CN111385582A - Hyperspectral remote sensing image lossless compression algorithm based on spectrum-space joint prediction - Google Patents
Hyperspectral remote sensing image lossless compression algorithm based on spectrum-space joint prediction Download PDFInfo
- Publication number
- CN111385582A CN111385582A CN202010284317.8A CN202010284317A CN111385582A CN 111385582 A CN111385582 A CN 111385582A CN 202010284317 A CN202010284317 A CN 202010284317A CN 111385582 A CN111385582 A CN 111385582A
- Authority
- CN
- China
- Prior art keywords
- prediction
- spectrum
- image
- remote sensing
- data
- 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.)
- Pending
Links
- 238000007906 compression Methods 0.000 title claims abstract description 55
- 230000006835 compression Effects 0.000 title claims abstract description 55
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 33
- 238000001228 spectrum Methods 0.000 claims abstract description 35
- 238000010219 correlation analysis Methods 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 abstract description 4
- 238000003860 storage Methods 0.000 abstract description 4
- 238000000034 method Methods 0.000 description 25
- BERDEBHAJNAUOM-UHFFFAOYSA-N copper(I) oxide Inorganic materials [Cu]O[Cu] BERDEBHAJNAUOM-UHFFFAOYSA-N 0.000 description 8
- LBJNMUFDOHXDFG-UHFFFAOYSA-N copper;hydrate Chemical compound O.[Cu].[Cu] LBJNMUFDOHXDFG-UHFFFAOYSA-N 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 6
- 230000003595 spectral effect Effects 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 5
- 238000002474 experimental method Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 238000009826 distribution Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 238000013144 data compression Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000013139 quantization Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000002441 reversible effect Effects 0.000 description 2
- 101100136092 Drosophila melanogaster peng gene Proteins 0.000 description 1
- 230000003139 buffering effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000009223 counseling Methods 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000003331 infrared imaging Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 238000005549 size reduction Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000012732 spatial analysis Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000012731 temporal analysis Methods 0.000 description 1
- 238000011426 transformation method Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/182—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a pixel
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/90—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using coding techniques not provided for in groups H04N19/10-H04N19/85, e.g. fractals
- H04N19/91—Entropy coding, e.g. variable length coding [VLC] or arithmetic coding
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
- Compression Of Band Width Or Redundancy In Fax (AREA)
Abstract
The invention discloses a hyperspectral remote sensing image lossless compression algorithm based on spectrum-space joint prediction, which comprises the following steps: establishing an optimal wave band sequence index table according to the correlation among the spectrums; determining a reference waveband according to the index table to predict on a three-dimensional space; obtaining a residual image through a predictor, and performing arithmetic coding on the residual image; and obtaining final compressed data after arithmetic coding. The hyperspectral remote sensing image lossless compression algorithm based on spectrum-space combined prediction adopts the optimal reference waveband to carry out three-dimensional space prediction, more fully utilizes the characteristics of the hyperspectral remote sensing image, realizes effective lossless compression on the precious data of the hyperspectral remote sensing image, can realize lossless compression more quickly on the basis of higher compression ratio, and is effectively applied to practice to solve the problem of huge pressure on transmission and storage caused by the current massive hyperspectral remote sensing data.
Description
Technical Field
The invention belongs to the technical field of image compression, and particularly relates to a hyperspectral remote sensing image lossless compression algorithm based on spectrum-space joint prediction.
Background
Since the 21 st century, the remote sensing platform technology, the sensor technology, the imaging spectrum technology and the space technology are developed at a high speed, so that the hyperspectral remote sensing images with large data volume originally are increased rapidly, great pressure is brought to the transmission and storage of data, and the compression of the data is particularly important. The data of the hyperspectral remote sensing image is required to be high in accuracy, so that more applications are to perform lossless compression processing on the data.
Currently, the hyperspectral remote sensing image compression methods can be summarized into three main categories: transform, vector quantization and prediction. For the transformation method, the compression effect of the plane common image is very obvious, but the data volume of the hyperspectral image is very large, so the time efficiency is low. The vector quantization method can greatly improve the compression effect, but the complexity and the calculation amount of the coding method increase exponentially along with the increase of the vector dimension, and the practical application of the method has certain difficulty. The prediction method is the minimum operation amount of the three algorithms, the principle of the prediction method is easy to understand in lossless coding, the method is easy to realize, and particularly, the redundancy among the spectral images can be effectively removed by using prediction aiming at the characteristic of large correlation among the spectra of the hyperspectral images (a Roche, a week sensitivity, a grand bud, a hyperspectral remote sensing image data compression [ M ]. the national defense industry publisher, 2011.3: 41).
