173 results about "Hyperspectral image classification" patented technology

The invention relates to a hyperspectral image classification method and device based on a state space model. The hyperspectral image classification method based on the state space model comprises the following steps: sequentially carrying out feature extraction and serialization processing on hyperspectral image data to obtain a shallow feature projection vector; performing global-local feature extraction on the shallow feature projection vector by adopting a neural network based on a state space model to obtain a fused feature projection vector; and carrying out pixel-by-pixel classification and dimension rearrangement on the hyperspectral image data in sequence to generate a classification result of the hyperspectral image data. According to the hyperspectral image classification method based on the state space model, long-range dependence modeling is achieved through the neural network based on the state space model with linear complexity, the calculation complexity is effectively reduced, and through feature fusion and residual error connection, the classification accuracy of the hyperspectral image is improved. And the perception capability of the neural network on different scale space-spectrum structures in the hyperspectral image is effectively enhanced.

The invention discloses a self-supervised hyperspectral image classification method suitable for a low-annotation sample scene, and relates to the technical field of hyperspectral remote sensing image processing, comprising a self-supervised category sensing network oriented to the low-annotation scene; in the pre-training stage, a grouping spectrum enhancement module, a spectrum self-attention module and mask reconstruction are adopted, and the model is guided to focus on category-sensitive space-spectrum features under the label-free condition by minimizing the difference between a reconstructed image and an original shielded area; in the fine tuning stage, pre-trained network parameters are used as initialization parameters, and feature expression is further refined through classification loss. Therefore, by adopting the self-supervised hyperspectral image classification method suitable for the low-label sample scene, the lossless transmission of difficult sample features is realized, the distinguishing feature expression of mixed pixels is enhanced, and the classification balance of few sample categories is improved.

The invention discloses a spectrum-space depth fusion hyperspectral image classification method oriented to a small sample condition, and the method comprises the steps: firstly carrying out the multi-scale hole convolution processing of input hyperspectral data through a range attention convolution SAC module, and extracting the multi-scale context features; then, a spatial normalization attention SNA mechanism is utilized to carry out adaptive weighting adjustment of spatial dimensions on the feature map, and spatial feature representation of the key area is enhanced; the method comprises the following steps: constructing a lightweight hybrid expert model LMOE, carrying out parallel processing and gating weighting through a multi-path expert network, carrying out efficient refining and mapping on features, finally fusing processed spectral features and spatial features, and carrying out pixel-level prediction through a classifier to obtain a terrain classification result map of a hyperspectral image. The method solves the problems that in the prior art, overfitting is prone to occurring under the small sample condition, the spectrum-space collaborative modeling capacity is insufficient, the long-range dependence obtaining efficiency is low, and the recognition precision is reduced under the class imbalance scene.

The invention discloses a hyperspectral image classification method for cross-domain small sample learning based on diffusion enhancement prototype knowledge distillation, and belongs to the technical field of hyperspectral image processing. The method comprises the following steps: extracting a neighborhood data cube, aligning spectrums, dividing a support set and a query set, and applying a mask and enhancing noise; executing domain adversarial denoising and reconstruction tasks, aligning feature distribution, and outputting a pre-training encoder; decoupling features, capturing spectrum-space global and local dependency relationships, and calculating similarity between a query set and a category prototype; constructing a distillation framework to realize knowledge migration; optimizing model parameters, and introducing a signal-to-noise ratio to enhance loss suppression noise; and performing feature extraction by using the optimized student model to generate a hyperspectral image classification result. According to the method, the problems of domain offset, intra-class feature dispersion, inter-class boundary fuzziness, noise interference and the like are solved, and the classification accuracy in a small sample scene is improved.

The invention provides a hyperspectral image classification method and system based on a Swin-Transform fusion network, and relates to the technical field of image classification, and the method comprises the steps: obtaining a hyperspectral data cube with a label, cutting a three-dimensional image block with a labeling pixel as a center, and constructing an initial training sample library; constructing a dynamic adversarial amplification system and a feature extraction classification network taking a mixed Swin-Transform as a trunk, and dynamically generating a pseudo sample with high classification difficulty to expand a training set through cooperative training in an alternate circulation mode; and cutting image blocks of pixels to be classified, inputting the image blocks into the trained classification network, outputting probability distribution and determining a final category. According to the method, the problem of small sample overfitting is effectively relieved, space and spectral features are captured through a double-branch structure, adaptive fusion is carried out, and the precision and robustness of hyperspectral image classification are improved.

