22results about How to "Achieve reconstruction" patented technology

This invention relates to the field of multimodal sentiment analysis technology, and in particular to a method and system for multimodal sentiment analysis based on sequential translation under uncertain and missing modalities. The method includes: acquiring multimodal sentiment data, including visual, textual, and auditory modalities; constructing a multimodal sentiment analysis model based on sequential translation, firstly by extracting self-attention encoding features of each modality using a Transformer encoder; and then designing a bidirectional sequential translation architecture centered on the textual modality. This invention improves the utilization of textual modal features in the feature fusion process, thereby enhancing the quality of multimodal fusion features and providing strong support for subsequent complex sentiment analysis.

The application provides a social network opinion reconstruction method and device based on graph movement smoothing, wherein the method comprises the following steps: determining a sampling user set according to a reverse greedy sampling algorithm based on a pre-constructed directed social network model; determining graph movement smoothing degrees corresponding to all users in the directed social network model; obtaining opinion observation values corresponding to each user in the sampling user set; and reconstructing social network opinions of the users in the directed social network model based on the graph movement smoothing degrees and the opinion observation values. The method is aimed at a directed social network structure, and the sampling user set is determined through the reverse greedy sampling algorithm. The method can not only effectively improve the accuracy of social network opinion reconstruction, overcome the defects that the existing user opinion reconstruction method is mainly aimed at a non-directed network structure and cannot be directly extended to a directed network structure and the reconstruction mean square error is large, and realize accurate reconstruction of social network opinions, but also effectively improve the running efficiency of the reconstruction process.

The application provides a physical constraint-based illumination adaptive spectral reflectance reconstruction system, an illumination adaptive spectral reflectance reconstruction framework (I-SRR) based on no known illumination, auxiliary hardware or multiple captures, which can accurately recover intrinsic spectral reflectance of a material from a single RGB image; a spectral imaging physical model is embedded into a deep unfolding framework, a differentiable optimization structure is constructed based on an alternating direction multiplier method (ADMM), a spectral formation and recovery process is implicitly integrated, a physically guided end-to-end learning is realized through deep unfolding optimization based on physical constraints; the application can realize robust reflectance recovery under unknown or complex illumination conditions, adaptively adjust features through a learned compensation mechanism, maintain consistent reconstruction accuracy in different color temperatures (4000K-11200K) and indoor and outdoor scenes, and realize robust reconstruction under dynamic illumination conditions.

The invention discloses an imaging method for passing through a random scattering medium in visible light based on a diffractive optical neural network, and the method comprises the following steps: 1, carrying out the optical coding of an input image through a coherent or low-coherent monochromatic plane wave, and enabling a coded light field to pass through a random phase diffuser for scattering; and step 2, continuous wavefront modulation of the scattered light field is obtained through a coherent or incoherent diffractive optical neural network, and finally, high-fidelity original image reconstruction is generated on an output plane. The invention finds that the coherent diffraction neural network obtains dynamic phase modulation by introducing a randomly generated diffuser in the training process, enhances the adaptive capacity of the network to the spatial coherence change of the light source, realizes the high-quality reconstruction of the scattering image under the visible light, has the robustness to the low-coherence light source, and is suitable for large-scale popularization and application. The method can be applied to low-power-consumption and real-time imaging of the dynamic scattering medium under natural light.

This invention discloses an underwater target reflectance inversion method based on an underwater transmission model, comprising: acquiring the original digital values ​​of the underwater target collected by a hyperspectral camera; combining the built-in radiometric calibration parameters of the hyperspectral camera; converting the original digital values ​​into radiance based on the sensor pre-calibration response characteristics; physically modeling the water absorption effect, scattering effect, and path radiation effect during underwater light propagation to determine the linear mapping relationship between radiance and target reflectance; constructing a training radiative transmission model using a non-negative linear regression model; using a calibrated target with known reflectance as a sample to train the radiative transmission model between the original digital values ​​and the target reflectance, and solving for the mapping parameters between the two; and based on the original digital values ​​and mapping parameters of the underwater target to be reconstructed, using the trained radiative transmission model to invert and obtain the true reflectance of the target, thus completing the spectral reconstruction.

