CN116563571B - Boltzmann entropy similarity-based hyperspectral image band selection method and system - Google Patents
Boltzmann entropy similarity-based hyperspectral image band selection method and system Download PDFInfo
- Publication number
- CN116563571B CN116563571B CN202310553197.0A CN202310553197A CN116563571B CN 116563571 B CN116563571 B CN 116563571B CN 202310553197 A CN202310553197 A CN 202310553197A CN 116563571 B CN116563571 B CN 116563571B
- Authority
- CN
- China
- Prior art keywords
- similarity
- band
- index
- wave
- hyperspectral image
- 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.)
- Active
Links
- 238000010187 selection method Methods 0.000 title claims abstract description 41
- 238000012545 processing Methods 0.000 claims abstract description 27
- 238000010606 normalization Methods 0.000 claims description 16
- 230000001174 ascending effect Effects 0.000 claims description 11
- 238000012163 sequencing technique Methods 0.000 claims description 6
- PMGQWSIVQFOFOQ-YKVZVUFRSA-N clemastine fumarate Chemical compound OC(=O)\C=C\C(O)=O.CN1CCC[C@@H]1CCO[C@@](C)(C=1C=CC(Cl)=CC=1)C1=CC=CC=C1 PMGQWSIVQFOFOQ-YKVZVUFRSA-N 0.000 claims description 3
- 238000000034 method Methods 0.000 description 36
- 238000004364 calculation method Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 5
- 238000000605 extraction Methods 0.000 description 4
- 238000004590 computer program Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000003595 spectral effect Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000019771 cognition Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/58—Extraction of image or video features relating to hyperspectral data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/761—Proximity, similarity or dissimilarity measures
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a hyperspectral image band selection method and a hyperspectral image band selection system based on Boltzmann entropy similarity, which belong to the technical field of image processing, and comprise the following steps: acquiring at least one wave band pair formed by each wave band in the hyperspectral image and other wave bands in sequence; determining the similarity of each band pair and the similarity threshold of the hyperspectral image based on the Boltzmann entropy difference of each band pair; determining a similarity index and a distinguishing index of each wave band based on the similarity of each wave band pair and the similarity threshold of the hyperspectral image; determining sequences of all wave bands of the hyperspectral image based on the similarity index and the distinguishing index of each wave band; based on the sequences of all wave bands of the hyperspectral image, a preset number of wave bands are selected, so that the accuracy of the wave band selection of the hyperspectral image is improved on the premise of ensuring the wave band selection efficiency.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a hyperspectral image band selection method and system based on Boltzmann entropy similarity.
Background
With the continuous improvement of the spectrum resolution, the development process of optical remote sensing can be divided into: panchromatic, color, multispectral, hyperspectral images consist of narrower bands. Hyperspectral images may have hundreds or thousands of bands and contain more information, so that dimension reduction processing is required on the hyperspectral images, and redundancy of the hyperspectral images is reduced.
The common hyperspectral image dimension reduction processing method comprises two methods of feature selection and feature extraction. Wherein feature selection, also called band selection, is an important component of image processing. The band selection can reduce the redundancy degree of hyperspectral, and is beneficial to the subsequent application of data. The method of band selection includes an optimal exponential method, an entropy method, and the like. The entropy method does not need to calculate the combination condition of all wave bands, so the applicability is stronger, and the shannon entropy is used for wave band selection in the prior art.
When the shannon entropy method is used for hyperspectral band selection, the spatial distribution is ignored due to the fact that the information quantity is measured through calculating proportion information, so that the result accuracy of hyperspectral band selection is low, and the hyperspectral band cannot be accurately selected on the premise that high efficiency is guaranteed in the prior art.
Disclosure of Invention
The invention provides a hyperspectral image band selection method and a hyperspectral image band selection system based on Boltzmann entropy similarity, which are used for solving the defect that hyperspectral image bands cannot be accurately selected on the premise of ensuring high efficiency in the prior art and improving the hyperspectral image band selection accuracy.
The invention provides a hyperspectral image band selection method based on Boltzmann entropy similarity, which comprises the following steps:
acquiring at least one wave band pair formed by each wave band in the hyperspectral image and other wave bands in sequence;
determining a similarity of each band pair and a similarity threshold of the hyperspectral image based on the boltzmann entropy difference of each band pair;
determining a similarity index and a distinguishing index of each wave band based on the similarity of each wave band pair and a similarity threshold of the hyperspectral image;
determining sequences of all wave bands of the hyperspectral image based on the similarity index and the distinguishing index of each wave band;
selecting a preset number of wave bands based on sequences of all wave bands of the hyperspectral image;
the similarity index is used for representing the similarity degree of each wave band and other wave bands, and the distinguishing index is used for representing the distinguishing degree of each wave band and other wave bands.
According to the method for selecting the band of the hyperspectral image based on the Boltzmann entropy similarity, which is provided by the invention, the similarity of each band pair and the similarity threshold value of the hyperspectral image are determined based on the Boltzmann entropy difference of each band pair, and the method comprises the following steps:
determining the similarity of each band pair based on the Boltzmann entropy difference of the band pairs;
and calculating a similarity threshold of the hyperspectral image based on the similarity of each band pair.
According to the hyperspectral image band selection method based on the Boltzmann entropy similarity, the similarity of each band pair is determined based on the Boltzmann entropy difference of each band pair, and the method comprises the following steps:
calculating absolute values of boltzmann entropy differences of the band pairs, and calculating absolute values of boltzmann entropy differences of the band pairs by using the formula (1) as similarity of the band pairs:
S ij ＝|BE(H i )-BE(H j )| (1)
wherein S is ij For the similarity of the band pairs composed of band i and band j, BE (H i ) Is Boltzmann entropy of band i, BE (H j ) Is the boltzmann entropy of band j.
According to the hyperspectral image band selection method based on Boltzmann entropy similarity, which is provided by the invention, the similarity threshold value of the hyperspectral image is calculated based on the similarity of each band pair, and the method comprises the following steps:
The similarity of each band pair is subjected to ascending order sorting, and a similarity sequence is obtained;
selecting a first preset number of similarity of the band pairs from the similarity sequence, calculating a similarity threshold of the hyperspectral image, and calculating the similarity threshold of the hyperspectral image by adopting a formula (2):
d＝mean(S x％ +…+S y％ ) (2)
wherein d is the similarity threshold of the hyperspectral image, S x％ Ordering the similarity sequence for the similarity of the band pairs of x% before the similarity sequence, S y％ Sorting the similarity sequences by y% beforeAnd the similarity of the band pairs.
