53 results about "Feature estimation" patented technology

The invention discloses a target tracking method, device and equipment in a multi-target multi-source information fusion scene, and relates to the field of target tracking. The method comprises the following steps: acquiring information data of a to-be-tracked target group; performing parallel processing on the information data by adopting a DBSCAN algorithm and a neural network to obtain a track association result; the track association result comprises a track measurement fragment and a target to which the track measurement fragment belongs; segmenting the track measurement segment based on a preset sliding window and step length, and determining a track residual error according to sub-track information obtained through segmentation by adopting an error feature estimation algorithm; based on the trained MsRFCFN, determining a fusion residual error according to the track residual error; and performing track reconstruction and fusion splicing processing according to the fusion residual error to obtain a fusion track for realizing target tracking. The invention aims to improve the fusion precision so as to improve the target tracking precision.

The invention discloses a grassland above-ground biomass estimation method and system based on semi-variance function coupling grazing heterogeneity, and belongs to the technical field of grassland feature estimation, and the method comprises the steps: obtaining a fixed resolution canopy height model CHM through the processing of collected grassland point cloud data, obtaining CGF and LAI raster data through the CHM, and carrying out the calculation of the grassland above-ground biomass. Fitting a semi-variance function model to obtain spatial heterogeneity characteristics of the grassland, obtaining an optimal inversion scale, screening out key independent variables with stability, and providing a reliable basis for grassland AGB inversion; through combination of CHM, CGF and LAI of different spatial scales and ground actually measured biomass data, inversion is carried out by using a linear regression model, a random forest and a BP neural network to obtain an optimal variable combination and an inversion model, and finally spatial heterogeneity characteristics of gAGB of different spatial scales in different management modes are analyzed, management measures based on local conditions are formulated, and the method is suitable for mass production. The grassland resource utilization efficiency is improved; and the ecological system is protected.

An audio encoding method and apparatus, an audio decoding method and apparatus, an electronic device, a computer-readable storage medium, and a computer program product. The audio decoding method comprises: acquiring an audio code stream encapsulation, wherein the audio code stream encapsulation comprises an audio code stream, the audio code stream is obtained by performing audio encoding on an audio signal on the basis of a target encoding mode and a target code rate mode, the target encoding mode is acquired from among a plurality of candidate encoding modes, and the target code rate mode is acquired from among a plurality of candidate code rate modes; in response to a decoding request for the audio code stream encapsulation, acquiring the target encoding mode and the target code rate mode from a frame header comprised in the audio code stream encapsulation; performing signal decoding on the audio code stream on the basis of the target encoding mode and the target code rate mode, so as to obtain an encoding feature estimation value corresponding to the audio code stream; and performing reconstruction on the encoding feature estimation value on the basis of the target encoding mode, so as to obtain a reconstructed audio signal corresponding to the audio code stream.

The invention discloses an ISAR image ship coarseness identification method and system based on scattering feature estimation, and relates to the technical field of target identification, and the method comprises the steps: constructing and training a target identification model; the target recognition model comprises a small grazing angle image feature enhancement module based on scattering characteristic estimation, a feature extraction fusion module based on motion attitude decoupling and a full connection module; the small grazing angle image feature enhancement module comprises a scattering center parameter estimation module, a complex waveform feature extraction module and a physical data feature fusion module; the feature extraction fusion module comprises a spatial frequency domain attitude mode representation module, an attitude adaptive deformation convolution module and a frequency domain space double-attention fusion module; obtaining an ISAR image of a to-be-identified marine ship; and inputting a to-be-identified marine ship ISAR image into the trained target identification model for ship type identification. According to the invention, the identification precision of the ship type under the conditions of small grazing angle and complex motion imaging is improved.

The invention relates to the technical field of image enhancement or restoration, in particular to an image restoration method based on diffusion mode and prior feature generation and a related device. In order to solve the technical problem of poor image restoration quality in the prior art, real data is utilized to train a priori feature restoration image model composed of a priori feature extraction network, a U-net-based mask combination restoration network and a mask restoration network; then, constructing a priori feature repairing and enhancing image model based on a training priori feature extraction network, an image noise adding network, an image denoising network, a priori feature estimation network and a training mask repairing network; a training image noise adding network, a training image denoising network, a training prior feature estimation network and a training mask repairing network are extracted to form an image repairing model, the model is used for processing a to-be-repaired image, and efficient and high-quality image repairing is carried out on a specific area.

