CN113111834A - Attention-based global local spectral weight waveband selection method - Google Patents
Attention-based global local spectral weight waveband selection method Download PDFInfo
- Publication number
- CN113111834A CN113111834A CN202110443963.9A CN202110443963A CN113111834A CN 113111834 A CN113111834 A CN 113111834A CN 202110443963 A CN202110443963 A CN 202110443963A CN 113111834 A CN113111834 A CN 113111834A
- Authority
- CN
- China
- Prior art keywords
- band
- pixel sample
- wave
- pixel
- neural network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000003595 spectral effect Effects 0.000 title claims abstract description 22
- 238000010187 selection method Methods 0.000 title claims abstract description 12
- 239000011159 matrix material Substances 0.000 claims abstract description 32
- 238000013528 artificial neural network Methods 0.000 claims abstract description 24
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 17
- 238000003064 k means clustering Methods 0.000 claims description 11
- 238000000034 method Methods 0.000 claims description 10
- 238000007781 pre-processing Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 4
- 239000000126 substance Substances 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 2
- 238000001228 spectrum Methods 0.000 abstract description 12
- 238000012549 training Methods 0.000 abstract description 3
- 230000008569 process Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 238000011176 pooling Methods 0.000 description 4
- 230000007423 decrease Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 241001270131 Agaricus moelleri Species 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000000701 chemical imaging Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/194—Terrestrial scenes using hyperspectral data, i.e. more or other wavelengths than RGB
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A40/00—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production
- Y02A40/10—Adaptation technologies in agriculture, forestry, livestock or agroalimentary production in agriculture
Abstract
The invention discloses an attention-based global local spectral weight waveband selection method, which is characterized in that a pixel sample in a preprocessed hyperspectral data set is input into an attention-based fully-connected neural network to obtain a weight matrix of the pixel sample. And then reconstructing the hyperspectral data through a convolutional neural network, obtaining the spectral weight of each pixel sample under the condition that the loss of the reconstructed spectrum and the loss of the original spectrum are not reduced any more, training the spectral weight of each pixel by utilizing a K-means algorithm, and selecting the wave band with the largest square difference in each cluster to obtain a final wave band selection result. Therefore, the selection of the optimal waveband subset can be carried out by simultaneously considering the global information and the local information of the spectrum.
Description
Technical Field
The invention relates to the technical field of hyperspectral imaging, in particular to a global local spectral weight waveband selection method based on attention.
Background
The hyperspectral camera imaging has higher spectral resolution and richer information content than a common standard RGB camera, and is widely applied to the conditions that the imaging environment is poor, the background is complex and the like and is difficult to distinguish by human eyes at the present stage, such as the fields of observing weather with satellite hyperspectrum, detecting targets with underwater hyperspectrum and the like. The hyperspectral technology has the characteristics of large data volume and high redundancy, so that a large amount of time is consumed in the data acquisition, transmission and processing processes, a Hughes phenomenon is easily caused, and the classification precision is increased firstly and then reduced along with the increase of the number of wave bands. Band selection has proven to be effective in avoiding such problems. The band selection is an effective dimension reduction mode under the condition of not damaging the original physical significance of the high spectrum, and is concerned by more and more researchers in recent years. The method for selecting the wave band by adopting the deep network is not critical, but mostly only the fully-connected network or the convolution network is utilized, if the fully-connected network is only used, the relation between the global spectrums can be captured, but the characteristic that each neuron in the fully-connected network is connected with all the neurons in the previous layer can cause the loss of information between the continuous wave bands, and on the contrary, only the convolution network which pays more attention to local information is used, which causes the loss of the global information.
Disclosure of Invention
The invention provides a global local spectral weight waveband selection method based on attention, which aims to overcome the technical problems.
The invention discloses a global local spectral weight waveband selection method based on attention, which comprises the following steps:
inputting a hyperspectral data set, and performing data preprocessing on the hyperspectral data set to enable all pixel samples to accord with the same distribution; a plurality of band values are included in the pixel sample;
inputting the preprocessed pixel samples into a fully-connected neural network to obtain a waveband weight matrix of each pixel sample;
performing Hadamard product on the weight matrix and the pixel point data of the pixel sample, and inputting the result into a convolutional neural network to obtain an optimized pixel sample;
solving the optimized pixel sample and the original pixel sample through a loss function, and obtaining a weight matrix of the optimized pixel sample when the fully-connected neural network and the convolutional neural network are both converged;
clustering the weight matrix of the optimized pixel sample by a K-means clustering algorithm; and obtaining the band index with the maximum variance in each band cluster according to the clustering result.