According to the hyperspectral image band selection method based on boltzmann entropy similarity, the similarity index and the distinguishing index of each band are determined based on the similarity of each band pair and the similarity threshold of the hyperspectral image, and the method comprises the following steps:
determining a similarity index of each wave band based on the similarity of each wave band pair and a similarity threshold of the hyperspectral image;
and determining the distinguishing index of each wave band based on the similarity index of each wave band.
According to the hyperspectral image band selection method based on boltzmann entropy similarity, the similarity index of each band is determined based on the similarity of each band pair and a similarity threshold of the hyperspectral image, and the method comprises the following steps:
Comparing the similarity of at least one band pair formed by each band with other bands in sequence with a similarity threshold of the hyperspectral image to obtain a comparison result, determining a similarity index of each band based on the comparison result, and determining the similarity index of each band by adopting a formula (3):
wherein max (S ij ) For the maximum value of the similarity of each band pair, alpha i Is the similarity index of the band i.
According to the hyperspectral image band selection method based on Boltzmann entropy similarity, the distinguishing index of each band is determined based on the similarity index of each band, and the method comprises the following steps:
determining a maximum similarity index of each wave band based on the similarity index of each wave band;
determining the maximum similarity index of each wave band by adopting a formula (4):
wherein,is the maximum similarity index of the wave band i;
determining a distinguishing index of each wave band based on the maximum similarity index of each wave band;
determining the distinguishing index of each wave band by adopting a formula (5):
wherein θ i Is a distinguishing index of the band i, Is the maximum value of the maximum similarity index for each band.
According to the hyperspectral image band selection method based on Boltzmann entropy similarity, the sequences of all bands of the hyperspectral image are determined based on the similarity index and the distinguishing index of each band, and the method comprises the following steps: normalizing the similarity index and the distinguishing index of each wave band;
multiplying the similarity index and the distinguishing index of each wave band after the normalization processing to obtain the product of the similarity index and the distinguishing index of each wave band;
and based on the product of the similarity index and the distinguishing index of each wave band, carrying out ascending arrangement on all wave bands of the hyperspectral image to obtain the sequence of all wave bands of the hyperspectral image.
According to the hyperspectral image band selection method based on Boltzmann entropy similarity, the normalization processing is carried out on the similarity index and the distinguishing index of each band, and the method comprises the following steps:
and (3) carrying out normalization processing on the similarity index and the distinguishing index of each wave band by adopting a formula (6):
Wherein alpha is i1 For the similarity index of the band i after normalization processing, theta i1 For the distinguishing index of the band i after normalization processing, alpha max Is the maximum value of the similarity index of all the wave bands, alpha min And theta is the minimum value of the similarity indexes of all the wave bands max θ is the maximum value of the distinguishing index of all the wave bands min Is the minimum value of the distinguishing index of all the wave bands.
The invention also provides a hyperspectral image band selection system based on Boltzmann entropy similarity, which comprises the following steps:
the band pair acquisition module is used for acquiring at least one band pair formed by each band and other bands in sequence in the hyperspectral image;
a first determining module, configured to determine a similarity of each band pair and a similarity threshold of the hyperspectral image based on boltzmann entropy differences of each band pair;
the second determining module is used for determining the similarity index and the distinguishing index of each wave band based on the similarity of each wave band pair and the similarity threshold value of the hyperspectral image;
the third determining module is used for determining sequences of all wave bands of the hyperspectral image based on the similarity index and the distinguishing index of each wave band;
The selection module is used for selecting a preset number of wave bands based on sequences of all the wave bands of the hyperspectral image;
the similarity index is used for representing the similarity degree of each wave band and other wave bands, and the distinguishing index is used for representing the distinguishing degree of each wave band and other wave bands.
The invention also provides electronic equipment, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor realizes the hyperspectral image band selection method based on Boltzmann entropy similarity when executing the program.
The present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a hyperspectral image band selection method based on boltzmann entropy similarity as any one of the above.
According to the hyperspectral image band selection method and system based on the Boltzmann entropy similarity, the band pairs formed by each band in the hyperspectral image and other bands are sequentially obtained, the information quantity difference between the band pairs is measured in a mode of taking absolute values by means of difference between every two bands, the similarity threshold value of the hyperspectral image is determined based on the Boltzmann entropy difference of each band pair, the similarity index and the distinguishing index of each band are obtained through comparison based on the similarity threshold value of the hyperspectral image as comparison references, the similarity degree of each band and other bands and the distinguishing degree of each band and other bands are respectively reflected, all bands in the hyperspectral image are finally ordered based on the similarity index and the distinguishing index of each band, a preset number of bands are selected and output, the information entropy of the bands of the hyperspectral image can be measured more accurately through the method, the number of output bands can be controlled through the preset number, repeated operation is not needed, and therefore the accuracy of band selection is improved on the premise that the band selection efficiency is ensured, and the practical application scene is expanded.
Drawings
In order to more clearly illustrate the invention or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of a hyperspectral image band selection method based on Boltzmann entropy similarity;
FIG. 2 is a second flow chart of the hyperspectral image band selection method based on Boltzmann entropy similarity provided by the invention;
FIG. 3 is a third flow chart of the hyperspectral image band selection method based on Boltzmann entropy similarity provided by the invention;
FIG. 4 is a flow chart of a method for selecting bands of hyperspectral images based on Boltzmann entropy similarity provided by the invention;
FIG. 5 is a schematic diagram of the structure of the hyperspectral image band selection system based on Boltzmann entropy similarity provided by the invention;
fig. 6 is a schematic structural diagram of an electronic device provided by the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The hyperspectral image band selection method and system based on Boltzmann entropy similarity of the invention are described below with reference to FIGS. 1-6.
The hyperspectral image band selection method based on Boltzmann entropy similarity can be applied to the technical field of image processing, and can be used for dimension reduction processing of hyperspectral images. Optionally, the hyperspectral image band selection method based on the boltzmann entropy similarity is realized by a hyperspectral image band selection system based on the boltzmann entropy similarity, and an execution subject can be any electronic equipment running the system.