The invention discloses a cross-domain hyperspectral image classification method based on semantic guidance and cross-domain alignment. The method comprises the following steps: 1) processing original hyperspectral remote sensing image data and labels thereof to construct a source domain data set; 2) extracting visual features of the processed hyperspectral image through a visual feature extraction module; according to the hyperspectral image classification method, semantic information is fully utilized to establish shared knowledge between a source domain and a target domain, and the accuracy of hyperspectral image classification is effectively improved.

The invention relates to a hyperspectral image classification method based on BoostFormer double-view perceptual feature fusion, and belongs to the technical field of information processing. The method comprises the following steps: S1, hyperspectral image acquisition and data preprocessing; s2, inputting the training sample into a BoostFormer dual-view perceptual feature fusion network for learning; s3, calculating loss and updating model parameters; s4, after training is completed, the test samples are classified, and a test result is obtained. The performance of the method provided by the invention on a hyperspectral image classification task is superior to that of other hyperspectral image classification algorithms.

The invention discloses a self-adaptive kernel selection fusion Transform network and method for hyperspectral image classification, and belongs to the technical field of hyperspectral image classification, and the network comprises three innovative modules: (1) a self-adaptive convolution module (ACBlock) improves the calculation efficiency while guaranteeing the feature extraction effect through a cascade structure of convolution and separable convolution; (2) a kernel selection fusion attention module (KSFA) adopts multi-scale convolution and dynamic attention fusion design, and feature enhancement and dynamic weighted integration are realized through a kernel selection strategy; and (3) the dynamic Tanh (DyT) replaces the traditional normalization layer, and two-dimensional convolution is introduced to compensate the information loss of the Transform layer. According to the innovative design, the AKSFormer keeps the advantages of Transform global modeling, and meanwhile, the problems that local feature extraction and small sample generalization are insufficient in ability and the like are effectively solved. Experiments on three common data sets and one homemade data set show that compared with the existing advanced method, the AKSFormer classification precision is improved.

The invention belongs to the field of hyperspectral image classification, and discloses a hyperspectral image classification method and system based on multi-scale spatial-spectral joint representation and dynamic context modeling, and the method comprises the steps: carrying out the feature dimension reduction processing of a hyperspectral image through principal component analysis; spatial spectrum collaborative information of hyperspectral data is deeply mined through a multi-scale spatial spectrum joint characterization module, and adaptive fusion and enhancement of spatial spectrum characteristics under different scales are realized; a dynamic context modeling strategy is introduced, and the perception ability of the model to context information is optimized by establishing a long-range dependency relationship between features; advanced feature integration and nonlinear transformation are carried out through a multi-layer perceptron, and precise classification of hyperspectral image ground objects is completed. According to the method, the performance superior to that of a current mainstream method is obtained on three public data sets, and the effectiveness and generalization ability of the method are verified.

The invention discloses an end-to-end polarization hyperspectral image classification method and system, belongs to the technical field of deep learning and optical imaging, and solves the technical problems of low reconstruction process speed, limited precision, low information utilization rate in a classification process and weak feature extraction capability in the prior art. The method comprises the following steps: carrying out target shooting based on a snapshot type space coding hyperspectral polarization imaging system, and carrying out system coding on a shot image to obtain two-dimensional aliasing data; reconstructing the two-dimensional aliasing data into a polarization hyperspectral data cube by using the trained reconstruction network; training the classification network based on the polarization hyperspectral data cube to obtain a trained classification network; and performing joint fine tuning on the trained reconstruction network and the trained classification network to obtain a polarization hyperspectral image classification model, and performing polarization hyperspectral image classification by using the polarization hyperspectral image classification model. The method is used for realizing high-quality reconstruction and accurate classification of the polarization hyperspectral image.

