Blind hyperspectral unmixing method and system based on end member query design and Transform attention mechanism

Abstract

The invention discloses a blind hyperspectral unmixing method and system based on end member query design and a Transform attention mechanism, and relates to the field of hyperspectral image processing. The problems that a current simplified mixing mechanism is mainly a linear mixing model, ground feature abundance and corresponding end member accuracy are low, a deep learning blind hyperspectral unmixing method depends on good end member initialization, and unmixing performance is seriously reduced during end member initialization are solved. The method comprises the following steps: constructing an abundance mapping module, extracting features in a hyperspectral image, and performing abundance decoding on the features to form an estimated abundance image; mapping the abundance graph into abundance embedding; constructing a Transform module fitted by a hybrid mechanism, and combining end member query, reconstruction query and abundance embedding through the Transform module fitted by the hybrid mechanism to obtain reconstruction query of end member information and abundance information; the original hyperspectral image is remodeled according to the reconstructed query. The method is applied to the field of blind hyperspectral image unmixing.

Description

Technical Field

[0001] The present invention relates to the field of hyperspectral image processing, and in particular to a blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism. Background Art

[0002] Blind hyperspectral unmixing methods based on deep learning have become a hot topic and have made significant progress. However, many existing research methods or experiments are mainly based on the simplified mixing mechanism hypothesis. By simplifying the mixing mechanism in the hyperspectral imaging process, the decoding structure of the autoencoder model is constructed, and the input spectrum is reconstructed to achieve simultaneous estimation of abundance and endmembers. The current simplified mixing mechanism is mainly a linear mixing model. It is generally believed that the ground objects are linearly mixed. The deep learning model built on this basis lacks the nonlinear mixing characteristics of the ground objects, and the estimated ground object abundance and corresponding endmembers are less accurate. In addition, blind hyperspectral unmixing methods based on deep learning generally rely on good endmember initialization, such as vertex component analysis. In the absence of good endmember initialization, the unmixing performance is severely degraded, which greatly limits the application of current deep learning methods in practice. Summary of the invention

[0003] Aiming at the problems that the current simplified mixing mechanism is mainly a linear mixing model, the accuracy of ground feature abundance and corresponding endmembers is low, and the blind hyperspectral unmixing method of deep learning generally relies on good endmember initialization, and the unmixing performance is seriously degraded when the endmembers are initialized, the present invention proposes a blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism, and the method comprises:

[0004] Construct an abundance mapping module, extract features from the hyperspectral image through the convolutional neural network in the abundance mapping module, and perform abundance decoding on the features to form an estimated abundance map;

[0005] Mapping abundance maps to abundance embeddings;

[0006] Construct a hybrid mechanism fitting Transformer module, combine the endmember query, reconstruction query and abundance embedding through the hybrid mechanism fitting Transformer module, and obtain the reconstruction query of endmember information and abundance information;

[0007] Reshape the original hyperspectral image according to the reconstruction query.

[0008] Furthermore, a preferred method is proposed, wherein the abundance mapping module is composed of a U-shaped structure, extracts features through a convolutional layer, and uses a Swish activation function for nonlinear transformation; after passing through the convolutional layer, the input hyperspectral image is mapped into an abundance map according to two downsampling and upsampling operations, and the output features of the encoder are optimized to the specified abundance space by using the abundance non-negative constraint ASC and the abundance sum-one ANC constraint:

[0009]

[0010] Among them, X (i,h,w) It represents the i-th feature that needs to be activated by Softmax at the spatial position (h, w), e represents the natural base, and ω represents the scale factor.

[0011] Furthermore, a preferred method is proposed, in which the abundance is mapped into an abundance embedding having the same size as the input hyperspectral image through a convolutional neural network, and the abundance embedding having the same size is obtained by retaining the original spatial size or by using any one of the downsampling operations.

[0012] Furthermore, a preferred embodiment is proposed, wherein the Transformer module of the hybrid mechanism fitting comprises:

[0013] Q represents the Query vector in the Transformer attention mechanism, K represents the Key vector in the Transformer attention mechanism, V represents the Value vector in the Transformer attention mechanism, and the output Z is expressed as:

[0014]

[0015] in, Used to control scale scaling;

[0016] Match the attention mechanism with the reconstructed query using a multi-head attention mechanism for optimization;

[0017] The Transformer module decoder is represented as:

[0018]

[0019] Among them, Z (l,1) represents the self-attention mechanism of the l-th decoder, Z (l,2) represents the cross attention mechanism of the l-th decoder;

[0020] The entire Transformer module uses a cascaded decoder structure to gradually optimize and reconstruct queries.

[0021] Furthermore, a preferred method is proposed, in which the structure of the cascade decoder is: taking the optimized reconstruction query, endmember query and abundance embedding of the previous layer as input to share the fully connected layer mapping and reshape it into a hyperspectral image, and gradually optimizing the generated reconstructed image from coarse to fine.

[0022] Furthermore, a preferred method is proposed, wherein the method further comprises constructing a loss function including three perspectives: reconstruction, abundance and end member:

[0023]

[0024] L final =L cas +L aux ,

[0025] Among them, L cas represents the cascade loss, Indicates that the focus of the entire model is the final result, represents the reconstructed image corresponding to the lth decoder, I represents the input hyperspectral image, represents the estimated abundance matrix, represents the estimated end-member matrix, N represents the number of cascaded decoders, represents the auxiliary output hyperspectral image under the linear mixed model, represents the transpose of the estimated endmember matrix, L aux represents additional linear mixing auxiliary constraints.

[0026] Furthermore, a preferred method is proposed, in which the end member is obtained as follows:

[0027]

[0028] in, represents the i-th end member, It represents the scale factor that controls the number of endmember queries corresponding to each endmember. represents endmember query; R represents the number of endmembers; L represents the number of bands of the image;

[0029] When K=1, the endmember query degenerates into the endmember.