For the development of remote sensing technology, the hyperspectral image lossless compression algorithm (CCSDS algorithm) standard issued by the spatial data system Counseling Committee (CCSDS) in 2012 mainly aims at that the design of a predictor predicts pixels by the sum of pixels of neighborhoods, and a better compression effect is realized. The method comprises the steps of carrying out multi-band prediction on a hyperspectral image by Huo C F, establishing a tree structure by utilizing correlation among bands, and searching an optimal reference band for a current band through the tree structure for prediction (Huo C F, ZHANG R, PENG TX. Lossless compression of hyperspectral images based on searching optimal temporal analysis for prediction [ J ]. IEEE Geo-science and removal sequencing Letters,2009,6(2): 339-.
On the basis of some advanced lossless prediction algorithms of hyperspectral remote sensing images, scholars in China improve the lossless prediction algorithms and obtain better experimental results. Setting a context window for a pixel to be detected based on the lossless compression of a hyper-spectral image reversely searched in the context window, calculating a prediction reference value of the pixel to be detected, carrying out reverse search prediction to obtain a candidate prediction value of the pixel to be detected, and selecting the candidate prediction value closest to the prediction reference value as a final prediction result of the pixel to be detected, wherein the algorithm is superior to single-band prediction (high-power, Liu, Guassa. the lossless compression of the hyper-spectral image based on the reverse search in the context window [ J ]. optical precision engineering, 2015,23(8): 2376-. A hyperspectral image compression method research adopting inter-spectral prediction provides a reference waveband optimization selection method, and the method can obtain a lower minimum mean square error, has a high operation speed and a certain practical value (Meijiang Tao, Liyong, research on a hyperspectral image compression method adopting inter-spectral prediction [ J ] computer engineering and application, 2011,47(4):188 + 190.). In order to reduce the calculation amount of a band sorting algorithm, a hyperspectral image lossless compression algorithm based on band grouping is grouped in advance according to the correlation size of adjacent bands, an optimal backward sorting algorithm is adopted to reorder the bands of each group, neighborhood pixels are used for performing optimal inter-spectral prediction on current prediction pixels, and the compression effect of the hyperspectral image lossless compression algorithm based on band grouping is improved compared with a JPEG-LS algorithm (J. modern electronic technology 2010,22:104 + 106).
The lossless compression ratio of the current hyperspectral remote sensing image is still in a limited level, and on the other hand, a complex conversion method is adopted for a higher compression ratio, so that the algorithm is long in consumed time, and the realization has certain limitation.
Disclosure of Invention
Based on the defects of the prior art, the invention aims to provide a spectrum-space joint prediction-based hyperspectral remote sensing image lossless compression algorithm, so that precious data of a hyperspectral remote sensing image can be effectively compressed in a lossless manner, and the lossless compression can be realized more quickly on the basis of a higher compression ratio, so that the method is effectively applied to practice and solves the problem of huge pressure brought to transmission and storage by the current massive hyperspectral remote sensing data.
In order to solve the technical problems, the invention is realized by the following technical scheme:
the invention provides a hyperspectral remote sensing image lossless compression algorithm based on spectrum-space joint prediction, which comprises the following steps of:
s1: establishing an optimal wave band sequence index table according to the correlation among the spectrums;
s2: determining a reference waveband according to the index table to predict on a three-dimensional space;
s3: obtaining a residual image through a predictor, and performing arithmetic coding on the residual image;
s4: and obtaining final compressed data after arithmetic coding.
Optionally, in step S1, the inter-spectrum global prediction is performed on the hyperspectral images, the inter-spectrum correlation is analyzed on the hyperspectral images, and an index of an optimal prediction reference waveband is obtained after the analysis.
Optionally, in step S2, inter-spectrum prediction is performed according to the index table, and the inter-spectrum prediction is to obtain a prediction residual according to the value of the reference band.
Further, in step S3, the arithmetic coding is entropy coding.
