95 results about "Multi resolution" patented technology

The application discloses a kind of based on adaptive boundary perception's spare part point cloud hierarchical super voxel segmentation method, it is related to three-dimensional point cloud super voxel segmentation technical field, this method calculates the multidimensional geometry feature of each point of point cloud first, constructs point cloud boundary weight model by weighted fusion and nonlinear transformation;Again, complete point cloud voxelization based on octree space division mechanism, according to boundary weight value, voxel is divided into boundary voxel and seed voxel, and the initial super voxel is generated by merging the point set in voxel through union-find set data structure;Then, relying on global energy optimization framework, combined with distance measurement function containing boundary weight difference item, hierarchical iterative fusion is carried out on the representative point of initial super voxel;Finally, the super voxel after fusion is corrected and the spatial connectivity is corrected, and the multi-resolution super voxel segmentation result is output.The application can solve the problem that traditional method is easy to over-segmentation, under-segmentation, and enhance the adaptive level of complex equipment spare part structure.

The application provides a large-scale satellite image automatic registration and enhancement method based on adaptive multi-source fusion, comprising: performing orthorectification and multi-resolution unified projection on a multi-source original data set obtained to obtain a coarse registration image set; calculating a local tensor structure of the coarse registration image set, performing repeated texture recognition, generating a final structure representation and a structure reliability; performing regional registration to obtain a full-image dense deformation field, then performing image transformation to be aligned and registration uncertainty estimation to obtain a high-precision registration image and structure registration uncertainty estimation; combining a pseudo-change probability of the high-precision registration image and the structure registration uncertainty estimation to obtain a change reliability weight; under the constraint of the change reliability weight, performing consistency enhancement on the high-precision registration image to output a quantitative and reliable enhanced image. The application realizes high-precision, low-false alarm and quantifiable and reliable automatic registration and enhancement of multi-source large-scale satellite images.

This invention belongs to the interdisciplinary field of marine remote sensing technology and artificial intelligence, specifically relating to a deep learning-based method for reconstructing a three-dimensional marine temperature field. The method includes: acquiring multi-source, multi-resolution satellite data and reanalysis data; constructing a sample library after preprocessing; building a three-branch MAUS Transformer model using a U-Net network as the framework, replacing convolutional blocks with Swing Transformer modules embedded in the encoding and decoding paths, and embedding a CBAM attention module to adaptively optimize feature representation. The three-branch model can extract and fuse features at resolutions of 1 / 100°, 1 / 8°, and 1 / 4° respectively; simultaneously, an adaptive loss function is designed to fuse horizontal partitioning and vertical layering, and a weighted constraint based on the real temperature gradient is introduced to reconstruct the thermocline. This invention can output a high spatiotemporal resolution three-dimensional marine temperature field, improving the accuracy, stability, and robustness of the reconstruction, and is of great significance for marine environmental protection, marine data development, marine engineering construction, and marine environmental safety assurance.

The application discloses a PCBA foreign matter detection method based on a multi-resolution subspace projection, acquires a to-be-detected board and a standard reference board image, determines an adaptive down-sampling multiple based on a minimum foreign matter physical area, performs image registration with the aid of a mask to eliminate geometric deformation, extracts a multi-resolution feature map through a pre-trained deep convolutional neural network, calculates feature differences, and uses a dynamic threshold to screen a preliminary candidate point set, constructs a local principal component subspace of a local neighborhood of a reference image, performs first-stage screening by calculating the orthogonal projection error of a test feature vector in the space, maps remaining candidate points to a multi-scale feature space and splices them, constructs a global multi-scale principal component subspace, performs second-stage screening through a global projection error, and outputs a final result selection, and through the implementation of the application, the problems of a high-definition image computing power bottleneck and a high false alarm in comparison are solved, the application has zero-sample generalization capability, and sample-free training can accurately detect tiny foreign matters.