[0030] Based on the same inventive concept, the present invention also proposes a blind hyperspectral unmixing system based on endmember query design and Transformer attention mechanism, the system comprising:

[0031] An abundance map acquisition unit is used to construct an abundance mapping module, extract features from the hyperspectral image through a convolutional neural network in the abundance mapping module, and perform abundance decoding on the features to form an estimated abundance map;

[0032] A mapping unit, for mapping the abundance map to an abundance embedding;

[0033] An information acquisition unit is used to construct a hybrid mechanism fitting Transformer module, and to combine the end member query, reconstruction query and abundance embedding through the hybrid mechanism fitting Transformer module to obtain the reconstruction query of the end member information and abundance information;

[0034] The hyperspectral image acquisition unit is used to reshape the original hyperspectral image according to the reconstruction query.

[0035] Based on the same inventive concept, the present invention also proposes a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism according to any one of the above-mentioned methods.

[0036] Based on the same inventive concept, the present invention also proposes a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism as described in any of the above items are executed.

[0037] The present invention is beneficial in that:

[0038] The blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism proposed in the present invention proposes the concepts of endmember query, reconstruction query and hybrid mechanism fitting on the basis of Transformer attention mechanism, decomposes the endmembers in the spectral dimension by means of endmember query, decomposes the accurate endmember estimation into the degree of attention of different queries in different bands, captures the mutual relationship between them, and realizes the training convergence and accurate estimation under random initialization of endmembers; learns the reconstruction query by cross-attention mechanism, realizes the inherent hybrid mechanism fitting of the hyperspectral image itself from the perspective of data fitting, and does not rely on any hybrid mechanism prior assumptions. The present invention can realize "end-to-end" unsupervised blind hyperspectral unmixing, improve the accuracy of ground feature abundance and endmember estimation, and provide more accurate sub-pixel information for analysis and interpretation of subsequent applications.

[0039] Traditional blind unmixing methods usually rely on the accuracy of endmember initialization. If the initialization is not good, the unmixing effect will be significantly affected. Through the endmember query design module, the present invention dynamically queries and optimizes the endmembers during the unmixing process, avoiding the limitations of relying on prior information in traditional methods. This not only reduces the error of endmember initialization, but also improves the flexibility and accuracy of blind unmixing. In the hyperspectral image unmixing task, the use of the Transformer module can better handle the complex relationship between different bands, thereby improving the accuracy of the unmixing results, especially under nonlinear and complex mixing models. The abundance mapping module extracts features from hyperspectral images through convolutional neural networks, and then generates abundance maps through abundance decoding. The use of convolutional neural networks (CNNs) can fully explore the spatial features and local information of images, provide high-quality input for subsequent abundance decoding, and thus improve the accuracy of abundance estimation during the unmixing process. By mapping the abundance map to abundance embedding, the model can better learn the abundance information of the features, and fit the mixing mechanism in combination with the Transformer module. This method can not only handle linear mixing problems, but also adapt to more complex nonlinear mixing situations, which helps to achieve more accurate and robust demixing.

[0040] Existing deep learning methods usually rely on good endmember initialization, while the endmember query design of the present invention can adaptively adjust the endmembers during the unmixing process, avoiding the initialization problem in traditional methods and making the unmixing process more robust. Even without accurate prior endmember information, the model can still achieve good unmixing effects.

[0041] The present invention is applied to the field of blind hyperspectral image unmixing. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a flow chart of the blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism described in implementation mode 1;

[0043] Figure 2 This is a schematic diagram of the abundance mapping module described in Implementation Method 2;

[0044] Figure 3 A schematic diagram of a Transformer module for hybrid mechanism fitting according to the fourth implementation mode;

[0045] Figure 4 This is a schematic diagram of the cascade decoding structure described in Implementation Mode 5;

[0046] Figure 5 The abundance estimation results of different methods described in Implementation Example 11 on the Jasper Ridge dataset are shown in FIG. Figure 5 (a) is the true value result, Figure 5(b) is the abundance estimation result obtained by FCLSU method. Figure 5 (c) is the abundance estimation result obtained by using the SCLSU method. Figure 5 (d) is the abundance estimation result obtained using the NUSAL method. Figure 5 (e) is the abundance estimation result obtained using the DAEU method. Figure 5 (f) is the abundance estimation result obtained using the SIDAEU method. Figure 5 (g) is the abundance estimation result obtained by MLAEM method. Figure 5 (h) is the abundance estimation result obtained using the DeepTrans method. Figure 5 (i) is the abundance estimation result obtained by MLM-1DAE method, Figure 5 (j) is the abundance estimation result obtained by A2SN method, Figure 5 (k) is the abundance estimation result obtained by A2SAN method, Figure 5 (l) is the abundance estimation result obtained using the UNTR method; Figure 5 (m) is the abundance estimation result obtained using the UNTR-ds method;

[0047] Figure 6 The endmember estimation results of different methods described in Implementation Example 11 on the Jasper Ridge data set are shown in the figure from top to bottom: trees, water bodies, soil, and roads; the red curve is the true value, and the blue curve represents the estimation results of each method; the horizontal axis represents the band, and the vertical axis represents the normalized reflectance; Figure 6 (a) is the estimation result using the VCA method; Figure 6 (b) is the estimation result using the DAEU method; Figure 6 (c) is the estimation result using the SIDAEU method; Figure 6 (d) is the estimation result using the MLAEM method; Figure 6 (e) Estimation results using the DeepTrans method; Figure 6 (f) is the estimation result using the MLM-1DAE method; Figure 6 (g) is the estimation result using the A2SN method; Figure 6 (h) is the estimation result using the A2SAN method; Figure 6 (i) is the estimation result using the UNTR method; Figure 6 (j) is the estimation result using the UNTR-ds method;