Therefore, the hyperspectral remote sensing image lossless compression algorithm based on spectrum-space joint prediction adopts the optimal reference waveband to carry out three-dimensional space prediction, so that the correlation among spectrums of hyperspectral image data is considered in the space prediction, the data of a better waveband can be used as the reference parameter of a predicted value in the prediction on the three-dimensional space, the characteristics of the hyperspectral remote sensing image are more fully utilized, the precious data of the hyperspectral remote sensing image can be effectively and losslessly compressed, the lossless compression can be more quickly realized on the basis of a higher compression ratio, the lossless compression can be effectively applied to the practice, and the huge pressure brought to transmission and storage by the current massive hyperspectral remote sensing data is solved.
The foregoing description is only an overview of the technical solutions of the present invention, and in order to make the technical means of the present invention more clearly understood, the present invention may be implemented in accordance with the content of the description, and in order to make the above and other objects, features, and advantages of the present invention more clearly understood, the following detailed description is given in conjunction with the preferred embodiments, together with the accompanying drawings.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings of the embodiments will be briefly described below.
FIG. 1 is a flow chart of a hyperspectral remote sensing image lossless compression algorithm based on spectrum-space joint prediction according to the invention;
FIG. 2 is a three-dimensional model diagram of a hyperspectral image of the invention;
FIG. 3 is a schematic representation of a Currite band 100 image of the present invention;
FIG. 4 is a graph of the correlation between the Cuprite adjacent bands of the present invention;
FIG. 5 is a histogram statistical chart of the correlation distribution of the present invention;
FIG. 6 is a gray scale graph showing the correlation between each two of all bands according to the present invention;
FIG. 7 is a graph of the distance of the optimal reference band from the current band in accordance with the present invention;
FIG. 8 is a histogram of the distance band number of the present invention;
FIG. 9 is a graph showing the gray scale of the residual image after inter-spectral prediction according to the present invention;
FIG. 10 is a partial display of residual image data according to the present invention;
Detailed Description
Other aspects, features and advantages of the present invention will become apparent from the following detailed description, taken in conjunction with the accompanying drawings, which form a part of this specification, and which illustrate, by way of example, the principles of the invention. In the referenced drawings, the same or similar components in different drawings are denoted by the same reference numerals.
The invention provides a hyperspectral remote sensing image lossless compression algorithm based on spectrum-space joint prediction. The method effectively utilizes the characteristic of high correlation among the hyperspectral spectrums and carries out three-dimensional space prediction, and compared with the common prediction method, the method obtains a high lossless prediction compression ratio for the hyperspectral remote sensing image.
As for the hyperspectral remote sensing images, because the same ground object reflects and images electromagnetic waves with different wavelengths, one hyperspectral remote sensing image generally comprises dozens of even hundreds of wave bands, and compared with a common two-dimensional image, the hyperspectral remote sensing image is equivalent to three-dimensional cubic data formed by superposing a plurality of two-dimensional images. As illustrated in fig. 2, is a three-dimensional model of a hyperspectral image.
The hyperspectral image is regarded as a three-dimensional data model, wherein f (x, y, z) is recorded as a value of one position in the three-dimensional data, and corresponds to a spatial row position of the hyperspectral image in the hyperspectral data, y is a spatial column position, and z is a frequency spectrum wave band position.
For a current pixel f (x, y, z), the predicted pixel that can be referred to is the value at a certain position in the previous band. For example, the same position value f (x, y, z-1) in the previous band can be used as a prediction reference value, so that the prediction residual g (x, y, z) ═ f (x, y, z) -f (x, y, z-1) is a simple prediction of the forward adjacent band.
The hyperspectral image is purely regarded as a three-dimensional cubic data body, the self characteristics of the hyperspectral image are ignored, and for a data source, the data can be subjected to multi-stage optimization prediction step by step from the self characteristics, so that the correlation among the spectrums can be more fully utilized. Firstly, performing integral prediction among spectrums on a hyperspectral image, analyzing the relevance among spectrums on the hyperspectral image, and obtaining an index of an optimal prediction reference waveband after analysis. Inter-spectral prediction is performed according to the reference band sequence of the index table.
The size of the data is 512 × 614 × 224 which is a hyperspectral image formed by shooting of an onboard hyperspectral imager, the image of the 100 th wave band is extracted and displayed as figure 3, the image is particularly complex in texture detail compared with a common plane image, the amplitude range of the data is much larger than that of the common image, and the data is expressed by 16 bits and is theoretically more than 200 times that of the common image.