The invention belongs to the technical field of feature recognition of underwater complex targets, and relates to a boundary element-based multi-bright-spot cluster target interference feature estimation method, which comprises the steps of geometric model construction and introduction, numerical calculation based on multi-physics field coupling, echo intensity representation by adopting an angle-frequency spectrum, and analysis of resonance generated by multi-body interference. According to the method, a numerical model for solving a three-dimensional double-gas-cylinder omnidirectional scattering sound field is established based on a boundary element method, and omnidirectional and broadband scattering characteristics of a target in horizontal and vertical attitudes are calculated; a multi-bright-spot dual-target strong scattering coupling mechanism is mainly analyzed, and a multi-body configuration and single target scattering interference fringe generation mechanism is disclosed; according to the method, the inherent attribute characteristics of the single target can still be effectively identified in a cluster target scattering field; the mapping relation among the geometric configuration, the spatial arrangement mode and the frequency response is obtained, and a theoretical formula suitable for double-target sound scattering forecast is deduced.

The invention provides an audio coding method and device, an audio decoding method and device, electronic equipment, a computer readable storage medium and a computer program product. The audio decoding method comprises the following steps: in response to a decoding request for audio code stream packaging, obtaining a target coding mode and a target code rate mode from a frame header included in the audio code stream packaging; wherein an audio code stream included in the audio code stream package is obtained by performing audio coding on the audio signal through a target coding mode and a target code rate mode, the target coding mode is obtained from a plurality of coding modes, and the target code rate mode is obtained from a plurality of code rate modes; performing signal decoding processing on the audio code stream through the target coding mode and the target code rate mode to obtain a coding feature estimation value corresponding to the audio code stream; and through the target coding mode, performing reconstruction processing on the coding feature estimation value corresponding to the audio code stream to obtain a synthetic audio signal corresponding to the audio code stream.

The invention relates to the technical field of wave feature estimation, in particular to a wave compensation-oriented wave feature estimation method, which comprises the following steps of: acquiring ship motion state data in real time by using a ship-borne inclination sensor, a gyroscope and an acceleration sensor, and carrying out data preprocessing through deep noise feature analysis and an accurate Kalman filtering algorithm; and noise interference and measurement errors are eliminated. On the basis of ship kinematics and dynamics principles, a wave feature estimation model is constructed in combination with machine learning algorithms such as a convolutional neural network and a long-short-term memory network, and a large amount of different sea condition data is used for training. Preprocessed data are input into the model, characteristic parameters such as the wave height, the wavelength, the period and the propagation direction of waves are rapidly and accurately calculated, and reliable information is provided for a wave compensation platform control system. Compared with a traditional method, the method is low in cost, high in real-time performance and high in accuracy, the control precision of the wave compensation platform is effectively improved, and the working capacity of a ship in a complex marine environment is enhanced.

An acoustic feature estimation method is an acoustic feature estimation method for estimating an acoustic feature of a space and includes acquiring data on the space, estimating situations in the space in accordance with the acquired data, correcting a provisional value of the acoustic feature in accordance with the estimated circumstances, and outputting the corrected provisional value.

Identify an object in various environments. 【Solution means】Communication terminal ( 2 ) is From the color image data using the first estimation model (30) 、 color image Colors that indicate the amount of facial features included a color feature estimation unit (40) that estimates a feature vector; From depth image data using the second estimation model (31) 、 depth image Depth of facial features included a depth feature estimation unit (41) that estimates a feature vector; From the color feature vector and depth feature vector using the third estimation model (32) 、 a three-dimensional image including a color image and a depth image Three-dimensional representation of facial features included in the image. a three-dimensional feature estimation unit (42) that estimates a feature vector; Color characteristics feature vector, depth degree characteristic feature vector, and three-dimensional Gentoku and an authentication unit (24) that performs face authentication using any of the feature vectors.

The application provides a multi-line structured light three-dimensional reconstruction frame point cloud real-time de-stacking and fusion method, and relates to the technical field of line structured light three-dimensional reconstruction. The distribution characteristics of the spatial position of the point cloud are utilized, a global grid structure is established through the position of the multi-view frame point cloud during three-dimensional reconstruction, the corresponding domain point set of each frame point cloud is solved, and the normal feature estimation is performed according to the distribution characteristics of the neighborhood points. The overlapping condition of the point cloud is detected according to the global grid structure, the cylindrical neighborhood structure of the overlapping frame point cloud is established, and the point cloud fusion is realized through the weighted sum of the point cloud vector direction of the target point and the fusion neighborhood point set and the normal vector feature. The algorithm can meet the system stability requirement in the three-dimensional reconstruction, improves the quality of the final point cloud model, greatly improves the measurement adaptability of the multi-line structured light three-dimensional reconstruction in the complex environment, and effectively improves the multi-line structured light three-dimensional reconstruction precision.