Fig. 1 is one of flow diagrams of a hyperspectral image band selection method based on boltzmann entropy similarity, provided by the invention, and as shown in fig. 1, the method comprises:
S1, acquiring at least one wave band pair formed by each wave band in a hyperspectral image and other wave bands in sequence;
because the band width of the hyperspectral image is narrower and the number is more, the hyperspectral image has richer spectral data information and higher spectral resolution than the multispectral image, and the information can be used for distinguishing and classifying to identify materials in the image. The hyperspectral wave bands have high correlation, representative wave bands are used as characteristic wave bands, the related technical means mainly comprise wave band extraction and wave band selection, the wave band extraction is to convert the original Gao Weibo section by using a linear or nonlinear mode so as to achieve the purpose of reducing the dimension of the hyperspectral image, the information in the characteristic wave bands is the combination of all original data information, the characteristic selection is to select a plurality of wave bands in the original high-dimensional wave bands through certain criteria or modes, and the emphasis is to select the mode in the original wave bands. Compared with the wave band extraction technology, the wave band selection method still keeps the physical meaning of the wave band representation, and is beneficial to improving the cognition of the object property.
In the field of hyperspectral image processing, the size of information entropy directly influences the quality of an image, the quality of the image is improved along with the increase of an entropy value, and the information quantity is increased.
In this step, a band pair is obtained in which each band in the hyperspectral image is sequentially combined with other bands, and if the hyperspectral image contains 10 bands in total, the result is thatAnd 45 band pairs.
S2, determining similarity of each band pair and a similarity threshold of the hyperspectral image based on Boltzmann entropy difference of each band pair;
in a specific implementation, the boltzmann entropy value of each band itself needs to be calculated first separately. Taking the Boltzmann entropy values of each band pair as differences, and then taking the absolute value of the Boltzmann entropy differences as the similarity of the two bands in the band pair, and in the step, comparing the similarity of each band pair to determine the maximum value of the similarity of all band pairs for subsequent processing; in this embodiment, 10 bands are shared in the hyperspectral image, 45 band pairs are formed together, boltzmann entropy of each of the 10 bands is calculated respectively, absolute values of boltzmann entropy differences of the 45 band pairs are sequentially calculated, the absolute values are used as similarity of the 45 band pairs, a band pair similarity matrix shown in table 1 is finally formed, and table 1 is specific data of the band pair similarity matrix in this embodiment:
TABLE 1
| Band information | Band 1 | Band 2 | Band 3 | Band 4 | Band 5 | Band 6 | Band 7 | Band 8 | Band 9 | Band 10 |
| Band 1 | / | 1241 | 7548 | 7620 | 23265 | 17515 | 4244 | 6656 | 21440 | 2890 |
| Band 2 | 1241 | / | 8789 | 8861 | 24506 | 18756 | 5485 | 7897 | 20199 | 1649 |
| Band 3 | 7548 | 8789 | / | 72 | 15717 | 9967 | 3304 | 891 | 28988 | 10437 |
| Band 4 | 7620 | 8861 | 72 | / | 15645 | 9895 | 3376 | 964 | 29060 | 10510 |
| Band 5 | 23265 | 24506 | 15717 | 15645 | / | 5750 | 19021 | 16609 | 44705 | 26155 |
| Band 6 | 17515 | 18756 | 9967 | 9895 | 5750 | / | 13271 | 10859 | 38956 | 20405 |
| Band 7 | 4244 | 5485 | 3304 | 3376 | 19021 | 13271 | / | 2412 | 25685 | 7134 |
| Band 8 | 6656 | 7897 | 891 | 964 | 16609 | 10859 | 2412 | / | 28097 | 9546 |
| Band 9 | 21440 | 20199 | 28988 | 29060 | 44705 | 38956 | 25685 | 28097 | / | 18551 |
| Band 10 | 2890 | 1649 | 10437 | 10510 | 26155 | 20405 | 7134 | 9546 | 18551 | / |
As shown in table 1, the hyperspectral image includes 10 bands in total, denoted as band 1 to band 10 in table 1; each band forms a band pair with the other remaining bands in sequence, 45 band pairs are formed together, and the numerical values in table 1 are the similarity of each band pair. Because the similarity has symmetry, when two wave bands in the wave band pair are combined into the same wave band, the similarity of the wave band pair is the same, the wave band pair is displayed as diagonal symmetry in a similarity matrix, and only half of data is needed to be compared in the actual comparison process.
Taking the similarity matrix of table 1 as an example, the similarity of the band pair consisting of the band 1 and the band 2 is 1241, and the similarity values of the remaining 44 band pairs can be obtained, as shown in table 1, in 45 band pairs, the maximum value of the band pair similarity is 44705.
Sequencing the similarity of all the wave band pairs according to an ascending order to obtain a wave band pair similarity sequence, selecting the similarity value of the first preset number of wave band pairs of the similarity sequence, and determining the similarity threshold value of the hyperspectral image as the basis for judging the similarity and the distinguishability of each wave band and other wave bands.
S3, determining a similarity index and a distinguishing index of each wave band based on the similarity of each wave band pair and a similarity threshold value of the hyperspectral image;
the similarity index is used for representing the similarity degree of each wave band and other wave bands, and the distinguishing index is used for representing the distinguishing degree of each wave band and other wave bands.
In this step, the similarity of all the band pairs formed by each band and other bands is compared with the similarity threshold of the hyperspectral image, and based on the comparison result, the similarity index of each band is determined, and the similarity index of each band is used for representing the similarity degree of the band and other bands.
Under the condition that the similarity of a certain wave band pair is smaller than a similarity threshold value, the similarity of the wave band pair is considered to be effective, the average value of all the effective similarities of the wave band pair formed by the wave band and other wave bands is taken as a similarity index of the wave band, under the condition that the similarity of all the wave band pair formed by the wave band and other wave bands is larger than or equal to the similarity threshold value, the similarity index of the wave band is set as the maximum value in all the similarities of the wave band pair formed by the wave band and other wave bands, and the similarity index of each wave band can be determined;
After the similarity index of each band is acquired, a distinction index of each band needs to be acquired, and the acquisition of the distinction index needs to be based on the maximum similarity index of each band.