According to the hyperspectral image classification method and system based on spectrum-space diffusion generation, a diffusion model is introduced to generate a hyperspectral generation sample, and a progressive denoising generation strategy of the diffusion model can generate a more diversified hyperspectral sample according with an actual spectrum physical rule while retaining spectrum details and a space structure; a category label embedding mechanism is introduced, a pre-trained diffusion model is finely adjusted, the generation capability of the diffusion model is optimized, and the problem of scarcity of small sample labeling is fundamentally relieved. Multiple screening standards including spectrum physical constraint, feature space alignment and semantic consistency verification are introduced, and authenticity and semantic consistency of generated samples in spectrum and space dimensions are ensured.

The invention discloses a hyperspectral image classification method based on frequency domain denoising and element gradient correction, and the method comprises the following steps: carrying out the preprocessing of all hyperspectral image data, and dividing an overall training sample set formed by the processed hyperspectral images into a training set and a verification set; constructing a sample weighting model based on frequency domain denoising and element gradient correction; in the training process, a parameterization frequency spectrum gating sensing transformation module is utilized to map features to a frequency domain through discrete Fourier transform, a learnable frequency spectrum response function is utilized to adaptively suppress spectrum jitter noise, and finally pure features are reconstructed. And automatically constructing a high-confidence pseudo-clean verification set based on a Gaussian mixture model and time domain consistency. According to the method, a time domain momentum updating mechanism is introduced, the variance of statistical estimation is effectively smoothed, random interference caused by training fluctuation is resisted, and the accuracy of pseudo clean set construction and the convergence stability of overall model training are further improved.

The invention provides a hyperspectral image classification enhancement method based on adaptive multi-scale superpixel constraint. The method comprises the following steps: establishing a main path and an auxiliary path; generating density information of an input feature map for the auxiliary path, and determining the number of superpixels; after the input feature map is compressed, smooth feature maps with different scales are generated; generating a corresponding number of superpixels in a corresponding kernel scale space according to the smooth feature map; for kernel scales, aggregating feature mean values in each superpixel region to obtain superpixel features of each kernel scale; calculating the weight of the corresponding kernel scale; obtaining a multi-scale super-pixel fusion feature map by using the weights and the aligned multi-scale features; total loss is constructed based on loss functions of the two classification paths; and training a neural network model of a main path and an auxiliary path by taking the neural network model as an optimization target to obtain an enhanced classification model for hyperspectral image classification. The method is compatible with an existing hyperspectral image classification network, and a lightweight enhancement scheme is provided for a deployed model.

The invention relates to an FFT-based unsupervised domain adaptive hyperspectral image classification method and system. The method comprises the steps of obtaining a remote sensing hyperspectral image, performing preprocessing, cross-domain alignment operation and edge filling operation, and performing alignment operation by using a local neighborhood; fourier transform is carried out on the source domain image and the target domain image; phase information of the source domain image is reserved, amplitude information is replaced with amplitude information of the target domain image, and the injection intensity of the amplitude of the target domain image is controlled; reconstructing the source domain image fused with the target domain amplitude information through inverse Fourier transform to obtain a data enhanced image; a ResNet model is used to extract joint features about a local space and a spectrum; performing feature extraction by adopting a channel attention mechanism and a space attention mechanism based on Fourier transform to obtain adaptive re-calibration features; unsupervised domain adaptive classification is realized by adopting multiple task heads and conditional adversarial training; and performing classification by using the trained classification model. According to the invention, the accuracy of image classification is improved.

The invention discloses a hyperspectral image classification method based on double-flow heterogeneous spatial feature collaboration. The method comprises the following steps: step 1, obtaining a hyperspectral image classification data set; step 2, designing an overall structure of a double-flow heterogeneous spatial feature collaborative modeling method (DHSFCM); step 3, setting experimental parameters; according to the hyperspectral image classification method based on double-flow heterogeneous spatial feature collaboration, noise reduction and feature enhancement are carried out through a spectrum transformation layer, then complementary features are extracted through a non-Euclidean feature extraction network (NEFEN) and an Euclidean feature extraction network (EFEN), and the hyperspectral image classification method based on double-flow heterogeneous spatial feature collaboration is obtained. And effective feature integration and interaction are realized by means of a cross attention fusion module (CAFM), and finally a classification result is output. According to the technology, accurate classification support can be provided for intelligent agriculture, resource exploration and ecological monitoring, and a new thought with engineering landing value is provided for intelligent analysis of hyperspectral images.

