25 results about "Fine resolution" patented technology

The invention provides a vegetation index spatio-temporal fusion method considering farmland spatio-temporal heterogeneity, which comprises the following steps: simultaneously acquiring high-resolution (fine-resolution) and low-resolution (coarse-resolution) satellite remote sensing data of a target area, preprocessing the satellite remote sensing data, carrying out field parcel segmentation on the high-resolution remote sensing data, and carrying out field parcel segmentation on the low-resolution remote sensing data; optimizing a selection strategy of similar pixels by taking the field parcel as a unit; vegetation index time series data of the high-quality coarse-resolution image are constructed through data reconstruction; fitting a crop vegetation index growth curve of the fine-resolution image by adopting a Double Logistic function, and generating a fine-resolution pixel initial value at a prediction moment; constructing an optimization model and optimizing a fine pixel vegetation index value in combination with coarse and fine resolution pixel data; weighted distribution is carried out based on the spatial consistency in the field, and a prediction image is optimized in combination with spatial filtering. According to the method, the problems of unreasonable assumption of a traditional space-time fusion model on crop spectrum linear growth in a high-heterogeneity farmland scene and inaccurate similar pixel matching are effectively solved, the fusion precision and robustness are improved, and the method has a wide application prospect.

The invention relates to a complex terrain air temperature refined downscaling method based on a TopoMW model, and belongs to the technical field of meteorological element downscaling, and the method comprises the steps: obtaining multi-source data of a target region, carrying out the preprocessing of the multi-source data, and fusing the multi-source data into a feature set; constructing a complex terrain multi-weighting matrix based on the preprocessed multi-source data; a TopoMW model improved based on UNet is established, a weighted MSE loss function is adopted to train the TopoMW model on the basis of a feature set in combination with a complex terrain multiple weighting matrix, an optimal downscaling model is obtained, and the contradiction between model statistic multiplexing and performance retention is solved; inputting a feature set to be predicted into the optimal downscaling model to obtain an initial prediction result; the initial prediction result is corrected based on the complex terrain multiple weighting matrix to obtain a final result, extreme deviation is eliminated, and accurate generation of the 30-meter high-resolution air temperature under the complex terrain is achieved.

Described herein are methods, systems and computer-program products for intra- and inter-gait classification for both symmetric and asymmetric gait patterns using images.

The embodiment of the invention provides a Transformer track prediction method based on improved track coding, and the method comprises the steps: obtaining track data: collecting and processing the dynamic information and navigation information of a ship and the static information of the ship, and obtaining the longitude, latitude, course and speed information of the ship; track information coding is improved, and obtained continuous observation data including longitude, latitude, course and navigational speed information is converted into a structured discrete form; track information embedding is carried out, embedding coding is carried out on each dimension of coded data, and the codes are mapped to a new multi-dimensional space from an original space to which the codes belong; the method comprises the following steps: performing track prediction, adopting a TrAS I former prediction model and cross entropy loss to replace mean square error, processing multi-modal characteristics of data, converting a track prediction task into a multi-classification problem, optimizing a loss function by combining data representation of fine resolution and coarse resolution, and generating a future position through recursion until the prediction of the whole track is completed.

The invention particularly relates to a hierarchical search three-dimensional airspace path planning method and system based on an adaptive resolution grid. The method comprises the following steps: constructing a corresponding multi-resolution grid model based on a three-dimensional grid map; global path planning is carried out on the coarse resolution grids, coarse paths of all coarse grid nodes are determined, and a coarse path sequence is constructed; performing non-overlapping segmentation on the coarse path sequence; planning local windows matched with the fine-resolution grids, and determining window entrance and exit points corresponding to the local windows; performing obstacle recognition on each local window based on the fine resolution grid, calling a corresponding path generation strategy according to an obstacle recognition result of the local window, and determining a local refined sub-path corresponding to each local window; and splicing the path information of each local window in sequence to obtain a path planning result. According to the method, efficient and reliable path planning of a large-scale and complex three-dimensional airspace environment can be realized.