In the above step, after obtaining the similarity index of each band, traversing the similarity indexes of the rest bands, and judging whether the similarity index of the band is the minimum value of the similarity indexes of all bands. If not, searching all wavebands smaller than the similarity index of the wavebands, calculating the similarity between all the wavebands meeting the requirements and the wavebands through absolute values of Boltzmann entropy differences, selecting the minimum value of the similarity value, and taking the minimum value as the maximum similarity index of the wavebands; if the similarity index of the wave band is the minimum value of the similarity indexes of all the wave bands, comparing the maximum similarity indexes of all the wave bands obtained through the calculation, selecting the maximum value of the maximum similarity index of all the wave bands, assigning the maximum value of the similarity index to the maximum similarity index value of the wave band, wherein the similarity index of the wave band is the maximum similarity index value of the wave band with the minimum value of the similarity index of all the wave bands, and the similarity index is the wave band maximum similarity index value under the condition of the minimum value. Since the similarity index and the distinguishing index of each band are oriented differently, the index forward conversion is needed before the subsequent processing, and the distinguishing degree of the band and other bands can be measured by calculating the distinguishing index of the band.
S4, determining sequences of all wave bands of the hyperspectral image based on similarity indexes and distinguishing indexes of each wave band;
in the step, the similarity index and the distinguishing index of each wave band are obtained based on the steps, and because the similarity index and the distinguishing index are different in magnitude, normalization processing is needed before subsequent processing is carried out, so that the similarity index and the distinguishing index of each wave band after normalization processing are obtained, then the product of the similarity index and the distinguishing index of each wave band after normalization processing is calculated, and based on the product result corresponding to each wave band, all wave bands are arranged in ascending order, namely the sequence of all wave bands of the hyperspectral image.
S5, selecting a preset number of wave bands based on sequences of all wave bands of the hyperspectral image;
in this step, a preset number of bands are selected as the output result of band selection according to the requirement. In a specific implementation, if the number of bands in the hyperspectral image is 100, the preset number of output bands is 10, and 10 bands are calculated according to the conventional method, but if the number of output bands is required to be replaced by 15, the whole set of method needs to be re-executed. According to the hyperspectral image band selection method based on Boltzmann entropy similarity, 100 bands are ordered through the steps, if the number of preset output bands is 10, the first 10 bands are taken, and if the number of the preset output bands is 15, 5 bands are selected backwards according to the ordering, the method is not required to be executed again, and processing time is saved.
According to the hyperspectral image band selection method based on Boltzmann entropy similarity, the band pairs formed by each band in a hyperspectral image and other bands are obtained, the information quantity difference between the band pairs is measured in a mode of taking absolute values by difference between every two bands, the similarity threshold value of the hyperspectral image is determined based on the Boltzmann entropy difference of each band pair, the similarity index and the distinguishing index of each band are obtained through comparison based on the similarity threshold value of the hyperspectral image as comparison references, the similarity degree of each band and other bands and the distinguishing degree of each band and other bands are respectively reflected, all bands in the hyperspectral image are finally ordered based on the similarity index and the distinguishing index of each band, a preset number of bands are selected and output, the information entropy of the hyperspectral image bands is measured more accurately, the number of output bands is controlled through the preset number, the accuracy of band selection is improved on the premise that the selection efficiency is met, and the explosion of calculated quantity is prevented under the condition that the number of bands is increased.
Based on the above embodiment, further, in the step S2, the specific steps of determining the similarity of each band pair and the similarity threshold of the hyperspectral image are as follows, and fig. 2 is a second schematic flow chart of the hyperspectral image band selection method based on boltzmann entropy similarity provided by the present invention, as shown in fig. 2, and the steps specifically include:
s21, determining the similarity of each band pair based on Boltzmann entropy differences of each band pair;
s22, calculating a similarity threshold value of the hyperspectral image based on the similarity of each band pair.
According to the hyperspectral image band selection method based on the Boltzmann entropy similarity, the Boltzmann entropy of all bands in the hyperspectral image is calculated, the similarity of band pairs is determined, and the similarity threshold of the hyperspectral image can be obtained based on the similarity of all band pairs, so that the calculation time consumption is reduced, and the operation efficiency of the method is guaranteed.
Based on the above embodiment, further, in S21, a specific method for determining the similarity of each band pair based on the boltzmann entropy difference of each band pair is as follows:
calculating absolute values of boltzmann entropy differences of the band pairs, and calculating absolute values of boltzmann entropy differences of the band pairs by using the formula (1) as similarity of the band pairs:
S ij ＝|BE(H i )-BE(H j )| (1)
Wherein S is ij For the similarity of the band pairs composed of band i and band j, BE (H i ) Is Boltzmann entropy of band i, BE (H j ) Is the boltzmann entropy of band j.
According to the hyperspectral image band selection method based on the Boltzmann entropy similarity, the calculated Boltzmann entropy of each band is subtracted from each other, the absolute value of the subtraction result is taken as the similarity of the two bands, and the degree of similarity between the two bands is measured through the absolute value of the Boltzmann entropy difference between the two bands.
Based on the above embodiment, further, in S22, a specific method for calculating the similarity threshold of the hyperspectral image based on the similarity of each band pair is as follows:
the similarity of each band pair is subjected to ascending order sequencing, and a similarity sequence is obtained;
selecting the similarity of a first preset number of band pairs from the similarity sequence, determining a similarity threshold of the hyperspectral image, and calculating the similarity threshold of the hyperspectral image by adopting a formula (2):
d＝mean(S x％ +…+S y％ ) (2)
wherein d is the similarity threshold of the hyperspectral image, S x％ Ordering the similarity sequence for the similarity of the band pairs of x% before the similarity sequence, S y％ Sequencing the similarity sequences by y% Is a similarity of the band pairs of (c).
Optionally, in this embodiment, the similarity of all band pairs is sorted according to the size, as S 1 <S 2 <S 3 ........S Y The above equation (2) shows that there are Y band pairs, and the similarity of the band pairs from the first 5% to the first 10% is selected to calculate the similarity threshold of the hyperspectral image, which is specifically shown in this embodiment as follows:
d＝mean(S 5％ +S 6％ +S 7％ +S 8％ +S 9％ +S 10％ )
according to the similarity matrix of table 1, the similarity calculation of the first 5% to 10% of the bands is selected, and the similarity threshold of the hyperspectral image is 927.