The application discloses a kind of variable-thickness curved surface weaving ceramic matrix composite three-dimensional reconstruction methods, belong to the field of microstructure identification and reconstruction.The application can quickly and accurately calculate the area of each cross section of the composite material by identifying the outer edge of the curved surface weaving composite material combined with the area of a single pixel;Meanwhile, by deleting the surface fibers arranged complexly, the classification accuracy of internal warp and weft can be improved, and it is more effective for subsequent reduction of human participation;By processing the three views of XCT data of the composite material respectively, the pore position can be accurately extracted from the original XCT image, the fiber bundles in different slices are numbered from the overhead view and the side view through the number of fiber columns and the spatial position relationship, and the spatial paths of warp yarns and weft yarns are respectively fitted, the fiber bundle cross section is represented by oblique elliptic function, the three-dimensional model reconstruction of curved surface weaving CMC can be quickly realized, the algorithm is practical, and the whole process is simple to operate.

The invention discloses a reconstruction method and a reconstruction device of a space protein expression map and a storage medium. Comprising the following steps: receiving a multi-channel high-resolution tissue image of a tissue region obtained based on tissue slices; receiving a polymerization true value of the expression quantity of each target protein on each first-direction tissue band and each second-direction tissue band, wherein the polymerization true value is obtained after the first-direction tissue band and the second-direction tissue band are cut on the first tissue slice and the second tissue slice; a deep learning network is trained based on the aggregation true values of the protein expression quantities on the tissue strips in the first direction and the second direction and the multi-channel high-resolution tissue image, and the trained deep learning network predicts the expression quantity of each target protein on each image unit based on the multi-channel high-resolution tissue image; and reconstructing a two-dimensional protein expression map. According to the method, the two-dimensional space protein map can be reconstructed with high precision only by using as few as two tissue slices and the aggregation value of the protein expression quantity on the tissue strip which is relatively easy to obtain without the protein expression quantity of each unit.

The invention relates to a real-time interactive hybrid ocean simulation method and simulation system based on unified spectrum control, and the method comprises the following steps: obtaining environment wind field parameters, and generating a unified ocean wave spectrum; based on the unified sea wave spectrum, generating a JONSWAP spectrum model by adopting fast Fourier transform, and simulating a global wave field by using the JONSWAP spectrum model; and determining the position of an interactive object, dynamically injecting wave particles at the boundary of a local area corresponding to the position based on the unified sea wave spectrum, generating near-field fluctuation, and superposing the near-field fluctuation to a global wave field to form a continuous mixed sea surface. Compared with the prior art, the method has the advantages of uniform spectrum control, efficient parallel computing, high physical consistency, high expandability and the like, and uniform simulation of the large-scale continuous sea surface and the local interaction area is realized.

This invention discloses a method and system for suppressing distributed cooperative interference on a motion platform. The method includes the following steps: Step S1. A transmitting node generates and transmits a desired signal; Step S2. A cooperative interference node generates and transmits a distributed cooperative interference signal; Step S3. An authorized receiving node reconstructs the distributed cooperative interference signal using the received signal; Step S4. The authorized receiving node defines the distributed cooperative interference signal estimation error, obtains the optimization objective using the root mean square value of the estimation error, and calculates the partial derivatives of the channel estimation parameters and frequency offset based on the optimization objective; Step S5. The authorized receiving node iteratively solves the channel estimation parameters and frequency offset using the LMS algorithm; Step S6. The authorized receiving node reconstructs the distributed cooperative interference signal, performs distributed cooperative interference suppression, and obtains the desired signal. This invention estimates the channel coefficients and frequency offset using parameter recursion, achieving distributed cooperative interference signal reconstruction and interference suppression in motion scenarios.