Here, description is given of inter-spectrum correlation of the hyperspectral remote sensing image, and for a pixel point (m, n) of a spectrum segment i image a and a spectrum segment i + t image B in an image space, where m is a row of the image space and n is a column of the image space, the correlation of the spectrum segment image i and i + t is defined as follows:
wherein A is a spectral band i image, B is a spectral band i + t image,
is the average value of the pixels of the image a,
is the average of the pixels of image B.
Firstly, the correlation analysis of adjacent bands is performed, and the result is shown in fig. 4.
It can be clearly seen that the correlation of most of the bands is above 0.9, and the statistics of the data distribution histogram can be carried out on the bands. As shown in fig. 5. The grouping can be obtained more accurately by statistical analysis, for example, by using Cuprite data, the number of bands in the range of the correlation above 0.9 is 194, and the number of bands in the range of [0.8,0.9) is 1, except that these bands can have the correlation between the front and back adjacent bands, some bands are special bands, and all the values of the pixels are 0.
Through correlation analysis of adjacent bands, some documents generally set a threshold value according to the correlation grouping of the adjacent bands, so as to include most of the bands with large correlation, and then perform some prediction methods of grouping, and through some improved inter-spectrum prediction algorithms, the compression ratio is improved to a certain extent. However, the grouping method does not consider the optimal band selection, selects the optimal band as a reference, does not perform grouping prediction according to the existing sequence of the bands, and performs full-combination correlation analysis on all the bands. And obtaining an upper triangular square matrix through the full calculation of the image correlation of each frequency spectrum section.
Taking Cuprite data as an example, which has 224 bands, all the bands are arranged and combined in a full manner, and the position information and the data are stored in an experiment to obtain an upper triangular square matrix 223 × 223, the data are subjected to gray scale display as shown in fig. 6, the larger the value of the gray scale image is, the higher the brightness is, and it can be more intuitively seen from fig. 6 that the band with large correlation with the current band is generally in the band close to the current band.
After the wave bands are analyzed, the position of the wave band with the maximum correlation with the current wave band is selected and recorded in an index table C (i). Wherein for some special bands, special values are recorded in the index table c (i).
The time complexity of the correlation calculated quantity after the full permutation and combination of every two wave bands is T (n)2). The number of bands apart from each other in the correlation optimum band index table is shown in fig. 7.
It can be seen that most of it is in several adjacent bands. For a more intuitive observation, a representation of its statistical histogram of the number of segments apart can be seen as shown in fig. 8.
As can be seen from the above figure, the optimal reference band for the Cuprite hyperspectral image is mostly within 8 bands away from the current band.
And obtaining an optimal index table C (i) of the band correlation after the analysis of the correlation among the spectrums. And performing inter-spectrum prediction according to the index table C (i), wherein the inter-spectrum prediction only needs to obtain a prediction residual according to the value of the reference waveband. The time complexity of this step algorithm is low, t (n). Temporary buffering of the residual values obtained by prediction requires another space of the whole data.
Some special bands in the prediction residual obtained at this time are not subjected to inter-spectrum prediction, for example, the first reference band, and some bands have all values of 0. These bands do not affect the prediction of the subsequent spatial dimension. Some of these bands may be even better spatially predicted from a practical point of view.
The residual image obtained after inter-spectral prediction is also a Cuprite image, and the residual image data obtained after inter-spectral prediction is displayed by selecting the residual image of the 100 th band, as shown in fig. 9.
The significant data size reduction of the inter-spectral predicted image can be seen from the figure, and the entropy of the inter-spectral predicted image is 0.3958 by entropy calculation.
If the correlation is still relatively large, inter-spectrum prediction can be performed again. It should be noted that the prediction is not necessarily better than the original one, because the prediction method is effective to compress the data when the magnitude of the residual value is much smaller than that of the original data.