The invention discloses an information transmission method and device based on adaptive coding. The method comprises the following steps: generating a detection information sequence; sending the detection information sequence to a receiving end from a sending end through a transmission channel, and receiving the detection information sequence by the receiving end to obtain a received information sequence; performing feature estimation processing on the detection information sequence and the receiving information sequence to obtain a channel estimation matrix, channel time-varying feature information and dimension information; and performing information transmission processing on sent information by using the channel estimation matrix, the channel time-varying feature information and the dimension information. According to the invention, for multiple types of wireless communication scenes with complex time varying, high-reliability data communication is realized.

The present disclosure provides a point cloud smoothing method and device based on local feature estimation, electronic equipment and storage medium. By introducing the noise point elimination of local density features, the normal estimation based on weighted local statistical analysis, and the neighbor update weight construction mechanism of fusing spatial distance, normal difference and normal distance, adaptive and structure-preserving smoothing processing of point cloud is realized. The method can effectively eliminate outliers and noise points, reduce the thickness of the point cloud, improve the stability of the normal estimation, and at the same time, preserve the object edges and geometric details in the smoothing process, significantly improving the overall quality and applicability of the point cloud.

The present disclosure presents a Multi-Input, Multi-Output artificial intelligence system that jointly processes video with other physiological signals during training to uncover shared representations across modalities and prevent the mislearning observed in existing methods. Once trained, the system requires video input at inference, yet remains capable of reconstructing physiological signals learned during training. By leveraging these crossmodal patterns, the system can use video alone to estimate complex signals such as blood pressure. This integration of video with complementary vital features during training therefore provides a pathway toward accurate, contactless estimation of a broad range of vital signs. A method for estimating vital features of a patient is provided. The method includes receiving a video segment showing a skin-exposed region of the patient, cropping the video segment, and estimating vital features by providing the crops as input to a neural network.

The application discloses an information transmission method and device based on adaptive coding, and the method comprises the following steps: generating a probe information sequence; sending the probe information sequence from a sending end to a receiving end through a transmission channel; receiving a receiving information sequence by the receiving end; performing feature estimation processing on the probe information sequence and the receiving information sequence to obtain a channel estimation matrix, channel time-varying feature information and dimension information; and performing information transmission processing on the sending information by using the channel estimation matrix, the channel time-varying feature information and the dimension information. The application realizes high-reliable data communication in complex time-varying multi-type wireless communication scenes.

The invention discloses an angle steel tower defect identification method based on a graph neural network. The method comprises the following steps: step 1, constructing an initial structure graph; 2, constructing visual structure joint features of dominant nodes to form a first node feature set; step 3, constructing label missing perceptual coding features of the hidden nodes to form a second node feature set; 4, performing feature estimation on the second node feature set by using a feature probability migration module to form a third node feature set; 5, correcting the third node feature set to obtain final hidden node features, and forming a fourth node feature set; 6, fusing the first node feature set and the fourth node feature set to generate a graph embedding vector set; step 7, performing defect type prediction on each node; and 8, outputting a defect identification result. According to the invention, the graph neural network and the feature probability migration module are combined to realize structural level defect intelligent identification of the unobservable area of the angle steel tower.

The present invention relates to the technical field of image enhancement or restoration, specifically an image restoration method and related device based on diffusion mode and prior features; to solve the technical problem of poor image restoration quality in the prior art, the present invention uses real data to train a prior feature restoration image model composed of a prior feature extraction network, a mask combination restoration network based on U-net, and a mask restoration network. Then, a prior feature restoration enhanced image model is constructed based on the training prior feature extraction network, image noise addition network, image denoising network, prior feature estimation network, and training mask restoration network. After training, an image restoration model composed of the training image noise addition network, training image denoising network, training prior feature estimation network, and training mask restoration network is extracted, and this model is used to process the image to be restored, performing efficient and high-quality image restoration for specific regions.