The application discloses a three-dimensional hair reconstruction method and system, an electronic device and a storage medium, relates to the technical field of computer graphics and computer vision, and the method comprises the following steps: acquiring an RGB portrait to be reconstructed; inputting the RGB portrait to be reconstructed into a three-dimensional voxel hair reconstruction model to obtain target three-dimensional voxel hair; the three-dimensional voxel hair reconstruction model is trained by using a training portrait set, a three-dimensional hair training set and corresponding text; the reconstruction network comprises an image encoder and a trained voxel decoder, the trained voxel decoder is determined according to a trained VAE-Patch_SNGAN network, and the trained VAE-Patch_SNGAN network is used for training the VAE-Patch_SNGAN network by using the three-dimensional hair training set; and a target three-dimensional hair image is reconstructed according to the target three-dimensional voxel hair. The application improves the reconstruction speed and accuracy of the three-dimensional hair.

This application provides a cargo loading control system, method, storage medium, and loading equipment based on a diffusion model, belonging to the field of automated cargo loading and unloading technology. It includes: a lidar for collecting three-dimensional point cloud data inside the cargo compartment; and an onboard industrial control computer, communicatively connected to the lidar and configured to perform the following steps: acquiring the three-dimensional point cloud data collected by the lidar; performing structured enhancement and completion on the three-dimensional point cloud data based on a preset point cloud enhancement diffusion model to obtain target point cloud data; generating multiple candidate cargo loading trajectories based on the target point cloud data, a virtual centerline, loading / unloading point positions, the attitude information of the loading equipment, and a preset trajectory diffusion model; selecting the target cargo loading trajectory from the multiple candidate trajectories and controlling the loading equipment to drive along the target cargo loading trajectory into the loading / unloading point position. The cargo loading control system based on the diffusion model of this invention can solve the problem of unstable recognition of the cargo compartment entrance by the loading equipment.

The invention discloses a DRF-Net-based small target detection method under a view angle of an unmanned aerial vehicle. Constructing an unmanned aerial vehicle aerial photography small target detection data set; a frequency domain guide feature enhancement module FCSA is introduced in the shallow layer stage of the backbone network; a multi-frequency reconstruction module MFRB is introduced in the high-level stage; constructing DRF-Neck in a neck structure, generating a saliency graph S based on MFRB output features, and introducing saliency-driven deformable convolution SAR-DCN, content aware recombination S-CARAFE and a dynamic feedback controller DFC to form a closed-loop optimization mechanism of saliency generation-self-routing convolution-dynamic feedback; and finally, the features are input into a decoupling detection head, network parameters are optimized through joint optimization of cross entropy loss and AHIoU loss, and a small target detection result in an unmanned aerial vehicle aerial photographing scene is output by using non-maximum suppression in a reasoning stage. Compared with the prior art, the method has the advantages that the detection precision on the unmanned aerial vehicle aerial small target data set is improved, and the method can be applied to the fields of unmanned aerial vehicle inspection, urban security monitoring, traffic target detection and the like.

This invention discloses a method for recovering intra-pulse relay interference signals based on a multi-residual codec network, relating to the field of radar technology. The method includes: receiving the interfered signal and preprocessing it to obtain input data; inputting the input data into a pre-trained multi-residual codec network, whereby the multi-residual codec network uses the encoder to extract features from the input data, compresses the feature dimensionality, increases the number of channels, and further uses a decoder connected to the encoder to perform signal recovery and dimensionality adjustment based on the input data to obtain the recovered signal. This method can reconstruct target echo signals in complex interference environments without relying on human experience, and exhibits good reconstruction performance under different types of relay interference.