The application discloses a kind of based on GiST spatial index multi-resolution fragmentation DEM database construction method, comprising: S1, with the resolution of fragmentation DEM image tile as hierarchical division standard;S2, existing DEM image tile is hierarchically divided according to S1 hierarchical division standard, to minimum outer rectangle as unit independent storage;S3, based on DEM image tile, the hierarchical structure rule of R tree is used to establish parent-child node relationship, and the initial R tree multi-resolution skeleton is constructed;S4, on the basis of initial R tree multi-resolution skeleton, uniform different hierarchical spatial relationship description rule is formulated, and according to the rule, the spatial correlation between DEM image tile is described, and the preliminary establishment of multi-resolution DEM database based on GiST spatial index is completed;S5, the DEM image tile to be stored is stored into database, and the database is improved.The application improves the performance of DEM data storage.

The present application provides a kind of quantitative evaluation method of the influence of wind tunnel inner surface roughness on airflow quality, belongs to the technical field of wind tunnel, the present application is collected by laser confocal scanner wall roughness multi-resolution point cloud data, high frequency, medium frequency and low frequency roughness component is extracted by wavelet decomposition, power spectral density function and fractal dimension are calculated to establish random roughness field probability distribution model, local velocity field and vorticity distribution are obtained by using lattice Boltzmann algorithm to simulate rough wall near flow field, the uniformity of airflow is calculated by generating equivalent random roughness field sample using Monte Carlo method for statistical analysis, the polynomial mapping function of roughness statistical characteristic parameter and airflow quality degradation index is established, when degradation index exceeds allowable threshold, wall roughness is over standard, solve the technical problems of low precision and insufficient calculation efficiency in quantitative evaluation of the influence of multi-scale random roughness on wind tunnel airflow quality.

A protein structure-density map fitting method based on feature point matching belongs to the field of bioinformatics and structural biology, constructs a protein complex structure dataset, converts the structure into a unified resolution point set through voxelization and uniform sampling, and adopts the farthest point sampling to construct a multi-resolution point set to depict scale geometry; a deep learning network is adopted to learn the rotation equivariant and invariant features of the structure and the density map under multi-resolution, and a geometric self-attention mechanism based on nearest neighbor layer-by-layer expansion is introduced at the neck to enhance the representation; a coarse-to-fine feature point matching strategy is adopted in the training stage, combined with a superpoint matching loss, a point-level matching loss and a contrast rotation loss optimization; in the inference stage, multi-scale and hybrid sampling are adopted, and candidate poses are generated through translation mask, the candidate poses are screened and optimized according to the structure-density map fitting score, and the final fitting result is output. The present application can realize high-precision and high-efficiency structure-density map fitting under complex conditions.

This application discloses a method and related equipment for evaluating microwave-assisted coal seam production enhancement based on fracture fractal parameters, relating to the field of coal and rock prediction technology. The method includes: constructing a multi-resolution digital core sample library based on sample data of coal and rock samples from the target mining area, and then extracting a fractal feature set of fractures; constructing a vector basis function mapping prediction algorithm based on the fractal feature set to predict the regional evolution of coal seam fracture fractal parameters during microwave mining; quantitatively characterizing fracture behavior using the fractal parameter evolution set, and then establishing a fractal characterization model coupling electromagnetic, thermal, seepage, and mechanical fields as a four-field cross-scale fractal coupling model; reducing the dimensionality of the fractal feature set to a low-dimensional embedding space, and then predicting and generating a 3D fractal parameter volume covering the target mining area; inputting the 3D fractal parameter volume into the four-field cross-scale fractal coupling model, and using the four-field cross-scale fractal coupling model to output production enhancement evaluation indicators. This application improves the prediction accuracy of coal seam fractures.