Further, the inputting a hyperspectral data set and performing data preprocessing on the hyperspectral data set to make all pixel samples conform to the same distribution includes:
inputting a PaviaU hyperspectral dataset, wherein the PaviaU hyperspectral dataset is represented by D, and D belongs to RN×LWherein N is the number of pixel samples, and L is the number of wave bands;
preprocessing a pixel sample in a PaviaU hyperspectral dataset by an equation (1), which is expressed as:
in the formula, a represents a pixel sample having a dimension of N × L.
Further, the inputting the preprocessed pixel samples into a fully-connected neural network to obtain a band weight matrix of each pixel sample includes:
obtaining a weight matrix of the pixel sample by equation (2), which is expressed as:
w＝softmax(f(x)) (2)
where f represents the fully-connected neural network, x represents the pixel samples, and softmax represents the softmax layer of the fully-connected neural network.
Further, the performing a hadamard product on the weight matrix and the pixel point data of the pixel sample, and inputting the result into a convolutional neural network to obtain an optimized pixel sample includes:
an optimized pixel sample is obtained by equation (3):
restrction＝f(x*w) (3)
where let z ═ x × w, f consist of five nested subfunctions, expressed as:
further, the clustering the weight matrix of the optimized pixel samples by the K-means clustering algorithm includes: dividing the wave bands into different wave band clusters by using a K-means clustering algorithm according to Euclidean distances among the wave bands; the K-means clustering algorithm comprises the following steps: randomly selecting K wave bands from L wave bands of the weight matrix of the optimized pixel sample as initial points; randomly selecting the positions of K wave bands as the mass center of the initial wave band cluster, and forming a wave band cluster of K wave band categories; calculating the distance between the wave bands except the K wave bands and the centroid of each initial wave band cluster, and marking the wave band class as the wave band cluster corresponding to the centroid closest to the centroid of the initial wave band cluster; if the wave band clusters to which the wave bands except the K wave bands belong are not changed, returning to the previous step, updating the positions of the mass centers in the wave band clusters, then recalculating the distances between the wave bands except the K wave bands and the mass center of each initial wave band cluster and classifying the wave bands to the corresponding wave band clusters again; and when the mass center of each band cluster does not change any more, finishing the algorithm.
Further, the obtaining of the band index with the largest variance in each band cluster according to the clustering result includes:
dividing all bands in the weight matrix of the optimized image sample into C1，C2，C3…CkAnd clustering to obtain the selected band subset, which is expressed as:
wherein the content of the first and second substances,
the band index represented by the largest variance in the kth band cluster.
The method comprises the step of inputting pixel samples in a preprocessed hyperspectral data set into a fully-connected neural network to obtain a weight matrix of the pixel samples. And then reconstructing the hyperspectral data through a convolutional neural network, obtaining the spectral weight of each pixel sample under the condition that the loss of the reconstructed spectrum and the loss of the original spectrum are not reduced any more, training the spectral weight of each pixel by utilizing a K-means algorithm, and selecting the spectrum with the maximum square difference in each cluster to obtain a final wave band selection result. Therefore, the selection of the optimal waveband subset can be carried out by simultaneously considering the global information and the local information of the spectrum.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a flow chart of a method of the present invention;
FIG. 2 is a schematic diagram of obtaining band values of different weights of pixel samples by a fully connected neural network according to the present invention;
FIG. 3 is a schematic diagram of spectral reconstruction by a convolutional neural network in accordance with the present invention;
FIG. 4 is a graph comparing the effect of the present invention on the selection of a waveform segment on a PaviaU data set in comparison to UBS, ONR algorithms;
FIG. 5 is a flow chart of the K-means algorithm of the present invention.
Detailed Description
In order to make the objects, 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 with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1, the present embodiment provides a band selection method based on attention global local spectral weight, including:
101. inputting a hyperspectral data set, and performing data preprocessing on the hyperspectral data set to enable all pixel samples to accord with the same distribution; the pixel sample comprises a plurality of wave band values;
specifically, the PaviaU dataset is data obtained by photographing the university of Pavia in 2002 with a ross sensor, and it contains 610 × 340 pixels, 9 categories of interest, and 103 wavelength bands, and the wavelength ranges from 0.43 to 0.86 μm, where the dataset PaviaU is taken as input, so N equals 610 × 340, and L equals 103.
The PaviaU hyperspectral data set is represented by D, and D belongs to RN×LWherein N is the number of pixel samples, and L is the number of wave bands; and randomly extracting 90% of pixel points for training and 10% of pixel points for verification. Preprocessing pixel samples in a PaviaU hyperspectral data set, aiming at enabling all the pixel samples to accord with the same distribution so as to accelerate convergence of an algorithm, and simultaneously, not losing the internal data relativity of each sample, and expressing as follows by an equation (1):
in the formula, a represents a pixel sample having a dimension of N × L.