The application relates to a vector polygon rasterization method of a spherical hexagonal grid system and belongs to the technical field of geographic information. The vector polygon area to be rasterized is rasterized from coarse resolution to fine resolution by using the hierarchical characteristics of the grid system and adopting a cross-layer subdivision mode. The subdivision mode can generate a subunit by spacing a plurality of grid levels of a parent unit, solves the position uncertainty problem of the subunit relative to the parent unit after multiple times of subdivision of an adjacent level subdivision scheme, greatly reduces the number of units needing spatial relation judgment with the vector polygon, avoids the rasterization realized by traversing and judging the spatial relation of units one by one, and greatly improves the efficiency of the vector polygon rasterization on the spherical hexagonal grid system.

A tritium-compatible pellet gas gun for a deuterium-tritium (D-T) pellet fueling apparatus is provided. In one aspect, the tritium-compatible pellet gas gun includes a pellet sizer assembly that provides in situ adjustment of pellet length during injector operation. The pellet sizer assembly includes a guillotine slide that is actuated by a pusher tube, the guillotine slide being operable to restrict the orifice of the extruder nozzle to reduce pellet length with fine resolution. The tritium-compatible pellet gas gun also includes dual flexible metal bellows seals that maintain hermetic separation between the D-T fuel region, a guard vacuum, and the ambient environment, while still permitting linear motion. This configuration allows the pellet length to be reduced by up to 50% without venting or warming the D-T pellet fueling apparatus.

The invention relates to a remote sensing spatio-temporal data fusion method based on phenological constraint and meteorological driving, and belongs to the technical field of data processing. A high spatial resolution remote sensing image and a high time resolution remote sensing image are subjected to regression modeling, spatial filtering and residual error compensation, a fine resolution image and a coarse resolution image are subjected to preliminary time-space fusion, and a meteorological driving model between meteorological variable change and vegetation index short-term response is established; extracting a time sequence of a sample region from the high-time-resolution data, performing nonlinear curve fitting on the time sequence by using a Gaussian model, establishing a phenological growth model, performing parameter solution and optimization on the model to obtain a regional representative phenological model of the crop type, and carrying out phenological constraint and meteorological drive correction on the preliminary time-space fusion result to generate a final high-temporal-spatial-resolution image sequence. According to the method, a high-temporal-spatial-resolution remote sensing product which is lower in noise, more stable in phenological characteristic and more credible can be generated.

The present invention relates to a method and system for fabricating customized orthodontic brackets using 3D printing technologies with ultra-fine resolution to produce the orthodontic bracket with multiple layer thicknesses, and each layer thickness ranges from 0.05 micrometers to 100 micrometers. The process involves accessing data related to a 3D CAD bracket structure model for the customized orthodontic bracket. The 3D printing technologies with adaptive slicing dynamically adjust layer thickness based on geometric complexity and surface roughness, ensuring the variable layer ranges from 0.05 micrometers to 100 micrometers. By applying variable layer thickness, the method optimizes surface roughness.

Systems and methods for configuring a hearing aid device may include automatically: selecting a sound file from a sound database, processing the sound file with a first set of parameter settings to generate two distinct coarse stimuli, playing the two distinct coarse stimuli to a user, recording a preferred coarse setting by the user, processing the sound file with a second set of parameter settings to generate two distinct shape stimuli, playing the two distinct shape stimuli to the user, recording a preferred shape setting by the user, processing the sound file with a third set of parameter settings to generate two distinct fine resolution stimuli, playing the two distinct fine resolution stimuli to the user, recording a preferred fine resolution setting by the user, and determining hearing aid parameter settings for the user based on the fine resolution setting.

Several embodiments include methods for performing hardware-assisted motion estimation while encoding blocks of a media image. A video encoder generates motion vector clues for multiple stages of motion estimation at progressively finer resolutions, culminating in the final motion estimation stage at the full resolution of the media image. The video encoder divides the media image into a grid of raster blocks, where the raster blocks are smaller than the standard encoding block size of the media image. At each stage, the video encoder generates a motion vector clue for each raster block at the selected resolution for that stage. For each motion vector clue, the video encoder generates forward-prediction clue data based on a reference media image that precedes the current media image.In addition, for each motion vector cue, the video encoder generates backward prediction cue data based on a reference media image that follows the current media image.