[0048] Figure 7 The abundance estimation results of different methods described in Implementation Example 11 on the Apex dataset are shown in the figure, which respectively represent roads, trees, roofs and water bodies from top to bottom; Figure 7 (a) is the true value result, Figure 7 (b) is the abundance estimation result obtained by FCLSU method. Figure 7 (c) is the abundance estimation result obtained by using the SCLSU method. Figure 7 (d) is the abundance estimation result obtained using the NUSAL method. Figure 7 (e) is the abundance estimation result obtained using the DAEU method. Figure 7 (f) is the abundance estimation result obtained using the SIDAEU method. Figure 7 (g) is the abundance estimation result obtained by MLAEM method. Figure 7 (h) is the abundance estimation result obtained using the DeepTrans method. Figure 7 (i) is the abundance estimation result obtained by MLM-1DAE method, Figure 7 (j) is the abundance estimation result obtained by A2SN method, Figure 7 (k) is the abundance estimation result obtained by A2SAN method, Figure 7 (l) is the abundance estimation result obtained using the UNTR method; Figure 7 (m) is the abundance estimation result obtained using the UNTR-ds method;

[0049] Figure 8 The endmember estimation results of different methods described in Implementation Example 11 on the Apex data set are shown in the figure from top to bottom: roads, trees, roofs and water bodies; the red curve is the true value, and the blue curve represents the estimation results of each method; the horizontal axis represents the band, and the vertical axis represents the normalized reflectance; Figure 8 (a) is the estimation result using the VCA method; Figure 8 (b) is the estimation result using the DAEU method; Figure 8 (c) is the estimation result using the SIDAEU method; Figure 8 (d) is the estimation result using the MLAEM method; Figure 8 (e) Estimation results using the DeepTrans method; Figure 8 (f) is the estimation result using the MLM-1DAE method; Figure 8 (g) is the estimation result using the A2SN method; Figure 8 (h) is the estimation result using the A2SAN method; Figure 8 (i) is the estimation result using the UNTR method; Figure 8 (j) is the estimation result using the UNTR-ds method;

[0050] Fig. 9The abundance estimation results of different methods described in Implementation Example 11 on the Urban dataset with four-terminal annotations are shown in the figure, where asphalt, lawn, tree and roof are represented from top to bottom respectively; Fig. 9 (a) is the true value result, Fig. 9 (b) is the abundance estimation result obtained by FCLSU method. Fig. 9 (c) is the abundance estimation result obtained by using the SCLSU method. Fig. 9 (d) is the abundance estimation result obtained using the NUSAL method. Fig. 9 (e) is the abundance estimation result obtained using the DAEU method. Fig. 9 (f) is the abundance estimation result obtained using the SIDAEU method. Fig. 9 (g) is the abundance estimation result obtained by MLAEM method. Fig. 9 (h) is the abundance estimation result obtained using the DeepTrans method. Fig. 9 (i) is the abundance estimation result obtained by MLM-1DAE method, Fig. 9 (j) is the abundance estimation result obtained by A2SN method, Figure 7 (k) is the abundance estimation result obtained by A2SAN method, Fig. 9 (l) is the abundance estimation result obtained using the UNTR method; Fig. 9 (m) is the abundance estimation result obtained using the UNTR-ds method;

[0051] Fig.10 The endmember estimation results of different methods described in Implementation Example 11 on the Urban data set with four endmember annotations, where asphalt, lawn, tree and roof are represented from top to bottom; the red curve is the true value, and the blue curve represents the estimation results of each method; the horizontal axis represents the band, and the vertical axis represents the normalized reflectance; Fig.10 (a) is the estimation result using the VCA method; Fig.10 (b) is the estimation result using the DAEU method; Fig.10 (c) is the estimation result using the SIDAEU method; Fig.10 (d) is the estimation result using the MLAEM method; Fig.10 (e) is the estimation result using the DeepTrans method; Fig.10 (f) is the estimation result using the MLM-1DAE method; Fig.10 (g) is the estimation result using the A2SN method; Fig.10 (h) is the estimation result using the A2SAN method; Fig.10 (i) is the estimation result using the UNTR method; Fig.10 (j) is the estimation result using the UNTR-ds method;

[0052] Fig.11 This is the end member optimization process of the proposed method on the Jasper Ridge dataset described in Implementation 11, wherein the horizontal axis represents Band, the vertical axis Reflectance represents reflectance, Fig.11 (a) indicates random initialization; Fig.11 (b) indicates 100 training times; Fig.11 (c) indicates 200 training times; Fig.11 (d) indicates 300 training times; Fig.11 (e) indicates 400 training times; Fig.11 (f) indicates training 500 times; Fig.11 (g) indicates 600 training times; Fig.11 (h) indicates 700 training times; Fig.11 (i) indicates 800 training times; Fig.11 (j) indicates 900 training times;

[0053] Fig.12 Schematic diagram of hyperspectral images reconstructed on the Urban dataset and their root mean square errors according to the eleventh embodiment. The first row is a color image synthesized by 28, 18, and 12 bands, and the second row is an error diagram between the reconstructed image and the original image, where: Fig.12 (a) represents the image and error map reconstructed by the linear mixed model. Fig.12 (b) represents the reconstructed image and error diagram of the first decoder under the nonlinear hybrid mechanism fitting method proposed by the present invention, Fig.12 (c) represents the reconstructed image and error diagram of the second decoder under the nonlinear hybrid mechanism fitting method proposed by the present invention, Fig.12 (d) shows the reconstructed image and error diagram of the third decoder under the nonlinear hybrid mechanism fitting method proposed by the present invention. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments.

[0055] Implementation method 1, see Figure 1 The blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism described in this embodiment includes:

[0056] Construct an abundance mapping module, extract features from the hyperspectral image through the convolutional neural network in the abundance mapping module, and perform abundance decoding on the features to form an estimated abundance map;

[0057] Mapping abundance maps to abundance embeddings;

[0058] Construct a hybrid mechanism fitting Transformer module, combine the endmember query, reconstruction query and abundance embedding through the hybrid mechanism fitting Transformer module, and obtain the reconstruction query of endmember information and abundance information;

[0059] Reshape the original hyperspectral image according to the reconstruction query.

