14 results about "Height field" patented technology

The invention discloses a three-dimensional indoor scene generation method and related equipment. The method comprises the steps that text description of an indoor scene is acquired; generating an indoor layout by the text description through a preset diffusion model; converting the indoor layout into a two-dimensional height field and a two-dimensional semantic graph; predicting a neural radiation field based on the two-dimensional height field and the two-dimensional semantic map; a plurality of viewpoints are sampled in the predicted neural radiation field, the plurality of viewpoints are rendered to obtain an RGB-D image corresponding to each viewpoint, RGB is color information of the two-dimensional image, and D is depth information of the two-dimensional image; and according to the RGB-D image corresponding to each viewpoint, constructing a truncated signed distance field body, and generating a mesh triangular patch based on the truncated signed distance field body to obtain an indoor three-dimensional scene. According to the method and the device, the high-quality three-dimensional indoor scene can be generated without real multi-view images, the efficiency of generating the three-dimensional scene is improved, and the method and the device can be widely applied to the technical field of computer vision.

The invention is suitable for the technical field of meteorological data processing, and provides a height field and wind field mutual smoothing method based on physical rule constraints, which comprises the following steps: preprocessing a two-dimensional height field lattice point matrix hgtgrid, a two-dimensional wind field horizontal component lattice point matrix ugrid and a two-dimensional wind field vertical component lattice point matrix vgrid to obtain a wind speed and a wind direction, and performing mutual smoothing processing based on a bidirectional coupling relationship between the height field and the wind field, and outputting the height field and the wind field which are subjected to mutual smoothing when the number of cycles reaches a preset maximum number of iterations maxter. Therefore, by establishing a bidirectional dynamic coupling mechanism of the wind field and the height field, visual drawing of the contour line of the weather map is supported, and the method has the advantages of being high in physical consistency and overall coordination.

The present disclosure relates to methods, devices, and systems for blending geographic data when combining geographic data sources. The methods, devices, and systems identify a blend region for transitioning between a first dataset and a second dataset. The methods, devices, and systems extrapolate geographic data from the second dataset to blend with the geographic data from the first dataset to create blended elevation data in the blend region. The methods, devices, and systems may generate an image for a geographic region with the first set of geographic data, the second set of geographic data, and the blended elevation data.

This invention discloses a method, system, and storage medium for LOD2 building reconstruction from single-view orthophotos. The method inputs a single-view orthophoto remote sensing image into a shared backbone network to generate a footprint mask probability map; constructs a first modulation factor, performs element-wise gating on the multi-scale feature map, and inputs it into a height decoder to output a height field; normalizes the height field to obtain a normalized height map and constructs a second modulation factor; fuses the second modulation factor with the first modulation factor to form a composite modulation factor, performs dual element-wise gating on the multi-scale feature map, and inputs it into a structure line decoder to output a distance field and orientation field; recovers the roof structure line based on the distance field and orientation field; and constructs the roof surface, walls, and ground by combining the roof structure line, height field, and footprint boundary to generate a LOD2 3D model of the building. This invention can automatically generate a topologically complete and geometrically accurate LOD2 3D model relying solely on a single orthophoto.

The invention discloses a grid reanalysis data-based whole-process objective identification method for a southwest low vortex, and a medium, and relates to the technical field of meteorological big data processing and numerical weather analysis. The method comprises the following steps: eliminating high-altitude grid points in a target identification area by using a terrain mask; searching a closed isoline according to an equivalence distance based on a preset air pressure horizon potential height field, extracting a closed low-pressure candidate body and determining a low-pressure center; calculating the attribution of a closed region, screening an effective candidate body, constructing a time sequence track by adopting adjacent time space adjacent matching, and introducing displacement constraint and duration criterion to confirm a low eddy process; and outputting starting and ending moments, duration time and a moving path, and supporting frequency, path and rainfall influence statistics. According to the method, automatic identification, stable tracking and consistency statistics of the southwest low vortex in a long-time sequence can be realized, subjective identification deviation and labor cost are reduced, and rechecking performance and service availability of an identification result under a complex terrain condition are improved.

This invention discloses a method, apparatus, device, and storage medium for determining wave parameters. The method includes: acquiring multiple historical water surface images from multiple water areas; determining wave surface height field labels corresponding to each historical water surface image based on a teacher network; inputting the historical water surface images and their corresponding wave surface height field labels into a student network; determining a multi-constraint joint loss function based on the wave surface height field labels corresponding to each historical water surface image, the historical wave surface height fields determined by the student network, and the confidence level; updating the student network based on backpropagation until the loss function converges, resulting in a trained student network; acquiring a sequence of water surface images from the current water area; inputting the left and right image sequences included in the water surface image sequence into the trained student network to enable the student network to output a wave surface height field sequence; reconstructing a continuous wave surface geometric model based on the wave surface height field sequence; and determining wave parameters by analyzing the continuous wave surface geometric model.

This invention discloses a three-dimensional measurement method for the surface morphology of grinding wheels, relating to the field of grinding inspection technology. The method involves fixing the grinding wheel to be tested on a rotatable grinding wheel clamping table, and acquiring a sequence of circumferential, multi-field, and multi-focal surface images of the grinding wheel's circumferential surface using an industrial camera that can move up and down and left and right. After image preprocessing, the pixel focus height is determined based on the sharpness response, and the grinding wheel's circumferential curved surface reference is subtracted to generate a three-dimensional height field of the grinding wheel surface. Then, by combining the three-dimensional height, grayscale texture, and topological connectivity, the abrasive grains, binder, and pore regions are separated, and the abrasive grain exposure, exposure volume, effective abrasive grain spacing, pore connectivity, and texture features are extracted. Finally, the results are compared with benchmark topographic data to output the grinding wheel surface topography evaluation results and suggestions for dressing or replacement. This method can be used for grinding wheel selection, dressing quality verification, and accurate judgment of grinding conditions.