This invention discloses a method for constructing a seamless land-water DEM based on multi-source satellite data fusion. The study area is divided into bare land, drawdown zone, and perennial waterlogged area based on the annual average water frequency of Sentinel-2 imagery. Using the ICESat-2 ATL08 ground point cloud during the dry season and contemporaneous Sentinel-2 imagery, the underwater topography of the drawdown zone is constructed using an XGBoost model. Using the ATL13 water surface elevation and depth during the wet season and contemporaneous Sentinel-2 imagery, the underwater topography of the perennial waterlogged area is constructed. Based on the water frequency, the results from the drawdown zone and the perennial waterlogged area are fused to generate a complete and continuous underwater topography. The elevation of the bare land ALOS DEM is then corrected using the ATL08 point cloud and fused with the topography of the drawdown zone and the perennial waterlogged area, achieving a seamless connection between land and water topography and constructing a high-precision integrated land-water DEM, thus improving the accuracy of the water area and shoreline topography.

This application discloses a three-dimensional spatial PM2.5 concentration estimation method, device, medium, and product, relating to the field of atmospheric environmental monitoring. The method includes: stitching together haze images from multiple perspectives within a sample area to obtain a three-dimensional spatial sample image; the three-dimensional spatial sample image and the corresponding actual PM2.5 concentration value constitute a training sample; inputting multiple training samples into an improved VIT model for training to obtain a PM2.5 concentration prediction model; the improved VIT model includes a multi-scale image feature embedding module and a Transformer encoder module arranged sequentially; inputting the three-dimensional spatial sample image corresponding to the area to be detected into the PM2.5 concentration prediction model to obtain the corresponding estimated PM2.5 concentration value. This application can obtain accurate estimated PM2.5 concentration values.

This application discloses a method, apparatus, and angle-resolved system for processing dispersive band diagrams. The method includes: performing nonlinear calibration on the image to be processed; detecting the number of image stitching segments; aligning the image stitching segments; and stitching and reconstructing the image stitching segments to obtain a complete image. The method, apparatus, and angle-resolved system disclosed in this application are compatible with dispersive band diagram testing of both photonic crystal structures without active regions and photonic crystal structures containing active regions. Through nonlinear calibration processing, measurement errors in large-angle regions are significantly reduced. Through the detection and alignment of the number of image stitching segments, image stitching errors are reduced. The method enables the stitching and reconstruction of multi-segment band structure test data, improving test accuracy.

This invention discloses a binocular reconstruction measurement method for the three-dimensional combustion surface of solid propellants: A binocular vision imaging system with a beam-splitting structure is established, and binocular images of the combustion surface during the solid propellant combustion process are acquired; the combustion surface region of the binocular image is extracted as a region of interest (ROI) image and enhanced; feature point identification and matching of the solid propellant combustion surface are performed on the enhanced ROI image to obtain matched feature points; a three-dimensional combustion surface point cloud is reconstructed based on the matched feature points obtained in step 3 using triangulation; the reconstructed three-dimensional surface point cloud is post-processed to obtain a three-dimensional combustion surface mesh of the solid propellant, thus obtaining the three-dimensional morphology information of the solid propellant combustion surface. This invention also provides a device for the binocular reconstruction measurement method of the three-dimensional combustion surface of solid propellants. This method and device can realize the reconstruction of the three-dimensional combustion surface of solid propellants, obtaining visualization of the three-dimensional combustion surface and three-dimensional morphology measurement results.

This invention provides a method, apparatus, electronic device, and storage medium for Gaussian splash reconstruction of 3D scenes based on virtual view completion. The method involves acquiring a context view image, processing the context view image based on a context view Gaussian prediction branch to predict camera parameters and a pixel-aligned context view 3D Gaussian primitive set; sampling a virtual viewpoint and generating a virtual view 3D Gaussian primitive set corresponding to the virtual viewpoint based on the virtual viewpoint Gaussian prediction branch; and fusing the context view 3D Gaussian primitive set with the virtual view 3D Gaussian primitive set to obtain a unified 3D Gaussian scene representation, thus completing the 3D scene reconstruction. This invention avoids the reconstruction quality degradation problem caused by dragging Gaussian primitives to cover occluded areas in pixel alignment methods by specifically constructing virtual view Gaussian primitives for occluded regions, achieving high-quality sparse view 3D scene reconstruction.