The invention belongs to the technical field of image processing, and discloses a hyperspectral image classification method based on multi-scale space and spectrum enhancement fusion, which comprises the following steps: carrying out dimension reduction processing on a hyperspectral image, then generating a series of image blocks, and carrying out feature extraction on the image blocks to obtain shallow layer features; the shallow layer features are input into a multi-scale spatial-spectral enhancement module, features are extracted through spectral branches and spatial branches, and spatial-spectral enhancement fusion features are output after final fusion; performing image block embedding processing on the spatial-spectral enhancement fusion features to obtain image block feature vectors, and sending the image block feature vectors into a context semantic association module for semantic enhancement coding to obtain semantic enhancement features; and the semantic enhancement features are mapped to a target category after layer normalization and full connection processing, and finally a classification thermodynamic diagram is generated. According to the method, the inter-class separability is improved, and the precise perception and enhanced expression ability of fine ground feature features can still be kept in a complex scene.

The invention discloses a remote sensing hyperspectral image classification method, system and device and a medium, belongs to the technical field of remote sensing image processing and mode recognition, and aims to solve the technical problem of how to improve the precision and efficiency of remote sensing hyperspectral image classification, reduce the calculation complexity, fully extract spectrum and spatial features and improve the classification efficiency of remote sensing hyperspectral images. According to the technical scheme, the method comprises the following steps: dividing a training set and a test set: dividing all pixels of a remote sensing hyperspectral image into the training set and the test set according to a set proportion; constructing a classification model: through a cross attention mechanism, constructing the classification model in combination with Transform of a spectral attention mechanism and KL divergence driven mutual learning classification; the test set is input into the classification model, a prediction result of the test set is obtained, a confusion matrix of the prediction result and a real label is calculated, and the final classification precision is obtained; and splicing a visual image.

The invention relates to the technical field of remote sensing image processing, in particular to a few-sample hyperspectral remote sensing image classification method and system based on multi-path evolution. The method comprises the steps of generating a spectral attention weight by utilizing spectral compression reconstruction according to acquired multi-scale deep feature representation, and performing three-dimensional convolution evolution on weighted features based on adaptive multi-path feature evolution to obtain information fusion features; performing weighted fusion on the information fusion features of the three paths through a dynamic gating mechanism; and obtaining a hyperspectral image classification result according to the fused features. According to the invention, self-adaptive evolution and cross-path selective fusion of multi-scale features are realized; the deep features subjected to multi-level feature coding are used for final classification reasoning, so that high-precision hyperspectral image classification is realized under the condition of extremely few samples.

The invention discloses a position-aware hyperspectral image classification method based on a subgraph convolutional network. The method comprises the following steps: carrying out data preprocessing on hyperspectral image data; establishing a hyperspectral image classification model by using the subgraph convolutional neural network and fusing pixel position codes; training a hyperspectral image classification model based on the preprocessed hyperspectral image data to obtain the hyperspectral image classification model; and classifying the pixel categories of the whole hyperspectral image by using the hyperspectral image classification model. The invention provides a sub-graph convolutional neural network method for efficiently fusing Laplacian-based symbol-invariant position coding, so that the diversity of pixel features can be enhanced, the classification deviation caused by spectral variation is reduced, and the classification precision is remarkably improved.

The invention discloses a grain hyperspectral image classification method based on incremental learning, and belongs to the field of image classification, and the method comprises the following steps: S1, the construction of an experimental data set: selecting a plurality of old-class rice grains and a plurality of new-class rice grains, collecting a fixed number of samples for each class, dividing a historical training set, a historical test set, a new category training set and a new category test set, converting the hyperspectral original data into a Tensor format, and normalizing the hyperspectral original data to [0, 1]; s2, initial model training; s3, new category incremental training: extracting historical category statistical parameters from a feature memory module to construct an anchoring pool, screening Top-3 similar historical categories, embedding a diffusion model noise adding process, injecting spectrum and space priori knowledge to generate virtual samples, performing dual dynamic verification, and combining real and virtual samples to form a new category training set; and calculating the distillation weight based on the spectrum similarity, and training the model through the dual-target distillation loss.