102. Inputting the preprocessed pixel samples into a fully-connected neural network to obtain a wave band weight matrix of each pixel sample;
specifically, the preprocessed pixel samples are input into the fully-connected neural network for processing, so as to filter out less important bands and simultaneously extract relatively important bands, the bands of interest can be given a larger weight, and the less important bands can be given a smaller weight. As shown in fig. 2, each pixel sample is input, and finally, a band value with different weight in each pixel sample is obtained (different weights of colors are different, and only different gray scales can be used for distinguishing due to the limitation of the drawing in the specification). Because the importance of each wave band is consistent for each sample when the spectral dimensions are consistent, jump connection is added to accelerate the convergence of the algorithm, a weight matrix w is obtained through a fully-connected neural network, and the whole process can be expressed as follows:
w＝softmax(f(x)) (2)
where f represents the fully-connected neural network, x represents the pixel samples, and softmax represents the softmax layer of the fully-connected neural network.
103. Performing Hadamard product on the weight matrix and pixel point data of the pixel sample, and inputting a result into a convolutional neural network to obtain an optimized pixel sample;
specifically, because the CNN network (convolutional neural network) has robustness to information processing of a context, and the hyperspectral image has a characteristic of high correlation of adjacent bands, the CNN network can easily capture a relationship between a band and upper and lower adjacent bands, and a network structure of a CNN reconstruction spectrum is as shown in fig. 3. The input of the CNN network is obtained by performing a Hadamard product on a weight matrix w obtained by a fully-connected neural network and original input data. In this reconstruction module, a pooling operation and an inverse pooling operation are inserted, the pooling operation is to reduce the dimensionality of the spectral raw data of the hyperspectral image, while keeping in mind the most characteristic spectrum (i.e. the band with the largest band value between every two bands), and in the subsequent inverse pooling operation, new spectral data can be reconstructed while maintaining the most characteristic spectrum. The process is represented as:
restrction＝f(x*w) (3)
let z be x w, f is nested from multiple functions, and since it contains jump connections, we consist of 5 sub-functions, which are expressed as:
obtaining the parameter of the formula (4) f, so that the reconstructed pixel point can be obtained after the product of the input weight and the original pixel point, the reconstructed pixel point is as close as possible to the original pixel point, and then a loss function is given: loss ═ r (failure-x)2The strategy of decreasing the learning rate and stopping the network early, used here to bring the loss as close to 0 as possible, decreases the learning rate by an order of magnitude of 0.1 when the verification set loss no longer decreases 5 consecutive times, and stops the network when it does not decrease 8 consecutive times.
104. Solving the optimized pixel sample and the original pixel sample through a loss function, and obtaining a weight matrix of the optimized pixel sample when the fully-connected neural network and the convolutional neural network are both converged;
specifically, the weight matrix can be regarded as a hyperspectral image with more obvious spectral characteristics than the original hyperspectral image. The purpose of continuously iterating the optimized pixel samples and the original pixel samples to make loss is to make the whole neural network converged, wherein the whole neural network comprises a fully-connected neural network and a convolutional neural network, and when the whole network is converged, the output of the fully-connected neural network in the convergence process, namely a w weight matrix, is taken out.
105. Clustering the weight matrix of the optimized pixel sample by a K-means clustering algorithm; and obtaining the band index with the maximum variance in each band cluster according to the clustering result.
Specifically, the bands are divided into different clusters according to the Euclidean distance between the bands by using a K-means clustering algorithm, which is shown in FIG. 5.
K-means clustering algorithm:
randomly selecting K wave bands from L wave bands of the weight matrix of the optimized pixel sample as initial points;
randomly selecting the positions of K wave bands as the mass center of the initial wave band cluster, and forming a wave band cluster of K wave band categories;
calculating the distance between the wave bands except the K wave bands and the centroid of each initial wave band cluster, and marking the wave band class as the wave band cluster corresponding to the centroid closest to the centroid of the initial wave band cluster;
if the wave band clusters to which the wave bands except the K wave bands belong are not changed, returning to the previous step, updating the positions of the centroids in the wave band clusters, then recalculating the distances from the wave bands except the K wave bands to the centroids of each initial wave band cluster and classifying the wave bands to the corresponding wave band clusters again;
when the centroid of each band cluster no longer changes, the algorithm ends.