[0060] The blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism proposed in this embodiment proposes the concepts of endmember query, reconstruction query and hybrid mechanism fitting on the basis of Transformer attention mechanism. The endmembers are decomposed in the spectral dimension by endmember query, so that the accurate endmember estimation is decomposed into the degree of attention of different queries in different bands, and the mutual relationship between them is captured to achieve training convergence and accurate estimation under random initialization of endmembers; the reconstruction query is learned by cross-attention mechanism, and the inherent hybrid mechanism fitting of the hyperspectral image itself is achieved from the perspective of data fitting, which does not rely on any prior assumptions of hybrid mechanism. The method proposed in this embodiment can realize "end-to-end" unsupervised blind hyperspectral unmixing, improve the accuracy of ground feature abundance and endmember estimation, and provide more accurate sub-pixel information for analysis and interpretation of subsequent applications.

[0061] The characteristics of the mixing mechanism in this embodiment are analyzed as follows:

[0062] Assumptions represents the hyperspectral image to be unmixed, where L, H and W represent the number of bands, height and width of the image respectively. represents the end member matrix, represents the corresponding abundance, where R represents the number of end members. The commonly used simplified mixing mechanism, the linear mixed model (LMM), can be expressed as:

[0063]

[0064] in, represents additive noise, and J represents a matrix of all 1s. Considering the physical meaning of abundance, They represent the abundance non-negative constraint (ANC) and the abundance sum is unity constraint (ASC) respectively. The linear mixed model assumes that the photons reaching the sensor interact with only one material, but in reality, the end members also interact with each other. The photons reaching the sensor may be affected by multiple different materials, which is not a simplified linear mixture. The more practical nonlinear mixed model can be shown as follows:

[0065]

[0066] Among them, φ represents the mixing mechanism existing in the hyperspectral image, which is a nonlinear function acting on the end member E and parameterized by the corresponding abundance M. Nonlinear mixing models such as PPNMM and MLM developed on this basis are simplifications of highly complex problems. Integrating them into deep learning methods may not fully utilize the advantages of data-driven methods. The advantage of deep learning is that it can use observed data to fit the required function. When applied to blind hyperspectral unmixing, the mixing mechanism φ can be fitted in a data-driven way.

[0067] The deep learning method commonly used for blind unmixing is mainly the autoencoder, which uses the output of the encoder as the abundance and the weights of the fully connected layer of the decoder as the endmembers to construct a self-supervised learning process in a linear mixing manner. Introducing the above-mentioned mixing mechanism fitting method into the autoencoder is a deepening of the general simplified linear mixing model. The output of the encoder is still considered as the abundance, but the parameters of the decoder no longer represent the endmembers but the mixing mechanism. The abundance output by the encoder and the additionally defined learnable endmember matrix are used as input to reconstruct the hyperspectral image to be unmixed, which can be specifically expressed as:

[0068]

[0069] Among them, ξ E represents the encoder, and ξ D represents the decoder and the mixing mechanism learned from the observed data. It is worth noting that this method itself does not correspond the abundance and endmembers or guarantee their physical meaning. Therefore, when constructing the loss function to reconstruct the input hyperspectral image, additional linear auxiliary constraints must be added to the constructed autoencoder:

[0070]

[0071] Among them, Θ E and θ D are the encoder and decoder parameters respectively. L n It is used to reduce the difference between the input hyperspectral image and the hyperspectral image reconstructed by the autoencoder, ensuring that the decoder parameters effectively fit the inherent mixing mechanism of the observed data. lIt is a linear mixing constraint, which is used to increase the similarity between the hyperspectral image reconstructed by the linear mixing mechanism and the original image, ensure the alignment of end members and abundance, maintain their physical meaning, and make the learning of the above nonlinear mixing mechanism practical.

[0072] The purpose of endmember query is to estimate the spectral curve of pure substances, i.e. endmembers, in hyperspectral images. By decomposing accurate endmember estimation into importance estimation of multiple endmember queries in different bands, the current deep learning method's need for good endmember initialization is solved, and accurate endmember estimation is achieved under random initialization, which is more suitable for practical use.

[0073] The purpose of the reconstruction query is to restore the hyperspectral image to achieve self-supervised learning of the entire unmixing process. The reconstruction query is optimized by capturing the endmember information and endmember query, so that it has rich endmember and abundance information and can restore the input hyperspectral image. Assume It represents the reconstructed query, which is initialized with 0 elements and will be gradually optimized by the subsequent cascaded Transformer decoders in the process of self-supervised learning.

[0074] Traditional blind unmixing methods usually rely on the accuracy of endmember initialization. If the initialization is not good, the unmixing effect will be significantly affected. Through the endmember query design module, this implementation method dynamically queries and optimizes endmembers during the unmixing process, avoiding the limitations of relying on prior information in traditional methods. This not only reduces the error of endmember initialization, but also improves the flexibility and accuracy of blind unmixing.

[0075] Implementation Method 2: See Figure 2 This embodiment is a further limitation of the blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism described in the first embodiment. The abundance mapping module is composed of a U-shaped structure, extracts features through a convolutional layer, and uses a Swish activation function for nonlinear transformation; after passing through the convolutional layer, the input hyperspectral image is mapped to an abundance map according to two downsampling and upsampling operations, and the output features of the encoder are optimized to the specified abundance space by using the abundance non-negative constraint ASC and the abundance sum-one ANC constraint:

[0076]

[0077] Among them, X (i,h,w) It represents the i-th feature that needs to be activated by Softmax at the spatial position (h, w), e represents the natural base, and ω represents the scale factor.

[0078] Specifically, the abundance mapping module constructed in this embodiment has two purposes. First, it converts the captured multi-scale features into an abundance map by using an abundance decoder to map the input hyperspectral image to the potential abundance space. Second, it forms an abundance embedding with the same dimensionality as the input from the estimated abundance, which contains the abundance information and maintains the same dimensionality as the input to provide abundance information input for subsequent decoding.