An end-to-end polarization hyperspectral image classification method and system, belonging to the field of deep learning and optical imaging technology, solves the technical problems of slow reconstruction speed, limited accuracy, low information utilization, and weak feature extraction capabilities in existing technologies. The method involves capturing images of a target using a snapshot-style spatially coded hyperspectral polarization imaging system. The captured images are then encoded by the system to obtain two-dimensional aliased data. A trained reconstruction network is used to reconstruct the two-dimensional aliased data into a polarization hyperspectral data cube. The classification network is then trained based on the polarization hyperspectral data cube to obtain a trained classification network. The trained reconstruction network and the trained classification network are jointly fine-tuned to obtain a polarization hyperspectral image classification model, which is then used to classify polarization hyperspectral images. This invention achieves high-quality reconstruction and accurate classification of polarization hyperspectral images.

In order to explore the effectiveness of the functional data analysis method in the hyperspectral image processing, the application proposes a classification method combining the Gaussian regression mixture model and the MRF hyperspectral function data; first, the polynomial regression is used to fit the hyperspectral image pixel spectrum curve, so as to express the pixel spectrum information in the form of function; then, the neighborhood relationship is introduced to establish the Markov random field model, and the neighborhood Gaussian regression mixture model is established in combination with the Gaussian regression mixture model; finally, according to the maximum posterior probability criterion, the final hyperspectral image classification result is obtained. Since the spatial-spectral information of the hyperspectral image is fully combined, the algorithm has high-precision classification result, and effectively improves the classification performance of the hyperspectral image.

The invention discloses a hyperspectral classification method and related equipment, and the method comprises the steps: obtaining a to-be-classified hyperspectral image, carrying out the region division processing based on the hyperspectral image, and obtaining a plurality of superpixel regions and the adjacent relation between the superpixel regions; constructing a multi-scale neighborhood based on the adjacent relationship, and performing relationship feature extraction on the super-pixel region based on the multi-scale neighborhood to obtain super-pixel-level features; performing multi-direction feature extraction processing on the hyperspectral image based on a convolution mode with a direction adjustment capability to obtain pixel-level features; performing multi-level feature fusion processing based on the super-pixel-level features and the pixel-level features to obtain fusion features; and based on the fusion features and a preset classifier, determining a target category of each pixel in the hyperspectral image. By performing multi-level fusion on the super-pixel-level features and the pixel-level features representing the regional relationship, regional context information and local details can be utilized at the same time, and the hyperspectral image classification precision can be improved.

The application relates to a dual-branch hyperspectral image classification method based on a graph convolutional neural network and an attention mechanism, which comprises the following steps: inputting a hyperspectral original image to data for pretreatment; inputting the data after dimension reduction into a CNN attention mechanism branch, and obtaining processed branch features through an attention mechanism module, a data processing module, an attention mechanism module and convolution operation; inputting the data after dimension reduction into a GCN branch, performing simple linear iterative clustering, segmenting into k superpixels, encoding the k superpixels, obtaining superpixel graph nodes, inputting the graph nodes into a GCN module for feature extraction, decoding the extracted superpixel node features into pixel-level branch features; performing feature fusion on the two branches, using a cross-entropy function as a loss function to train the model, and using a softmax function to obtain a class label probability. The application takes limited hyperspectral images as research objects, improves high classification precision, and ensures classification speed.

The present application belongs to the field of image classification, and particularly relates to a hyperspectral image classification method. The hyperspectral image classification method of the present application obtains Retinex features of different scales by selecting different standard deviation values of Gaussian functions, and fuses the Retinex features to form fused space-spectrum features for hyperspectral image classification, so as to take into account the global features and the detail features of the hyperspectral image, and thus the fused features used for classification contain more image information, thereby improving the classification accuracy. Moreover, the standard deviation values of the Gaussian functions capable of making the corresponding single-scale Retinex features reach the maximum classification accuracy and the adjacent standard deviation values thereof are selected to obtain Retinex features and fusion, so as to obtain feature values with higher classification accuracy. Finally, through comparison with the classification results of the spectral features, the single-scale Retinex features and other classification methods, it is verified that the classification method of the present application has higher classification accuracy compared with other methods.

This invention relates to a hyperspectral image classification method based on unsupervised domain adaptation, conditional alignment, and pseudo-label generation. The purpose of this invention is to address the low accuracy problem in existing cross-scene hyperspectral image classification methods. The specific process of the hyperspectral image classification method based on unsupervised domain adaptation, conditional alignment, and pseudo-label generation is as follows: 1. Acquire source and target domain data; 2. Construct a pixel-patch unified network model; the pixel-patch unified network model includes an encoder, a global average pooling generator, and a classifier; 3. Train the pixel-patch unified network model based on the source and target domain data from step 1 to obtain a trained pixel-patch unified network model; 4. Input the target domain data into the trained pixel-patch unified network model, and the trained pixel-patch unified network model outputs the classification result. This invention is applicable to the field of hyperspectral image classification.

The present application significantly improves the comprehensive performance of hyperspectral image classification. By effectively suppressing the noise and redundant information in the spectral and spatial dimensions, the method adaptively focuses on the wavebands and texture regions with strong discriminability, significantly enhancing the feature selection and representation ability of the model. The method achieves leading overall accuracy, average accuracy and Kappa coefficient on multiple public datasets, especially in scenes with unbalanced sample classes and complex ground object distribution, showing superior robustness and generalization ability. The generated classification map has clearer class boundaries and higher regional consistency, significantly reducing the misclassification caused by the "same object different spectrum" and "same spectrum different object" phenomenon, and the classification result is closer to the real ground object distribution. At the same time, the model has moderate computational complexity while maintaining high accuracy, achieving a good balance between performance and efficiency, providing reliable support for the accurate interpretation and practical deployment of hyperspectral images.

The application discloses a kind of position-aware subgraph convolution network-based hyperspectral image classification methods, comprising: data preprocessing is carried out to hyperspectral image data;Utilize subgraph convolution neural network and fuse pixel position coding to establish hyperspectral image classification model;Based on the hyperspectral image data after pre-processing, the hyperspectral image classification model is trained, and the hyperspectral image classification model is obtained;The overall hyperspectral image pixel class is classified using the hyperspectral image classification model.The application proposes a subgraph convolution neural network method that efficiently fuses laplace-based sign-invariant position coding, which can enhance pixel feature diversity, reduce classification bias caused by spectral variation, and significantly improve classification accuracy.

The present application relates to the technical field of remote sensing image processing, and particularly relates to a hyperspectral image classification method and device, electronic equipment and storage medium, wherein the method comprises: obtaining a hyperspectral image dataset with classification labels, determining all classifications and pixel point data corresponding to each classification; determining the data volume of a training sample set; obtaining the training sample set based on the data volume of the training sample set, the total number of all classifications and the pixel point data corresponding to each classification, and forming a test sample set with the remaining pixel point data; constructing a classification model; performing uniform sampling with replacement on the training sample set through a bagging method according to the total number of voting channels to obtain a new training sample set; training each voting channel in the classification model based on the training sample set; testing the trained classification model; obtaining a hyperspectral image data and inputting the classification model for identification and classification to obtain a classification result. The present application can improve the accuracy and stability of hyperspectral image classification.

The invention provides a domain adaptive hyperspectral image classification method and system based on multi-scale attention, and relates to the technical field of computer vision, and the method comprises the steps: obtaining and mapping data of a source domain and a target domain; a multi-scale spatial spectrum feature extraction module is utilized to capture multi-level spatial spectrum features through cascade expansion convolution; a cross-domain image adaptive module is utilized to construct and update a cross-domain graph of which the similarity is measured by KL divergence, so that structured domain self-adaption is realized; and the model is optimized by combining the small-sample classification loss and the domain self-adaptive loss. According to the method, the technical problem of poor model generalization ability caused by insufficient feature extraction and non-uniform cross-domain distribution under the condition of few samples is effectively solved, and high-precision and high-robustness hyperspectral image classification under the scene of scarce marks and domain offset is realized.