105 results about "Sparse regularization" patented technology

The invention belongs to the technical field of equipment integration, and discloses a multi-modal equipment integrated management system and method based on intelligent AI, and the system comprises an autonomous sensing module which carries out the real-time causal modeling of multi-modal data through embedded causal analysis; in combination with a self-supervision anomaly detection mechanism, outputting a device cause and effect graph and a data stream with a confidence label; the liquid feature factory module is used for extracting time sequence features by utilizing a gating time sequence convolutional network based on an equipment causal graph and a data stream with a confidence label, reconstructing a feature path through a dynamic feature molten pool mechanism, and performing compression selection on high-dimensional features in combination with a sparse regularization method so as to generate a feature tensor; and accurate perception, efficient decision making, accurate execution and intelligent optimization management of the equipment can be realized.

The invention discloses a sports video commentary generation method and system based on a multi-modal large language model, and the method comprises the steps: obtaining a multi-modal data set which comprises a sports video, and an audio and a commentary text corresponding to the sports video; constructing a multi-modal large language model, encoding the sports video, the audio and the explanation text so as to project corresponding video frames, audio waveforms and metadata to a shared embedding space, and determining a multi-modal embedding vector; a multi-modal clustering memory unit is set, multi-modal embedded vectors are grouped, and feature alignment between modals is optimized through comparative learning and information entropy regularization; based on a retrieval enhanced context learning mechanism, a historical instance is retrieved through sparse regularization distance measurement to serve as reference input of a current input multi-mode embedded vector; and jointly inputting the current multi-modal embedded vector and the reference input into the multi-modal large language model to obtain a sports video commentary. According to the method and the device, the problems of insufficient multi-modal information integration and insufficient context utilization are solved.

The invention provides a power big data-based sparsity feature adaptive compression method, which comprises the steps of adaptively optimizing a sensing matrix through a dictionary learning method aiming at a screened multi-granularity feature subset, introducing a sparse regularization item to enhance the sparsity of data reconstruction, and according to a missing mode of a missing value and context information, carrying out adaptive compression on the sparse feature of the power big data. Establishing a mapping relationship among the compression ratio, the reconstruction error and the feature subset, and determining a self-adaptive compression strategy; based on a self-adaptive compression strategy, a compressed sensing method is adopted to carry out self-adaptive compression on the multi-time-granularity power data, the compression ratio is dynamically adjusted according to reconstruction error feedback, the data compression and reconstruction quality is balanced, and the reconstructed multi-granularity power data is synchronized according to timestamp precision.

The invention discloses a deep learning assisted acceleration fracturing construction parameter intelligent real-time optimization method, and belongs to the field of unconventional oil and gas field development, and the method comprises the steps: constructing a structured sample database covering geological attributes, engineering parameters and fracturing effect responses, and carrying out the unified preprocessing of multi-source data; constructing a multi-modal space-time collaborative prediction model, and combining a multi-expansion time permissible convolutional network, a three-dimensional residual network and a cross attention mechanism; supervised learning training is carried out based on a database, sparse regularization and learning rate scheduling are introduced, and the model precision and generalization ability are improved; a hierarchical collaborative optimization framework of outer-layer Bayesian search-inner-layer CMA-ES refinement is provided, and reservoir transformation volume maximization and construction feasibility are achieved under complex constraints; compared with an existing method, optimization parameters of the method include the cluster distance and the section distance and further include the displacement, the sand concentration, the fracturing fluid type and other construction parameters, and real-time optimization of the fracturing construction parameters can be achieved through the machine learning agent model.