And searching the wave band with the maximum variance value in each cluster in the result of each cluster to obtain the final result. The larger the variance of a band is, the larger the data range is, and the larger the information amount is. The specific process is as follows:
dividing all bands in the weight matrix of the optimized image sample into C1，C2，C3…CkAnd clustering to obtain the selected band subset, which is expressed as:
wherein the content of the first and second substances,
the band index represented by the largest variance in the kth band cluster.
As shown in fig. 4, the algorithm of the present invention is superior to the UBS, ONR algorithm when the number of selected bands is less than 15.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.
Claims (6)
1. An attention-based global local spectral weight band selection method, comprising:
inputting a hyperspectral data set, and performing data preprocessing on the hyperspectral data set to enable all pixel samples to accord with the same distribution; a plurality of band values are included in the pixel sample;
inputting the preprocessed pixel samples into a fully-connected neural network to obtain a waveband weight matrix of each pixel sample;
performing Hadamard product on the weight matrix and the pixel point data of the pixel sample, and inputting the result into a convolutional neural network to obtain an optimized pixel sample;
solving the optimized pixel sample and the original pixel sample through a loss function, and obtaining a weight matrix of the optimized pixel sample when the fully-connected neural network and the convolutional neural network are both converged;
clustering the weight matrix of the optimized pixel sample by a K-means clustering algorithm; and obtaining the band index with the maximum variance in each band cluster according to the clustering result.
2. The attention-based global local spectral weight band selection method according to claim 1, wherein the inputting a hyperspectral dataset and performing data preprocessing on the hyperspectral dataset to make all pixel samples conform to the same distribution comprises:
inputting a PaviaU hyperspectral dataset, wherein the PaviaU hyperspectral dataset is represented by D, and D belongs to RN×LWherein N is the number of pixel samples, and L is the number of wave bands;
preprocessing a pixel sample in a PaviaU hyperspectral dataset by an equation (1), which is expressed as:
in the formula, a represents a pixel sample having a dimension of N × L.
3. The attention-based global local spectral weight band selection method according to claim 2, wherein the inputting the preprocessed pixel samples into a fully-connected neural network to obtain a band weight matrix of each pixel sample comprises:
obtaining a weight matrix of the pixel sample by equation (2), which is expressed as:
w＝softmax(f(x)) (2)
where f represents the fully-connected neural network, x represents the pixel samples, and softmax represents the softmax layer of the fully-connected neural network.
4. The attention-based global local spectral weight band selection method of claim 3, wherein the Hadamard product is performed on the weight matrix and the pixel point data of the pixel sample, and the result is input into a convolutional neural network to obtain an optimized pixel sample, and the method comprises:
an optimized pixel sample is obtained by equation (3):
restruction＝f(x*w) (3)
where let z ═ x × w, f consist of five nested subfunctions, expressed as:
5. the attention-based global local spectral weight band selection method according to claim 4, wherein the clustering the weight matrix of the optimized pixel samples by a K-means clustering algorithm comprises:
dividing the wave bands into different wave band clusters by using a K-means clustering algorithm according to Euclidean distances among the wave bands; the K-means clustering algorithm comprises the following steps:
randomly selecting K wave bands from L wave bands of the weight matrix of the optimized pixel sample as initial points;
randomly selecting the positions of K wave bands as the mass center of the initial wave band cluster, and forming a wave band cluster of K wave band categories;
calculating the distance between the wave bands except the K wave bands and the centroid of each initial wave band cluster, and marking the wave band class as the wave band cluster corresponding to the centroid closest to the centroid of the initial wave band cluster;
if the wave band clusters to which the wave bands except the K wave bands belong are not changed, returning to the previous step, updating the positions of the mass centers in the wave band clusters, then recalculating the distances between the wave bands except the K wave bands and the mass center of each initial wave band cluster and classifying the wave bands to the corresponding wave band clusters again;
and when the mass center of each band cluster does not change any more, finishing the algorithm.