The invention discloses an anti-time-varying sensitive remote sensing image space-time fusion method and system, and belongs to the technical field of remote sensing image processing, and the method comprises the steps: obtaining a priori moment fine resolution image, a priori moment coarse resolution image and a prediction moment coarse resolution image; extracting multi-scale prior detail semantic features from the prior moment fine resolution image through a resolution reduction encoder; multi-scale semantic features are extracted from the coarse resolution images at the prior moment and the prediction moment through a super-resolution encoder, and time-varying semantic features are obtained through subtraction operation; inputting the two types of features into a double-orientation three-attention fusion decoder, and sequentially generating a preliminary local-global cross fusion feature and a decision fusion feature; the reconstruction unit reconstructs the decision fusion feature into a prediction moment fine resolution image; and performing an adversarial learning optimization generation process on the predicted image and the real image. According to the method, the prediction robustness of time-space fusion on time-varying information is improved.

This application relates to the field of X-ray detection technology, specifically to a high-throughput X-ray multi-energy spectral detection system and method. The detection method includes preparing n filters Z with different K absorption limits. i , i∈(1,n), n≥15; the incident X-rays, after collimation, pass through each filter Z i The electrical signal S was detected after attenuation. i S i =∫η×[φ(E)×C i (E)]dE,(E min E max ); for electrical signal S i The incident X-ray energy spectrum is obtained by discretizing the spectrum and then solving for the least squares solution. Based on the detection principle of energy integration, this application achieves flexible modulation and demodulation of high-throughput X-ray energy spectra through mathematical combinations of filters with different linear attenuation coefficients, balancing detection cost with obtaining high-resolution X-ray energy spectra.

The application discloses a fire smoke environment life detection method based on 80GHz millimeter wave radar multi-receiving antenna joint voting. In the application, the false alarm caused by random noise and environmental clutter of a single antenna channel is effectively inhibited through voting confirmation after independent detection of multiple antennas. Since real vital signs can be consistently detected on multiple antennas, and interference sources are difficult to simultaneously meet the consistency condition of multiple channels, the reliability of the detection result is significantly improved. Meanwhile, the mechanism is not limited by the number of antennas and array configuration, has good universality and scalability, is convenient to deploy and implement on different hardware platforms, and the spatial coherence accumulation of beam forming enhances the signal-to-noise ratio of weak breathing echoes behind the shelter, so that the trapped personnel sheltered by sofas, tables and chairs and the like can still be effectively detected. Meanwhile, the fine resolution capability in the angle dimension improves the accuracy of target direction estimation, and provides more accurate positioning information for fire rescue.

The embodiment of the invention provides a robot positioning loss self-checking and repositioning method, device and equipment, and is applied to the technical field of robot autonomous navigation and positioning. The method comprises the following steps: acquiring an estimated pose corresponding to a current track node and current laser scanning point cloud data acquired by a laser radar in real time; constructing a local subgraph candidate set in a credible radius range by taking the current estimated pose as a center; the current laser scanning point cloud data are matched with each sub-graph in the local sub-graph candidate set, and if all matching scores are lower than a preset similarity threshold value, it is judged that positioning loss happens to the robot; after it is judged that positioning is lost, coarse resolution matching is executed in parallel in a global map, and invalid candidate areas are removed through an upper bound pruning strategy; and carrying out fine resolution matching on the reserved candidate sub-images, and finally outputting an accurate relocation pose. In this way, the continuity and reliability of the navigation task of the robot in a complex environment are effectively guaranteed.

A low dropout regulator with fine loop preset including a coarse resolution regulator and a fine resolution regulator. The coarse resolution regulator adjusts a digital control value in response to an output transient to control a coarse output driver to regulate an output voltage with coarse resolution. The coarse resolution regulator eventually enters a limit-cycle oscillation mode with a corresponding duty cycle. The fine resolution regulator is disabled when the digital control value initially changes, is preset based on the duty cycle of the digital control value when in the limit-cycle oscillation mode, and is re-enabled after being preset to drive at least one fine output driver to regulate the output voltage with fine resolution. The preset provides a closer estimate of the supplemental current needed to meet the output load level when turned on after an output load transient to achieve less output ripple voltage.

The present invention discloses a device and method for measuring the flow field around scour pile piers using stereo PIV. The device of the present invention can eliminate scattering and reflection between scour pits and different media through a stereo PIV system combined with an optical refraction component, thereby ensuring the accuracy of the flow field in the scour pit to the greatest extent. Combined with PIVlab software, based on the fine resolution of the field of view, a three-dimensional joint grid is generated according to the discrete data position of the PIV particles, and the Reynolds decomposition data of the velocity vector is interpolated onto the three-dimensional joint grid to assimilate the overlapping measurement areas in the relevant field of view to the data set of the same particle displacement. The present invention quantifies the turbulent structure of the flow field around the scour pile pier; determines the vertical velocity distribution pattern of the characteristic position in the scour pit; analyzes the water flow separation process at the scour edge and its interaction with the main vortex; determines the spatiotemporal evolution of the shear stress of the water flow wall during the scour process, etc.

Various embodiments include techniques for performing hardware-assisted motion estimation when encoding blocks of a media frame. A video encoder generates motion vector hints for multiple phases of motion estimation at increasingly finer resolutions until the final motion estimation phase at the full resolution of the media frame. The video encoder divides the media frame into a grid of grid blocks, where the grid blocks are smaller than the standard encoding block size of the media frame. At each phase, the video encoder generates one motion vector hint for each grid block at the selected resolution for that phase. For each motion vector hint, the video encoder generates forward prediction hint data based on a reference media frame previous to the current media frame. Further, for each motion vector hint, the video encoder generates backward prediction hint data based on a reference media frame that follows the current media frame.

This application relates to the field of remote sensing hydrological monitoring technology and discloses a method for estimating evapotranspiration based on data fusion and a physics-guided neural network. This method acquires multi-source satellite remote sensing and meteorological data, utilizes an enhanced spatiotemporal adaptive reflectance fusion model and a data mining sharing algorithm to fuse coarse and fine resolution images, generating daily seamless high-resolution surface reflectance and surface temperature data. A physics-guided neural network model is constructed, comprising a parametric subnetwork and a main network embedding the surface energy balance physical equations. The parametric subnetwork is used to predict the thermodynamic roughness length, driving the main network to iteratively calculate sensible and latent heat fluxes in conjunction with meteorological data. The model is trained based on a loss function containing physical constraints. This invention, through deep coupling of data-driven and physical mechanisms, resolves the spatiotemporal resolution contradiction of remote sensing data and improves the model's generalization ability and physical consistency in data-scarce regions.

The application discloses a cloud computing space-time fusion method considering coarse and fine resolution scale conversion error correction, which comprises the following steps: synthesizing a cloud-free reference image pair based on the principles of time consistency and space similarity, and determining a coarse resolution remote sensing image and a fine resolution remote sensing image at a predicted time; expressing the spatial heterogeneity of the earth's surface; calculating a scale conversion error term; calculating a residual term; correcting the scale conversion error term and the residual term based on the result of superpixel segmentation, and compensating to the fine resolution remote sensing image in the cloud-free reference image pair, so as to obtain a final fine resolution remote sensing image fusion result. The application improves the fusion precision of heterogeneity and changing landscapes by introducing the scale conversion error term; stable fusion results can still be obtained even in the scene of thick cloud and continuous cloud pollution; the method is not only suitable for the reconstruction of NDVI time series, but also suitable for the fusion of multispectral reflectivity.