[0079] like Figure 2 As shown. In this embodiment, a simplified U-shaped network structure is used to form an abundance mapping module, including two downsampling and upsampling operations to extract rich features to generate an abundance map. This convolution structure effectively integrates multi-scale local features and enhances feature representation. Considering that the activation function has a greater impact on feature extraction, the present invention uses the Swish activation function to activate the features after convolution:

[0080]

[0081] Where x represents the feature to be activated. Mapping the input hyperspectral image to an abundance map involves optimizing the output features of the encoder to the specified abundance space using ASC and ANC constraints. The Softmax activation function is extremely suitable for ASC and ANC constraints to ensure abundance sparsity, so it is used in this implementation to generate the final abundance map:

[0082]

[0083] Among them, X (i,h,w) It represents the feature that needs to be activated by Softmax at the spatial position (h, w). Considering that the abundance map is often sparse in reality, the scale factor ω is used here to achieve the sparse abundance feature, and the default setting is 3.5.

[0084] This implementation adopts a simplified U-Net structure, extracts the features of the hyperspectral image through the convolution layer, and then restores the spatial resolution through downsampling and upsampling operations, which can effectively retain the spatial information of the input image and help accurately generate the abundance map of the ground objects. U-Net combines the high-level features of the encoder stage with the low-level features of the decoder stage through jump connections, which can effectively combine the global information of the image with the local details and improve the accuracy of abundance decoding. In this method, the convolution layer is used to extract the features of the input hyperspectral image, and the Swish activation function is used to perform nonlinear transformations on these features. When modeling complex features in hyperspectral images (such as the nonlinear relationship between different ground objects and mixed backgrounds), the performance of the unmixing results can be improved.

[0085] In this implementation, downsampling helps compress input data and extract high-level features, and the upsampling operation restores image resolution, generates a detailed abundance map, and maintains the spatial consistency of the image. This operation helps the model capture the feature hierarchy of the hyperspectral image through the encoding and decoding process, thereby improving the unmixing accuracy. Due to the jump connection in the U-Net structure, after downsampling and upsampling operations, low-level features and high-level features can be effectively combined, which helps to maintain spatial details in the unmixing process and avoid information loss.

[0086] During the unmixing process, the abundance of the objects usually needs to satisfy non-negativity (i.e., the abundance value should be greater than or equal to zero) and the sum of the abundances of all objects must be one (i.e., the sum of the abundances of all objects must be 1). This embodiment optimizes the unmixing process by adding an abundance non-negative constraint (ASC) and an abundance sum-to-one constraint (ANC). The non-negative constraint ensures that the abundance estimation results of the objects are physically reasonable and avoids the situation where the abundance has negative values. The abundance sum-to-one constraint ensures that the sum of the abundances of multiple objects is 1, further ensuring that the unmixing results are consistent with the physical characteristics of hyperspectral imaging data. These constraints help to optimize the network learning process, making the model more consistent with the abundance distribution of actual objects during training, avoiding overfitting of the model or learning results that are inconsistent with reality, thereby improving the accuracy of unmixing.

[0087] In the final stage of the model, the Softmax activation function is used to activate the i-th feature in the spatial position to ensure that the value of the abundance estimate is within a reasonable range. Softmax converts the output of the network into a probability distribution, thereby ensuring that the abundance value of each feature falls within the interval [0, 1] and that the sum of the abundances of each feature is 1. This processing method is highly interpretable in practical applications and facilitates the analysis and verification of the unmixing results. The output of the Softmax activation can clearly determine the abundance of different features for each pixel, especially in a complex mixed background, which can effectively reduce uncertainty and improve the accuracy and robustness of the unmixing process.

[0088] Implementation method three: This implementation method is a further limitation of the blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism described in implementation method one. The abundance map is mapped to an abundance embedding of the same size as the input hyperspectral image through a convolutional neural network. The abundance embedding of the same size is obtained by retaining the original spatial size or by using any one of the downsampling operations.

[0089] This embodiment is explained in conjunction with the first embodiment. Since the self-attention and cross-attention mechanisms of the Transformer are used to fit the inherent mixing mechanism in the data after the abundance map is output. Since the cross-attention mechanism requires the dimensions of multiple input data to be consistent, the abundance map obtained by the abundance mapping module must be converted into a form suitable for subsequent Transformer decoding. In order to achieve this goal, the concept of abundance embedding is proposed in this embodiment, and the abundance is mapped to a feature with the same number of bands as the input image and containing abundance information through a simple convolutional neural network. Taking into account the sparsity and local similarity of the image, two methods are used to form pixel-by-pixel abundance embedding: one method retains the original spatial size, and the other is to use a downsampling operation, setting the step size to 2 during convolution to form a downsampled sparse abundance embedding. Subsequent experiments show that both methods perform well in blind hyperspectral unmixing. The key difference is that the downsampling method reduces the amount of floating-point operations. Therefore, when the amount of floating-point operations is not a priority, the present invention recommends using the first method. If the amount of operations needs to be reduced, the second downsampling method can be used.

[0090] Implementation method 4: See Figure 3 This embodiment further defines the blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism described in Embodiment 1, and the Transformer module fitted by the mixing mechanism includes:

[0091] Q represents the Query vector in the Transformer attention mechanism, K represents the Key vector in the Transformer attention mechanism, V represents the Value vector in the Transformer attention mechanism, and the output Z is expressed as:

[0092]

[0093] in, Used to control scale scaling;

[0094] Match the attention mechanism with the reconstructed query using a multi-head attention mechanism for optimization;

[0095] The Transformer module decoder is represented as:

[0096]

[0097] Among them, Z (l,1) represents the self-attention mechanism of the l-th decoder, Z (l,2) represents the cross attention mechanism of the l-th decoder;

[0098] The entire Transformer module uses a cascaded decoder structure to gradually optimize and reconstruct queries.