According to the hyperspectral image band selection method based on Boltzmann entropy similarity, the similarity of all the calculated band pairs is sequenced in an ascending order, the average value of the band pairs from the first 5% to the first 10% after the sequencing in the ascending order is calculated as the similarity threshold value of the hyperspectral image, the screening quantity of the similarity is reduced by improving the distinguishing degree of the similarity threshold value, and the selection time of the method operation is shortened to a certain extent.
Based on the above embodiment, further, in S3, the specific steps of determining the similarity index and the distinguishing index of each band are as follows, and fig. 3 is a third schematic flow chart of the hyperspectral image band selection method based on boltzmann entropy similarity, as shown in fig. 3, where the specific steps include:
S31, determining a similarity index of each wave band based on the similarity of each wave band pair and a similarity threshold of the hyperspectral image;
s32, determining the distinguishing index of each wave band based on the similarity index of each wave band.
According to the hyperspectral image band selection method based on Boltzmann entropy similarity, which is provided by the invention, based on the similarity of each band pair and the similarity threshold value of hyperspectral images obtained through the embodiment, the similarity index and the distinguishing index of each band are determined, the information quantity of the band is thinned, and the accuracy of a final band selection result is improved.
Based on the above embodiment, further, in S31, a specific method for determining the similarity index of each band based on the similarity of each band pair and the similarity threshold of the hyperspectral image is as follows:
comparing the similarity of at least one band pair formed by each band with other bands in sequence with a similarity threshold of the hyperspectral image to obtain a comparison result, determining a similarity index of each band based on the comparison result, and determining the similarity index of each band by adopting a formula (3):
wherein max (S ij ) For the maximum value of the similarity of each of the band pairs, alpha i Is the similarity index of the band i.
In the method for selecting the hyperspectral image wave band based on the Boltzmann entropy similarity, in the concrete implementation, under the condition that the similarity of a certain wave band pair is smaller than a similarity threshold, the wave band pair is considered to be effective in similarity, the average value of all effective similarities of the wave band pair formed by the wave band and other wave bands is taken as a similarity index of the wave band, under the condition that the similarity of all the wave band pair formed by the certain wave band and other wave bands is larger than or equal to the similarity threshold, the fact that the effective similarity does not exist in all the wave band pairs formed by the wave band and other wave bands is considered, the similarity index of the wave band is set to be the maximum value of all the similarities of the wave band pair formed by the wave band and other wave bands, and the similarity index of each wave band and other wave bands is calculated to measure the similarity degree of the wave band and other wave bands.
Based on the above embodiment, further, in S32, the method for determining the distinguishing index of each band based on the similarity index of each band is as follows:
determining a maximum similarity index of each wave band based on the similarity index of each wave band;
Determining the maximum similarity index of each band by adopting a formula (4):
wherein,is the maximum similarity index of the wave band i;
based on the maximum similarity index of each wave band, the distinguishing index of each wave band is determined by the following specific modes: after the similarity index of each wave band is obtained, traversing the similarity indexes of the rest wave bands, and judging whether the similarity index of the wave band is the minimum value of the similarity indexes of all the wave bands. If not, searching all wavebands smaller than the similarity index of the wavebands, calculating the similarity between all the wavebands meeting the requirements and the wavebands through absolute values of Boltzmann entropy differences, selecting the minimum value of the similarity value, and taking the minimum value as the maximum similarity index of the wavebands; if the similarity index of the wave band is the minimum value of the similarity indexes of all the wave bands, comparing the maximum similarity indexes of all the wave bands obtained through the calculation, selecting the maximum value of the maximum similarity index of all the wave bands, giving the maximum value of the similarity index to the maximum value of the similarity index of the wave band which is the maximum similarity index of the wave band with the minimum value of the similarity index of all the wave bands, wherein the similarity index is the maximum similarity index of the wave band under the condition of the minimum value.
Determining the distinguishing index of each wave band by adopting a formula (5):
wherein θ i Is a distinguishing index of the band i,is the maximum value of the maximum similarity index for each band.
Based on the above embodiment, further, in S4, the specific steps of determining the sequence of all bands of the hyperspectral image based on the similarity index and the distinguishing index of each band are as follows: fig. 4 is a schematic flow chart of a hyperspectral image band selection method based on boltzmann entropy similarity, and as shown in fig. 4, the method specifically includes:
s41, normalizing the similarity index and the distinguishing index of each wave band;
s42, multiplying the similarity index and the distinguishing index of each wave band after normalization processing to obtain the product of the similarity index and the distinguishing index of each wave band;
s43, based on the product of the similarity index and the distinguishing index of each wave band, all wave bands of the hyperspectral image are arranged in an ascending order, and sequences of all wave bands of the hyperspectral image are obtained.
Based on the above embodiment, further, in S41, the specific way of normalizing the similarity index and the distinguishing index of each band is as follows:
And (3) carrying out normalization processing on the similarity index and the distinguishing index of each wave band by adopting a formula (6):
wherein alpha is i1 For the similarity index of the band i after normalization processing, theta i1 For the distinguishing index of the band i after normalization processing, alpha max For the maximum value of similarity indexes of all wave bands, alpha min For the minimum value of similarity indexes of all wave bands, theta max For the maximum value of the distinguishing index of all the wave bands, theta min Is the minimum value of the distinguishing index of all the wave bands.
Table 2 is specific data of the similarity index, the maximum similarity index, the distinguishing index, the normalized similarity index, the normalized distinguishing index, the index product, and the band selection result provided in this embodiment, table 2 is the final result of calculation based on the data in the foregoing embodiment and table 1, the hyperspectral image in this embodiment has ten bands in total, the preset number of bands is 3 bands, that is, the first three bands of the sequence of all bands of the hyperspectral image are selected to be output, as shown in table 2:
TABLE 2
As shown in table 2, the above embodiment is used to obtain the similarity index, the maximum similarity index, the distinguishing index, the normalized similarity index, the normalized distinguishing index and the index product of 10 bands, where the index product refers to the product of the normalized similarity index and the normalized distinguishing index of each band, and finally, the ascending order is performed on all bands according to the index product of each band of the bands 1-10, in this embodiment, the preset number of bands is 3 bands, that is, the first three bands of the sequence of all bands of the hyperspectral image are selected, that is, the index product is the bands corresponding to 0.0000 and 0.0092, that is, the bands 3, 4 and 9.