The invention relates to the technical field of electric power task adjustment, and provides an electric power task model fine tuning method and system, terminal equipment and a storage medium, and the method comprises the steps: inserting a shared low-rank expert module in a pre-training model trunk, dynamically calculating the activation weight of each expert according to the input characteristics through a gating network, and carrying out the fine tuning of an electric power task model; and a sparse regularization constraint activation number is introduced. During training, expert output is subjected to weighted fusion to generate fine adjustment parameters, multi-task joint training is adopted, balance optimization is achieved by dynamically adjusting task weights, and during reasoning, only part of experts with the highest weights are activated, and sparse calculation is achieved. According to the method, a plurality of low-rank expert modules are inserted into a large model trunk, a trainable gating routing mechanism is introduced, the optimal expert combination weight is dynamically calculated according to semantic features of input task samples, and knowledge fusion, parameter multiplexing and reasoning sparse activation among multiple tasks are achieved.

The invention discloses a sparse regularization direction of arrival estimation method based on a risk minimization principle, and belongs to the technical field of array signal processing and underwater acoustic signal processing. The method comprises the steps of receiving array signals and establishing an observation model; constructing a sparse representation and over-complete dictionary; establishing and initializing a regularization optimization model; carrying out adaptive weight updating and risk-driven parameter selection; after regularization parameters are determined, a fast iterative shrinkage threshold algorithm FISTA is adopted to carry out optimization solution, and dictionary refinement is carried out on the detected direction after each iteration convergence so as to reduce off-grid errors; and after a small amount of outer layer iteration is repeated, outputting a final DOA estimation result and corresponding power. According to the method, self-adaptive selection of regularization parameters and noise levels can be realized, and the problem of precision degradation under complex conditions of low signal-to-noise ratio, limited snapshot number, signal source correlation, power imbalance and the like is effectively solved without manual parameter adjustment, so that the robustness and practicability of estimation are remarkably improved.

The invention provides a marine visibility observation data quality optimization method based on multiple elements, and belongs to the technical field of marine visibility observation.The marine visibility observation data quality optimization method comprises the steps that abnormal values are screened and marked by establishing a dynamic threshold value judgment mechanism, and two-layer optimization processing is conducted through a second Lora fine adjustment model; a visibility deviation parameter and a confidence coefficient parameter after correction are calculated through a sparse regularization-based feature selection enhancement mechanism and meteorological physical constraint conditions, three-layer optimization processing is carried out by adopting a third Lora fine tuning model, and a residual error is finely corrected through a gradient acceleration algorithm based on momentum optimization. A multi-element time sequence fusion verification mechanism is established, a time sequence consistency index and a space continuity index are calculated, quality identification grading marking is carried out, and finally optimized seaborne visibility observation data and a quality control report are output. The technical problems that the precision of the marine visibility observation data is insufficient in a complex marine environment and an effective multi-element collaborative correction mechanism is lacked are solved.

The invention discloses an epilepsy abnormal brain network identification method based on a multi-scale static-dynamic fusion network, and belongs to the field of brain image analysis. The method comprises the following steps: firstly, constructing a static function connection weighted graph and a dynamic function connection graph; and fusing the static and dynamic representations by adopting a cross attention module. In order to describe a multi-scale spatial relationship, performing lexical meta-processing on brain connection according to anatomical partition and a functional network; and the local-global fusion module is used for integrating the fine granularity and the macroscopic relationship, so that the brain region with diagnostic significance is highlighted. In the training stage, cross entropy, reverse contrast loss and sparse regularization based on contrast graph adjacency matrix entropy are jointly used. The method is verified on multi-center functional magnetic resonance data, compared with other mainstream depth models, the classification accuracy, generalization and interpretability are remarkably improved, an abnormal brain region consistent with an epilepsy network can be positioned, and brain image markers with biological significance can be connected and recognized.