[0099] Specific:

[0100] Assuming that Q, K, and V represent the Query, Key, and Value vectors in the Transformer attention mechanism, the output Z can be expressed as:

[0101]

[0102] in, Used to control scale scaling. When the attention mechanism corresponds to the optimization of the present invention for reconstructing queries, a decoder in the proposed Transformer module can be expressed as:

[0103]

[0104] Among them, Z (l,1) and Z (l,2) They represent the self-attention mechanism and cross-attention mechanism of the l-th decoder respectively. The self-attention mechanism is used for the initial combination of the end-member query and the reconstruction query, while the cross-attention mechanism further fuses the end-member query with the abundance information to obtain the optimized reconstruction query. In actual use, the multi-head attention mechanism (MHA) is used to implement it. (l,1) and T (l,2) represents the l-th decoder summation and layer normalization operation. When l is 1, Initialized with all 0 elements, updated It can be expressed as:

[0105]

[0106] In the formula, MLP stands for multi-layer perceptron, which also stands for fully connected layer. In addition, the specific schematic diagram of a single decoder in the Transformer module is as follows Figure 3 In order to obtain sufficiently accurate reconstructed queries, the entire Transformer module adopts a cascade decoder structure to gradually optimize the reconstructed queries.

[0107] Implementation method 5: See Figure 4 This embodiment further defines the blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism described in Embodiment 4, and the structure of the cascade decoder is as follows: the optimized reconstruction query, endmember query and abundance embedding of the previous layer are used as input to share the fully connected layer mapping and reshape it into a hyperspectral image, and the reconstructed image generated is gradually optimized from coarse to fine.

[0108] This embodiment is explained in conjunction with the fourth embodiment. In order to obtain sufficiently accurate reconstructed queries, the entire Transformer module adopts a cascade decoder structure to gradually optimize the reconstructed query, such as Figure 4 As shown in Figure 2. This approach allows the reconstruction query generated by each decoder to restore multiple input hyperspectral images through a shared fully connected layer, achieving step-by-step optimization from coarse to fine. Therefore, the loss function can be composed of the reconstruction loss of multiple intermediate reconstruction queries and the reconstruction loss of the final reconstruction query, which only differs in the degree of refinement of the restored image.

[0109] Implementation 6. This implementation is a further limitation of the blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism described in Implementation 4. The method also includes constructing a loss function including three angles of reconstruction, abundance and endmember:

[0110]

[0111] L final =L cas +L aux ,

[0112] Among them, L cas represents the cascade loss, Indicates that the focus of the entire model is the final result, represents the reconstructed image corresponding to the lth decoder, I represents the input hyperspectral image, represents the estimated abundance matrix, represents the estimated end-member matrix, N represents the number of cascaded decoders, represents the auxiliary output hyperspectral image under the linear mixed model, represents the transpose of the estimated endmember matrix, L aux represents additional linear mixing auxiliary constraints.

[0113] Blind hyperspectral image unmixing does not rely on true value model training, and only uses a self-supervised approach, making the construction of the loss function crucial. In the present invention, the loss function is constructed from three perspectives: reconstruction, abundance, and endmembers. The reconstruction loss ensures the consistency between the input hyperspectral image and the output hyperspectral image, so that the model can fully learn the hyperspectral image feature representation. The abundance constraint takes into account the sparsity of the abundance of ground objects, while the endmember constraint ensures that the estimated endmembers are numerically reasonable. Considering that the mean square error (MSE) captures numerical differences, while the spectral angular distance (SAD) captures scale invariance, the present invention simultaneously uses both to construct the basic reconstruction loss, as shown below:

[0114]

[0115] where (h,w) represents the spatial position, ||·|| 2 It means l 2 In addition, α represents the weight of the spectral angle distance, which is set to 1 in the experiment of the present invention.

[0116] For the abundance constraint, the Softmax activation function can ensure the ASC and ANC constraints. At present, many studies generally believe that the abundance is sparse. Therefore, this implementation adopts l 1 / 2 Regularization to achieve abundance sparsity:

[0117]

[0118] Considering that endmember estimation in blind hyperspectral unmixing often leads to unreasonable values ​​in practical applications, the present invention proposes an endmember loss L endm To improve this problem, the endmember value is constrained to be the average value of the input hyperspectral data:

[0119]

[0120] Then the adopted loss function can be expressed as:

[0121]

[0122] Where β and γ are the weights of the sparse constraint and the endmember constraint respectively. Figure 4 As shown in Figure 1, the decoding structure is carried out step by step from coarse to fine in the process of reconstructing the input hyperspectral image. In addition to the final reconstructed image, N-1 intermediate restored images are also generated. This allows the adopted loss function to be expanded to a cascade loss L cas :

[0123]

[0124] L cas With L n Corresponding. It represents the reconstructed image corresponding to the lth decoder. Since the focus of the entire model is the final result Therefore, the weight of the intermediate reconstruction result is reduced. In order to ensure the physical meaning of end members and abundances and to correspond the two, an additional linear mixing auxiliary constraint L aux It can be expressed as:

[0125]

[0126] Among them, L aux and L in formula (1.4) l Correspondingly, the final loss function L final It can be expressed as:

[0127] L final =L cas +L aux .

[0128] Implementation VII: This implementation is a further limitation of the blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism described in Implementation 1. The endmember acquisition method is:

[0129]

[0130] in, represents the i-th end member, It represents the scale factor that controls the number of endmember queries corresponding to each endmember. represents endmember query; R represents the number of endmembers; L represents the number of bands of the image;

[0131] When K=1, the endmember query degenerates into the endmember.

[0132] In this embodiment, the endmember query from i·K to (i+1)·K is used to obtain the i-th endmember. Through this indirect acquisition method, a relatively accurate endmember estimate can be decomposed into the average of multiple endmember queries in different bands, reducing the demand for well-initialized endmembers.

[0133] Embodiment 8: The blind hyperspectral unmixing system based on endmember query design and Transformer attention mechanism described in this embodiment comprises:

[0134] An abundance map acquisition unit is used to construct an abundance mapping module, extract features from the hyperspectral image through a convolutional neural network in the abundance mapping module, and perform abundance decoding on the features to form an estimated abundance map;

[0135] A mapping unit, for mapping the abundance map to an abundance embedding;

[0136] An information acquisition unit is used to construct a hybrid mechanism fitting Transformer module, and to combine the end member query, reconstruction query and abundance embedding through the hybrid mechanism fitting Transformer module to obtain the reconstruction query of the end member information and abundance information;

[0137] The hyperspectral image acquisition unit is used to reshape the original hyperspectral image according to the reconstruction query.