Digital step attenuator (DSA) and digital phase shifter (DPS) multi-stage circuit architectures that provide for high resolution. Embodiments use a dithering approach to weight bit positions to provide a much finer resolution than the lowest-valued individual stage. Bit position weights for stages are determined so as to enable selection of combinations of n bit positions that provide a desired total attenuation or phase shift range while allowing utilization of the large number of states (2n) available to produce fractional intermediate steps of attenuation or phase shift. The fractional intermediate steps have a resolution finer than the lowest-valued stage. Bit position weights may be determined using a weighting function, including weightings determined from a linear series, a geometric series, a harmonic series, or alternating variants of such series. In some embodiments, at least one bit position has a fixed value that is not determined by the bit position weighting function.

A method of simulating additive manufacturing, including determining a first displacement field on a first adaptive finite element mesh of an object; determining a second adaptive finite element mesh by adding an additional element group to the first adaptive finite element mesh, and when coarsening criteria is met, generating at least one coarse element by coarsening two or more elements of the first adaptive finite element mesh; calculating, from the first displacement field, a second displacement field on the determined second adaptive finite element mesh; determining an error between the first and second displacement fields; calculating a displacement field change on the determined second adaptive finite element mesh using a finite-element analysis model; and calculating a fine displacement field on a fine finite element mesh by adding together the second displacement field, the determined error, and the calculated displacement field change.

The invention discloses a soil water content downscaling method based on terrain and vegetation indexes, which comprises the following main steps of: firstly, extracting terrain data of 10km and fine spatial resolution of a watershed, and obtaining a boundary file of the watershed; acquiring topographic data of coarse resolution for downscaling and normalized differential infrared index raster data; selecting an SMOS soil water content satellite product as the soil water content data, and extracting the soil water content data of the watershed according to the watershed mask; determining the soil water content weight of each fine resolution grid based on the terrain humidity index TWI and the terrain humidity index NDII; and multiplying the weight coefficient of each grid by the soil water content of the coarse resolution to obtain the soil water content of the fine spatial resolution. According to the method, the soil water content is downscaled by considering the spatial heterogeneity of the underlying surface, and the influence of topographic factors and vegetation distribution on the spatial distribution of the soil water content is fully considered, so that the downscaling of the coarse-resolution soil water content is realized.

The invention discloses a diffusion model space-time fusion method based on optical flow guidance. The method comprises the following steps: acquiring a fine spatial resolution image and a coarse spatial resolution image of the same observation area at a reference moment and a coarse spatial resolution image at a target moment; performing up-sampling on the coarse spatial resolution image to obtain an up-sampling image with the same scale as the fine spatial resolution image, and constructing a pseudo fine resolution image; estimating a pixel-level displacement optical flow for the fine spatial resolution image and the pseudo fine resolution image based on an optical flow estimation network to obtain a target scale optical flow; the gradient consistency of the fine spatial resolution image and the pseudo fine resolution image is calculated, and an optical flow confidence map is constructed; establishing a conditional diffusion model to obtain an initial fusion image at a target moment; and splicing the initial fusion image and the upsampled coarse spatial resolution image along the channel dimension, and outputting a high-resolution fusion image at the target moment. According to the method, the target moment high-resolution fusion image which is consistent in structure, clear in detail and good in radiation consistency can be generated.

This invention discloses a point cloud instance segmentation method and system based on coarse-to-fine adaptive clustering. First, it acquires LiDAR point cloud data of the target scene and obtains semantic segmentation labels using a semantic segmentation model, separating the point cloud into background and foreground object classes. A multi-level bird's-eye view BEV pyramid is constructed from coarse to fine resolution. At each resolution level, the clustering kernel size is adaptively adjusted based on the distance from the point cloud to the sensor for iterative clustering. Geometric center estimation and bounding box detection based on symmetry completion are performed on the candidate instances obtained from clustering. Oversegmentation and undersegmentation of candidate instances are geometrically verified using full coverage constraints and spatiotemporal mutual exclusion constraints. Undersegmented instances are passed to the next higher resolution level for processing, while oversegmented instances are fused. Finally, the background point cloud is combined to generate a panoramic segmentation result. This invention's method does not require additional training for instance segmentation, improving the accuracy and robustness of panoramic segmentation.