The invention provides a prostate cancer focus accurate positioning method based on a fusion navigation technology, and relates to the technical field of data processing, and the method comprises the steps: collecting a plurality of prostate images; selecting a control point from the prostate image in combination with a sparse regularization algorithm; determining an optimal deformation field according to the control points, and performing elastic registration on the prostate image based on the optimal deformation field; differential multi-resolution fusion is carried out on the registered prostate images, and a prostate three-dimensional model is reconstructed; constructing a lesion dynamic extraction model comprising a three-dimensional convolutional network, a pooling layer, an interested layer and a bounding box regression layer which are connected in sequence; a loss function of the focus dynamic extraction model is established, and the loss function comprises an intersection-to-union ratio loss item and a distributed focus loss item; under the supervision of the loss function, training the focus dynamic extraction model; and positioning a lesion point in the prostate three-dimensional model through the trained lesion dynamic extraction model. And the focus point positioning speed and accuracy are improved.

The invention relates to a transformer abnormal sound source positioning method and system, and belongs to the technical field of power equipment state evaluation, and the method comprises the steps: constructing a Bayesian neural network embedded with a voiceprint physical mechanism, and enabling a sound wave propagation equation to serve as a physical constraint to be embedded into the Bayesian neural network; reconstructing the voiceprint signal by adopting a compressed sensing technology to obtain a reconstructed voiceprint field; designing a multi-task objective function including data fitting, physical constraint and positioning loss, and optimizing data fitting, physical constraint and positioning precision to obtain a trained Bayesian neural network; based on a gradient sound source inversion positioning algorithm and the trained Bayesian neural network, sparse regularization is combined to obtain a prediction result of accurate positioning; and a prediction result is visualized to a three-dimensional model of the transformer, and the position of an abnormal sound source is visually displayed. According to the method, the limitation of a traditional method in a complex environment is overcome, and high-precision and high-robustness transformer abnormal sound source positioning is realized.

The invention belongs to the technical field of operation safety monitoring, and relates to an intelligent monitoring system and method for a high-altitude operation safety rope, and the method comprises the steps: collecting strain signals of all key stress points of the safety rope, and carrying out the self-adaptive noise reduction preprocessing of the strain signals through a composite noise reduction method combining sliding window dynamic statistics and variational mode decomposition; a machine learning model is constructed and trained, during training, a weight initialization strategy based on strain extreme value distribution is adopted, a dynamic sparse regularization method is adopted to optimize model parameters, real-time strain signals are collected, and after self-adaptive noise reduction preprocessing, the qualified machine learning model is input to predict and predict a strain value; and dynamically adjusting a safety threshold reference according to the current operation height and the motion state, comparing the predicted strain value with the safety threshold reference, and sending out an early warning signal. The adaptive capacity to complex loads and extreme working conditions is improved, and the monitoring and early warning precision is improved.

The invention discloses a portrait file gathering method and device based on sparse regularization decoupling non-local attention, and relates to the field of image processing, and the method comprises the steps: constructing and training a portrait file gathering model, and obtaining a trained portrait file gathering model; a sparse regularization decoupling non-local attention module in the model receives an output feature from a previous structure as an input feature, and generates a query feature, a key feature, a value feature and a unary feature respectively; generating a pairwise attention weight matrix for representing the similarity between the query features and the key features by querying the related pairwise context streams; processing the unary features by querying the irrelevant unary context stream to generate a unary attention weight matrix; a sparse regularized decoupling feature is generated based on the pairwise attention weight matrix, the unary attention weight matrix, the value feature, and the input feature. The problem that non-local attention is difficult to effectively and independently learn the contextual information of the pairwise relation items and the contextual information of the unary items is solved.

The invention discloses an electroencephalogram and myoelectricity heterogeneous graph action recognition method based on dynamic graph optimization and functional adjacency fusion, which comprises the following steps: firstly, constructing a weighted fusion adjacency matrix of electroencephalogram (EEG) and myoelectricity (sEMG) channels by utilizing multiple functional connection indexes (including a phase locking value, a weighted phase lag index and mutual information) to form a heterogeneous graph structure; a learnable channel weight mechanism is introduced, redundant channel pruning is realized through sparse regularization, the expression efficiency of a graph structure is optimized, then local and global features of a fusion graph are extracted based on graph attention, and a comparative learning guide model is utilized to learn consistency and discriminative features between modes. The method can effectively improve the generalization ability and action recognition precision of electroencephalogram and myoelectricity joint modeling, and has good robustness and interpretability in a multi-channel brain-computer interface system.