The hyperspectral image band selection system based on the boltzmann entropy similarity provided by the invention is described below, and the hyperspectral image band selection system based on the boltzmann entropy similarity described below and the hyperspectral image band selection method based on the boltzmann entropy similarity described above can be correspondingly referred to each other.
Fig. 5 is a schematic structural diagram of a hyperspectral image band selection system based on boltzmann entropy similarity, and as shown in fig. 5, the hyperspectral image band selection system based on boltzmann entropy similarity provided by the invention includes:
a band pair acquisition module 51, configured to acquire at least one band pair formed by each band in the hyperspectral image and other bands in sequence;
a first determining module 52, configured to determine a similarity of each band pair and a similarity threshold of the hyperspectral image based on the boltzmann entropy difference of each band pair;
a second determining module 53, configured to determine a similarity index and a distinguishing index of each band based on the similarity of each band pair and a similarity threshold of the hyperspectral image;
a third determining module 54, configured to determine a sequence of all bands of the hyperspectral image based on the similarity index and the distinguishing index of each band;
A selection module 55, configured to select a preset number of bands based on a sequence of all bands of the hyperspectral image;
the similarity index is used for representing the similarity degree of each wave band and other wave bands, and the distinguishing index is used for representing the distinguishing degree of each wave band and other wave bands.
According to the hyperspectral image band selection system based on Boltzmann entropy similarity, through mutual coordination among the modules, the preset number of bands which are finally required to be output are selected from a hyperspectral band sequence through the selection module, the band pairs formed by each band in the hyperspectral image and other bands are sequentially obtained, the information quantity difference between the band pairs is measured in a mode of taking absolute values by pairwise difference, the similarity threshold of the hyperspectral image is determined based on Boltzmann entropy difference of each band pair, the similarity threshold of the hyperspectral image is used as a comparison benchmark, the similarity index and the distinguishing index of each band are obtained through comparison, the similarity degree of each band and other bands and the distinguishing degree of each band and other bands are respectively reflected, all bands in the hyperspectral image are finally sequenced based on the similarity index and the distinguishing index of each band, the information of the hyperspectral image is more accurately measured, the quantity of the bands which are output is not required to be repeatedly calculated through quantity control, and accordingly, on the premise that the accuracy of the selection efficiency of the bands is improved is ensured, and the quantity of the selected bands is prevented from increasing is calculated.
Fig. 6 illustrates a physical schematic diagram of an electronic device, as shown in fig. 6, which may include: processor 610, communication interface (Communications Interface) 620, memory 630, and communication bus 640, wherein processor 610, communication interface 620, and memory 630 communicate with each other via communication bus 640. The processor 610 may invoke logic instructions in the memory 630 to perform a hyperspectral image band selection method based on boltzmann entropy similarity, the method comprising: acquiring at least one wave band pair formed by each wave band in the hyperspectral image and other wave bands in sequence; determining a similarity of each band pair and a similarity threshold of the hyperspectral image based on the boltzmann entropy difference of each band pair; determining a similarity index and a distinguishing index of each wave band based on the similarity of each wave band pair and a similarity threshold of the hyperspectral image; determining sequences of all wave bands of the hyperspectral image based on the similarity index and the distinguishing index of each wave band; selecting a preset number of wave bands based on sequences of all wave bands of the hyperspectral image; the similarity index is used for representing the similarity degree of each wave band and other wave bands, and the distinguishing index is used for representing the distinguishing degree of each wave band and other wave bands.
Further, the logic instructions in the memory 630 may be implemented in the form of software functional units and stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
In yet another aspect, the present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, is implemented to perform the method for selecting hyperspectral image bands based on boltzmann entropy similarity provided by the methods described above, the method comprising: acquiring at least one wave band pair formed by each wave band in the hyperspectral image and other wave bands in sequence; determining a similarity of each band pair and a similarity threshold of the hyperspectral image based on the boltzmann entropy difference of each band pair; determining a similarity index and a distinguishing index of each wave band based on the similarity of each wave band pair and a similarity threshold of the hyperspectral image; determining sequences of all wave bands of the hyperspectral image based on the similarity index and the distinguishing index of each wave band; selecting a preset number of wave bands based on sequences of all wave bands of the hyperspectral image; the similarity index is used for representing the similarity degree of each wave band and other wave bands, and the distinguishing index is used for representing the distinguishing degree of each wave band and other wave bands.
The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (7)
1. The hyperspectral image band selection method based on Boltzmann entropy similarity is characterized by comprising the following steps of:
acquiring at least one wave band pair formed by each wave band in the hyperspectral image and other wave bands in sequence;
determining a similarity of each band pair and a similarity threshold of the hyperspectral image based on the boltzmann entropy difference of each band pair;
determining a similarity index and a distinguishing index of each wave band based on the similarity of each wave band pair and a similarity threshold of the hyperspectral image;
determining sequences of all wave bands of the hyperspectral image based on the similarity index and the distinguishing index of each wave band;
Selecting a preset number of wave bands based on sequences of all wave bands of the hyperspectral image;
the similarity index is used for representing the similarity degree of each wave band and other wave bands, and the distinguishing index is used for representing the distinguishing degree of each wave band and other wave bands;
the determining the similarity index and the distinguishing index of each waveband based on the similarity of each waveband pair and the similarity threshold of the hyperspectral image comprises the following steps:
determining a similarity index of each wave band based on the similarity of each wave band pair and a similarity threshold of the hyperspectral image;
determining a distinguishing index of each wave band based on the similarity index of each wave band;
the determining the similarity index of each waveband based on the similarity of each waveband pair and the similarity threshold of the hyperspectral image comprises the following steps:
comparing the similarity of at least one band pair formed by each band with other bands in sequence with a similarity threshold of the hyperspectral image to obtain a comparison result, determining a similarity index of each band based on the comparison result, and determining the similarity index of each band by adopting a formula (3):
Wherein max (S ij ) For the maximum value of the similarity of each band pair, alpha i Is the similarity index of the wave band i;
the determining the distinguishing index of each wave band based on the similarity index of each wave band comprises the following steps:
determining a maximum similarity index of each wave band based on the similarity index of each wave band;
determining the maximum similarity index of each wave band by adopting a formula (4):
wherein,is the maximum similarity index of band i；
Determining a distinguishing index of each wave band based on the maximum similarity index of each wave band;
determining the distinguishing index of each wave band by adopting a formula (5):
wherein θ i Is a distinguishing index of the band i,is the maximum value of the maximum similarity index for each band.