An exemplary real-time simulation framework for a mixture of fluid and particulate matter is disclosed. This framework may be based on modeling the altitude fields and horizontal velocities of different material phases, which in the context of a water-sand system may include sand, water, and mixed water. This framework balances simulation fidelity and performance and provides real-time computation for interactive applications. The exemplary framework formulates external frictional and elastoplastic internal forces of sand based on a horizontal grid and further addresses the water / sand bonding through diffusion and momentum exchange. Time updates for the simulation are efficiently performed using a semiimplicit operator partitioning discretization scheme and an asynchronous scheme for the fluid and particulate matter.

The application discloses a plane prescription lens phase coding method and system applied to the technical field of plane prescription microstructure design, and the method comprises the following steps: obtaining prescription parameters, wearing posture parameters and manufacturing assembly constraints, performing equivalent compensation on the prescription and executing axis position rotation in a plane lens coordinate system, further generating a target wavefront, a two-dimensional phase field and a height field, performing local power continuity checking and manufacturability checking on the phase field, and outputting a phase diagram, a height diagram, a manufacturing file and a parameter list, and the system comprises an input module, a coordinate transformation module, a phase coding module, a manufacturability evaluation module and an output module. Through the above scheme, the spherical lens, the cylindrical lens and the axis position can be written into the plane microstructure, and the local image quality continuity and the manufacturing realizability can be controlled simultaneously.

The invention relates to the technical field of industrial visual inspection, in particular to an electric connector detection method and system based on image recognition. The method comprises the following steps: acquiring a pure luminosity component through narrowband multispectral time-sharing stroboscopic acquisition and channel decoupling; constructing an anisotropic highlight confidence field based on the second derivative features to suppress texture noise; a global energy function containing smooth constraint is introduced, and the balance between denoising and edge preservation is realized through global optimization; and finally, calculating a three-dimensional comprehensive skewness index based on height field segmentation. According to the method, stamping texture artifacts can be effectively eliminated, the space attitude anomaly of the pin is accurately identified while micron-order edge features are reserved, and the detection robustness is remarkably improved.

The invention discloses a machine vision detection method and system based on a phase deflection technology. The method comprises the following steps: collecting a plurality of light stripe images with different phases formed on an object to be detected; constructing a photometric model according to the light stripe image, and solving the photometric model to obtain a surface normal vector and an albedo of each pixel point in the light stripe image; calculating a height field of each pixel point according to the surface normal vector of each pixel point, and generating a three-dimensional reconstruction image of the to-be-detected object based on the height field; calculating an average curvature according to the albedo and the height field of each pixel point, and performing contrast stretching on the three-dimensional reconstructed image based on the average curvature to generate a stretched image; and inputting the stretched image into a trained defect detection model, so that the defect detection model determines whether the to-be-detected object has defects according to the stretched image. By implementing the method, imaging of defects such as weak guide scratches and the like can be realized, and real defects and false defects can be distinguished.

The invention discloses a chuck flatness detection method based on surface texture image analysis, and the method comprises the steps: collecting a plurality of frames of texture images of the surface of a chuck through a program-controlled array light source at different illumination angles, and constructing a multi-view illumination information sequence; inputting the image sequence into a photometric stereoscopic vision reconstruction model based on total variation regularization, establishing a Lambert reflection constraint equation, introducing a total variation regularization item to carry out spatial smooth constraint, and carrying out iterative optimization to solve a surface normal vector field; mapping the gradient field converted from the normal vector field to a frequency domain through Fourier transform by using a Frankot-Chellappa frequency domain integral operator, constructing a global integrability constraint to solve a height field, and performing inverse transformation to output a three-dimensional morphology height map; and fitting a reference plane by adopting a least square method, calculating a vertical distance of discrete points, and extracting peak and valley values as flatness indexes. According to the invention, non-contact, high-efficiency and high-precision chuck flatness detection is realized, and the quality detection requirement in the precision manufacturing field is met.

The present application relates to the technical field of industrial vision detection, and more particularly to a kind of detection method and system of electric connector based on image recognition.The method comprises the following steps: obtaining pure photometric component by narrowband multispectral time-sharing frequency flash acquisition and channel decoupling;Anisotropic highlight confidence field is constructed based on second derivative feature to suppress texture noise;Global energy function containing smoothing constraint is introduced, and the balance of denoising and edge preservation is realized by global optimization;Finally, three-dimensional comprehensive skewness index is calculated based on height field segmentation.The present application can effectively eliminate stamping texture artifacts, accurately identify the spatial posture anomaly of pin while retaining micron-level edge features, and significantly improve the detection robustness.

The application discloses a terrain special effect processing method and device, a storage medium, equipment and a program product. The method comprises the following steps: acquiring terrain height field data of a target model, determining terrain features of the target model based on the terrain height field data; based on the terrain features, determining target region information of the target model that meets preset terrain features, wherein the target region information comprises spatial feature information and modification parameter information of the target region; based on the target region information, selecting a target modification tool from preset terrain modification tools, and controlling the target modification tool to modify the terrain of the target region, so that the terrain of the target region matches the terrain of the target modification tool. The target modification tool can make the terrain form, terrain height and terrain material of the target region consistent with the preset details of the target modification tool, and the modification can be quickly completed through the target modification tool, thereby improving the modification efficiency. The terrain features quantize the terrain form (such as high curvature for ridges and high slope for steep slopes), ensure that the terrain feature recognition conforms to the natural law, and avoid the deviation of artificial judgment.