The invention is suitable for the technical field of archive management query, and particularly relates to a large model archive question-answering system based on two-parameter sparse regularization parameter fine tuning, and the system comprises a query generation module which is used for converting a question input by a user into semantic query data through natural language processing, and generating a semantic vector; the knowledge base retrieval module is used for screening out file contents queried by the user; the generation enhancement module is used for outputting answers; and the optimization module is used for expanding the semantic coverage range of the knowledge base, analyzing the picture content, preprocessing the OCR document and finely adjusting the large model. According to the large model file question-answering system based on the two-parameter sparse regularization parameter fine tuning, aiming at the common problems of messy codes, information loss and structure confusion in an OCR document, a qwen2-7b-struct model is adopted for intelligent preprocessing, through the model, the system can automatically clean redundant line feed characters and reconstruct paragraph structures, and therefore the efficiency is improved. Therefore, the logicality and continuity of the text are improved.

The invention relates to an aerospace target radar image data generation method based on low illumination enhancement correction. The method comprises the following steps: designing a content delivery decomposition network, decomposing low light characteristics into illumination-independent reflection components and adaptive illumination components through Hadamard product constraint in a submerged space, proposing a learnable intensity compression function, and dynamically generating a submerged space mask through gradient consistency constraint and sparse regularization; embedding a mask into a reverse denoising process to realize directional repair of a degraded region, and combining the generation capability of a diffusion model with the guidance of physical prior; a diffusion path is reconstructed based on a bidirectional diffusion principle, physical consistency of a generation process is constrained through end point binding, a diffusion step length is adaptively adjusted, a cyclic implicit iteration mechanism is introduced, a generation result is gradually refined through a cyclic neural network module, and accurate optical-ISAR image translation can be realized.

The invention discloses a hyperspectral image waveband selection method and system based on a three-dimensional convolution auto-encoder. The method comprises the following steps: carrying out block processing on an original hyperspectral image; designing a three-dimensional space-spectrum reconstruction network based on a global spectrum sensing module; constructing a combined loss function of a reconstruction error term and a sparse regularization constraint term, and performing iterative updating on the network by using an Adam optimizer; constructing a wave band comprehensive evaluation criterion based on triple constraints and a dynamic fusion mechanism; and successively selecting an optimal wave band by adopting a greedy algorithm according to a comprehensive evaluation criterion, and outputting a wave band selection result of the hyperspectral image. Through fusion of a three-dimensional spatial spectrum reconstruction module and a multi-head self-attention mechanism, local and global association of hyperspectral data is comprehensively mined, and deep joint representation of spectrum and spatial information is realized. Meanwhile, sparse regularization constraint is introduced into a loss function, and the network is guided to focus the most valuable key wave band.

The invention discloses a deep network video saliency detection method based on RPCA, and aims to solve the problems of high singular value decomposition (SVD) calculation complexity and tedious parameter selection in a traditional RPCA method. By introducing the Schatten-p norm constraint in a matrix decomposition form, the bottleneck of SVD calculation is avoided; designing a piecewise linear threshold function with a learnable parameter set, and realizing adaptive learning of sparse regularization terms; and constructing an end-to-end deep network framework, and optimizing all parameters in a data-driven manner. Experimental results show that the method shows higher detection precision and efficiency in a video saliency detection task, and has a wide application prospect.

The invention discloses an intelligent management method and system for cloud assets, and relates to the technical field of data management, and the method comprises the steps: defining cloud asset state features, carrying out the data collection, forming state vectors according to the state features, carrying out the density matrix mapping, and calculating a total outer product matrix; calculating feature traces and determining a density matrix according to the quantized value of the overall state of the state features, setting a random observation matrix and combining the random observation matrix with the density matrix to determine a random observation value, analyzing the degree of freedom of the matrix to determine the number of observation times, constructing a total set corresponding to the number of observation times, and optimizing and reconstructing the density matrix through sparse regularization. According to the method, a symmetric matrix calculation method is introduced through state feature set definition and density matrix mapping, extension from a local cloud asset state to global asset state feature representation can be realized, and energy expression of a global state is simplified through calculation of a total outer product matrix and introduction of feature traces.