[0138] Embodiment 9. A computer device described in this embodiment includes a memory and a processor, wherein a computer program is stored in the memory. When the processor runs the computer program stored in the memory, the processor executes the blind hyperspectral unmixing method based on end-member query design and Transformer attention mechanism described in any one of embodiments 1 to 7.

[0139] Implementation method ten, a computer-readable storage medium described in this implementation method, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism as described in any one of implementation methods one to seven are executed.

[0140] Implementation Method 11: See Figures 5 to 10 This embodiment provides a specific example of the blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism described in the first embodiment, and is also used to explain the second to seventh embodiments, specifically:

[0141] The abundance mapping module and the hybrid mechanism fitting Transformer module proposed in this embodiment constitute a blind unmixing autoencoder network UNTR, which involves multiple parameter settings. The settings of general hyperparameters are shown in Table 1 below.

[0142] Table 1 General hyperparameter settings

[0143]

[0144] Among them, ω represents the weight of the Softmax activation function, α is the weight of the spectral angular distance loss, γ is the weight of the endmember loss, and β is the weight of the abundance sparsity.

[0145] The number of end-member queries and the number of cascaded decoders for different data sets are set as shown in Table 2 below.

[0146] Table 2 Endmember query and cascade decoder number settings of the proposed method on different datasets

[0147]

[0148]

[0149] In Table 2, UNTR-ds refers to the way of downsampling the abundance embedding. K and N represent the number of endmember queries and the number of cascaded decoders, respectively.

[0150] In order to quantitatively evaluate the performance of each blind hyperspectral image method, the root mean square error (RMSE) and spectral angular distance (SAD) are used as evaluation indicators for abundance estimation and endmember estimation of each feature category in this implementation. In addition, the average root mean square error (aRMSE) and average spectral angular distance (aSAD) are used as global evaluation indicators, as shown below:

[0151]

[0152] The proposed UNTR and UNTR-ds algorithms are compared with the traditional methods FCLSU, SCLSU and NUSAL, the pixel-level deep learning methods DAEU, SIDAEU, MLAEM and MLM-1DAE, and the full-image-level deep learning methods DeepTrans, A2SN and A2SAN. Except for the proposed method and A2SN and S2SAN, which do not require endmember initialization, the other methods all use vertex component analysis (VCA) to initialize the endmembers.

[0153] Table 3 Experimental results on the Jasper Ridge dataset

[0154]

[0155] Table 3 shows the unmixing results of different methods on the Jasper Ridge dataset. The UNTR method proposed in the present invention achieved the lowest aRMSE and aSAD, which were 0.0470 and 0.0360, respectively. In terms of abundance, the estimation error of each category was the lowest, which was significantly reduced compared with other methods. The most significant error reduction was observed in the water body category, which was reduced by 65%. For end members, reliable results were obtained for each category. Even if it was not the best estimate, a suboptimal estimate was usually obtained, and the deviation from the best estimated end member was very small. On the other hand, the proposed UNTR method does not use any end member initialization method. Compared with A2SN and A2SAN, which also have no good end member initialization, the proposed UNTR achieves such a high accuracy under random initialization, which shows the effectiveness of the method.

[0156] In addition, the abundance maps and endmember spectra estimated by different methods are shown in Figure 2. Figure 5 and Figure 6 As shown. Visually, the abundance map of UNTR is very similar to the true value and maintains consistency in local details. For end members, the spectrum estimated by UNTR (blue curve) and the true spectrum (red curve) tend to change in a consistent manner, with significantly smaller deviations in the road category, while other methods often show larger errors in this category.

[0157] Table 4 Experimental results on the Apex dataset

[0158]

[0159] Table 4 shows the unmixing results of different methods on the recently proposed Apex dataset. Quantitative analysis shows that DeepTrans performs best in abundance and endmember estimation on this dataset, followed closely by the UNTR method proposed in the present invention. For specific endmembers, recent methods such as DeepTrans, MLM-1DAE, and A2SN show high accuracy in certain categories, while the UNTR of the present invention performs well in abundance estimation of roads and rooftops, and performs well in all endmember estimations. Considering that the proposed method has no endmember initialization, it has achieved a significant improvement compared with A2SN and A2SAN without endmember initialization, which once again proves the feasibility and effectiveness of the proposed method. The abundance maps and endmember spectra estimated by different methods on the Apex dataset are shown in Figure 4. Figure 7 and Figure 8 shown.

[0160] Table 5 Experimental results on Urban dataset with four end members

[0161]

[0162] Considering the comparison of different blind hyperspectral image unmixing methods on a larger spatial range and more texture information dataset, Table 5 gives the experimental results on the Urban dataset with four-endmember annotations. It can be seen that the proposed UNTR method achieves the lowest aRMSE and aSAD in all four categories. Among them, the aSAD is significantly lower than that of other methods, and is reduced by 61% compared with the second-best method A2SN. The abundance maps and endmember spectra estimated by different methods on the Urban dataset with four-endmember annotations are shown in Figure 5. Fig. 9 and Fig.10 shown.

[0163] Furthermore, in order to more clearly show the convergence process of the proposed method of random endmember initialization during training and verify the feasibility of the proposed endmember query, Fig.11 The changes and convergence of end members during the training process are given.

[0164] from Fig.11 It can be seen that since the endmember query is randomly initialized, the endmembers formed are also disordered. However, as the number of training times increases, the spectral curves of different endmembers gradually become separable until they are subsequently stable. Such endmember visualization of the training process further verifies the effectiveness of the method proposed in the present invention, which can obtain endmember spectra under random initialization and is a deep learning estimation method that is more suitable for practical use.