Entropy coding is generally required for storing data, and the data can be effectively compressed and stored after being coded. Entropy coding is a compression theory in information theory. The amount of information that a source can express is represented in entropy. The entropy coding method mainly includes huffman coding, arithmetic coding, run length coding, golomb coding, etc. Huffman coding belongs to a mode of variable length coding, which was originally proposed by Huffman in 1952, and encodes data with different code lengths according to the frequency of occurrence of the data. Run-length coding is a method for data compression by eliminating spatial redundancy, and is greatly influenced by a data source. Run-length coding is mainly represented by using data having the same value consecutively and using the data and its run length. Golomb coding is efficient Huffman coding, and Golomb coding is used in a JPEG-LS compression method.
Huffman coding allows for optimal coding when the probability of occurrence of a signal for the data is negative integer powers of 2. Arithmetic coding is a coding mode close to the information entropy, unlike Huffman coding, which does not require the use of an integer number of codewords. The core idea of arithmetic coding is to represent the data stream to a sub-interval on interval [0,1), recursively partitioning the lower bound and length of the sub-interval according to the probability of each signal of data. The more data signals, the smaller the subintervals generated in the encoding, which means that more bits are needed for representation. The data with high probability has smaller interval length than the data with low probability, and occupies less code words.
Fig. 10 shows a partial screenshot of the data distribution of the Cuprite band 100 prediction residual image.
As can be seen from the partial data display in the above figure, a large number of 0 values appear in the data values of the predicted residual image, and the change amplitude of the values is greatly reduced and concentrated. In this case, arithmetic coding and JPEG-LS coding are used.
The experiments of the algorithm are all performed in the same environment, and the configuration of a computer adopted in the experimental environment is that a processor is Intel Core (TM) i5-4210M @2.60 GHz; the memory size is 12.00 GB; the operating system is Windows10(64 bits); the hard disk size is 1 TB; the processing software used by the algorithm was MATLAB R2017 a.
The experiment introduced by the invention mainly adopts AVIRIS (aircraft Visible Infrared imaging spectrometer) data, wherein the AVIRIS data are data shot by an onboard Visible light/infrared imager and can provide a hyperspectral remote sensing image with 20m spatial resolution and 10nm spectral resolution, the AVIRIS data comprise 224 spectral bands, the spectral coverage range is 200-2400 nm, MATLAB is used for reading the image and obtaining three-dimensional data, the data in the experiment mainly comprise two data of Cuprite (512 × 614 × 224) and Sandiego (400 × 400 × 224), the first data is data of which the size is 512 × 614 × 224, and the second data is data of which the size is 400 × 400 × 224, which are obtained from the Cuprite hyperspectral image.
The information entropy ratio of the residual image of the image after three-dimensional space prediction is shown in table 2.1.
TABLE 2.1 entropy of three-dimensional prediction residual image information
From table 2.1, it can be seen that the information entropy of the image predicted by the three-dimensional space with reference to the optimal band is improved to a certain extent, and the high entropy value enables the predicted image to contain more information. The arithmetic coding and the JPEG2000 coding are selected for the coding of the residual image, and the following table shows the compression ratio of the coded image.
TABLE 2.2 compression ratio of images after encoding
From the data comparison in table 2.2, it can be understood that the result obtained by performing arithmetic coding on the residual image after three-dimensional spatial prediction has a significantly improved compression ratio and a significantly higher compression ratio for the data image with high correlation, compared with the arithmetic coding without spatial prediction. The compression ratio of the image after JPEG2000 coding is also improved to a great extent, and the compression ratio of the image also has high compression ratio for some good data.
The optimal reference waveband is adopted in the design of the spectrum-space joint predictor for three-dimensional space prediction, so that the correlation among spectrums of hyperspectral image data is considered in the space prediction, the data of a better waveband can be used as a reference parameter of a predicted value for the prediction in the three-dimensional space, and the characteristics of the hyperspectral remote sensing image are more fully utilized. The residual image obtained after prediction by the spectrum-space joint predictor adopts arithmetic coding due to the data characteristics, and the compression ratio reaches 5.8879. Experimental results show that the method has a certain degree of improvement on the compression of the hyperspectral image, the time consumption is increased by a few, and the method has no decisive influence on the algorithm implementation. In the experiment, according to the actual situation, most of the optimal wave bands are within 10 adjacent wave bands, and the optimal wave bands can be searched in ten adjacent wave bands, so that the time complexity is greatly reduced, but the compression ratio higher than or equal to the index reference of the optimal lookup table cannot be obtained.