The invention discloses a causal graph inference method and a causal graph inference system which are oriented to strong nonlinear continuous variable distribution and based on RealNVP and micro NOTEARS constraints. According to the method, a structure parameter matrix is constructed to represent candidate causal connection, parent variable condition input is constructed for each variable based on structure parameters, and accurate likelihood modeling is performed on the condition density of each variable by adopting a condition RealNVP normalization flow model. By constructing an objective function containing a conditional log-likelihood item, a sparse regular item and a NOTEARS style differentiable acyclic constraint item, a sparse causal structure meeting acyclic constraint is obtained by utilizing gradient optimization, and a causal graph is output. Further, based on the learned causal graph and a conditional RealNVP model, anti-fact inference is executed under a given intervention condition, and an anti-fact result is output. The method is suitable for strong nonlinear relation and complex continuous distribution scenes, and has the advantages of being stable in structure inference, high in interpretability and capable of supporting anti-fact analysis.

The application discloses a deep learning assisted accelerated fracturing construction parameter intelligent real-time optimization method, and belongs to the field of unconventional oil and gas field development, and the steps are as follows: a structured sample database covering geological properties, engineering parameters and fracturing effect responses is constructed, and unified pretreatment is carried out on multi-source data; a multi-modal space-time collaborative prediction model is constructed, combined with multi-expansion time convolution network, three-dimensional residual network and cross attention mechanism; supervised learning training is carried out based on the database, sparse regularization and learning rate scheduling are introduced, and the model precision and generalization ability are improved; a hierarchical collaborative optimization framework of outer layer Bayesian search-inner layer CMA-ES refinement is proposed, reservoir reconstruction volume maximization and construction feasibility are realized under complex constraints; compared with the existing method, the optimization parameters of the application not only include cluster distance and section distance, but also include construction parameters such as displacement, sand concentration and fracturing fluid type, and the application can realize real-time optimization of fracturing construction parameters by using a machine learning agent model.

The application discloses a hyperspectral anomaly detection method and device based on factor group sparse regularization and belongs to the technical field of remote sensing image processing. The method reduces three-dimensional hyperspectral data into a two-dimensional matrix form, decomposes the matrix into a background part and an anomaly part based on a background dictionary; utilizes a Schatten-p norm to regularize and constrain the low-rank characteristics of the background part, utilizes a 2,1 norm to regularize and constrain the column sparse characteristics of the anomaly part; converts the Schatten-p norm into a factor group sparse regularization form, introduces an auxiliary matrix to replace variables after performing singular value decomposition on a coefficient matrix; iteratively solves the replaced variables by using an alternating direction multiplier method to obtain an anomaly matrix and calculates an anomaly detection value of each pixel. The application avoids a complex calculation process, guarantees detection accuracy and significantly improves detection speed.

The invention discloses a nonlinear causal discovery method (KAN-NOTEARS) based on a Kolmogorov-Amold network, and belongs to the technical field of artificial intelligence and causal inference. The method comprises the following steps: acquiring an observation data set containing a plurality of variables; a causal model with the learnable weighted adjacency matrix and a set of KAN functions as parameters is constructed, the KAN functions are based on the Kolmogorov-Arnold representation theorem, and a causal mechanism between variables is modeled through a learnable one-dimensional nonlinear function; constructing a target function containing structural equation model fitting loss and a sparse regular term; converting a directed acyclic graph constraint into a continuous differentiable algebraic constraint; solving the constraint optimization problem by adopting an augmented Lagrangian method to obtain an optimal adjacent matrix and function parameters; and performing threshold processing on the adjacent matrix to obtain a final cause and effect graph structure. According to the method, the theoretical advantages of the KAN network in the aspect of function approximation are utilized, the problems of insufficient accuracy and low calculation efficiency when a traditional method is used for processing a complex nonlinear causal relationship are solved, and the causal discovery precision and efficiency are remarkably improved.