A method of audio signal processing, a hearing system and a hearing device. A primary input audio signal is obtained using an audio input unit of a hearing device. A secondary input audio signal is obtained using a second device of the hearing system. The secondary input audio signal is transmitted from the second device to the hearing device. A level feature is determined based on the primary and secondary input audio signals using a feature estimation unit of the hearing device. An output audio signal is obtained by an audio processing unit, wherein the level feature is used for steering at least one audio processing routine of the audio processing unit. The output audio signal is outputted by an audio output unit of the hearing device.

The application belongs to the technical field of polarization imaging, and particularly relates to a water image despeckling model construction method based on polarization imaging, which comprises the following steps: using a feature extraction encoder to map four polarization angle images and calculated intensity images to a low-dimensional latent feature space; introducing a diffusion model network, a forward process of which is used to gradually add Gaussian noise until the noise data conform to a standard Gaussian distribution, starting from clear latent features extracted based on a clear water target intensity image and a turbid water target polarization image; a reverse process of the diffusion model network is used to denoise the noise data under the guidance of polarization features of the turbid water target, generate clear latent features of the underwater target, and obtain clear latent feature estimation results; using a reconstruction module network, based on the estimation results and the intensity image of the underwater target in the corresponding sample, a corresponding clear underwater target image is generated. The application can adapt to scattering environments of water bodies with different turbidity, and has high generalization and high robustness.

The invention discloses a dual-channel signal denoising method using a noise correlation threshold, and belongs to the technical field of automation. The method comprises the following steps: framing a dual-channel signal, extracting a power spectrum and a cross-power spectrum by using fast Fourier transform, and updating frequency domain feature estimation through a dynamic smoothing coefficient; and deducing an NCT dynamic calculation formula in combination with the coherence function model and theoretical noise power, and generating a frequency adaptive NCT curve through processing. And mapping the frequency domain NCT into a time-frequency domain wavelet threshold parameter, and processing a signal by adopting multi-scale wavelet decomposition and a soft threshold function to realize frequency domain noise feature and time-frequency domain denoising processing closed-loop linkage. According to the method, through the steps of framing processing, frequency domain analysis, NCT calculation, wavelet denoising and the like, local mutation and detail information of signals can be reserved while noise is removed, calculation efficiency is improved, and denoising robustness is enhanced.

The present invention belongs to the technical field of integrated detection and communication, and specifically relates to a micro-motion target detection method based on OFDM sounding integrated signal. The present invention extracts the peak position sequence of broadband HRRP of micro-motion target, and roughly estimates the micro-motion information as prior information; then performs time-frequency analysis on narrowband echo to obtain the micro-Doppler information it carries, and performs extreme point screening on the time-frequency distribution result, and uses the dynamic programming Viterbi algorithm to search for the shortest extreme path at the cost of gradient change, and obtains the secondary estimation result after curve fitting, and performs joint feature estimation and extraction on the target micro-motion information, and more comprehensively utilizes the target micro-motion feature information carried in the echo, so as to obtain higher-precision target micro-motion parameters, and can realize parameter extraction and estimation of multiple micro-motion targets, while ensuring accuracy with smaller calculation amount.

This invention relates to the field of random sound source localization and feature estimation technology, specifically to a method for random sound source localization and feature estimation in layered media, comprising the following steps: establishing a random sound source model in the layered media; at each observation point in the upper layer of the layered media, for multiple different frequencies corresponding to wavenumbers, using sensors to perform multiple independent sound wave field measurements to obtain sound wave field measurement data; dividing the region where the random sound source is located into a two-dimensional uniform grid, and determining the initial values ​​of the numerical solutions for the mean and variance of the random sound source based on the grid points; based on the sound wave field measurement data, and based on the kernel matrix singular value decomposition, solving the random sound source model in the layered media, iteratively updating the numerical solutions for the mean and variance of the random sound source to obtain the final numerical solutions for the mean and variance; this invention can achieve stable and high-precision reconstruction of the spatial distribution characteristics of random sound sources in layered media.

The invention relates to the technical field of image processing, in particular to an underwater image enhancement method based on a color channel interactive attention mechanism. Comprising the following steps: performing color channel separation on an original underwater image to obtain R, G and B single-channel images; extracting sparse features of each channel through a sparse feature estimation module; realizing cross-channel feature interaction and information fusion by using a channel interaction attention module; fusion parameters are generated through an environment sensing module, and the interaction feature fusion effect is optimized; and reconstructing the fused features into corresponding channel images, and splicing to obtain an enhanced underwater image. According to the method, the color balance, the structure consistency and the detail retention capability of the underwater image are effectively improved, and the method is suitable for underwater vision application scenes such as underwater robots and ocean exploration.