2. The boltzmann entropy similarity-based hyperspectral image band selection method according to claim 1, wherein the determining the similarity of each band pair and the similarity threshold of the hyperspectral image based on the boltzmann entropy difference of each band pair includes:
determining the similarity of each band pair based on the Boltzmann entropy difference of the band pairs;
And calculating a similarity threshold of the hyperspectral image based on the similarity of each band pair.
3. The boltzmann entropy similarity-based hyperspectral image band selection method according to claim 2, wherein the determining the similarity of each of the band pairs based on boltzmann entropy differences of the band pairs includes:
calculating absolute values of boltzmann entropy differences of the band pairs, and calculating absolute values of boltzmann entropy differences of the band pairs by using the formula (1) as similarity of the band pairs:
S ij ＝|BE(H i )-BE(H j )| (1)
wherein S is ij For the similarity of band pairs of band i and band j,BE(H i ) Is Boltzmann entropy of band i, BE (H j ) Is the boltzmann entropy of band j.
4. The boltzmann entropy similarity-based hyperspectral image band selection method according to claim 2, wherein the calculating a similarity threshold of the hyperspectral image based on the similarity of each of the band pairs includes:
the similarity of each band pair is subjected to ascending order sorting, and a similarity sequence is obtained;
selecting a first preset number of similarity of the band pairs from the similarity sequence, calculating a similarity threshold of the hyperspectral image, and calculating the similarity threshold of the hyperspectral image by adopting a formula (2):
d＝mean(S x％ +…+S y％ ) (2)
Wherein d is the similarity threshold of the hyperspectral image, S x％ Ordering the similarity sequence for the similarity of the band pairs of x% before the similarity sequence, S y％ And sequencing the similarity of the band pairs of y% before sequencing the similarity sequence.
5. The boltzmann entropy similarity-based hyperspectral image band selection method as claimed in claim 1, wherein the determining the sequence of all bands of the hyperspectral image based on the similarity index and the distinguishing index of each band includes:
normalizing the similarity index and the distinguishing index of each wave band;
multiplying the similarity index and the distinguishing index of each wave band after the normalization processing to obtain the product of the similarity index and the distinguishing index of each wave band;
and based on the product of the similarity index and the distinguishing index of each wave band, carrying out ascending arrangement on all wave bands of the hyperspectral image to obtain the sequence of all wave bands of the hyperspectral image.
6. The boltzmann entropy similarity-based hyperspectral image band selection method as claimed in claim 5, wherein normalizing the similarity index and the distinguishing index of each band comprises:
And (3) carrying out normalization processing on the similarity index and the distinguishing index of each wave band by adopting a formula (6):
wherein alpha is i1 For the similarity index of the band i after normalization processing, theta i1 For the distinguishing index of the band i after normalization processing, alpha max Is the maximum value of the similarity index of all the wave bands, alpha min And theta is the minimum value of the similarity indexes of all the wave bands max θ is the maximum value of the distinguishing index of all the wave bands min Is the minimum value of the distinguishing index of all the wave bands.
7. A hyperspectral image band selection system based on boltzmann entropy similarity, the system comprising:
the band pair acquisition module is used for acquiring at least one band pair formed by each band and other bands in sequence in the hyperspectral image;
a first determining module, configured to determine a similarity of each band pair and a similarity threshold of the hyperspectral image based on boltzmann entropy differences of each band pair;
the second determining module is used for determining the similarity index and the distinguishing index of each wave band based on the similarity of each wave band pair and the similarity threshold value of the hyperspectral image;
the third determining module is used for determining sequences of all wave bands of the hyperspectral image based on the similarity index and the distinguishing index of each wave band;
The selection module is used for selecting a preset number of wave bands based on sequences of all the wave bands of the hyperspectral image;
the similarity index is used for representing the similarity degree of each wave band and other wave bands, and the distinguishing index is used for representing the distinguishing degree of each wave band and other wave bands;
the second determining module is specifically configured to:
determining a similarity index of each wave band based on the similarity of each wave band pair and a similarity threshold of the hyperspectral image;
determining a distinguishing index of each wave band based on the similarity index of each wave band;
the determining the similarity index of each waveband based on the similarity of each waveband pair and the similarity threshold of the hyperspectral image comprises the following steps:
comparing the similarity of at least one band pair formed by each band with other bands in sequence with a similarity threshold of the hyperspectral image to obtain a comparison result, determining a similarity index of each band based on the comparison result, and determining the similarity index of each band by adopting a formula (3):
wherein max (S ij ) For the maximum value of the similarity of each band pair, alpha i Is the similarity index of the wave band i;
the determining the distinguishing index of each wave band based on the similarity index of each wave band comprises the following steps:
determining a maximum similarity index of each wave band based on the similarity index of each wave band;
determining the maximum similarity index of each wave band by adopting a formula (4):
wherein,is the maximum similarity index of the wave band i;
determining a distinguishing index of each wave band based on the maximum similarity index of each wave band;
determining the distinguishing index of each wave band by adopting a formula (5):
wherein θ i Is a distinguishing index of the band i,is the maximum value of the maximum similarity index for each band.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310553197.0A CN116563571B (en) | 2023-05-16 | 2023-05-16 | Boltzmann entropy similarity-based hyperspectral image band selection method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310553197.0A CN116563571B (en) | 2023-05-16 | 2023-05-16 | Boltzmann entropy similarity-based hyperspectral image band selection method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116563571A CN116563571A (en) | 2023-08-08 |