The application discloses a sparse regularization DOA estimation method based on risk minimization principle, and belongs to the technical field of array signal processing and underwater acoustic signal processing. The application receives array signals and establishes an observation model; constructs sparse representation and an overcomplete dictionary; establishes and initializes a regularization optimization model; adaptively updates weights and selects parameters in a risk-driven manner; after determining the regularization parameter, a fast iterative shrinkage threshold algorithm (FISTA) is used for optimization solving, and after each iteration convergence, dictionary refinement is performed on the detected direction to reduce off-grid errors; after a small amount of outer iteration is repeated, the final DOA estimation result and the corresponding power are output. The application can realize adaptive selection of the regularization parameter and the noise level, effectively overcome the precision degradation problem under complex conditions such as low signal-to-noise ratio, limited snapshot number, signal source correlation and power imbalance without manual parameter adjustment, and thus significantly improve the robustness and practicability of estimation.

The invention relates to the technical field of image noise processing, and provides an image denoising processing method and system fusing a wavelet frame and a sharpening operator, and the method comprises the steps: carrying out the sharpening processing of an obtained to-be-processed image; wavelet domain transformation is carried out on the sharpened image, and a denoising preprocessing mathematical model is constructed with the purpose of minimizing the weighted sum of a deblurring error item, an image sharpening consistency item and a wavelet domain sparse regular item; solving of the denoising preprocessing mathematical model is decomposed into a plurality of sub-problems, an alternating direction multiplier method is adopted for solving, and a restored image after denoising processing is obtained. According to the method, wavelet multi-scale analysis and sharpening operator edge enhancement are fused, an optimization model for minimizing the deblurring error, sharpening consistency and sparse regularization is constructed, and an alternating direction multiplier method and soft threshold processing are adopted, so that the image denoising effect and the detail retention capability are effectively improved; the problems of detail loss, fuzzy aggravation and low calculation efficiency in the image denoising processing process are solved.

The application discloses a hyperspectral anomaly detection method and device based on factor group sparse regularization and belongs to the technical field of remote sensing image processing. The method reduces three-dimensional hyperspectral data into a two-dimensional matrix form, decomposes the matrix into a background part and an anomaly part based on a background dictionary; utilizes a Schatten-p norm to regularize and constrain the low-rank characteristics of the background part, utilizes a 2,1 norm to regularize and constrain the column sparse characteristics of the anomaly part; converts the Schatten-p norm into a factor group sparse regularization form, introduces an auxiliary matrix to replace variables after performing singular value decomposition on a coefficient matrix; iteratively solves the replaced variables by using an alternating direction multiplier method to obtain an anomaly matrix and calculates an anomaly detection value of each pixel. The application avoids a complex calculation process, guarantees detection accuracy and significantly improves detection speed.

The invention provides a topological map expression method based on multi-model AGV scheduling. The topological map expression method comprises the following steps: carrying out first-order difference on ground penetrating radar echo data, taking an absolute value, calculating an image mean value, setting a weighted threshold value, inhibiting background clutter and noise, and highlighting a texture change region; thirdly, constructing a random dictionary, iteratively optimizing an expression coefficient, removing unmatched clutter components through sparse regularization, retaining key features and enhancing local consistency of the image; then, singular value decomposition is carried out on the regularized data, singular values with large adjacent derivative differences are screened for reconstruction, key textures are purified, and void defect features are highlighted; and finally, performing cubic interpolation processing on the data after clutter suppression processing, and inputting the data into a YOLOv7 target detection model to realize automatic detection of the bonding state of the external thermal insulation layer of the building. Clutter interference can be effectively inhibited, key texture features are reserved, the detection efficiency is greatly improved through automatic detection, and technical support is provided for quality evaluation and safety guarantee of the building external thermal insulation layer.