Provided are a method and a system for comprising cross-modal similarity on the basis of contrastive learning. Unlike existing methods based on Center Loss, the method for comparing cross-modal similarity according to an embodiment of the present invention performs cross-modal similarity comparison which is relatively robust to changes in performance according to the amount of hyperparameters and learning data by applying a comparison method based on similarity-based reference contrastive learning between several modals, and performs feature estimation which is robust to various changes required in a real environment by applying a data augmentation technique. Accordingly, by performing the cross-similarity comparison between modals for objects expressed in various modals such as images, meshes, and point clouds, similarity comparison is performed for objects of various modals that are randomly acquired in the process of implementing and configuring a 3D virtual environment, thereby enabling extensive object recognition, classification, detection, and search.

The invention relates to the technical field of motor fault detection, discloses a motor turn-to-turn fault feature observation method based on an improved sliding-mode observer, and aims to solve the technical problem that the fault feature estimation precision is reduced due to the buffeting problem of a traditional sliding-mode observer (SMO). An improved fault characteristic observer is designed by taking a back electromotive force residual error under a two-phase static coordinate system as a fault characteristic quantity, the observer is optimized by adopting a linear superhelix algorithm (STA) and a hyperbolic tangent function (tanh), high-frequency buffeting is effectively inhibited, and the convergence of the observer is proved through a Lyapunov stability theory. According to the method, the counter electromotive force estimation precision can be remarkably improved, reliable input is provided for motor fault detection, and the method is suitable for online monitoring and diagnosis of PMSM stator turn-to-turn short circuit faults.

The invention discloses a modulation mode estimation enhancement method based on data set expansion. The method comprises the following steps: expanding a signal training data set by using a model data dual-drive mode; constructing a long-short memory network model oriented to background information learning, training and generating a network model on the expanded data set to extract background information of the signals, and realizing signal-to-noise ratio estimation; a signal modulation mode feature estimation network based on a deep residual shrinkage network is constructed, a dynamic channel threshold updating structure is formed by constructing a channel attention threshold function, applying an attention mechanism and introducing a residual shrinkage module unit, and estimation of signal sample modulation modes in different environments is achieved. According to the invention, the problem of inaccurate signal modulation mode estimation when the signal-to-noise ratio is low and the data set is less is effectively solved, and high-precision modulation mode estimation is realized.

The present application belongs to the technical field of anti-radiation target sorting, and particularly relates to an anti-radiation target sorting method based on a data stream clustering algorithm. The anti-radiation target sorting method comprises: direction finding, direction finding is performed on full pulses, and a target line-of-sight angle calculated by an anti-radiation seeker is taken as real-time angle data stream; signal feature estimation, the real-time angle data stream is clustered to obtain pulse clusters with similar spatial relationships, signal feature estimation and threat determination of electromagnetic targets are performed on the pulse clusters after clustering, and target sorting is realized. The target sorting method provided by the present application can obtain high-quality and stable target sorting results, so that the anti-radiation seeker has stronger adaptability to interference and deception in complex environments, thereby improving the combat effectiveness of anti-radiation weapons.

The invention relates to the technical field of computer vision, and provides a blurred image enhancement method based on feature estimation and multi-scale fusion detail enhancement and related equipment. The method comprises the following steps: step 1, acquiring a to-be-enhanced blurred image and preprocessing the to-be-enhanced blurred image to obtain a high-frequency information image; 2, inputting the blurred image and the high-frequency information image into a blurred image enhancement model to obtain an enhanced image of the blurred image; wherein the blurred image enhancement model comprises a feature estimation module and a multi-scale fusion detail enhancement module; the feature estimation module is used for estimating features of the blurred image and performing preliminary enhancement, and the multi-scale fusion detail enhancement module is used for fusing and further enhancing the preliminarily enhanced blurred image and the high-frequency information image to obtain a final enhanced image.

The invention discloses a laser welding system, a molten pool depth function obtaining method and a molten pool morphological characteristic estimation method. The molten pool morphological characteristic estimation method comprises the following steps: providing a molten pool depth function taking voltage as a variable; irradiating a workpiece group with a first laser beam to form a first molten pool; sensing a first coaxial reflected light reflected from the first molten pool, and correspondingly outputting a first voltage signal; and substituting the first voltage signal into the molten pool depth function to instantly estimate the depth of the first molten pool. Therefore, the welding yield can be improved, and the overall production cost and time can be reduced.