The application discloses a mechanical seabed garbage intelligent cleaning robot and a seabed garbage cleaning method.The robot adopts a modular integrated design and comprises a main frame, a buoyancy material, a track chassis, a material cabin, a mechanical arm, an electric control cabin, a conveying belt, an underwater sensing system, a water quality monitoring module and an optoelectrical composite cable.A mechanical grabbing-transporting-storing integrated operation chain is used to realize the garbage cleaning concept of accurate grabbing and minimum disturbance.The mechanical arm has multiple degrees of freedom and a grabbing and shearing integrated function.The track chassis is adapted to complex seabed topography.The double-layer material cabin realizes efficient garbage storage and rapid unloading, and the multi-modal sensing system supports intelligent autonomous operation.The application greatly improves the seabed garbage cleaning efficiency, significantly reduces the disturbance to the seabed ecology and is suitable for large-scale garbage cleaning operation in different water depths and various bottom environments.

The application belongs to the technical field of historical building repair, and discloses a damaged historical building decorative component original appearance reasoning reconstruction method and system. The method comprises the following steps: S1, constructing a database of texture symbols and connection state attributes of the connection part of the historical building decorative component; S2, fitting the shape contour of the damaged part, and constructing the connection state attributes of the damaged part; S3, comparing the obtained connection state attributes with the database, obtaining the texture symbol matched with the connection state attributes, constructing the point cloud model of the matched texture symbol, registering the point cloud model with the point cloud model of the historical building decorative component to be repaired, and obtaining the complete point cloud model after repair; and S4, using the complete point cloud model to perform surface reconstruction on the historical building decorative component to be repaired, so as to realize the original appearance reasoning reconstruction of the historical building decorative component to be repaired. Through the application, the problem that the original appearance of the damaged component is difficult to reason and repair in the prior art is solved.

The invention discloses an ultrasonic image defect segmentation method based on a memory bank-patch denoising diffusion probability model, and belongs to the field of ultrasonic nondestructive testing. The method is an ultrasonic image unsupervised automatic segmentation method which is based on a memory bank and does not need labels, a defect-free image is adopted to train an unsupervised anomaly detection model, and the trained model is used to carry out automatic defect segmentation on a defect-containing image. According to the method, the anomaly detection capability is improved through the memory bank module, automatic segmentation of ultrasonic image defects can be realized on the premise that data does not need to be labeled, and excellent segmentation performance is obtained. According to the method, the dependence on a large amount of manual annotation data can be effectively relieved, and a more accurate solution is provided for automatic segmentation of ultrasonic image defects.

The invention relates to the technical field of genital tract repair instruments, and discloses an integrated collagen degradable stent for repairing fine parts of fallopian tubes, which comprises an isthmus adaptive section, an interstitial section and a uterus end adaptive section which are sequentially connected into a whole and are coaxial, and meanwhile, the isthmus adaptive section, the interstitial section and the uterus end adaptive section form a through gradually-changed drainage channel; isthmus anti-slip anchoring protrusions and uterus end anti-slip anchoring protrusions are arranged on the outer surface of the isthmus adaptation section and the outer surface of the uterus end adaptation section correspondingly, and a double-end anchoring structure is formed; the isthmus adaptive section is of a tubular structure, and the inner wall of the isthmus adaptive section is a roughened surface; the interstitial section is of a conical structure, and a plurality of through holes are formed in the surface. The uterus end adaptation section is configured to be of a double-layer structure with a horn-shaped opening, the outer layer is a compact supporting layer, and the inner layer is a mucous membrane induction layer. The integrated collagen degradable stent is good in size adaptability, stable in anchoring, accurate in positioning and safe in conveying, and tissue compression damage or displacement and falling caused by improper size are avoided.