A structural damage identification method based on random sparse regularization comprises the following steps: S1, establishing a finite element model of a to-be-identified damage structure, dispersing the structure into a finite number of units, setting a damage working condition, and simulating the damage working condition; s2, performing modal analysis, calculating a stiffness matrix and a mass matrix of the structure, adding boundary conditions, and solving feature values and feature vectors; s3, constructing an objective function of damage identification; s4, introducing an l1 / 2 regularization optimization method, and establishing a structural damage identification equation; s5, introducing a random class algorithm, optimizing and solving the structural damage identification target equation, and obtaining a solution vector; and S6, determining the damage position and the damage degree in the structure according to the solution vector of the damage identification equation. According to the method, the damage position can be quickly positioned, the damage degree can be accurately quantified, the accuracy of a damage identification result is improved, and the identification time of damage identification is shortened.

The application particularly relates to a UAV target main body and micro-motion component echo signal separation method based on factor decomposition group sparse regularization, which comprises the following steps: firstly, constructing a radar echo time-frequency representation into a Hankel matrix; secondly, using the characteristic that the echo Hankel matrix of the target main body uniform motion has low rank, establishing a signal separation model containing a low rank term, a sparse term and a noise term; thirdly, using a factor decomposition group sparse regularization method to effectively relax the rank function in the model, so as to improve the robustness of the algorithm; finally, using a linearized alternating direction multiplier method to iteratively solve the optimization model, so as to realize effective separation of the target main body echo and the micro-motion component echo in the time-frequency domain. The application is particularly suitable for narrow-band radar and short coherent accumulation time detection conditions, can effectively overcome the problems of model mismatch and low calculation efficiency existing in traditional methods, and simulation and measured data verify that the method has good separation precision and robustness in a noise environment.

The invention relates to the technical field of image processing and remote sensing analysis, and provides a robust depth non-negative matrix factorization method for hyperspectral image unmixing, which comprises the following steps of: decomposing hyperspectral data into a product of multiple layers of non-negative low-rank matrixes by adopting a multi-layer non-negative matrix factorization method; expressing similarity and difference of data in each layer of non-negative low-rank matrix based on a double-graph adversarial learning mechanism; structural sparse regularization based on an inner product is adopted to constrain a Gramb matrix in the decomposition process; and constructing a comprehensive optimization model to obtain an unmixing result of the hyperspectral data. According to the method, hyperspectral data are expressed as a plurality of non-negative low-rank matrixes, and meanwhile, an adversarial graph regular term, a hierarchical sparse constraint and a truncation activation mechanism are introduced to improve the robustness and expression ability of the model. And finally, performing efficient optimization by using an alternating direction multiplier method (ADMM), so that the provided method can accurately extract end members and abundance in a complex noise environment, and stable hyperspectral image unmixing is realized.

The invention discloses a robust ellipse positioning method and device based on sparse regularization and ADMM. The method comprises the following steps: acquiring known parameters required when a target is positioned by using an ellipse positioning method; the known parameters are substituted into the BR measurement abnormal value number optimization problem; according to the BR measurement abnormal value quantity optimization problem, under the constraint condition that dense inner circumference noise energy is bounded, the sparse regularization optimization problem of the target position and sparse abnormal noise is solved with the purpose of minimizing the sparse abnormal noise quantity; the sparse abnormal noise is a measurement error generated by external environment interference and multipath propagation; and solving the BR measurement abnormal value quantity optimization problem based on an ADMM algorithm, and finally obtaining a target positioning result. According to the method, the influence of the BR measurement abnormal value is effectively inhibited, the robustness and the target positioning precision of the ellipse positioning method are remarkably improved, and the method is more efficient and more suitable for real-time online processing.