The application provides a physical mechanism coupling data-driven fast prediction method, and belongs to the technical field of disaster prediction.The application extracts an invariant topological feature vector through topological data analysis driven phase space reconstruction and inputs a physical constraint identification model to obtain a conservation law deviation vector and a physical feasible region boundary parameter, adopts a multi-resolution adaptive grid technology to perform multi-scale decomposition, extracts long-term evolution trend features and spatial local features through a full connection layer network and a convolution layer network, combines the physical feasible region boundary parameter to perform projection gradient descent iterative optimization to generate a preliminary prediction field, performs Bayesian uncertainty quantization, adjusts an artificial intelligence model regularization coefficient according to a disaster precursor identification function value, and outputs a graded early warning result, so that the technical problem of insufficient prediction accuracy caused by insufficient coupling of a physical constraint and a data-driven model in a disaster prediction system is solved.

The application discloses a single-view three-dimensional texture consistency generation method based on geometric structure, pre-processes single-view image input data of a to-be-processed object and corresponding view angle parameters to obtain training input images, a view angle parameter set and supervision information, carries out geometric structure feature extraction and weight mapping processing on the training input images to obtain a geometric structure weight map, carries out sampling density distribution and position sampling on three-dimensional space texture query positions according to the geometric structure weight map to obtain a sampling point set, carries out multi-resolution coding on each sampling point in the sampling point set, and constructs a continuous texture representation through a multi-resolution texture hash grid to obtain texture attributes corresponding to each sampling point, and then carries out differentiable rendering, loss calculation and parameter updating to obtain a three-dimensional texture generation result. Compared with the prior art, the application improves the structure consistency, detail integrity and stability of three-dimensional texture generation under single-view conditions.

The present disclosure relates to a method, device, electronic equipment and storage medium for digital terrain model data, the method comprising: obtaining digital elevation model data in a target operation area; generating the digital elevation model data into target digital elevation model data with multiple resolutions; obtaining target terrain factors of the target digital elevation model data at different resolutions; and obtaining digital terrain model data with multiple resolutions based on the target digital elevation model data and the target terrain factors. Since the embodiments of the present disclosure can generate digital terrain model data containing target terrain factors and having multiple resolutions, the safety and efficiency of driving can be greatly improved when an unmanned vehicle operates in a target operation area based on a map containing the digital terrain model data.

A system for generating a real-time object-focused video using minimal camera arrays with pre-computed sports-field-optimized spatial mapping. The system positions virtual cameras to maintain tracked objects in focused, straight-ahead orientations while supporting one-dimensional movement between two cameras using geometric interpolation and two-dimensional movement within three-camera triangular configurations using barycentric coordinates. Computer spatial mapping with discretized depth information optimized for fast-moving object tracking in sports environments eliminates real-time depth calculation overhead, avoiding latency bottleneck and enabling ultra-low latency processing suitable for live sports broadcasting. The system includes predictive camera set switching using mathematical positioning variables, multi-object tracking capabilities with distributed processing frameworks, and intelligent 2D occlusion handling optimized for broadcast video output with parallax-induced occlusion management. Video synthesis techniques including adaptive geometric transformation, and multi-resolution image processing achieve rapid processing performance for live broadcasting applications while preventing discrete camera switching artifacts through continuous interpolation coefficients that eliminate abrupt perspective transitions.

This application discloses a multi-scale spatiotemporal fusion method and system for phenological identification based on remote sensing images. It utilizes a multi-scale spatiotemporal fusion deep learning framework, constructing continuous time series based on high-frequency acquired baseline resolution images as time anchors. This is then spatially enhanced by combining these with sparse high-resolution images. By introducing a missing-aware gating fusion mechanism, multi-resolution features on unaligned time axes are dynamically fused. The resulting fused feature sequence is then input into a long short-term memory network to achieve phenological stage identification of crop germplasm resources. This application can stably identify the jointing-booting stage, heading-flowering stage, milk stage, waxy ripening stage, and maturity stage under conditions of sparse high-resolution images and asynchronous sampling times, balancing identification accuracy and operational efficiency. It is suitable for large-scale germplasm resource phenotypic monitoring, critical period early warning, and precision breeding applications, providing a flexible, scalable, and cost-effective solution for high-throughput phenotypic analysis.