The key point of the method is that the full-combination correlation analysis is carried out on all the wave band images, and an upper triangular square matrix can be obtained through the full calculation of the correlation of each frequency spectrum band image. After the inter-spectrum correlation analysis, obtaining an optimal index table C (i) of the band correlation, and performing inter-spectrum prediction according to the index table C (i), wherein the inter-spectrum prediction only needs to obtain a prediction residual according to the value of a reference band. And an arithmetic coding mode is adopted for the characteristics of the residual image, so that the lossless compression degree of the hyperspectral image is improved. The algorithm fully considers the image characteristics of each link in the spectral-spatial analysis of the hyperspectral image and the entropy coding processing of the residual image, and adopts an effective algorithm to realize an ideal result.
While the foregoing is directed to the preferred embodiment of the present invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof, and the scope thereof is determined by the claims that follow.
Claims (4)
1. A hyperspectral remote sensing image lossless compression algorithm based on spectrum-space joint prediction is characterized by comprising the following steps:
s1: establishing an optimal wave band sequence index table according to the correlation among the spectrums;
s2: determining a reference waveband according to the index table to predict on a three-dimensional space;
s3: obtaining a residual image through a predictor, and performing arithmetic coding on the residual image;
s4: and obtaining final compressed data after arithmetic coding.
2. The lossless compression algorithm for hyperspectral remote sensing images based on spectrum-space joint prediction as claimed in claim 1, wherein in step S1, the hyperspectral images are subjected to spectrum-to-spectrum global prediction, the hyperspectral images are subjected to spectrum-to-spectrum correlation analysis, and an index of an optimal prediction reference band is obtained after the analysis.
3. The lossless compression algorithm for hyperspectral remote sensing images based on spectrum-space joint prediction as claimed in claim 1, wherein in step S2, inter-spectrum prediction is performed according to an index table, and the inter-spectrum prediction is to obtain a prediction residual according to the value of a reference waveband.
4. The lossless compression algorithm for hyperspectral remote sensing images based on spectral-spatial joint prediction according to claim 1, wherein in the step S3, the arithmetic coding is entropy coding.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010284317.8A CN111385582A (en) | 2020-04-13 | 2020-04-13 | Hyperspectral remote sensing image lossless compression algorithm based on spectrum-space joint prediction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010284317.8A CN111385582A (en) | 2020-04-13 | 2020-04-13 | Hyperspectral remote sensing image lossless compression algorithm based on spectrum-space joint prediction |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111385582A true CN111385582A (en) | 2020-07-07 |
Family
ID=71217513
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010284317.8A Pending CN111385582A (en) | 2020-04-13 | 2020-04-13 | Hyperspectral remote sensing image lossless compression algorithm based on spectrum-space joint prediction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111385582A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113613001A (en) * | 2021-09-02 | 2021-11-05 | 沈阳航空航天大学 | Image high-speed compression method and system based on FPGA under CCSDS standard |
| CN115174697A (en) * | 2022-06-30 | 2022-10-11 | 上海航天电子通讯设备研究所 | Spaceborne hyperspectral image lossless compressor |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1632479A (en) * | 2005-01-20 | 2005-06-29 | 北京工业大学 | Lossless compression method for high spectrum image based on three-dimensional prediction |
| US20080219575A1 (en) * | 2003-12-17 | 2008-09-11 | Andreas Wittenstein | Method and apparatus for faster-than-real-time lossless compression and decompression of images |
| CN101720043A (en) * | 2009-11-20 | 2010-06-02 | 北京工业大学 | Imaging spectrum image compression method based on multi-mode prediction |
| CN101883274A (en) * | 2009-05-08 | 2010-11-10 | 中国科学院沈阳自动化研究所 | Spatial-spectral associated prediction-based hyperspectral image lossless compression method |
| CN104270640A (en) * | 2014-09-09 | 2015-01-07 | 西安电子科技大学 | Lossless spectrum image compression method based on support vector regression |
| CN104618718A (en) * | 2014-12-30 | 2015-05-13 | 华中科技大学 | Space-time multi-prediction mode based lossless compression method and system |
| CN107133992A (en) * | 2017-04-17 | 2017-09-05 | 东北大学 | Based on spectrum correlation adaptive grouping high spectrum image distributed associating compression method |
-
2020
- 2020-04-13 CN CN202010284317.8A patent/CN111385582A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080219575A1 (en) * | 2003-12-17 | 2008-09-11 | Andreas Wittenstein | Method and apparatus for faster-than-real-time lossless compression and decompression of images |
| CN1632479A (en) * | 2005-01-20 | 2005-06-29 | 北京工业大学 | Lossless compression method for high spectrum image based on three-dimensional prediction |
| CN101883274A (en) * | 2009-05-08 | 2010-11-10 | 中国科学院沈阳自动化研究所 | Spatial-spectral associated prediction-based hyperspectral image lossless compression method |
| CN101720043A (en) * | 2009-11-20 | 2010-06-02 | 北京工业大学 | Imaging spectrum image compression method based on multi-mode prediction |
| CN104270640A (en) * | 2014-09-09 | 2015-01-07 | 西安电子科技大学 | Lossless spectrum image compression method based on support vector regression |
| CN104618718A (en) * | 2014-12-30 | 2015-05-13 | 华中科技大学 | Space-time multi-prediction mode based lossless compression method and system |
| CN107133992A (en) * | 2017-04-17 | 2017-09-05 | 东北大学 | Based on spectrum correlation adaptive grouping high spectrum image distributed associating compression method |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113613001A (en) * | 2021-09-02 | 2021-11-05 | 沈阳航空航天大学 | Image high-speed compression method and system based on FPGA under CCSDS standard |
| CN113613001B (en) * | 2021-09-02 | 2023-05-19 | 沈阳航空航天大学 | Image high-speed compression method and system based on FPGA under CCSDS standard |
| CN115174697A (en) * | 2022-06-30 | 2022-10-11 | 上海航天电子通讯设备研究所 | Spaceborne hyperspectral image lossless compressor |
| CN115174697B (en) * | 2022-06-30 | 2024-04-12 | 上海航天电子通讯设备研究所 | Satellite-borne hyperspectral image lossless compressor |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | An efficient reordering prediction-based lossless compression algorithm for hyperspectral images | |
| Motta et al. | Compression of hyperspectral imagery | |
| US8615138B2 (en) | Image compression using sub-resolution images | |
| Mielikainen et al. | Lossless compression of hyperspectral images using a quantized index to lookup tables | |
| CN102683149B (en) | Mass analysis data processing method and mass analysis data treatment system | |
| KR101565265B1 (en) | Coding of feature location information | |
| US9123091B2 (en) | Basis vector spectral image compression | |
| CN111385582A (en) | Hyperspectral remote sensing image lossless compression algorithm based on spectrum-space joint prediction | |
| WO2004019003A2 (en) | Image processing of mass spectrometry data for using at multiple resolutions | |
| KR20180077060A (en) | Method and apparatus for encoding and decoding lists of pixels | |
| US20030081852A1 (en) | Encoding method and arrangement | |
| Pereira et al. | A low-cost hardware accelerator for ccsds 123 predictor in fpga | |
| CN112887713B (en) | Picture compression and decompression method and device | |
| Jain et al. | Edge-based prediction for lossless compression of hyperspectral images | |
| CN101720043A (en) | Imaging spectrum image compression method based on multi-mode prediction | |
| CN111107360B (en) | Spectrum-space dimension combined hyperspectral image lossless compression method and system | |
| CN113240761A (en) | High bit depth image lossless compression method suitable for remote sensing satellite | |
| Kong et al. | Lossless compression codec of aurora spectral data using hybrid spatial-spectral decorrelation with outlier recognition | |
| Wenbin et al. | The hyper-spectral image compression based on k-means clustering and parallel prediction algorithm | |
| Wang et al. | A lossless compression of remote sensing images based on ANS entropy coding algorithm | |
| CN113068044B (en) | Iterative hyperspectral image lossless compression method based on low-rank representation | |
| Boddu et al. | Quantum-dot Cellular Automata Based Lossless CFA Image Compression Using Improved and Extended Golomb-rice Entropy Coder | |
| Ni et al. | Onboard lossless compression of hyperspectral imagery based on hybrid prediction | |
| CN103442236A (en) | Remote sensing signal compressed encoding method of multilevel and fractal dimension vector quantization | |
| Song et al. | Differential prediction-based lossless compression with very low-complexity for hyperspectral data |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200707 |
|
| WD01 | Invention patent application deemed withdrawn after publication |