| CN116563571B true CN116563571B (en) | 2023-11-21 |
Family
ID=87494300
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310553197.0A Active CN116563571B (en) | 2023-05-16 | 2023-05-16 | Boltzmann entropy similarity-based hyperspectral image band selection method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116563571B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105354584A (en) * | 2015-08-24 | 2016-02-24 | 西安电子科技大学 | Waveband dissimilarity based high-spectral data waveband representation selection method |
| CN106022391A (en) * | 2016-05-31 | 2016-10-12 | 哈尔滨工业大学深圳研究生院 | Hyperspectral image characteristic parallel extraction and classification method |
| CN106778680A (en) * | 2017-01-06 | 2017-05-31 | 杭州电子科技大学 | A kind of hyperspectral image band selection method and device extracted based on critical bands |
| CN107607203A (en) * | 2017-09-08 | 2018-01-19 | 武汉大学 | Conspicuousness band selection method based on structural similarity |
| CN110232694A (en) * | 2019-06-12 | 2019-09-13 | 安徽建筑大学 | A kind of infrared polarization thermal imagery threshold segmentation method |
| AU2020103887A4 (en) * | 2020-12-04 | 2021-02-11 | kale, Karbhari Vishwanath DR | A method for automated endmember identification, selection and extraction from hyperspectral imagery |
| CN112380367A (en) * | 2020-10-27 | 2021-02-19 | 中南大学 | Entropy-based remote sensing image data screening method |
| CN112525346A (en) * | 2020-12-03 | 2021-03-19 | 安徽理工大学 | Method and system for selecting optimal band of spectral image based on improved OIF and storage medium |
| CN114048810A (en) * | 2021-11-10 | 2022-02-15 | 东华大学 | Hyperspectral image classification method based on multilevel feature extraction network |
| CN114398948A (en) * | 2021-12-13 | 2022-04-26 | 西安邮电大学 | Multispectral image change detection method based on space-spectrum combined attention network |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8406469B2 (en) * | 2009-07-20 | 2013-03-26 | The United States Of America As Represented By The Administrator Of The National Aeronautics And Space Administration | System and method for progressive band selection for hyperspectral images |
| JP2023509236A (en) * | 2020-09-08 | 2023-03-07 | 深▲せん▼市海譜納米光学科技有限公司 | Method and apparatus for reconstructing light source spectrum based on hyperspectral image |
-
2023
- 2023-05-16 CN CN202310553197.0A patent/CN116563571B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105354584A (en) * | 2015-08-24 | 2016-02-24 | 西安电子科技大学 | Waveband dissimilarity based high-spectral data waveband representation selection method |
| CN106022391A (en) * | 2016-05-31 | 2016-10-12 | 哈尔滨工业大学深圳研究生院 | Hyperspectral image characteristic parallel extraction and classification method |
| CN106778680A (en) * | 2017-01-06 | 2017-05-31 | 杭州电子科技大学 | A kind of hyperspectral image band selection method and device extracted based on critical bands |
| CN107607203A (en) * | 2017-09-08 | 2018-01-19 | 武汉大学 | Conspicuousness band selection method based on structural similarity |
| CN110232694A (en) * | 2019-06-12 | 2019-09-13 | 安徽建筑大学 | A kind of infrared polarization thermal imagery threshold segmentation method |
| CN112380367A (en) * | 2020-10-27 | 2021-02-19 | 中南大学 | Entropy-based remote sensing image data screening method |
| CN112525346A (en) * | 2020-12-03 | 2021-03-19 | 安徽理工大学 | Method and system for selecting optimal band of spectral image based on improved OIF and storage medium |
| AU2020103887A4 (en) * | 2020-12-04 | 2021-02-11 | kale, Karbhari Vishwanath DR | A method for automated endmember identification, selection and extraction from hyperspectral imagery |
| CN114048810A (en) * | 2021-11-10 | 2022-02-15 | 东华大学 | Hyperspectral image classification method based on multilevel feature extraction network |
| CN114398948A (en) * | 2021-12-13 | 2022-04-26 | 西安邮电大学 | Multispectral image change detection method based on space-spectrum combined attention network |
Non-Patent Citations (2)
| Title |
|---|
| 一种新的高光谱图像波段选择方法;何元磊等;光电工程;126-130 * |
| 面向图像空间信息度量的玻尔兹曼熵;高培超;测绘学报;136 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116563571A (en) | 2023-08-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8675989B2 (en) | Optimized orthonormal system and method for reducing dimensionality of hyperspectral images | |
| US8660360B1 (en) | System and method for reduced incremental spectral clustering | |
| Xia et al. | Identifying recurring patterns with deep neural networks for natural image denoising | |
| JP2008542911A (en) | Image comparison by metric embedding | |
| CN110766708B (en) | Image comparison method based on contour similarity | |
| CN107679539B (en) | Single convolution neural network local information and global information integration method based on local perception field | |
| US11734802B2 (en) | Image processing apparatus, method, and storage medium for patch-based noise reduction | |
| CN111369450A (en) | Method and device for removing Moire pattern | |
| CN111127316A (en) | Single face image super-resolution method and system based on SNGAN network | |
| CN109800815B (en) | Training method, wheat recognition method and training system based on random forest model | |
| CN112785441A (en) | Data processing method and device, terminal equipment and storage medium | |
| CN108537752B (en) | Image processing method and device based on non-local self-similarity and sparse representation | |
| Wei et al. | Effects of lossy compression on remote sensing image classification based on convolutional sparse coding | |
| Omari et al. | A statistical reduced-reference method for color image quality assessment | |
| CN109657083B (en) | Method and device for establishing textile picture feature library | |
| CN111079930A (en) | Method and device for determining quality parameters of data set and electronic equipment | |
| CN112990339B (en) | Gastric pathological section image classification method, device and storage medium | |
| CN116563571B (en) | Boltzmann entropy similarity-based hyperspectral image band selection method and system | |
| CN111753921B (en) | Hyperspectral image clustering method, device, equipment and storage medium | |
| CN116071625B (en) | Training method of deep learning model, target detection method and device | |
| CN116309364A (en) | Transformer substation abnormal inspection method and device, storage medium and computer equipment | |
| CN112861965B (en) | Image matching method based on multi-feature cross consistency model | |
| CN110570376A (en) | image rain removing method, device, equipment and computer readable storage medium | |
| CN111753723B (en) | Fingerprint identification method and device based on density calibration | |
| CN108021874A (en) | A kind of EO-1 hyperion Endmember extraction preprocess method combined based on sky-spectrum |
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 |