An embodiment provides a method comprising the steps of: receiving an activation signal for accessing data in at least one memory; accessing a data structure in the at least one memory according to the reception of the activation signal, wherein the data structure includes a diagnostic image; loading the diagnostic image from the at least one memory; generating, by at least one processor, at least one response with respect to the diagnostic image by using at least one artificial intelligence model using the diagnostic image as an input, wherein the at least one artificial intelligence model is pre-trained to perform image analysis on the basis of a feature representation for a causal dependency relationship between a low-magnification feature and a high-magnification feature of the diagnostic image; and inputting (ingesting) the at least one response to at least one subsequent processing component.

A hierarchical contact search algorithm applicable to multi-resolution particle methods is disclosed. The algorithm includes the following steps: reading the physical signals of large and small particles in a fluid model, and reading the geometric information of high-precision and low-precision regions; dividing the entire computational region into multiple large grids by combining the overall geometric information of the fluid model and the size of the large particles; dividing the high-precision region into multiple small grids by combining the geometric information of the high-precision region and the size of the small particles; the large and small grid regions can overlap; positioning large particles belonging to the low-precision region to the large grid region, and positioning small particles belonging to the high-precision region to the small grid region; determining the grid to which the computational particle belongs and all neighboring particles within the influence radius of adjacent grids, forming contact pairs with the computational particle, calculating the interaction forces, and improving the computational efficiency of the entire computational region.

The application discloses a single-end protection method based on wavelet transform and pulse neural network, which comprises the following steps: collecting transient voltage and current signals of a single end of a direct-current distribution network in real time; performing multi-resolution decomposition on the collected transient waveforms by using wavelet transform to obtain detail coefficients and high-frequency energy spectrum under different scales; taking an input layer of the pulse neural network as an encoding layer, and converting the extracted continuous analog wavelet coefficients into discrete binary pulses; and finally inputting the generated pulse sequence into the pulse neural network to complete accurate classification of fault types and output corresponding relay protection tripping instructions. The application proposes a single-end protection method based on wavelet transform and pulse neural network for problems such as uncertain energy flow direction, complex mapping relationship between fault characteristics and fault types and the like caused by large access of distributed energy, and avoids strong dependence on accurate fault mathematical models and fixed protection thresholds.

This application discloses a method, device, and storage medium for correcting digital pathological fluorescence images, relating to the field of digital pathology technology. The method includes: downsampling the original fluorescence image to obtain a sampled fluorescence image; solving for initial path parameters and initial quenching field parameters of the sampled fluorescence image according to a joint optimization objective function; upsampling the initial path parameters and mapping them to the image coordinates of the original fluorescence image to obtain original path parameters; solving for the target quenching field of the original fluorescence image based on the original path parameters, the initial quenching field parameters, and the joint optimization objective function; and correcting the original fluorescence image according to the target quenching field to obtain the target fluorescence image. This application solves the problem of distortion in the results of quantitative pathological analysis by using a multi-resolution joint optimization scheme, first solving the initial path and quenching field parameters at a coarse scale, and then mapping them to a high-resolution solution for the target quenching field, thereby improving the iteration convergence speed while balancing correction accuracy and pathological signal fidelity.

The application discloses a deep recursive medical image fusion method based on self-supervised generative distillation, comprising three modules of a teacher network, a student network and a fusion network; the teacher network module comprises a feature extraction stage and a feature fusion stage; self-supervised training is carried out to generate a reconstructed image to provide guidance for the student network; the student network comprises a feature extraction stage and a feature fusion stage, implicit neural is used to obtain prior knowledge from a source image, and generative distillation is combined to improve the efficiency of knowledge transfer; the fusion network integrates the specific features learned by the student network in multiple modes, and finally outputs a fusion image; the teacher network of the application adopts a dense multi-resolution preservation strategy, so that the semantic information transmitted to the fusion image is more rich, the local interaction of the cross-space feature block is refined, and a high-quality fusion image is realized.

The application provides a weather forecast optimization method and system based on complex terrain, and relates to the technical field of weather forecast.The method comprises the following steps: collecting historical meteorological data by using a meteorological data acquisition device, establishing a terrain influence evaluation model, then obtaining N terrain influence factors, combining real-time meteorological data to construct a multi-level multi-resolution weather forecast model, and then generating weather forecast information through the weather forecast model.The application mainly solves the problem that the traditional weather forecast method ignores the specific influence of terrain on weather, especially in complex terrain areas, which can cause large errors.Through analyzing a large amount of geographic information data and combining the data with a weather forecast model, the technical effect of improving the accuracy of the forecast is achieved.

This invention relates to the field of point cloud component segmentation technology, specifically to a general lightweight component segmentation method for dense point clouds, comprising the following steps: Step 1: Let the returned point cloud scan data be P = {p i =(x i ,y i ,z i Step 1: Perform centering and scaling operations on point cloud Q to obtain point cloud Q; Step 2: Perform adaptive multi-scale voxel mesh downsampling on point cloud Q; Step 3: Perform data augmentation operations on point cloud D, performing random scaling, random translation, slight rotation, and random noise addition to obtain point cloud DP; Step 4: Perform component segmentation on point cloud DP. This invention maximizes the preservation of local features through adaptive multi-scale voxel mesh downsampling, introduces a multi-resolution domain feature extraction module, and incorporates feature Euclidean distance weights, effectively improving the component segmentation accuracy in dense point cloud domains and significantly reducing the requirements for GPU memory and computing power.

The application discloses a distributed adaptive framing output method based on large-range remote sensing images, and relates to the technical field of remote sensing image processing. The preprocessed large-range remote sensing images are stored in a COG format, the framing size and quantity are dynamically adjusted in combination with the target area range and resolution, the tasks are distributed by a data locality perception scheduler after Z-order curve sorting, the computing nodes load the framing data with a buffer area, the framing data is interpreted by a random forest, and the buffer area is cropped, finally, the virtual data set is aggregated, efficient distributed parallel processing of the large-range images is realized, the I / O bottleneck and the computing resource limitation are effectively relieved, the context missing problem on the framing boundary is solved, the intelligent interpretation efficiency and precision are improved, and the multi-resolution on-demand rendering and elastic expansion are supported to adapt to the data scale growth.

The invention discloses an infrared weak and small target detection method based on a multi-stage series full-resolution fusion network, and belongs to the technical field of target tracking. A decoder of a multi-resolution dense interaction U-shaped module receives a feature map subjected to interactive attention fusion as input; the multi-resolution dense interaction U-shaped modules are connected in series to form a cascaded full-resolution fusion U-shaped network; the same-level feature maps of the preorder module are spliced by the execution channels and then are sent to the encoder of the next-level module, and are fused by using interactive attention and then are sent to the decoder of the next-level module; the encoder of each sub-network is normalized to obtain a weak and small target segmentation prediction map, SoftIoU loss is calculated with a true value, and model parameters are optimized; the interactive attention takes a plurality of to-be-fused feature maps as input, respective corresponding attention heat maps are calculated, and the to-be-fused feature maps are multiplied by the corresponding attention heat maps and then added. According to the method, the problem that the weak and small target in the infrared image is relatively weak in gray feature and texture feature and is easy to submerge is solved.

The application relates to the field of computer vision, and particularly discloses an image segmentation method and system, wherein a boundary distribution map is predicted by a multi-scale boundary distribution generation module as prior knowledge, a space-frequency domain joint supervision strategy is adopted, the boundary distribution map simultaneously approximates a real boundary in a spatial position and a spectral structure, false edge responses caused by local high-frequency interference are effectively inhibited, and the continuity and stability of the boundary are improved; an effective receptive field is expanded by a state space semantic enhancement module, and global semantic consistency is enhanced; a boundary-guided cross-scale decoder is combined with a boundary feature enhancement module, boundary-related responses are strengthened in multi-resolution fusion, and collaborative optimization of region positioning and boundary refinement is realized, so that accurate and complete polyp segmentation results are obtained under complex imaging conditions such as low contrast, various morphologies, and interference such as reflection and wrinkles, and reasonable calculation overhead is maintained.

This invention relates to a trajectory recognition method based on distributed acoustic sensing, belonging to the field of vehicle trajectory recognition technology. The trajectory recognition method based on distributed acoustic sensing utilizes a distributed acoustic sensing system to achieve high-precision quantitative measurement of external mechanical disturbances by demodulating the phase change of Rayleigh backscattered light. Its unique feature lies in that the trajectory recognition method includes the following steps: Step 1, acquiring low SNR vibration signals collected by the DAS system within a unit acquisition time; Step 2, in VTR Net, employing the Global-Local Feature Collaborative Modeling Mechanism (GL-FCM) and the Multi-Resolution Attention Feature Map Recovery Architecture (AG-FMR); Step 3, using the CEH-FA model robustness enhancement method for vehicle trajectory recognition to identify and finally output the vehicle trajectory.

This invention relates to the field of digital image processing technology, and particularly to a method and system for encrypting watermarks in 2D digital human rendering. The method includes acquiring encrypted watermark information; acquiring key feature blocks of a digital human rendered image; decomposing the key feature blocks into low-frequency and high-frequency layers using a multi-resolution transformation method; and allocating and embedding the encrypted watermark information into the low-frequency and high-frequency layers of the key feature blocks. This invention combines illumination analysis, complex texture analysis, and dynamic change recognition to effectively extract key feature blocks from digital human rendered images, including facial, limb, and clothing detail areas. This ensures that the watermark embedding can be accurately positioned in important visual areas of the image. By using a multi-resolution transformation method to decompose the key feature blocks into low-frequency and high-frequency layers and rationally allocating the encrypted watermark information to each layer according to sequential intervals, the watermark achieves good stability in the low-frequency layer.

This invention discloses a method and system for predicting the remaining life of equipment by integrating mechanistic information and degradation perception, relating to the field of big data technology. The method includes: extracting electrochemical and physical features from the raw operating data of the equipment under test to obtain multiple interpretable features; constructing an equipment remaining life prediction model based on a degradation stage perception soft-gating mechanism, an encoder with multi-resolution Mamba units, and a projection layer; training the equipment remaining life prediction model using machine learning algorithms based on preset historical equipment data to obtain a trained equipment remaining life prediction model; and using the multiple interpretable features as input, predicting the remaining life of the equipment under test using the trained equipment remaining life prediction model to obtain the prediction result. This invention alleviates the technical problem of the difficulty in accurately predicting the remaining life of equipment in existing technologies.

This application relates to a method and apparatus for segmenting road target instances in remote sensing images. The method includes: acquiring remote sensing images of the target to be detected by a remote sensing satellite; preprocessing the remote sensing images through an input module; extracting features from the preprocessed image using a backbone network; fusing the extracted multi-resolution features using a dual-stream feature pyramid in the neck region; dividing the head region into two branches: a detection head and a prototype; and performing category prediction and instance segmentation refinement tasks on the image entering the head region; outputting the high-resolution remote sensing image road target instance segmentation result; using a DMH-YOLO model to perform instance segmentation on the target remote sensing image to determine the target objects in the target remote sensing image; and using the Darknet-53 architecture framework in the backbone network and incorporating a C2f module to implement a residual learning mechanism. This invention can accurately identify and segment small target objects.

A data processing method suitable for multi-resolution particle methods includes the following steps: reading initial data and establishing an original model array to start calculation; detecting and judging all particles separately to determine whether particles split or merge, with the judgment condition being whether particles of different sizes flow across precisions; marking and recording the corresponding number of particles that split or merge; constructing two intermediate arrays to store particle information after splitting and merging; updating particle information in conjunction with time nodes, and updating the final particle data and corresponding particle information to the original model array for subsequent calculations, thereby improving overall computational efficiency.