[0165] In order to verify the nonlinear capability of the proposed UNTR method, Fig.12The reconstructed hyperspectral image and root mean square error are given under the hybrid mechanism of estimated abundance, endmember query and Transformer cascade decoder fitting.

[0166] It can be found that if the reconstructed hyperspectral image ( Fig.12 (a) obviously loses detail information and has a large reconstruction error. The proposed UNTR uses reconstruction queries with different optimization levels to restore the hyperspectral images. Fig.12 As shown in (b)-(d), the restored hyperspectral image maintains better detail information and has a lower RMSE. In addition, as the number of decoders increases, the reconstruction error of the restored hyperspectral image gradually decreases, indicating the effectiveness of the cascade structure. The error of the hyperspectral image restored by the first decoder has been greatly reduced compared to the image restored by the linear mixed model. Although the error of the subsequent cascade decoders has decreased, the decrease is not obvious, which shows the importance of decoding in the first layer. The structure of the subsequent cascade decoders can ensure the stability of the algorithm.

[0167] Those skilled in the art will appreciate that the embodiments of the present disclosure may be provided as methods, systems or computer program products. Therefore, the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present disclosure may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0168] The present disclosure is described with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present disclosure. It should be understood that each process and / or block in the flowchart and / or block diagram and the combination of processes and / or blocks in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing device to work in a specific way, so that the instructions stored in the computer-readable memory produce a product including an instruction device, which implements the functions specified in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0169] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0170] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present disclosure rather than to limit its protection scope. Although the present disclosure has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that after reading the present disclosure, those skilled in the art can still make various changes, modifications or equivalent substitutions to the specific implementation methods of the invention, but these changes, modifications or equivalent substitutions are all within the protection scope of the disclosed claims to be approved.

Claims

1. A blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism, characterized in that: The method comprises: Construct an abundance mapping module, extract features from the hyperspectral image through the convolutional neural network in the abundance mapping module, and perform abundance decoding on the features to form an estimated abundance map; Mapping abundance maps to abundance embeddings; Construct a hybrid mechanism fitting Transformer module, combine the endmember query, reconstruction query and abundance embedding through the hybrid mechanism fitting Transformer module, and obtain the reconstruction query of endmember information and abundance information; Reshape the original hyperspectral image according to the reconstruction query.

2. The blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism according to claim 1 is characterized in that: The abundance mapping module is composed of a U-shaped structure, extracts features through a convolutional layer, and uses a Swish activation function for nonlinear transformation; After passing through the convolutional layer, the input hyperspectral image is mapped to an abundance map according to two downsampling and upsampling operations, and the output features of the encoder are optimized to the specified abundance space by using the abundance non-negative constraint ASC and the abundance sum-one constraint ANC: Among them, X (i,h,w) It represents the i-th feature that needs to be activated by Softmax at the spatial position (h, w), e represents the natural base, and ω represents the scale factor.

3. The blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism according to claim 1 is characterized in that: The abundance map is mapped to an abundance embedding having the same size as the input hyperspectral image through a convolutional neural network, and the abundance embedding having the same size is obtained by retaining the original spatial size or by using any one of downsampling operations.

4. The blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism according to claim 1 is characterized in that: The Transformer module of the hybrid mechanism fitting includes: Q represents the Query vector in the Transformer attention mechanism, K represents the Key vector in the Transformer attention mechanism, V represents the Value vector in the Transformer attention mechanism, and the output Z is expressed as: in, Used to control scale scaling; Match the attention mechanism with the reconstructed query using a multi-head attention mechanism for optimization; The Transformer module decoder is represented as: Among them, Z (l,1) represents the self-attention mechanism of the l-th decoder, Z (l,2) represents the cross attention mechanism of the l-th decoder; The entire Transformer module uses a cascaded decoder structure to gradually optimize and reconstruct queries.

5. The blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism according to claim 4 is characterized in that: The structure of the cascade decoder is as follows: the optimized reconstruction query, endmember query and abundance embedding of the previous layer are taken as input to share the fully connected layer mapping and reshape into a hyperspectral image, and the reconstructed image generated is gradually optimized from coarse to fine.

6. The blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism according to claim 4 is characterized in that: The method also includes constructing a loss function including three perspectives: reconstruction, abundance, and endmembers: L final =L cas +L aux , Among them, L cas represents the cascade loss, Indicates that the focus of the entire model is the final result, represents the reconstructed image corresponding to the lth decoder, I represents the input hyperspectral image, represents the estimated abundance matrix, represents the estimated end-member matrix, N represents the number of cascaded decoders, represents the auxiliary output hyperspectral image under the linear mixed model, represents the transpose of the estimated endmember matrix, L aux represents additional linear mixing auxiliary constraints.

7. The blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism according to claim 1 is characterized in that: The end member acquisition method is: in, represents the i-th end member, It represents the scale factor that controls the number of endmember queries corresponding to each endmember. represents endmember query; R represents the number of endmembers; L represents the number of bands of the image; When K=1, the endmember query degenerates into the endmember.

8. A blind hyperspectral unmixing system based on endmember query design and Transformer attention mechanism, characterized by: The system comprises: An abundance map acquisition unit is used to construct an abundance mapping module, extract features from the hyperspectral image through a convolutional neural network in the abundance mapping module, and perform abundance decoding on the features to form an estimated abundance map; A mapping unit, for mapping the abundance map to an abundance embedding; An information acquisition unit is used to construct a hybrid mechanism fitting Transformer module, and to combine the end member query, reconstruction query and abundance embedding through the hybrid mechanism fitting Transformer module to obtain the reconstruction query of the end member information and abundance information; The hyperspectral image acquisition unit is used to reshape the original hyperspectral image according to the reconstruction query.

9. A computer device, characterized in that: It includes a memory and a processor, wherein a computer program is stored in the memory, and when the processor runs the computer program stored in the memory, the processor executes the blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed by a processor, executes the steps of the blind hyperspectral unmixing method based on endmember query design and Transformer attention mechanism as described in any one of claims 1 to 7.

Images