6. The method according to claim 5, wherein the obtaining the band index with the largest variance in each band cluster according to the clustering result comprises:
dividing all bands in the weight matrix of the optimized image sample into C1，C2，C3…CkAnd the cluster processing is carried out on the band clusters to obtain a selected band subset, and the cluster processing is represented as:
wherein the content of the first and second substances,
the band represented by the largest variance in the Kth band clusterAnd (4) indexing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110443963.9A CN113111834B (en) | 2021-04-23 | 2021-04-23 | Global local spectrum weight band selection method based on attention |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110443963.9A CN113111834B (en) | 2021-04-23 | 2021-04-23 | Global local spectrum weight band selection method based on attention |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113111834A true CN113111834A (en) | 2021-07-13 |
| CN113111834B CN113111834B (en) | 2024-04-05 |
Family
ID=76719696
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110443963.9A Active CN113111834B (en) | 2021-04-23 | 2021-04-23 | Global local spectrum weight band selection method based on attention |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113111834B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017066825A1 (en) * | 2015-10-19 | 2017-04-27 | National Ict Australia Limited | Spectral reconstruction |
| CN111160273A (en) * | 2019-12-31 | 2020-05-15 | 北京云智空间科技有限公司 | Hyperspectral image space spectrum combined classification method and device |
| CN111191514A (en) * | 2019-12-04 | 2020-05-22 | 中国地质大学（武汉） | Hyperspectral image band selection method based on deep learning |
| WO2021003951A1 (en) * | 2019-07-08 | 2021-01-14 | 南京信息工程大学 | Hyperspectral image classification method based on label-constrained elastic network graph model |
| CN112580480A (en) * | 2020-12-14 | 2021-03-30 | 河海大学 | Hyperspectral remote sensing image classification method and device |
-
2021
- 2021-04-23 CN CN202110443963.9A patent/CN113111834B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017066825A1 (en) * | 2015-10-19 | 2017-04-27 | National Ict Australia Limited | Spectral reconstruction |
| WO2021003951A1 (en) * | 2019-07-08 | 2021-01-14 | 南京信息工程大学 | Hyperspectral image classification method based on label-constrained elastic network graph model |
| CN111191514A (en) * | 2019-12-04 | 2020-05-22 | 中国地质大学（武汉） | Hyperspectral image band selection method based on deep learning |
| CN111160273A (en) * | 2019-12-31 | 2020-05-15 | 北京云智空间科技有限公司 | Hyperspectral image space spectrum combined classification method and device |
| CN112580480A (en) * | 2020-12-14 | 2021-03-30 | 河海大学 | Hyperspectral remote sensing image classification method and device |
Non-Patent Citations (1)
| Title |
|---|
| 李道纪;郭海涛;卢俊;赵传;林雨准;余东行;: "遥感影像地物分类多注意力融和U型网络法", 测绘学报, no. 08 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113111834B (en) | 2024-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | Adaptive spectral–spatial multiscale contextual feature extraction for hyperspectral image classification | |
| CN112232280B (en) | Hyperspectral image classification method based on self-encoder and 3D depth residual error network | |
| CN110287869B (en) | High-resolution remote sensing image crop classification method based on deep learning | |
| CN112836773B (en) | Hyperspectral image classification method based on global attention residual error network | |
| CN110210313B (en) | Hyperspectral remote sensing image classification method based on multi-scale PCA-3D-CNN (principal component analysis-three dimensional-CNN) space spectrum combination | |
| CN110472682B (en) | Hyperspectral remote sensing image classification method considering spatial and local characteristics | |
| Kong et al. | Spectral–spatial feature extraction for HSI classification based on supervised hypergraph and sample expanded CNN | |
| CN110533077B (en) | Shape adaptive convolution depth neural network method for hyperspectral image classification | |
| CN112101271A (en) | Hyperspectral remote sensing image classification method and device | |
| Cai et al. | Residual-capsule networks with threshold convolution for segmentation of wheat plantation rows in UAV images | |
| CN111161362B (en) | Spectral image identification method for growth state of tea tree | |
| CN110309811A (en) | A kind of hyperspectral image classification method based on capsule network | |
| Huang et al. | Spectral-spatial masked transformer with supervised and contrastive learning for hyperspectral image classification | |
| CN112580480B (en) | Hyperspectral remote sensing image classification method and device | |
| CN111914909B (en) | Hyperspectral change detection method based on space-spectrum combined three-direction convolution network | |
| CN109190511B (en) | Hyperspectral classification method based on local and structural constraint low-rank representation | |
| CN115512162A (en) | Terrain classification method based on attention twin network and multi-mode fusion features | |
| CN111160392A (en) | Hyperspectral classification method based on wavelet width learning system | |
| Rajendran et al. | Hyperspectral image classification model using squeeze and excitation network with deep learning | |
| CN112990314A (en) | Hyperspectral image anomaly detection method and device based on improved isolated forest algorithm | |
| Yan et al. | Application and interpretation of deep learning methods for the geographical origin identification of Radix Glycyrrhizae using hyperspectral imaging | |
| CN110516648B (en) | Ramie plant number identification method based on unmanned aerial vehicle remote sensing and pattern identification | |
| Song et al. | Using dual-channel CNN to classify hyperspectral image based on spatial-spectral information | |
| CN112784907A (en) | Hyperspectral image classification method based on spatial spectral feature and BP neural network | |
| Hamouda et al. | Framework for automatic selection of kernels based on convolutional neural networks and ckmeans clustering algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |