a node scheduling strategy according to the normal relationship between the view point and the view cone and the rendering node, and respectively calling a reading thread and a rendering thread for simultaneous reading and rendering according to the current scheduling strategy. According to the method, the rendering nodes are constrained by adding the normal information, the problem of invalid rendering of a large number of shielded point clouds, scene back point clouds and the like in a large-scale point cloud rendering process is solved, the number of points of actual rendering point clouds is reduced, and the authenticity of point cloud display is ensured.

The invention discloses a method for fusing multiple paths of videos and a three-dimensional GIS scene, which comprises the following steps: defining a data structure in a video object, and endowing each parameter of each video object with an initial value; judging the spatial position information and the attitude information of the video object in the scene and the observable area information ofthe camera, and abstracting the video object into a view cone geometric object according to the information; according to the attribute information of the camera, classifying all the view frustum geometric objects in the scene to form a plurality of video layers; establishing r-tree index information of all video objects under each video layer in the scene; entering the visible range of the three-dimensional scene, storing rendering objects in the visible range in real time, and generating a view frustum of the three-dimensional scene in the visible range in real time. The method has the advantages that the topological information of the video and the three-dimensional scene is established, so that the quick positioning and fusion efficiency of the fusion object can be effectively improved, and the method is suitable for three-dimensional scene virtual fusion of multiple (more than 4-5) videos.

A film construction (320) includes a broad band reflective polarizing film (312) that may be immersed in an ultra low refractive index medium (322, 324). The reflecting polarizing film is characterized by a pass axis and a block axis, and its reflectivity for white light of the pass state polarization increases with increasing incidence angle to provide a compressed or narrowed viewing cone in one or two planes.

The invention provides a three-dimensional video fusion method and system based on WebGL. According to the invention, there is no need to process a video source, and the method comprises the steps: accessing an HTTP video stream, updating a video texture object based on WebGL, updating and setting a near cutting surface, a far cutting surface and a camera position and orientation of a view cone ofan observation camera, then updating the scene depth of an observation view angle, projecting and restoring to an observer camera coordinate system, fusing with the live-action model, performing distortion correction on the camera, and finally realizing a video area cutting effect by adopting masking. The problems in the prior art are solved, three-dimensional video fusion is achieved on the basis of WebGL, the projection area is cut, adjacent videos can be prevented from being displayed in an overlapped mode, distortion correction is conducted on the cameras, and therefore the good display effect can be achieved for the cameras with large distortion and the situation that the installation positions are low.

The invention relates to a visualization method and device for a photogrammetry model in an unreal engine, and computer equipment. The method comprises the following steps: acquiring OSGB data of a photogrammetry model, converting the OSGB data into Uasset format data in an unreal engine, reconstructing a virtual quadtree, and then performing binary coding on nodes in the reconstructed quadtree; in a current frame, realizing tile hierarchical loading, cutting through a view cone, screening tile hierarchies for a tile tree in a view field range through a threshold value, judging a set membership between a current node and a node requested to be loaded in a previous frame through binary coding, and loading the current node; and asynchronously loading or unloading the request data nodes having the father-son-child-father relationship with the current node in the request data node set according to a preset rule. Visualization of the photogrammetry model is achieved in the unreal engine, the real-time visualization fluency of the oblique photography model can be improved, and memory use is reduced.